Promotor: Prof. dr. ir. W. Verbeke Department of Agricultural Economics, Ghent University Dean: Prof. dr. ir. G. Van Huylenbroeck Rector: Prof. dr. P. Van Cauwenberge Filiep Vanhonacker SOCIETY VERSUS FARM ANIMAL WELFARE: PUBLIC CONCEPTION, ATTITUDE AND CITIZEN-CONSUMER AMBIVALENCE Thesis submitted in fulfillment of the requirements for the degree of Doctor (PhD) in Applied Biological Sciences Dutch translation of the title: De maatschappij versus het welzijn van landbouwhuisdieren: publieke opvatting, houding en burger-consument ambivalentie Way of citation: Vanhonacker, F. 2010. Society versus farm animal welfare: public conception, attitude and citizen-consumer ambivalence. Doctoral thesis. Ghent University. ISBN-number: 978-90-5989-411-2 The author and the promotor give the authorisation to consult and to copy parts of this work for personal use only. Every other use is subject to the copyright laws. Permission to reproduce any material contained in this work should be obtained from the author. Members of the jury Prof. dr. ir. Wim Verbeke (promotor) Department of Agricultural Economics, Ghent University Prof. dr. ir. Christian Stevens (chairman) Department of Organic Chemistry, Ghent University Dr. ir. Volkert Beekman Section Sustainable Development, Agriculture & Entrepreneurship, LEI – WUR, the Netherlands Prof. dr. ir. Stefaan De Smet Department of Animal Production, Ghent University Prof. dr. Xavier Gellynck Department of Agricultural Economics, Ghent University Prof. dr. Dirk Lips Department Agro- and Biotechnology, KaHo-Sint Lieven Centre for Science, Technology and Ethics, K.U. Leuven Dr. Frank Tuyttens Animal Sciences Unit, Institute for Agricultural and Fisheries Research (ILVO) I Acknowledgements Dit doctoraat is het resultaat van een dikke vijf jaar onderzoek, waarin ik heb kunnen proeven van verschillende onderzoeksthema’s. Dierenwelzijn, traditionele voeding en aquacultuur maakten daarvan het grootste deel uit. Met dierenwelzijn heb ik uiteindelijk gekozen voor het thema dat mij het meeste boeide. Het is ergens grappig vast te stellen dat dit doctoraat een bundeling is van hetgeen mij als kind reeds sterk interesseerde, namelijk cijfers, schrijven en dieren. Mijn ouders zullen beamen dat cijfertjes steeds een soort fascinatie zijn geweest voor mij. Als ik daarnaast terug denk aan de opstellen in de lagere school met als titel ‘Wat ik later worden wil’ koos ik steevast voor ‘journalist’ of ‘zoöloog’. Ik kan dan ook na vijf jaar zeggen dat ik dit werk graag gedaan heb en graag doe. Aangezien het schrijven van een doctoraat niet de verdienste van één persoon is, wil ik graag deze mogelijkheid aangrijpen om enkele personen te bedanken. Eerst en vooral wil ik mijn promotor prof. Wim Verbeke bedanken. Wim, bedankt voor de kansen die ik gekregen heb, de amicale manier van samenwerken, de vrijheid en het vertrouwen die ik steeds heb genoten. Jouw brede kennis, je vele contacten, je snelle analyses en kritische kijk op de zaken hebben zeker positief bijgedragen tot dit finaal resultaat. Daarnaast wil ik ook de voorzitter en de leden van de examencommissie bedanken voor de tijd die ze genomen hebben voor het kritisch nalezen van mijn manuscript. Frank, extra woordje voor jou: bedankt voor de goede, leerrijke en vooral fijne samenwerking met jou en je groep doorheen de jaren. Daarnaast wil ik ook een woordje van dank richten aan alle collega’s die doorheen deze vijf jaar mijn pad zijn gekruist, hier aan de vakgroep of via de verschillende onderzoeksprojecten waarin ik betrokken was. Zuzia en Stijn, als mijn ‘ervaren’ eerste office-mates kon ik bij jullie steeds terecht voor vragen, zowel werk- als privégerelateerd. Ook wil ik onze onderzoeksgroep-unit niet vergeten. Zuzia, Armando, Christine, Lynn en Pieter, niet zelden zitten we eens bij elkaar in de bureau voor een korte en deugddoende babbel. Ook bedankt aan de rest van de vakgroep. De cafébezoekjes de vrijdag na het werk, het kaarten over de middag, de barbeques, ... waren een leuke manier om elkaar ook naast het werk beter te leren kennen en bevorderlijk voor de werksfeer. Ten slotte wil ik ook Elza hierbij niet vergeten. Het eerste project dat ik uitvoerde was er één samen met jou. Een kort, maar zeer leuk project waarin we samen heel wat werk hebben verzet. De lange telefoondiscussies, de vele treintrips doorheen Vlaanderen om te enquêteren, op congres naar Oslo zijn maar enkele zaken die mij zullen bij blijven. Het afleggen van een doctoraatsparcours is verder ook niet mogelijk zonder een goed evenwicht tussen werk en ontspanning. Die ontspanning vond ik niet enkel in de sport, maar ook in het samenzijn met vrienden. Bedankt dus, Mathieu en Riet, mini-landbouw, poker-, pingpong-, squash-, tennis-, en racketlonvrienden, oud-kotgenootjes, oud-studiegenootjes en al dezen die ik hierbij geheel onterecht vergeet. III Ik heb ook steeds kunnen terugvallen op een hechte familie en schoonfamilie. Ma en pa, bedankt voor de financiële input en ons de mogelijkheid te geven om verder te studeren. Ook bedankt, net als de schoonoudjes, voor alle praktische en technische hulp met huis, tuin, Timon, .... (Schoon)broers en (schoon)zussen, al zijn de zondagse bezoekjes vaak wat hectisch met al dat rondlopend wild (en belooft het nog heel wat hectischer te worden), het is toch iets waar we steeds naar uitkijken. Afsluiten wil ik met mijn eigen gezinnetje. Joke, bedankt voor de mooie jaren samen en de extra dimensie die je aan mijn leven hebt gegeven. Ik zou het niet anders willen. Timon is mijn grote held. Ik ben ervan overtuigd dat we hem samen een warm nestje zullen geven en kijk vol nieuwsgierigheid uit naar onze tweede spruit. IV Table of contents Members of the jury I Acknowledgements III Table of contents V List of tables XI List of figures XVI List of abbreviations XVIII PART I General introduction, objectives and outline of the thesis 1. General introduction 3 1.1. Relevance of the topic 3 1.2. Scope of the thesis 8 1.2.1. Public conception 9 1.2.2. Ambivalence and attitude variety: need for a segmented and targeted approach 2. 12 1.2.3. Piglet castration 14 Conceptual framework 17 2.1. Theory of reasoned action 18 2.2. Modifiers of farm animal welfare attitude 19 2.2.1. Social structural variables – Place-related variables 20 2.2.2. Social structural variables – Other variables 21 2.2.3. Individual animal-related experiential variables 21 2.3. Factors influencing the ambivalence between attitude and behavioural intention 22 2.3.1. Perceived consumer effectiveness (PCE) 23 2.3.2. Information 24 2.3.3. Availability 25 2.3.4. Cost 26 2.3.5. Dissociation 27 3. Research objectives and hypotheses 28 4. Research design and data sources 29 4.1. Study 1: Secondary data sources 30 4.2. Study 2: “Defining the concept farm animal welfare” 31 4.3. Study 3: “Farm animal welfare and consumer behaviour” 33 4.4. Study 4: “Surveying Flemish pig producers about their attitude towards surgical castration without anaesthesia and its alternatives” 4.5. 34 Study 5: “Consumer perception towards immunocastration as alternative practice for surgical castration of piglets” 34 V 4.6. Study 6: Immunocastration versus surgical castration with anaesthesia/analgesia 35 5. Thesis outline 36 PART II Public conception of farm animal welfare CHAPTER 1. Public conception of farm animal welfare – qualitative approach 41 1. Introduction 42 2. Material and methods 43 2.1. Survey (Study 1) 43 2.2. Focus group discussions (part of Study 2) 45 3. Results 3.1. Citizens’ conceptualisation of farm animal welfare 46 3.2. Personal relevance attached to animal welfare 48 3.2.1. Survey 48 3.2.2. Focus group discussions 49 Information and animal welfare 50 3.3.1. Survey 50 3.3.2. Focus group discussions 51 3.3. 4. 46 Discussion and conclusion 53 CHAPTER 2. Public conception of farm animal welfare – quantitative approach 55 1. Introduction 56 2. Material and methods 57 2.1. Qualitative research – focus group discussions and literature search 58 2.2. Quantitative research – survey 59 2.2.1. Study design and subjects 59 2.2.2. Survey 59 2.2.3. Exploratory factor analysis 59 2.2.4. Confirmatory factor analysis and construct validity 60 In depth interviews 60 2.3. 3. 4. VI Results 61 3.1. Exploratory factor analysis 61 3.2. Confirmatory factor analysis and construct validity 61 3.3. In depth interviews 66 Discussion and conclusion 69 PART III Public attitudes, perceptions and concerns towards farm animal welfare & attitude modifiers CHAPTER 3. Concept of farm animal welfare: citizens versus farmers 75 1. Introduction 76 2. Material and methods 77 2.1. Study design and subjects 77 2.2. Survey 78 2.3. Statistical analyses 79 3. Results 79 3.1. Perceived importance of animal welfare aspects 79 3.2. Evaluative beliefs of the aspects’ practice in Flemish animal production 82 3.3. Discordance between citizens and farmers 83 4. Discussion 84 5. Conclusion 87 CHAPTER 4. Welfare Quality® definition: opinion of citizens, farmers and vegetarians 89 1. Introduction 90 2. Material and methods 90 2.1. Survey 90 2.2. Statistical analyses 91 3. 4. Results 91 3.1. Importance of the 12 Welfare Quality® criteria 91 3.2. Relative weights of the four Welfare Quality® principles 92 Discussion and conclusion 94 CHAPTER 5. Piglet castration – Flemish citizen sample 97 1. Introduction 98 2. Material and methods 99 2.1. Survey 99 2.2. Statistical analyses 101 3. 4. Results 102 3.1. Consumer awareness of piglet castration 102 3.2. Consumer concerns about pig production practices 102 3.3. Impact of the information message related to immunocastration 104 3.4. Consumer attitude towards immunocastration 105 3.5. Self-reported willingness to pay for pork from immunocastrated pigs 105 Discussion and conclusion 106 VII CHAPTER 6. Piglet castration – European citizens 109 1. Introduction 110 2. Material and methods 112 2.1. Measurement and scaling 112 2.2. Statistical analyses 113 3. Results 115 3.1. Descriptive analyses 115 3.2. Segmentation analysis 116 3.3. Between-group comparisons 119 3.4. Acceptance of both alternative methods 119 3.5. Preference for the vaccine method over physical castration 121 3.6. Associations with claimed pork purchasing intentions 121 4. Discussion 123 5. Conclusions 125 CHAPTER 7. Piglet castration – Flemish producers 127 1. Introduction 128 2. Materials and methods 129 2.1. Survey 129 2.2. Statistical analyses 130 3. 4. Results 130 3.1. Sample profile 130 3.2. Surgical castration without anaesthesia 131 3.3. Castration with anaesthesia and/or analgesia 132 3.4. Immunocastration 133 3.5. Sperm sexing 135 3.6. Production of entire males 135 3.7. Ranking the alternatives 136 3.8. Influencing factors 137 Discussion and conclusion 138 CHAPTER 8. Modifiers of societal concerns related to stocking density 141 1. Introduction 142 2. Material and methods 143 2.1. Sample and procedure 143 2.2. Dependent variables 144 2.3. Statistical analyses 146 3. VIII Results 146 3.1. Relative importance of stocking density, pen size and group size 146 3.2. Influence of socio-structural variables 147 3.3. Relation between perceptions of stocking density, pen size and group size 149 4. Discussion and conclusion 150 CHAPTER 9. Acknowledging consumer variety: a segmentation analysis 153 1. Introduction 154 1.1. Background 154 1.2. Public attitude toward animal welfare 154 1.3. Animal welfare related consumer behaviour 155 1.4. Scope and objectives 156 2. 3. Materials and methods 157 2.1. Measurement of constructs 157 2.2. Statistical analyses 158 Results 3.1. 3.2. 4. 159 Segmentation analysis 159 3.1.1. Segmentation variables 159 3.1.2. Cluster analysis 159 Profiling of the clusters 161 3.2.1. Determinants of animal welfare perception 161 3.2.2. Meat and meat substitute consumption 163 3.2.3. Knowledge 164 3.2.4. Information variables 164 Discussion and conclusion 165 PART IV Factors influencing the citizen-consumer ambivalence CHAPTER 10. Citizen-consumer ambivalence: influencing factors 171 1. Introduction 172 1.1. State of the art 172 1.2. Rationale and scope 172 2. 3. Material and methods 173 2.1. Study design and subjects 173 2.2. Survey 174 2.3. Statistical analyses 176 Results and discussion 176 3.1. Attitude towards animal welfare and pro-welfare behaviour 176 3.2. Socio-demographic characteristics and meat consumption 178 3.3. Role of product attribute perceptions 179 3.4. Role of perceived consumer effectiveness 181 3.5. Outcome variables – interest in welfare labelling 182 3.6. Impact of having a pet on attitudes towards farm animal welfare 184 IX PART V Discussion and conclusion 1. Recapitulation 189 2. General discussion and policy recommendations 191 2.1. Public conception of farm animal welfare 191 2.2. Determinants of public attitudes towards farm animal welfare 192 2.3. Citizen attitudes versus consumer behaviour 194 2.4. Piglet castration 198 Limitations and future research 200 3. ANNEX 1. Improving animal welfare – solutions and stakeholder roles 1. 2. Selected overview of solutions proposed to improve the level of animal welfare 203 1.1. Four ways to progress (Aerts and Lips, 2010) 203 1.2. Defra report (McInerney, 2004) 203 1.3. EU FAIR CT98-3678 (Harper and Henson, 2001) 204 1.4. LEI-report a (Ingenbleek et al., 2006a) 204 1.5. LEI-report b (de Bakker and Dagevos, 2010) 204 Role of different actors 205 2.1. European Union 205 2.2. National governments 207 2.3. Non-governmental organisations 209 2.4. Retail sector 210 2.5. Veterinarians 210 2.6. Scientists 211 2.7. Industry / Animal production chain 211 2.8. Media 212 2.9. Citizens and consumers 212 Summary 213 Samenvatting 215 References 217 Scientific curriculum vitae 240 X List of tables Table 1.1. Impacts of improvements in production efficiency on farm animal welfare (source: Winter et al., 1998, p. 310) 7 Table 1.2. The Five Freedoms as described by the Farm Animal Welfare Council (1993) 8 Table 1.3. A summary of different understandings of animal welfare (source: Fisher, 2009) 9 Table 1.4. Welfare Quality’s® operational definition of farm animal welfare with 12 criteria aggregated into 4 principles (Botreau et al., 2008) 11 Table 1.5. Sample characteristics of the survey respondents within Study 1 31 Table 1.6. Sample characteristics of the survey respondents within Study 2 32 Table 1.7. Sample characteristics of the survey respondents within Study 3 33 Table 1.8. Sample characteristics of the survey respondents within Study 5 (%, n=225) 34 Table 1.9. Sample characteristics of the survey respondents within Study 6 35 Table 2.1. Characteristics of the focus group participants 46 Table 2.2. List of aspects that reflect how people think about or conceptualise farm animal welfare. The list is based on the focus group discussions, and aspects are arbitrarily assigned to a dimension (in capitals), and possible sub-dimensions (in italic) Table 2.3. 47 Comparing animal welfare to other product attributes. Mean values with standard deviations on five-point Likert scales are provided, together with the share of the sample that attributes a higher importance (scores 1 and 2; % less important), an equal importance (score 3), and a lower importance (scores 4 and 5; % more important) to animal welfare Table 2.4. 48 Citizens’ view on information provision and monitoring in relation to animal welfare. Mean value and standard deviation on five-point Likert agreement scales. Sample share disagreeing (scores 1 and 2), answering neutral (score 3), and agreeing (scores 4 and 5) Table 2.5. 51 Full list of aspects used for framing the public conception of farm animal welfare. Between brackets the aspect coding is given that will be used in the syntax of the confirmatory factor analysis and in further reporting 58 XI Table 2.6. Output of the exploratory factor analysis on the full list of aspects. Only the factor loadings above 0.30 are presented for the twelve factors with Eigenvalue above 1 62 Table 2.7. Results of the confirmatory factor analysis: measurement model and goodness of fit statistics for the consecutive models Table 2.8. 63 Standardised factor loadings of the retained aspects to the final model (M4) with 4 environment-based and 3 animal-based dimensions 65 Table 2.9. Correlation matrix of the animal welfare dimensions 66 Table 2.10. Aspects that were suggested to be removed or added by experts 69 Table 3.1. Perceived importance and evaluative belief scores on 72 aspects related to animal welfare by citizens and farmers. Aspects are classified per dimension and ranked based on citizens’ mean perceived importance score (5-point Likert scales) Table 3.2. 81 Structure of the Welfare Quality® operational definition of farm animal welfare with 12 criteria aggregated into four principles. For each of the 12 criteria the mean weights (for the different animal types) derived by Welfare Quality® are presented (WQ weight) as well as the mean (+SD) importance score (on a scale from 1 to 10) allocated by adult Flemish citizens, farmers and vegetarians. Significant differences between these types of respondents (rows) are indicated by a different letter (x, y, z) in superscript before the score. Within type of respondent (columns), importance scores without a common superscript (a-n) after the number differ significantly (p<0.05) Table 3.3. 93 Average (+SD) relative weights (scored as %) allocated by adult Flemish citizens, farmers and vegetarians to the four principles of animal welfare as defined by Welfare Quality®. Significant differences between types of respondents (rows) are indicated by a different letter (x-z) in superscript before the figure. Within type of respondent (columns), significant differences are indicated by a different superscript (a-d) after the figure Table 3.4. Message conditions tested in the survey: message content and number of respondents (n) exposed to each message Table 3.5. 94 101 Factor analysis: Flemish consumer concerns related to pig production. Factor loadings from principal component analysis. The right column reports the mean concern score and standard deviation (SD) for the total sample (n=225) XII 103 Table 3.6. Consumer evaluation of immunocastration relative to surgical castration. Figures are mean values on a seven-point scale, that ranges from “much worse (than surgical castration)” (1) to “much better” (7). Mean values are presented for the full sample and for the different message conditions 104 Table 3.7. Textual information provided to participants 114 Table 3.8. Descriptive analyses for the different countries 117 Table 3.9. Profile of consumer segments on segmentation variables; relative (sum across attributes equals 4) and absolute importance scores Table 3.10. Profile of the different segments in terms of socio-demographics, pork consumption frequency and absolute product importance scores Table 3.11. 121 Relation between claimed behaviour and attitude measures related to piglet castration Table 3.14. 120 Intention to seeking out and avoidance of pork from physically castrated and vaccinated male pigs, % of consumers Table 3.13. 119 Combination of the responses on acceptance of physical castration method and acceptability of vaccine method Table 3.12. 118 122 Attitude of Flemish pig farmers with regard to surgical castration without anaesthesia. Percentage of responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option Table 3.15. 132 Attitude of Flemish pig farmers with regard to surgical castration with anaesthesia and/or analgesia. Percentage of responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option Table 3.16. 133 Attitude of Flemish pig farmers with regard to immunocastration (IC). Percentage of responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option Table 3.17. 134 Attitude of Flemish pig farmers with regard to sperm sexing (SS). Percentage of responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option 135 XIII Table 3.18. Attitude of Flemish pig farmers with regard to the production of entire males (EM). Percentage of responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option Table 3.19. 136 Mean rankings of the different practices to eliminate boar taint. A lower score corresponds with a more positive ranking (SCN: surgical castration without anaesthesia; SCA: surgical castration with anaesthesia; IC: immunocastration; SS: sperm sexing; EM: production of entire males) Table 3.20. 137 Influencing factors on the rankings of the different practices to eliminate boar taint. Figures present mean rankings, where a lower score corresponds with a more favourable ranking (SCN: surgical castration without anaesthesia; SCA: surgical castration with anaesthesia; IC: immunocastration; SS: sperm sexing; EM: production of entire males) Table 3.21. 138 Perceived importance, evaluative belief, and concern scores for 23 animal welfare aspects in dataset 1 (n=521); mean scores with standard deviation on 5-point scale; aspects are ranked according to their (descending) concern score Table 3.22. 146 Perceived importance (PerImp), evaluative belief (EvalBel) and concern (Concern) scores of the 72 animal welfare aspects listed in dataset 2 (n=459); mean (standard deviation) on 5-point scale; aspect are ranked according to their (descending) concern score Table 3.23. 148 Influence of socio-demographic characteristics on the perceived importance, the evaluative belief and the concern towards stocking density; mean scores on 5-point scales Table 3.24. 149 Regression analysis with stocking density as dependent variable and pen size and group size as explanatory variables; dataset 2 (n=459) 150 Table 3.25. Profile of the segments on the segmentation variables (n=459) 160 Table 3.26. Determinants of animal welfare for the different segments (n=429), frequency distributions (%); total sample characteristics are mentioned between brackets 163 Table 3.27. Table 3.28. XIV Profiling of the segments based on meat and meat substitute consumption frequency; mean scores 163 Profiling of the segments based on knowledge and information variables 165 Table 4.1. Attitude towards animal welfare in terms of welfare importance and welfare evaluation for consumer groups with a different level of pro-welfare behaviour as a reflection of an possible attitude-behaviour gap Table 4.2. 178 Socio-demographic profile of segments with different levels of pro-welfare behaviour (n=400). The non-shaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat 178 Table 4.3. Meat consumption frequency of segments with different levels of pro-welfare behaviour. The non-shaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat Table 4.4. 179 Drivers in choosing animal products. Attributes are ranked with descending mean per column (n=400). The non-shaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat 180 Table 4.5. Possible barriers and motives for engagement in pro-welfare behaviour. The nonshaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat 180 Table 4.6. Predictor values for the level of pro-welfare behaviour 181 Table 4.7. Pro-welfare behaviour and perceived consumer effectiveness (PCE) 181 Table 4.8. Pro-welfare behaviour and perception on welfare label issues. The non-shaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat Table 4.9. 182 Attitudes towards farm animal welfare. Impact of having a pet in the household 185 Table A1.1. Policy options for indicating animal welfare related information on products of animal origin 207 XV List of figures Figure 1.1. Consumption of meat in Belgium per capita per year, in kg carcass weight, 19502000 (source: Niesten et al., 2003, p.178) 5 Figure 1.2. Conceptions of animal welfare 10 Figure 1.3. Citizen attitude versus consumer opinion. Percentage of respondents that agree on statement “thinking of animal welfare in general” versus “thinking of animal welfare when buying” (source: Welfare Quality® Conference Proceedings, 17/18 November 2005, Brussels, Belgium, p.31) Figure 1.4. 14 Estimates of the percentage of male pigs (conventional production) surgically castrated per country (source: Fredriksen et al. (2009)) 15 Figure 1.5. Vaccination timing of piglet castration (www.berengeur.com) 16 Figure 1.6. Conceptual framework used to discuss modifiers of public attitudes, perceptions and concerns towards farm animal welfare, and factors that influence the relationship between attitude and behavioural intention 17 Figure 1.7. Outline of the thesis in relation to the conceptual framework 36 Figure 2.1. Association of higher welfare standards with other product attributes. Mean scores above 3 correspond with positive associations, mean scores below 3 with negative associations 49 Figure 2.2. Methodological summary 57 Figure 2.3. Final structure of farm animal welfare conception 68 Figure 3.1. Mean scores of perceived importance of 72 aspects in obtaining an acceptable level of farm animal welfare according to farmers (Y-axis) and citizens (X-axis). The line in the figure is the bisector, indicating equal mean scores. Colours correspond to aspects classified in the same dimensions. For the aspects corresponding with the numbers, please see Table 3.1. Mind that the scales of the X and Y axes are not equal Figure 3.2. 80 Mean scores of evaluative belief of 72 aspects according to both farmers (Y-axis) and citizens (X-axis). The line in the figure is the bisector, indicating equal scores. Colours correspond to aspects classified in the same dimensions. For the aspects corresponding with the numbers, please see Table 3.1. Mind that the scales of the X and Y axes are not equal XVI 82 Figure 3.3. Discordance in perception between citizen and farmer, depicted for each dimension separately. The X-axis corresponds with the perceived importance at farmer level; the Y-axis corresponds with the gap (perceived importance – evaluative belief) at citizen level. For the aspects corresponding with the numbers, see Table 3.1. Figure 3.4. 84 General attitude of consumers towards immunocastration (IC) relative to attitude towards the traditional practice of surgical castration. Answers were registered on a seven-point scale ranging from immunocastration perceived as (1) “much worse” to (7) “much better” Figure 3.5. 105 Self-reported willingness to pay (WTP) a price premium (expressed as percentage over the price conventional pork) for meat from immunocastrated pigs. Selfreported WTP is measured on a seven-point scale ranging from “very unlikely to buy” (1) to “very likely to buy” (7) Figure 3.6. 106 General attitude of farmers towards immunocastration (IC) relative to attitude towards the routine practice of surgical castration without anaesthesia. Answers are registered on a seven-point scale ranging from IC perceived as “much worse” to “much better” Figure 3.7. Conceptualisation of concern as a combination of perceived importance (PI) and evaluative belief (EB) Figure 3.8. 134 145 Mapping of the cluster centres according to relative importance (RI z-score) and evaluation (EV z-score) of farm animal welfare; the size of markers reflects cluster size Figure 4.1. Buying behaviour of higher welfare eggs and higher welfare chicken meat (%, nonvegetarians: n=400) Figure 4.2. 177 Self-reported willingness to pay for higher welfare eggs, registered on a five-point Likert scale ranging from “very unlikely” (1) and “very likely” (5) Figure 4.3. 160 183 Self-reported willingness to pay for higher welfare chicken meat, registered on a five-point Likert scale ranging from “very unlikely” (1) and “very likely” (5) 183 Figure A1.1. Different country approaches (source: Kjaernes et al. (2007)) 208 Figure A1.2. Supply chain funnel on European level (Ingenbleek et al., 2004, p.58) 210 XVII List of abbreviations ALT Afdeling Land- en Tuinbouw (Vlaamse overheid) AMS Afdeling Monitoring en Studie (Vlaamse overhead) ANI Animal Needs Index ANOVA Analysis of Variance CAFO Concentrated Animal Feeding Operations CAP Common Agricultural Policy CFI Comparative Fit Index DF Degrees of Freedom DG SANCO Directorate General for Health and Consumer Affairs EEC European Economic Community EFSA European Food Safety Authority EM Entire Males EU European Union EWBL Economie, Werkgelegenheid, Binnenlandse Aangelegenheden en Landbouw FAWC Farm Animal Welfare Council FOD Federale Overheidsdienst GAIA Global Action in the Interest of Animals GnRH Gonadotropin-Releasing Hormone GmbH Gesellschaft mit beschränkter Haftung IC Immunocastration IGD The Institute of Grocery Distribution LEI Landbouw Economisch Instituut LNV Ministerie van Landbouw, Natuur en Voedselkwaliteit (NL) MAX Maximum MIN Minimum NA Not Applicable NGO Non-Governmental Organisation NNFI Non-Normed Fit Index XVIII OIE World Organisation for Animal Health (Office International des Epizooties) OIVO Onderzoeks- en Informatiecentrum van de Verbruikersorganisaties PCE Perceived Consumer Effectiveness RMSEA Root Mean Square Error of Approximation SCA Surgical Castration with anaesthesia and/or analgesia SCN Surgical Castration without anaesthesia or analgesia SD Standard Deviation SPSS Statistical Package for the Social Sciences SS Sperm sexing VILT Vlaams Infocentrum Land- en Tuinbouw WTO World Trade Organization WTP Willingness to pay WQ Welfare Quality® WWF World Wildlife Fund XIX XX PART I General introduction, objectives and outline of the thesis General introduction 1. General introduction 1.1. Relevance of the topic Farm animal welfare from a societal or public perspective entails a topical, challengeable and valuable research theme. “Topical” as a stand-alone issue as well as an inherent and significant part of the wider concept of sustainability (Boogaard et al., 2008); sustainability which is a vogue word in the current consumption society. Korthals (2005) defined issues relating to animal welfare in a list of the seven most common substantive consumer concerns by European consumers. Next to widespread public concerns about animal welfare (Mejdell, 2006; Serpell, 1999), its importance is acknowledged by all stakeholders along the animal production chain (Bracke et al., 2005). It has been recognised as one of the key dimensions in a responsible food chain (Pouta et al., 2010) and is the more and more required for market access. “Challengeable” first because of the complexity of the concept farm animal welfare (Appleby, 1999b; Fraser, 1995; Lund et al., 2006; Mason and Mendl, 1993), because of the debate about what good or acceptable welfare exactly is (Duncan, 2005) and because of the specific properties of animal welfare as a food product attribute. Second because of the complexity associated with research into public1 attitudes, perceptions and behaviour or behavioural intentions (Engel et al., 1990). “Throughout the development of the sociology of consumption, there has been very little consensus on consumers and their characteristics” (George Ritzer (2001) in Dagevos (2005, p.32)) “Valuable” because of the role that the public can play in improving the farm animal’s welfare; improvements which have not (yet) been proportional to the amount of scientific research for animal welfare (De Jonge and Goewie, 2000). Heleski et al. (2006) indicated economics, lack of consumer willingness to pay, tradition, producer attitudes and inadequate welfare science research as the five most named obstacles to improve the animals’ welfare. So far, research on animal welfare has been too much expert-oriented, while involving the public in a constructive dialogue is deemed highly desirable and necessary (Boogaard et al., 2010). The importance of insights in public attitudes and values is nicely reflected in the following quote: “…no amount of scientific evidence will ever be sufficient to bring about improvements in animal welfare unless this evidence also speaks to public attitudes and values…” (Serpell, 2004, p.145) 1 I deliberately choose to use the word ‘public’, as a kind of umbrella term that covers both the individual in his/her role of ‘consumer’ and ‘citizen’. The opposition between consumer and citizen will be part of this PhD dissertation and is discussed in more detail in subsequent parts. 3 Part I In the first part of this general introduction, I will provide a short overview of some evolutions and events, which have been – in my opinion – crucial for the emerging topicality of farm animal welfare. The second part will be on the current knowledge available about the conceptualisation of farm animal welfare and about public attitudes, perceptions and behaviours in relation to farm animal welfare. As these will be central themes of this PhD dissertation, this part will be more extended and approached as the prelude for the conceptual framework (Part I.2). Humans have lived together with animals for as long as they exist. In that sense, human-animal relationships have always been an issue of interest for humans and for scientific research. Whereas scientific literature has mainly focused in the past on animal welfare in relation to laboratory animals, companion animals, and circus, zoo and/or wild animals, research in relation to farm animals is rather new and has strongly boomed over the last decades (Bayvel, 2005; Hewson, 2004). The societal topicality is reflected in the rise of terminology like sustainable behaviour (e.g. Vermeir and Verbeke, 2006), ethical or green consumerism (e.g. Manning et al., 2006; McEachern and Schröder, 2002; Shrum, 1995), corporate social responsibility (e.g. Ingenbleek and Immink, 2010) and studies explicitly referring to an increased interest for farm animal welfare as a public issue (e.g. Krystallis et al., 2009; Verbeke, 2002; Verbeke, 2009) and as a product attribute (e.g. Ingenbleek et al., 2006a; Verbeke and Viaene, 2000). The scientific topicality is reflected in the high number of recent studies and the multitude of disciplines that cover the issue (Hewson, 2004). From the 687 hits (articles and review papers only) that appear in the ISI Web of Science (08/09/2010) for the search term “farm animal welfare” (starting year 1955), 79.2 percent has been published from the year 2000 on. An even higher percentage (83.3 percent) appears when it concerns the search term “animal welfare AND consumer”. The institutional topicality, present on EU-level, is exemplified by the following selection. The Amsterdam Treaty (1999) has introduced a protocol2 on the protection and welfare of animals and officially recognised that animals are sentient beings. It further required the European institutions and Member States to give full regard to the welfare requirements of animals in formulating and implementing community legislation. The European Commission has further financed the Welfare Quality® project (2004-2009), an integrated project within the 6th Framework Programme that aimed to accommodate societal concerns and market demands, to develop reliable on-farm monitoring systems, product information systems and practical speciesspecific strategies to improve animal welfare (www.welfarequality.net). The Community Action Plan on the Protection and Welfare of Animals 2006-2010 is another reflection of topicality on EU-level. This plan aimed to realise the integration of animal welfare in European legislation. Also, funds are given to projects dealing with a specific animal welfare problem, such as PIGCAS for the case of 2 The protocol reads as follows: “The high contracting parties, desiring to ensure improved protection and respect for the welfare of animals as sentient beings, have agreed upon the following provision, which shall be annexed to the Treaty establishing the European Community, in formulating and implementing the Community’s agricultural, transport, internal market and research policies, the Community and the Member States shall pay full regard to the welfare requirements of animals, while respecting the legislative or administrative provisions and customs of the Member States relating in particular to religious rites, cultural traditions and regional heritage.” 4 General introduction piglet castration. Finally the European Commission has devoted a Special Eurobarometer study on the attitudes of consumers towards the welfare of farmed animals (European Commission, 2005). I have categorised the evolutions and events that account for the emerging (societal) relevance of farm animal welfare under three main points: the rise of intensive animal production; changing public attitudes and values; and some specific isolated events. Over the past century, Western animal production has changed dramatically and has resulted in a profound change of the wider agricultural panorama. The hunger experienced during and after two World Wars and the aim to never feel hunger again led to a strong reorientation of agriculture towards productivity goals (Boogaard et al., 2006). Farms underwent a process of reform, yielding larger-scaled and more specialised units, a process that was necessary in order to survive economically (Winter et al., 1998). This restructuring was supported by increased scientific research (to maximise production and profitability), technological improvements (to increase mechanisation and motorisation), and the development of more efficient logistic systems including refrigeration. The level of intensification, specialisation and mechanisation was further expedited by the trade internationalisation, increasing competition, decreasing margins, high land prices, low transport costs, high labour costs, the availability of cheap animal feed and an increased demand for meat caused by population growth and increased prosperity in Western societies. Meat changed from a luxury product for the wealthy to a daily and accessibly staple food for everybody (Niesten et al., 2003, p.178). The resulting increase in meat consumption for Belgium is presented in Figure 1.1. 60 kg carcass weight / capita / year 50 40 30 20 10 0 1950-1955 Beef & Veal 1965-1970 Pork 1985-1990 Chicken 1995-2000 Horsemeat, rabbit & game Figure 1.1. Consumption of meat in Belgium per capita per year, in kg carcass weight, 1950-2000 (source: Niesten et al., 2003, p.178) 5 Part I The focus on maximising productivity, production efficiency and profitability led to the development of intensive animal rearing conditions (e.g. indoor housing to maximise production within the minimum space) that rarely took animal welfare into account. Next to some benefits compared to other systems (e.g. hygiene and health control compared to extensive outdoor systems) more intensive production systems are also associated with some negative impacts on the animal’s welfare (e.g. Tuyttens et al., 2008; Winters et al., 1998, see Table 1.1). This negative association is especially vivid in the consumer perception point of view in a stigmatised way. In addition to the intensification of animal production, some changes at consumer and citizen level have contributed to the current significance of farm animal welfare. Prosperity in Western society strongly increased during the second half of the past century. Food security was achieved, real incomes raised, education levels increased and the population moved away from the countryside to urban places. All this has positively contributed to an increased interest for animal welfare (Hughes, 1995; Seamer, 1998) and an increasing importance of qualitative product traits. Quality is a very broad concept. From a consumer perspective, Grunert et al. (2000) defines four dimensions: sensory characteristics, healthiness, convenience and process characteristics. This last dimension fits in with the emergence of sustainability and ethical issues as important themes in food production and consumption debates (Vermeir and Verbeke, 2006). Within the concept of sustainability and agricultural ethics, animal welfare arises next to issues like environmental friendliness, fair trade, landscape management, biodiversity and a viable and profitable rural livelihood for producers. Food security further triggered a reversal of the food chain, with a shift from a supply orientation to a demand orientation. As such, public concern for, and interest in animal welfare has its impact throughout the animal and meat production chain. This was also uttered by Grunert (2006) who identified individuals’ concerns for animals and the environment as one of four key areas of relevance to future patterns of livestock and meat production. This tendency is already reflected in the consumption of organic food (Makatouni, 1999), the rediscovery of a healthy and natural food diet (Miele, 1998) and a growing interest in a vegetarian diet (Harper and Henson, 2001; VILT, 22 September 2010). Probably, part of the increased interest in, and critical attitude towards the way our food is produced can be related to the fact that the present generation of shoppers and meat consumers has not experienced the Wars and the related food rationing (Eastwood, 1995), and to some recent food scares (see further). The present food consumer is more conscious about the relation between food consumption on the one hand and health and well-being on the other hand and associates naturally produced food with health benefits. The Eurobarometer survey performed in 2005 (European Commission, 2006b) supports the European citizens’ concerns for animal welfare in figures. The survey indicated that 82.3 percent of Europeans evaluated the overall welfare of farm animals as being somewhere between moderate and very bad, with 78.3 percent strongly believing that more should be done to improve the welfare and protect the living conditions of farm animals in the EU. 6 General introduction Table 1.1. Impacts of improvements in production efficiency on farm animal welfare (source: Winter et al., 1998, p. 310) Change Likely negative impact on animal welfare More housing of animals, including individual Confinement and deprivation housing Higher stocking density (in terms of individuals per Greater disease risk unit land area or per housing area/volume) Larger management units (in terms of number of Greater disease risk; less individual care animals) Lower ratios of stockmen to animals Problems may be missed Less veterinary time per animal Disease may not be treated Fewer, larger and faster throughput abattoirs Longer journeys, poorer care Special markets (e.g. white veal, foie gras) Confinement, anaemia, suffering Further, more urbanised societies are frequently mentioned in literature as an important factor for the rising importance of animal welfare (Harper and Henson, 2001; Hughes, 1995; Kendall et al., 2006). In short, urbanisation has coincided with a lower familiarity with, and decreased practical knowledge level of animal production practices. Consequently perceptions became more emotional and empathy-driven (termed “sentimental anthropomorphism” by Taylor and Signal (2009)). In many studies, pet ownership is also mentioned in this perspective, shifting the attitudes towards animals from a more utilitarian vision to a more humanitarian approach. Further elaboration on this will be provided in the discussion of the conceptual framework. Finally a selection of specific events that have contributed to the significance of farm animal welfare will be highlighted in brief. First, Ruth Harrisons’ Animal Machines (1964), in which the downsides of intensive animal production (also termed factory farming or CAFOs (Concentrated Animal Feeding Operations) are stressed, has been very influential as an impulse for animal welfare related research in many research domains. It has also been the direct cause for the Brambell’s report3 (1965), which has been the prelude of the Five Freedoms (FAWC, 1993) (Table 1.2). Second, the public requirements with respect to the minimum standards for animal welfare have changed because of world-wide food scares and alarming signals about food from animal origin, and because of the increased media-interest in these issues (Quintili and Grifoni, 2004). Food scares and food crises and the way it is presented in the media, have strongly damaged the image of, and trust in animal production and animal production practices among the public. A lowered confidence in production systems further resulted from the increasing gap between consumer and producer, following on the urbanisation (Kanis et al., 2003). 3 In the Brambell’s report, it was formulated that an animal should at least have sufficient freedom of movement to be able without difficulty to turn around, groom itself, get up, lie down and stretch its limbs. 7 Part I Table 1.2. The Five Freedoms as described by the Farm Animal Welfare Council (1993) Freedom from hunger and thirst – by ready access to fresh water and a diet to maintain full health and vigour Freedom from discomfort – by providing an appropriate environment including shelter and a comfortable resting area Freedom from pain, injury or disease – by prevention or rapid diagnosis and treatment Freedom to express normal behaviour – by providing sufficient space, proper facilities and company of the animal’s own kind Freedom from fear and distress – by ensuring conditions and treatment which avoid mental suffering 1.2. Scope of the thesis In this research, I will approach farm animal welfare mainly from the public’s perspective. Broadly speaking, I will focus on three themes, mainly to provide a knowledge base that can be applied to support a constructive dialogue, and to communicate and inform more effectively about farm animal welfare to the public. The first theme covers the public conception of farm animal welfare. The second theme deals with determinants of attitudes, perceptions and concerns towards farm animal welfare. The third research theme investigates factors that impede or support an adequate translation of attitude into behaviour or behavioural intention, captured in the thesis title as the “ambivalence” between people’s sayings and doings. This term was used by Te Velde et al. (2002) in the same context, and related to the discrepancies between perception and behaviour. All three themes will be introduced hereunder for their relevance in the scope of this PhD dissertation. The research is primarily focused on Flanders, the Northern Dutch-speaking part of Belgium, which is a densely populated region, with a high density of intensive animal (mainly pork and poultry) production. Farm animal welfare is approached as a general concept and applies to the region’s economic most important non-aquatic production animals. One specific dwell on the general approach is introduced, in terms of the extension to a specific and current welfare-related case, i.e. piglet castration. The choice for piglet castration has multiple reasons. It became a hot topic during the course of my PhD dissertation, with several campaigns from GAIA (a Belgian animal welfare organisation), and a Belgian bill for a Royal decree that bans castration without anaesthesia, drawn up by the then Minister of Public Health Laurette Onkelinx, following the advice of the National Council of Animal Welfare. In addition, my research will reveal “pain caused by human intervention” as one of the major public concerns in relation to farm animal welfare (see Chapter 3). Also this topic is interesting to research from a public point of view for some more reasons. Flemish citizens are only poorly aware of the practice, which necessitates information provision before attitudes can be properly measured, but allows at the same time studying the impact of different types of information provision. Also public acceptance is assigned a crucial role for a successful market introduction of immunocastration, an alternative practice for physical or surgical piglet castration that will be approached in more detail. 8 General introduction 1.2.1. Public conception In this section, the relevance of clarifying the way the public perceives the concept of farm animal welfare is highlighted, together with an overview of efforts done so far in conceptualising farm animal welfare. “Animal welfare” is increasingly referred to as important by different stakeholders along the food production chain (Verbeke, 2009), yet its meaning and conception differ depending on who is using the concept (Hewson, 2003a). In a review paper from Fisher (2009), a summary of different understandings of the concept of farm animal welfare is provided (Table 1.3). Table 1.3. A summary of different understandings of animal welfare (source: Fisher, 2009) Definition The absence of, or freedom from, pain and suffering (anxiety, fear, pain and distress) and sometimes the presence of positive states or pleasures (see Fraser and Duncan, 1998; Tannenbaum, 1991) The absence of suffering Fit, healthy, well and happy or feeling good (see Nordenfelt, 2006; Tannenbaum, 1991; Webster, 2005a) Fit and happy or feeling good Needs are fulfilled (Bracke et al., 1999) A state of being, in which at least basic needs are met and suffering is minimised Physical and mental well-being, including not being fearful, frustrated or deprived (Dawkins, 2006) Animals are healthy and they have what they want The ability to, or the degree to which the animal is able to, adapt or cope (Broom, 1991; Tannenbaum, 1991) The state of an animal as it attempts to cope with its environment The ability to respond to changes (Korte et al., 2007) A broad predictive physiological and behavioural capacity to anticipate environmental challenges A natural state (Hewson, 2003b; Tannenbaum, 1991) A state of complete mental and physical health, where the animal is in harmony with its environment The quality of an animal’s life (McMillan, 2000; Tannenbaum, 1991) A state that includes some measure of a successful life Humane interactions with animals (Banks, 1982; Kilgour, 1978) The use of methods for handling and management that impose the least amount of stress or distress The stance that humans have a responsibility or duty to care for or protect animals (http://en.wikipedia.org/wiki/Animal_welfare) The viewpoint that is morally acceptable for humans to use non-human animals provided unnecessary suffering is avoided An economic or socio-political issue (McInerney, 1991) Animal welfare is ultimately an economic or socio-political issue, primarily a subjective matter of human perceptions, a subset of human welfare since people’s preferences determine their actions 9 Part I In general, literature distinguishes between three conceptions or perspectives, whether or not used in combination. A first conception expresses animal welfare in terms of the body and the physical environment (“if an animal is healthy and producing well, it is faring well”) (e.g. Broom, 1991; Broom, 2001; Broom, 1986). A second conception relates animal welfare to the mind or to feelings and emotions (“if an animal is feeling well, it is faring well”) (e.g. Dawkins, 2006; Duncan, 1996). The third view focuses on living a natural life (“animals fare best if they can live according to their nature and perform their full range of natural behaviours”) (e.g. Kiley-Worthington, 1989). Most properly, the three conceptions should not be seen independently from each other but they are interrelated (“animal welfare comprises the state of the animal’s body and mind, and the extent to which its nature is satisfied”) (Appleby, 1999a; Duncan and Fraser, 1997; Fraser, 1995) (Figure 1.2). Feeling well Elimination of fear, pain & suffering Functioning well Good health, growth & survival Natural living Ability to move, exercise & perform natural behaviours Figure 1.2. Conceptions of animal welfare (http://www.afac.ab.ca/insights/07summer/babel.pdf) Nonetheless, Hewson (2003b) and Duncan (2005) indicate and illustrate that the different aspects (body, mind and nature) of welfare sometimes are in conflict. Many of the conceptions, especially the ones developed from an animal science perspective, are criticised for being too limited, too expert-oriented and not sufficiently addressing the issues of actual public concern (Fraser, 2008; Lund et al., 2006; Rushen, 2003; Rushen and Depassille, 1992), whereas animal welfare is, by nature, a social concept that reflects societal values (Fisher, 2009; McInerney, 1991). “although science has an important role in providing sound defensible information on how animals respond to a specific practice, ultimately it is an ethical decision by the general community that will determine the acceptable standards for farm animals” (Boogaard et al., 2006, p.2) Fisher (2009) suggests that the concept should be considered as a judgment to be undertaken by combining varied understandings, and to acknowledge underlying values and assumptions, both moral and scientific, rather than risking to exclude valid viewpoints through pursuing a prescriptive definition. Peter Singer’s (1975) consequentialist/utilitarian and Tom Regan’s (1983) deontological approach of animal welfare have been the offset of many moral, ethical and/or philosophical discourses on animal welfare (e.g. Adams and Donovan, 1995; Nussbaum, 2006; Rollin, 1981; 10 General introduction Sandøe and Simonsen, 1992). The multitude of research dedicated to the conceptualisation of animal welfare has very recently resulted in some review articles summarising the efforts (e.g. Carenzi and Verga, 2009; Fisher, 2009) and in several publications that criticise some of the existing conceptions (e.g Degrazia, 1998; Haynes, 2008; Koch, 2009; Rushen, 2003). Fisher (2009) indicates that the essential reason for defining a term or concept is to communicate. An accepted definition could act as a focus for debate and function as a target for improving animal welfare. In addition, Rushen (2003) stressed the relevancy of research into societal views on animal welfare, and their values and norms, to shed light on the factors that steer the behaviour of people. Moreover it is acknowledged that the concerns of the public dictate the need for animal welfare standards and animal welfare legislations (Caporale et al., 2005; Edwards and Schneider, 2005). In this perspective the goal of this PhD dissertation is to investigate the societal conception of farm animal welfare, which is, compared to conceptions from other perspectives and despite its accredited relevancy (Fraser, 2001; Rushen, 2003), not yet extensively covered in scientific literature. The main effort done so far in this perspective relates to the Welfare Quality® project (www.welfarequality.net), which presents a framework built upon consensus between different stakeholders, among them also citizens. This conception involves four principles which are further subdivided into twelve criteria (Table 1.4). Investigating public opinion, in particular with ethical loaded concepts like animal welfare, is interesting, though very challenging. The debate is often impeded by the criticism on the paradox between the sayings of citizens and the doings of consumers (Dagevos and Sterrenberg, 2003; Honkanen and Olsen, 2009; Korzen and Lassen, 2010; McEachern and Schröder, 2002) (see further). On top the public is poorly knowledgeable and aware of animal production (Harper and Henson, 2001), which contributes to a discordance between the conception of animal welfare at the producer and the public level (Lassen et al., 2006; Te Velde et al., 2002). These arguments and discordance justify the need for a societal definition and conception of animal welfare. Table 1.4. Welfare Quality’s® operational definition of farm animal welfare with 12 criteria aggregated into 4 principles (Botreau et al., 2008) Principles Good feeding Good housing Good health Appropriate behaviour Criteria 1. Absence of prolonged hunger 2. Absence of prolonged thirst 3. Comfort around resting 4. Thermal comfort 5. Ease of movement 6. Absence of injuries 7. Absence of disease 8. Absence of pain induced by management procedures 9. Expression of social behaviours 10. Expression of other behaviours 11. Good human-animal relationship 12. Absence of general fear 11 Part I 1.2.2. Ambivalence and attitude variety: need for a segmented and targeted approach In this section I will motivate the importance of acknowledging differences in society with regard to attitudes and reported pro-welfare behaviour4. The market for higher welfare products is rather small. Only non-cage eggs have a considerable market share (86.2 percent in Belgium (VILT, 8 april 2009)). The market share of organic meat, for instance, is much smaller and estimated at 0.4 percent (Belga, 01/04/2010). As a consequence, the market for higher welfare animal products should be considered a niche market that attracts consumers with a rather specific profile. A review paper of Verbeke (2009) on stakeholder, citizen and consumer interests in farm animal welfare indicates the need for appropriate market segmentation when analysing and addressing animal welfare concerns. A segmented and targeted strategy of information provision is most likely to be efficient and effective when it meets the specific needs of a target audience (Wilson, 1981). Salaün and Flores (2001) support this and state that much of today’s information about food is ignored, since much of it does not address its audiences’ needs and expectations. Thus, segmentation is a necessary tool in order to understand how to make higher welfare food relevant to different consumer profiles and how to position these products in a competitive food marketing environment. In literature, very few efforts have been done to segment the market based on prowelfare behaviour. A British report ordered by Freedom Food (IGD, 2007) revealed four shopper profiles based on the purchase frequency of higher welfare products, resulting in four groups that were termed “no welfare” (36 percent), “little welfare” (20 percent), “some welfare” (34 percent), and “high welfare” (10 percent). Most of the segmentation efforts however were based on segmentation variables in a broader food context, among which a segment was identified with higher interests in higher welfare products. Hansman (1999), in a Dutch research report, defined four consumer segments based on general food consumption patterns: the “cooperating consumer”, with a traditional food pattern; the “responsible consumer”, who feels highly responsible for the environment, health and animal welfare, and who has mainly a vegetarian and ecological consumption pattern; the “competitive consumer”, who likes to eat exclusive; and the “rational consumer”, who is considered as a mainstreamer as he/she cannot be differentiated from other consumers. Meuwissen and van der Lans (2004) identified six consumer segments in The Netherlands: Environmentalists, Ecologists, Animal Friends, Health Concerned, Unpronounced and Economists. The second-to-fourth of these groups showed a significantly higher willingness to pay for pork produced with attention to animal welfare. Together with the Environmentalists, these segments ranked animal welfare as one of the three most important product attributes. Grunert and Valli (2001), in a pan-European study, defined four consumer segments based on the way in which consumers subjectively link quality attributes of beef products to consequences and values. A “concerned segment” (17 percent) was identified next to a “moderate”, “health conscious” and “knowledgeable” segment. This concerned segment highly valued a fair treatment of animals amongst other issues. Similarly, Nilsson et al. (2006) defined in a U.S. study “concerned shoppers” 4 I use the term ‘pro-welfare behaviour’ to refer to the purchase of higher welfare products, analogous to ‘sustainable behaviour’ 12 General introduction (43 percent) next to “attribute conscious” and “price conscious” shoppers based on a choice experiment that included conventional and certified pork chops. The importance of segmentation is further amplified by the opposition between individuals in their role of consumer versus citizen. Despite the increased importance of ethical considerations, norms and values in consumers’ food choices (Codron et al., 2006; Cohen et al., 2007; Honkanen and Olsen, 2009; Robinson and Smith, 2003; Storstad and Bjørkhaug, 2003; Van Calker et al., 2005; Verbeke et al., 2010), attitudes – measured in questionnaires and reflecting the citizens’ idealistic opinions – still poorly match with actual consumer preferences and purchasing behaviour. “Only some walk their talk” (Demeritt, 2005) In many cases, consumers choose for cheap, convenience and tasty, are motivated by selfish shortterm interests and an immediate satisfaction of their needs. They are called individualistic, egoistic, materialist, realistic and rational, profit maximising in their food choice, guided by costs and benefits. Citizens have society-oriented long-term interests and put ahead issues like animal welfare, environmental friendliness and justice. They are driven by morality and an idealistic opinion. Yet, many researchers do not agree with the stated ambivalence between citizens and consumers. Market figures and market shares of higher welfare animal products are considered a poor reflection of public concerns, which is related to the specific characteristics of the food buying decision process, the specificity of animal welfare as a product attribute, limitations of survey data in measuring attitudes relating to ethical-loaded issues (hypothetical, self-selection and social desirability bias), and diversity within the citizen and consumer population. Figure 1.3, depicted from results within the Welfare Quality® project, reveals how nuancing a survey question unravels the opposition between citizens and consumers. 13 Part I 100 90 80 70 % 60 50 40 30 20 10 0 HU IT FR Thinking of animal welfare in general GB NL NO SE Thinking of animal welfare when buying Figure 1.3. Citizen attitude versus consumer opinion. Percentage of respondents that agree on statement “thinking of animal welfare in general” versus “thinking of animal welfare when buying” (source: Welfare Quality® Conference Proceedings, 17/18 November 2005, Brussels, Belgium, p.31) 1.2.3. Piglet castration In this section, the issue of boar taint, piglet castration and its alternatives will be introduced5. Boar taint is the unpleasant odour or flavour that consumers may detect if pork meat from male pigs that have reached puberty is cooked and/or heated. The accumulation of the compounds skatole (Vold, 1970; Walstra and Maarse, 1970) and androstenone (Patterson, 1968) in the fat tissue of male pigs is identified as the major cause of the development of boar taint in pigs. Humans are most sensitive to skatole (faecal odour and flavour). Elevated levels of androstenone (urine or sweat odour) are more prevalent in entire male pigs. Despite being harmless for human health (Clarke et al., 2008), the off-flavour strongly impairs the meat quality and is prohibited by food quality regulations in most countries. In the vast majority of European countries, male piglets are surgically castrated (most often without anaesthesia or analgesia) to avoid boar taint, and safeguard the product quality (Figure 1.4). 5 This section is adapted from the introduction from: Vanhonacker, F., Verbeke, W. (2010). Consumer reactions towards the possible use of a vaccine method to control boar taint versus physical piglet castration with anaesthesia: A quantitative pan-European study. Animal, in press. 14 General introduction 100 90 % male pigs surgically castrated 80 70 60 50 40 30 20 10 Ireland UK Portugal Spain Cyprus Greece Poland Lithuania Sweden Slovakia Slovenia Denmark Hungary Switzerland France Finland Belgium Norway Netherlands Austria Estonia Germany Latvia Luxembourg Italy Czech Republic 0 Figure 1.4. Estimates of the percentage of male pigs (conventional production) surgically castrated per country (source: Fredriksen et al. (2009)) Though surgical castration without anaesthesia is very effective in eliminating boar taint, it is very much contested and subject to public concern because of its negative impact on the animal’s welfare and integrity. As such it is very unlikely to be tenable as a future practice within the European Union (de Roest et al., 2009). For instance, Norway (since 2002), Switzerland (since 2010) and the organic farming sector in the Netherlands (since July 2007) have already banned surgical castration without anaesthesia. In addition, market forces and marketing actions have emerged such as large retail companies and well-established assurance schemes (e.g. Qualitätssicherung scheme) refusing meat from pigs not given pain relief before castration. Recently, the largest Belgian supermarket chain Colruyt even decided to sell only pork from non-surgically castrated but vaccinated pigs from 2011 on. The discussion on the practice and the public concern has been the direct reason for a review report by the European Food Safety Authority (EFSA) on the welfare aspects of piglet castration (EFSA, 2004). Following on the report, the PIGCAS consortium (http://w3.rennes.inra.fr/pigcas) aimed to provide information on piglet castration and the possible alternatives to support EU policy development. A special issue of the journal Animal (2009, volume 11) has been devoted to some results of this project. This has been the prelude of more scientific research for alternative methods that should be equally efficient in eliminating boar taint, that enhance animal welfare and that are economically feasible (for a review see von Borell et al. (2008)). 15 Part I In general, three important alternatives are distinguished: immunocastration (also referred to as vaccine method or immunovaccination), surgical castration with anaesthesia and/or analgesia; and no castration (e.g. through sperm sexing, detection systems at the slaughter line, genetic control for boar taint) (Zamaratskaia and Squires, 2009). The latter, i.e. no castration, has also been the specific focus of ALCASDE, an EU-funded initiative (http://www.alcasde.eu/). In this dissertation, the focus will be on immunocastration. This practice pertains to a two-shot vaccine that stimulates the pig’s immune system to produce specific antibodies against gonadotropin-releasing hormone (GnRH) (Adams, 2005; Zamaratskaia et al., 2008). This inhibits the function of the testicles and thus the accumulation of the compounds skatole and androstenone. The moment for the first injection is flexible, but should be at least four weeks before the second injection. The second injection is done four to six weeks before slaughter (Mackinnon and Pearce, 2007) (Figure 1.5). Figure 1.5. Vaccination timing of piglet castration (www.berengeur.com) 16 Conceptual framework 2. Conceptual framework This section will discuss the conceptual framework that is proposed in this PhD dissertation to reveal the determinants of public attitudes, perceptions and concerns towards farm animal welfare, and some factors that influence the translation of attitudes and concerns into market place behaviour. Research on the public conception of farm animal welfare is a general research objective, and is as such extraneous to the framework (Figure 1.6). The framework is depicted from existing literature and will be discussed and complemented with relevant information from additional sources in literature within this section. First I will comment on Fishbein and Ajzen’s Theory of Reasoned Action (1975), and the limitations it faces in being applied in the context of prowelfare behaviour. Second the structure suggested by Kendall and co-authors (2006) to explain variation in attitude towards animal welfare from a stratification theory perspective will be integrated and complemented with insights from other studies. Finally factors that influence (either positively or negatively) the translation of welfare related attitudes in pro-welfare behaviour, discussed in the FP6 project “Consumer concerns about animal welfare and the impact on food choice” (EU-FAIR-CT98-3678) (Harper and Henson, 2001) will be discussed and extended. THEORY OF REASONED ACTION (Fishbein and Ajzen, 1975) (CITIZEN) (CONSUMER) ATTITUDE BEHAVIOURAL INTENTION Farm animal welfare: attitudes, perceptions and concerns Attitude modifiers Social structural variables place-based variables other social structural variables Individual animal-related experiential variables (based on Kendall et al. (2006)) Ambivalence Influencing factors Perceived consumer effectiveness Cost Information Availability Dissociation Product attribute importance Animal welfare versus others Altruistic versus Egoistic (based on Harper and Henson (2001) and Miele (2001)) Figure 1.6. Conceptual framework used to discuss modifiers of public attitudes, perceptions and concerns towards farm animal welfare, and factors that influence the relationship between attitude and behavioural intention 17 Part I 2.1. Theory of reasoned action The theory of reasoned action assumes that given the assumptions people make, their intentions are the result of a reasoned process, thus suggesting that cognition would seem to be implicated as the primary determinant of behavioural intentions. Different sources already stressed that attitude is not a good predictor of behaviour or behavioural intention in relation to animal welfare (Ajzen, 2001; Harper and Henson, 2001; Spaargaren, 2003; Vermeir and Verbeke, 2006). I will discuss in brief some specificities related to the food buying decision process and to animal welfare as a product attribute that hinder the cognitive process. In addition I will give some examples of theories, mentioned in literature, that explain motives used by consumers to explain why their behaviour does not follow from their attitudes. This is mainly used by humans as a motive to avoid feelings of guilt, but does imply that one cannot simply consider a duality between consumers and citizens. First the food buying decision process is a routine process. Everyday consumption practices are typically driven by convenience, habit, value for money, personal health concerns, hedonism and individual responses to social and institutional norms. They are associated with a direct needs satisfaction and are likely to be resistant to change (Casimir and Dutilh, 2003; Vermeir and Verbeke, 2006). This has two implications in relation to the degree of pro-welfare behaviour. First, a routine process is associated with low information processing and time expenditure. Given that animal welfare is a credence attribute and that higher welfare products are limited available (see further), buying higher welfare requires an active information and product search. Second, rational consumers balance costs and benefits in a time perspective, where the present weighs disproportionally much more than the future (Diederen, 2003). Related to the aim of a direct satisfaction of the needs, a price premium that has to be paid at the supermarket’s cash desk, to contribute to a higher level of animal welfare in the indefinable future, does not associate with the idea of a direct needs’ satisfaction. This corresponds with Coleman (2007), who indicates that prowelfare behaviour should not only be related to consumption frequency of higher welfare products, but also to the willingness to engage in community behaviours such as donating to animal welfare groups. Moreover, there is no such thing as “the consumer” (Te Velde et al., 2002; Verbeke, 2009). Segmentation is a prerequisite to approach the pro-welfare consumer and supports the search for determinants of attitudes and of an adequate translation of attitudes in behavioural intention and behaviour. In order to deal with the ambivalence between attitude and behaviour, people use several mechanisms, mainly to reduce the feelings of guilt that originate from using animals for food purposes. Several authors refer to this as “cognitive dissonance” (Heleski et al., 2004; Mayfield et al., 2007; McEachern and Schröder, 2002). The same is defined by Te Velde et al. (2002) as “coping strategies” (~social psychology), as a combination of dissonance reduction and distancing devices. Dissonance reduction is further divided into “adding consonants to behaviour”, “eliminating dissonance”, “amplifying consonants” and “trivialising dissonance” (Festinger, 1964). Also four distancing devices can be identified (Serpell, 1996): “detachment”, “shifting responsibility”, “concealment” and “misrepresentation”. 18 Conceptual framework 2.2. Modifiers of farm animal welfare attitude The next section is used to elucidate the conceptual approach from Kendall et al. (2006) to determinants of animal well-being6. Several reasons were present to motivate the choice for the conceptual approach from this study. The study is performed under the assumption that attitudes towards animal welfare are not a conscious part of citizens’ day-to-day thinking. In addition it is framed based on findings from existing empirical studies and theoretical sources and it indicates that different social groups think differently about animal welfare. These issues match with the reasoning in this PhD dissertation. Kendall and co-authors make use of stratification theories to conceptualise determinants of animal welfare. They further acknowledged that this is not the only possible approach. In response, where possible and relevant, I will incorporate additional visions to explain determinants. In this introductory part hypotheses will be formulated based on theories from literature. Back coupling to results from previous consumer and citizen studies are restricted to the results and discussion chapters only. The stratification theory7 postulates that individuals’ attitudes are shaped by social structural positions, i.e. members of the same group share similar attitudes that reflect their social positions. Variation within social groups exists and is linked to differences in individualised experiences. Social structural positions are subdivided in relation to urban versus rural living environment (place-based variables), and positions including socio-economic status, gender, age, and other socio-demographic variables (other social structural variables). In rural and farm settings greater use is made of animals for food production. This suggests an impact on attitudes and behaviours that reflect that use (Bourdieu, 1984, 1990; Giddens, 1986; Tilly, 1999). Proximity to place-based resources affects everyday cultural practices, and place itself provides a structural context for maintaining those practices. This suggests that rural and urban places provide not only different opportunities for contact and relationships with animals, but also different cultural experiences that shape and reinforce attitudes about animals. For the other social structural variables, the study applies the underdog hypothesis (Elder et al., 1998; Griffith et al., 2002; Nibert, 1994). This hypothesis suggests that those lower in the stratification hierarchy may have greater concern with animal welfare. It is further recognised that individuals have their own life trajectory and experiences. Therefore, distinct individual animal-related experiences are expected to also affect attitudes towards animal welfare. I will separately discuss the different determinant variables, including also insights from other studies. 6 Kendall and co-authors use the term ‘animal well-being’ to capture a broader meaning than animal welfare. It is also not limited to attitudes towards farm animals but to animals as a whole. In order to avoid confusing word usage, I will further speak about (farm) animal welfare. 7 References explaining the stratification theory are obtained from Kendall et al. (2006) 19 Part I 2.2.1. Social structural variables – Place-related variables Living environment in terms of rural or urban residence, together with rural experiences is indicated as one of the most pronounced determinants of attitudes towards farm animal welfare. This is related to differences in familiarity with, awareness of, and knowledge about animal production and animal production practices. These concepts are affiliated to the involvement concept. Involvement refers to personal relevance and importance attached to issues, based on inherent needs, values and interests (Zaichkowsky, 1985). This is also illustrated in Te Velde et al. (2002), where differences between consumers and farmers are discussed based on differences in convictions, norms, values, knowledge and interests. In the same context, higher involvement results in a higher level of active reasoning about a topic (Petty and Cacioppo, 1986). This suggests that differences in familiarity, awareness and knowledge involve a different vision towards farm animals and a different weight of ratio versus emotion or empathy in attitude formation. Plous (1993) states that the dissociation between the way people perceive animals and the way they treat them is maintained through spatial distances. Farmers or rural inhabitants have a more direct reliance on animals for food, family livelihood and protection. In this way they hold a more utilitarian vision. This utilitarian vision appears in many different formulations in literature. Descartes, in his time, defined non-humans as “nothing but complex automata, with no souls, minds or reasons”. Related terminology is found in “objectification of the animals” (Hills, 1993), “instrumentality” (Hills, 1993), “commoditisation” (Te Velde et al., 2002), “commodification” (Kaiser, 2005) and “counter-anthropomorphising” the animals with increasing rural or farm experience (Paul and Podberscek, 2000). In different sources, authors speak about a “domestic or ancient or social contract” between farmers and their animals (e.g. MacArthur Clark, 2008; Oma, 2010; Safran Foer, 2009). All these visions relate to the specific use of the animals for food production. Contrary, with the expansion of urbanisation (and the intensification, involving farm animals disappearing from the landscape view), the sight of food animal slaughter moved out of public view. In addition, pet keeping in the city replaced the more utilitarian relationships people previously shared with animals (Jasper and Nelkin, 1992) and triggered a more humanitarian vision towards animals (anthropocentrism versus anthropomorphism). The relation of these different visions with affect and cognition is also reflected in the distinction Hills (1993) makes between “instrumentality” and “empathy”, and in Serpell’s model (2004) that distinguishes between two motivational dimensions: “Affect” (emotional responses) and “Utility” (perceptions of animals’ instrumental value). Boogaard et al. (2006) define emotional experiences with farming (measured as urbanisation degree and connection to agriculture) and factual knowledge as determinants of the perception of farm animal welfare. In summary, it is expected that rural background is related to less concern about animal well-being. Rural background is considered beyond geography and current residence, and is extended to childhood residence, ties to farmers and rural lifestyle experiences. 20 Conceptual framework 2.2.2. Social structural variables – Other variables Regarding other social structural variables, Kendall et al. (2006) define assumptions based on the underdog hypothesis and indicate females, lower socio-economic positions, younger people and people without dependent children to be higher concerned about animal welfare. The relationship between higher concerns for animal welfare among women from a social perspective is related to the caretaking role of women in a family context that possibly extends to animals, and from a higher likelihood of being engaged in household tasks that put them in contact with animals. This might increase sensitivity to the needs of, and conditions surrounding animals. Other explanations relate the higher concern among females to a higher health consciousness of females (Beardsworth et al., 2002; Kubberod et al., 2002; Verbeke and Vackier, 2005), to a moralistic, humanistic female view versus a utilitarian, dominionistic male view (Wells and Hepper, 1997), and to a higher level of empathy among females (Heleski et al., 2004; Paul and Podberscek, 2000). The higher concern among younger people is related to the life-cycle stage, and is associated with other social structural variables such as marriage, childbearing and employment. Wells and Hepper (1997) associate younger people with more naturalistic and humanistic versus utilitarian. The underdog hypothesis is further applied to associate a higher animal welfare concern for individuals in lower income categories (Nibert, 1994; Uyeki and Holland, 2000) and those with less education (Nibert, 1994; Peek et al., 1997), two variables that reflect socio-economic status. Finally, family status in terms of marital status or presence of children is reported to determine animal welfare related attitudes. People with children are thought to be foremost concerned for children’s issues (Schlozman et al., 1995). This suggests that people without children may direct more attention to other issues, such as concern for animals. The lower animal welfare concerns among married people are associated to the tendency of higher incomes and greater wealth, and to the higher adherence to conventional norms and values. 2.2.3. Individual animal-related experiential variables Investigated factors that could explain differences within social structural groups in Kendall and coauthors’ study pertain to hunting, having a pet in the household, being vegetarian, cooking and food shopping, and concerns about the environment and food. Non-hunting, having a pet in the household, vegetarianism, less involvement in daily food shopping and cooking, food conscious attitudes and behaviours, and concerns about the environment were related to greater concerns with animal welfare. For this PhD dissertation, interest goes to hypothesised positive relationship between concerns for farm animal welfare on the one hand, and having a pet in the household, being vegetarian, and interest in information related to animal welfare as an ethical issue and as a product attribute on the other hand. 21 Part I 2.3. Factors influencing the ambivalence between attitude and behavioural intention In the next section I will elaborate on some specificities of the food purchasing process, and on some barriers (based on Harper and Henson, 2001) that prevent consumers from an adequate translation of their reported ethical preferences in actual food choice. These insights should contribute to a better understanding of the so-called citizen-consumer ambivalence. “De treurige constatering luidt dat de consument dubbelhartig is. Terwijl zijn rechterhand een enquête invult over scharrelvlees vult zijn linkerhand de kar bij de kiloknaller, de Aldi en de Lidl. Niet dierenwelzijn, het behoud van het landschap, betrokkenheid bij onze eigen boeren of duurzaamheid bepalen ons aankoopbeleid, maar de prijs.” (Marc van Dinther, “Liever goedkoop dan goed”. De Volkskrant, 26 april 2003: p.17) The classical decision making process distinguishes between five steps: problem recognition, information search, evaluation of the alternatives, choice and the consequences of the choice (Engel et al., 1990). These stages, however, are not fully completed for every single purchase decision. Dependent on the consumer knowledge of the product, the nature of the product and the importance of the purchase, three types of purchase behaviour are distinguished: extended, limited and routine purchase behaviour. Food is highly substitutable, low-priced, purchased repeatedly, and is primarily satisfying basic physiological needs (Verbeke, 2005). As such food purchase decision making typically belongs to the third category (Binnekamp et al., 2005). Correspondingly, foodrelated decisions are most often based on heuristics or follow peripheral routes of information processing (Binnekamp and Ingenbleek, 2006; Frewer et al., 1997b; Verbeke, 2005). Animal welfare is a credence attribute, which necessitates external information search, and imposes some limitations that can help to explain the gap between citizens’ survey responses expressing high interest in animal welfare issues and consumers actual food purchasing behaviour. This will be discussed in more detail in the following paragraphs and pertains mainly to the challenge of transferring the right information, in the right format, to the right audience. Present failure on these issues is referred to by Verbeke (2005) as “information asymmetry” and by Prabha et al. (2007) as “information inadequacy”. Information providers should tackle the reported information unavailability in relation to animal welfare against a context of information overload, bounded rationality of the consumer, the risk of rational ignorance, consumer confidence in information and the fact that consumer are neither a uniform group, nor experts in the matter of animal welfare. In addition, despite its increasing relevance as a product attribute (reflected in the inclusion of ethical motives in the food choice questionnaire (Lindeman and Väänänen, 2000)), animal welfare is counterbalanced by other, more important motives, as there are quality, safety, health, convenience, price and taste (Bernués et al., 2003b; McCarthy et al., 2003; Nocella et al., 2010; Verbeke and Viaene, 2000). These attributes typically relate to the product itself and its performances in the meal. Related, McEachern and Schröder (2002) indicated that tangible quality 22 Conceptual framework attributes are of high importance in any food purchasing situation, while intangible attributes are only taken into account by the more involved consumer. The annual Consumer Behaviour Monitor study executed in Belgium by the OIVO (Onderzoeks- en Informatiecentrum van de Verbruikersorganisaties) confirms this, and reports in the synthesis of the 2009 report that price, quality and freshness are the most important choice criteria for a food product. Taste, safety, respect for the environment and for the animal’s welfare appear in 2009 as four important elements for the consumer. It is further important to distinguish between animal-centred (or altruistic) and human-centred (or anthropocentric) concerns in relation to higher welfare products. Animal-centred concerns involve that the interest for higher welfare products is mainly motivated by the interest and concerns for the animal itself. To the contrast human-centred concerns involve that the consumer interest for higher welfare products is in the main related to a perceived higher product quality and unique sensory characteristics (Harper and Henson, 2001; Ingenbleek et al., 2006a; Miele, 2001), thus beneficial for the consumer. Following on McEachern and Schröder (2002), who related ethical motivations to altruistic concerns for other living things, it is hypothesised that animal-centred concerns support a more adequate translation of attitudes in actual pro-welfare behaviour, as will do human-centred concerns. In the following, and partly related to the previous section, I will discuss the five barriers indicated by Harper and Henson (2001) that prevent an adequate translation of attitudes into behaviour: perceived consumer effectiveness; information; availability; cost; and dissociation. 2.3.1. Perceived consumer effectiveness (PCE) Perceived consumer effectiveness is defined as “the evaluation of the self in the context of the issue” (Berger and Corbin, 1992; Kim and Choi, 2005). More concrete, it refers to the extent to which individuals believe that their actions make a difference in solving a problem (Ellen et al., 1991). This construct has already been used in literature to explain more general concepts such as green consumerism (e.g. Laskova, 2007; Shrum, 1995), sustainable food consumption (e.g. Vermeir and Verbeke, 2006), societal responsible consumerism (e.g. Meulenberg, 2003) and proenvironmental behaviour (e.g. Nelson, 2004; Van de Velde et al., 2010). This PhD dissertation will verify whether the PCE construct is also applicable in the context of pro-welfare behaviour, i.e. does PCE positively influence the relation between animal welfare concerns and pro-welfare behaviour. The PCE construct is to some extent affiliated with the question about who is responsible for animal welfare. Interestingly stakeholders indifferently delegate the responsibility to other stakeholders (Serpell, 1999; Sundrum, 2007), imposing strong limitations on an effective improvement of the animal’s welfare. Animal protectionists, similar to retailers, delegate the responsibility to politicians and commercial farmers. Consumers refer to politicians and retailers. Veterinarians delegate the responsibility to farmers; and farmers delegate the responsibility to politicians, veterinarians and consumers. The consumer reluctance to accept responsibility for animal welfare is also concluded 23 Part I by Miele (1998), Dagevos and Sterrenberg (2003, p11) and Skarstad et al. (Skarstad et al., 2007). This type of consumer behaviour, where consumers only want to act if the others also act, is referred to as the “prisoner’s dilemma” (or in Dutch: “na-u-effect”) (Diederen, 2003). Consumers want animal welfare to be arranged and expect available products to meet certain minimum standards. This is reflected in the low market share of higher welfare products, and is critically phrased by Diederen (2003) as: “the market gets the best out of products and the worst out of people”. This closely matches with the discussion whether to consider animal welfare as a private or a public good, and is in depth discussed by McInerney (2004). “Considering [the ways pigs are raised] makes me say that I am not going to buy an awful lot of that pork! But then again, it is hard to be a consumer, seeing a good offer, isn’t it? They look the same – and if you have a large family it might not be enough. (…) They shouldn’t produce all the ugly stuff. If they didn’t produce it, we weren’t forced to buy it!” (Quote from focus group participant in a Danish study, available on: http://www.ncrc.fi/files/5211/PP-lassen2.pdf) 2.3.2. Information Verbeke (2005) discusses, in a review paper on information in the food sector, some reasons that hamper a rational decision making process, thus imposing limitations on Fishbein and Ajzen’s Theory of Reasoned Action (1975), and on the neo-classical micro-economics that assume individuals to be always aware of the full range of options, and to be capable of weighing up all the pros and cons. First reason mentioned pertains to the information availability. The public is poorly informed about animal welfare conditions during rearing, transport and slaughter. In order for information provision to be more effective, it will be important that trusted information is passed on to the persons who are actually interested in the information, in the right format and through the right channel. Consumers are not all alike and not all consumers are interested in animal welfare related information. In this perspective, Verbeke (2005) suggests to segment consumers and to provide targeted information, instead of overloading the whole population with information, with the risk to induce consumer indifference or loss of confidence (see further). Information should also address specific information needs and expectations. In this perspective, it is important to construct a definition for farm animal welfare that incorporates the public view, and to gain insights in the values that shape people’s beliefs about and attitudes towards farm animal welfare, acknowledging the existing gap between consumers and producers. The latter is exemplified by Fraser (2007), who indicates that it is more likely that the public will be favourable towards a non-intensive, outdoor system that has high levels of parasitism and neo-natal mortality, compared to an intensive system that achieves a high level of health and hygiene but prevents virtually all natural behaviour. A similar example is given by Abeni and Bertoni (2009), applied on dairy production. The authors stressed the positive public image about cows that graze in the pasture, while this actually involves 24 Conceptual framework some problems with the five freedoms. A third example of the need for a good consumer understanding is taken from Binnekamp and Ingenbleek (2006), who mentioned a low success of the introduction of 56-days chickens. Consumers were not aware that in conventional systems, chickens only lived for 42 days, and perceived 56 days as being too short. Information providers should take into account the limited practical knowledge base and the limited consumer awareness in relation to production practices. Also, consumers and citizens mainly obtain information on animal welfare through mass media and advertising. Characteristic for mass media is the preference to cover negative news items over positive news (“bad news hypothesis”). Although criticised, this is mainly driven by the demand of the public and links with prospect theory (Kahneman and Tversky, 1979) and the endowment effect (Kahneman et al., 1991, 1990). People remember the massive slaughter of animals due to a food crisis or scandal (Niesten et al., 2003), pictures that create an biased image of animal production among the lowly aware public. This negative perception is also reflected in the more negative image people have of farmed (“industrial”) fish versus wild (“natural”) fish (Verbeke et al., 2007b). Advertising on the other hand also contributes to the information asymmetry by misrepresenting the sector making use of romantic and idyllic pictures (Te Velde et al., 2002; van Bruchem, 2003). This induces consumer uncertainty and a loss of consumer trust and confidence (Harper and Henson, 2001; Passantino et al., 2008). In this perspective, Worsly and Lea (2003) revealed the honesty of food labels in general as one of the major food concerns, and the importance of transparency is commonly acknowledged. Second reason mentioned pertains to the consumer ability and willingness to process information. The present society is an information society. The consumers are overwhelmed with information and can not see the wood for the trees anymore. Also in relation to farm animal welfare, an explosion of often vague, confusing and misleading schemes for welfare standards occurs (Dawkins, 2008; Lewis et al., 2008; Ransom, 2007; Thompson et al., 2007; Verbeke, 2009). Verbeke (2009, p331) terms food labels as “information cue high density areas”, while van Bruchem (2003, p117) indicates that supermarkets get the more and more the character of a library, and labels that of books. Consumers are mainly interested in search attributes like price, product weight and expiry date. They are only able to process a certain amount of information (“bounded rationality”). In this perspective, the information overload often induces an inverse effect. Following transaction cost theory, when the opportunity cost of processing information is too high compared with the marginal benefits from information, consumers can just ignore the information (“rationally ignorant consumer hypothesis”), and base their decision on heuristics (Mayfield et al., 2007). This rationally ignorance also applies for animal food products in another context, namely that of averting feelings of guilt (“what you don’t know, you can can’t be responsible for”) (LNV Consumentenplatform, 2005; Sundrum, 2007). 25 Part I 2.3.3. Availability The barrier “availability” relates both to information and product availability. Information availability mainly relates to information asymmetry. Product availability as a barrier relates to the inconsistency between the interest in, and concern for higher welfare products and their (perceived) availability, for example due to a low visibility at the point of purchase, and an inadequate promotion (Vermeir and Verbeke, 2006). This barrier is further amplified by the low consumer awareness about higher welfare products (European Commission, 2005; Vermeir and Verbeke, 2006), the limited knowledge on, and confusion about labels indicating higher welfare production standards. 2.3.4. Cost The importance of cost as a barrier for higher welfare foods is minimised by the public that claims a positive willingness to pay for higher welfare products in survey studies. However, price still is a strong determinant of consumers’ food choice. Interestingly, resulting from the annual Consumer Behaviour Monitor study of OIVO, only few respondents were able to estimate the price of the food products they buy. The notion of a high price for a food product was concluded to depend on the comparison with other products from the same category available at the point of purchase, and is very sensitive to promotions. As such, consumers should be considered more calculating, rather than price conscious individuals. In this perspective, an important role is linked to the World Trade Organisation (WTO), to governments, to supermarket policies and to the way of communication. The WTO considers animal welfare a non-trade concern, meaning that countries cannot refuse foreign meat, produced under lower standards of animal welfare (read: cheaper meat). Governments could for instance subsidise farmers that produce against higher welfare standards. Supermarkets from their side have the aim to appeal to as many consumers as possible. Not all consumers are interested in higher welfare products, and as a consequence, the cheaper products are also offered, and negatively influence the market share of higher welfare products (De Tavernier et al., 2010). In addition, price-wars between supermarkets and the success of hard discounters negatively contribute to the adoption of higher welfare animal food products. Finally, communication can also make use of the consumers’ relative interpretation of price. Currently, communication about higher welfare products mainly relates to the higher (quality) standards of these products. However, given that individuals are more interested in avoiding a “bad” versus seeking out a “good” (prospect theory), it could be expected that communicating about the low standards of mainstream product would be a more effective marketing strategy as compared to highlighting the benefits from higher welfare products (within the limits imposed by legislation on comparative advertising). This nuance is also exemplified by Diederen (2003, p27-28), with regard to the importance of the anchor point used. If the conventional meat is used as anchor point, then the consumer has to pay a price premium to make the animal’s life better. Opposite, when the higher welfare product is the anchor point, consumers can save through choosing the cheaper meat, but this is at the expense of the animal. 26 Conceptual framework 2.3.5. Dissociation Dissociation can be seen in two ways. First in relation to the population’s urbanisation, and second in terms of detaching the food product from the animal. The present public is only lowly aware of the animal production practices on farm. In addition, there is currently a high interest for television programmes in relation to animals, though mainly in the role of companion, zoo or wild animals rather than production animals. A more humanitarian reflection of animals has its impact on the consumers’ perception towards production animals (Bayvel, 2005). The second interpretation pertains to the detachment of the product from the animal (e.g. Hoogland et al., 2005; Skarstad et al., 2007). The consumer does not want to be reminded to the animal when he or she buys meat (termed “de-animalisation” (Bock et al., 2007; Buller and Cesar, 2007; Kennedy et al., 2004), “concealment” (Te Velde et al., 2002, p214) or “pervasive ignorance of modern production techniques” (Harper and Henson, 2001)). This is related to reducing feelings of guilt (Te Velde et al., 2002), to the reluctance of bearing responsibility (Skarstad et al., 2007; Sundrum, 2007) and to a certain degree of disgust (Hamilton, 2006; Kubberød et al., 2002). The latter is described by Hopkins and Dacey (2008) as “yuck factor”, and is further exemplified in a review essay from Birke (1993), where beef was considered “dead cattle” for vegetarians, while this association was not made by meat-eaters (“most people consume meat only as long as they can distance themselves from the nasty facts of its production”). This tendency is also noticeable in the increased share of processed meat products (Eastwood, 1995), in the nomenclature and in commercialisation strategies (Buller and Cesar, 2007), in which the live animal is no longer recognisable (Phan-Huy and Fawaz, 2003). Specifically for Belgium, GfK consumption data illustrate that meat consumption becomes the more and more dominated by meat mixtures, like minced meat, sausages, skewers, pita and gyros, burgers, and cordon blues, accounting for 33.4 percent of the total fresh meat purchases in 2009 (VILT, 18/03/2010). In the UK meat industry, it is even a rule of thumb that pictures of the live animal do not appear on the packaging of processed meat products (Hughes, 1995). In this perspective it is also believed that eggs and milk are subject to a better perception as compared to meat (Miele, 2001). All together, detaching the product from the animals hampers a conscious and rational food buying behaviour (van Bruchem, 2003). “Modern American society loves to watch television cooking shows – the creativity, the sensuousness, the clever technique. But chances are, if a lamb were dragged in and killed at the beginning of the program, most of the viewers would find themselves less interested in the lamb chop recipes. They would be horrified or disgust to enjoy the rest of the program. And yet, if the lamb’s flesh is brought in already killed and sliced, almost all sense of horror and sympathy is muted enough to be nearly unfelt” (Hopkins and Dacey, 2008, p. 579-580) 27 Part I 3. Research objectives and hypotheses The overall objective of this PhD-dissertation is to contribute to a better understanding of the public conception of, and perception towards farm animal welfare. This is relevant for better aligning the different stakeholders’ visions and to support a constructive dialogue along the animal production chain. As such more effective policies that improve the welfare level of farm animals could be developed and achieved. Also the findings should allow gaining insights in to how to educate, communicate with, and inform the public in order to support demand-led welfare improvements. Study results are mainly based on Flanders and for farm animal welfare as a general concept. Where possible the research theme is narrowed to a specific and topical animal welfare case, namely piglet castration. More specifically, this research will try to meet this purpose by focusing on four research objectives: (1) on the conception of farm animal welfare from a public’s point of view (Part II); (2) on revealing determinants of the societal perception of farm animal welfare (Part III); (3) on getting a better understanding of some influencing factors that contribute to the poor reflection of the public’s interest in, and concern for animal welfare in actual food purchase behaviour (Part IV); (4) on highlighting the issue of piglet castration and the role of the public in the search for alternatives. Additionally, specific research objectives are included in the subsequent thesis chapters. Furthermore, ten hypotheses are advanced in this research. Verification of these research hypotheses will yield valuable insights for improved understandings of public perception of farm animal welfare, and for more efficiently addressing the chain end users’ expectations in future policies and information provisions. H1: Place-based variables induce differences in attitudes, perceptions and concerns with respect to farm animal welfare, and are related to societal variation in familiarity, awareness and practical knowledge levels. Rural residence (past and present) and familiarity with animal production will associate with a more utilitarian interpretation of the concept, and lower level of concerns about farm animal welfare. H2: Attitudes, perceptions and concerns in relation to farm animal welfare are influenced by socio-demographic factors. Concerns will be higher (a) among females and (b) families without children, and will be inversely related with (c) age, and (d) socio-economic status. H3: Attitudes, perceptions and concerns in relation to farm animal welfare are affected by individual animal-related experiential variables. Higher concerns exist among (a) vegetarians, (b) individuals with a pet and (c) individuals with a higher interest in animalwelfare related information. 28 Research objectives and hypotheses H4: Perceived consumer effectiveness influences the relationship between attitudes and concerns for animal welfare and pro-welfare behaviour, i.e. individuals with high perceived consumer effectiveness will be more likely to engage in pro-welfare behaviour. H5: Consumers/citizens are not satisfied with the availability of welfare related information and the way it is provided. H6: The price premium for higher welfare products is a barrier to engage in pro-welfare behaviour. H7: Higher welfare products are perceived to be too less available. H8: Consumers de-animalise animal food products. In this perspective, given that eggs are derived animal products, egg consumption will be less affected by animal welfare concerns as compared to chicken meat consumption. H9: A high reported attribute importance of animal welfare, relative to other product attributes, associates positively with pro-welfare behaviour. H10: Consumer segments that differ in level of engagement in pro-welfare behaviour can be identified on the Flemish market. This last hypothesis is a general hypothesis which does not directly relate to a specific research objective. It is incorporated in this dissertation for the importance of targeted marketing of higher welfare products, especially in an open and global food market. 29 Part I 4. Research design and data sources Information required to meet the research objectives and to test the hypotheses is gathered through a combination of qualitative and quantitative research procedures, applied on both primary and secondary data sources. The data for this dissertation stems from six studies that are executed independently from each other, including different sets of respondents, and on different time occasions. The studies, the survey questions relevant in the scope of this dissertation and the applied data analysis procedures are considered in more detail in the methodology sections of the forthcoming thesis chapters. I will briefly present the different studies in this section in chronological order, with reference to the time period, the research procedure (sample selection and data collection), the number of respondents, the sample characteristics and the context in which the study was executed. 4.1. Study 1: Secondary data sources For my dissertation, I was able to use secondary data that were collected through the distribution of postal questionnaires in three consecutive years (2000-2001-2002). Identical questionnaires were developed to obtain insights in the self-reported impact of farm animal welfare issues on meat consumption decisions. A convenience sampling procedure was used among individuals that are responsible for food purchasing within their household, with age as single quota control variable. A valid net response of 521 respondents was obtained with respectively n=179, n=185 and n=175 respondents in the consecutive years. The sample distribution in terms of gender, age, living environment, family status and education level is presented in Table 1.5. 30 Research design and data sources Table 1.5. Sample characteristics of the survey respondents within Study 1 Year 2000 2001 2002 Total Valid response 179 185 157 521 Male 40.2 40.1 33.1 38 48.9 Female 59.8 59.9 66.9 62 51.1 Children in Yes 48.6 63.9 50.4 54.5 40.4 household (%) No 51.4 36.1 48.8 45.5 59.6 Age of children in ≤ 12 years 16.2 20 38.1 22.3 household (%) 12 to 18 years 11.7 16.2 16.5 14.5 > 18 years 72.1 63.8 45.4 63.1 Living environment Urban 36.3 51.9 34.4 41.2 36.3 (%) Rural 63.7 48.1 65.6 58.8 63.7 Education level (%) ≤ 18 years 50.3 36.1 47.1 44.3 > 18 years 49.7 63.9 52.9 55.7 <25 years 29.6 21.6 38.2 29.4 29.3 25-45 years 34.1 34.6 22.9 30.9 28.4 >45 years 36.3 43.8 38.9 39.7 42.4 Mean (years) 37.29 40.63 37.43 38.51 40.2 Gender (%) Age categories (%) Age * Flanders* Source: FOD Economie, Algemene Directie Statistiek en Economische Informatie, Dienst Demografie (http://statbel.fgov.be/nl/statistieken/cijfers/index.jsp) 4.2. Study 2: “Defining the concept farm animal welfare” The first primary data for my dissertation was gathered within a research that was financed by the Ministry of the Flemish Community – Department EWBL – Administratie Land- en Tuinbouw – Afdeling Monitoring en Studie, for a period of six month (1st January – 30 June 2006) (ALT/AMS/2005/1). The general aim of the project was to develop an interpretation for the concept farm animal welfare from a public point of view. The study combined qualitative and quantitative research techniques. The first part of the study consisted of four focus group discussions that were organised in March 2006. The participants (29 in total) were selected based on different meat consumption profiles and on gender. Findings from this qualitative stage were applied to construct a questionnaire for the quantitative part of the study. Additional to the study’s aim, the questionnaire covered the relative importance of animal welfare in the food purchasing process and the interplay between information and animal welfare. 31 Part I Postal and web-based questionnaires were distributed in April 2006, applying a combination of convenience, snowball and stratified sampling. The sampling procedure yielded a gross response of 1,081 respondents and after eliminating incomplete or double questionnaires, a net response of 834 respondents, of which 609 (73 percent) and 225 (27 percent) from web-based and postal questionnaires respectively. Of this sample, 204 respondents were farmers or farmers’ offspring and will be used as a separate dataset (“farmer sample”). A quota sampling procedure was used to select a sample of 459 respondents (“citizen sample”) that is representative of the Flemish adult population with regard to gender, age, living environment (urban versus rural) and region (province distribution) as control variables. Table 1.6 provides further socio-demographic details on both samples. Table 1.6. Sample characteristics of the survey respondents within Study 2 Citizen Sample (n=459) Gender (%) Population (n=204) Men 48.5 49.3 64.5 Women 51.5 50.7 35.5 37.8 (14.1) 40.2 35.9 (13.4) Age (years) Mean (S.D.) Living environment (%) Urban 38.9 36.3 8.9 Rural 61.1 63.7 91.1 Flemish Brabant 16.8 17.0 11.9 Antwerp 21.7 27.7 20.6 Limburg 8.9 13.3 13.8 West-Flanders 25.2 19.0 15.6 East Flanders 27.5 23.0 38.1 ≤ 18 years 32.3 38.6 > 18 years 67.7 61.4 6.1 100 Province (%) Educational level (%) Connection to agriculture (%) * Farmer * Source: FOD Economie, Algemene Directie Statistiek en Economische Informatie, Dienst Demografie (http://statbel.fgov.be/nl/statistieken/cijfers/index.jsp) In the final stage of the project, stakeholder representatives and interest groups were contacted and interviewed about their expert opinion on the resulting conception. In total, seven in depth interviews were carried out. Three persons were recruited to represent the producer side, one person represented the retail side, one represented animal welfare organisations and finally two persons with a scientific background in animal welfare or veterinary sciences were interviewed. 32 Research design and data sources 4.3. Study 3: “Farm animal welfare and consumer behaviour” This data was gathered in the scope of a students’ project (Colman et al., 2007). The project had the aim to compare the public’s evaluation of the welfare of the different farm animals and to determine the impact animal welfare has on the purchase of different animal product categories (dairy, beef, pork, eggs and chicken meat). Cross-sectional survey data were collected through selfadministered web-based questionnaires in Flanders during a five-week period in March and April 2007. Consumers were selected using a non-probability snowball sampling procedure, with the aim to obtain a sample that matches with the age distribution of the adult population and that includes a high degree of variation in other demographic variables. To account for sample deviations from the age distribution of the population, weighing coefficients were applied during analyses. In total, the sample comprised 469 valid consumer questionnaires. Sample characteristics are presented in Table 1.7. Table 1.7. Sample characteristics of the survey respondents within Study 3 Gender (%) Population* Male 39.6 49.3 Female 60.4 50.7 40.7 (13.5) 40.2 Age Mean (SD) Province (%) Flemish Brabant 8.3 17.0 Antwerp 24.5 27.7 Limburg 2.5 13.3 West-Flanders 37.0 19.0 East-Flanders 27.7 23.0 Urban 39.6 36.3 Rural 60.4 63.7 Past living environment Urban 41.4 (%) Rural 58.6 Education (%) ≤18yrs 29.7 >18yrs 70.3 Yes 14.6 No 85.4 Living environment (%) Vegetarian (%) * Sample (n=469) Source: FOD Economie, Algemene Directie Statistiek en Economische Informatie, Dienst Demografie (http://statbel.fgov.be/nl/statistieken/cijfers/index.jsp) 33 Part I 4.4. Study 4: “Surveying Flemish pig producers about their attitude towards surgical castration without anaesthesia and its alternatives” In the context of a master thesis (Verhille, 2008), 300 Flemish pig farmers were randomly selected from the national SANITEL record, which includes the addresses and telephone numbers of the Belgian pig producers. Four addresses were double, two were Dutch addresses and two were incorrect, leaving 292 pig farmers that were contacted through a postal questionnaire from October 2007 until February 2008. If questionnaires had not been returned (a pre-stamped envelope was provided) within two weeks, the farmers were contacted by telephone. In total, 160 completed questionnaires were received and judged suitable for further analyses. 4.5. Study 5: “Consumer perception towards immunocastration as alternative practice for surgical castration of piglets” Also within the context of a master thesis (Furniere, 2008), survey data were collected through selfadministered web-based questionnaires in Flanders during January and February 2008. Participants were recruited through non-probability snowball-sampling. This involves that initial contact persons were asked to complete the questionnaire, together with the request to send the web-link to their acquaintances, and so on. This sampling method is an efficient way of gathering a substantial amount of data in a short time’s notice with a limited budget, ideally for a preliminary qualitative type of data collection. It should be noted that this sampling method does not yield a statistically representative sample. Hence, findings mainly apply within the characteristics of the sample and generalisation to the overall population is not appropriate. Nonetheless, the contact with respondents was steered in such a way to obtain a wide diversity in gender, age, family size and living environment (Table 1.8). Table 1.8. Sample characteristics of the survey respondents within Study 5 (%, n=225) Gender (%) Male 55.0 Age < 26 years 38.6 Female 45.0 categories 26 – 40 years 24.5 (%) 41 – 54 years 23.6 > 55 years 13.2 Rural 43.4 Urban 36.7 Neutral 19.9 Family size (%) 1 * 5.4 2 28.1 3 14.9 4 28.1 5+ 23.5 Living environment * Living environment was self-assessed by the respondents on a seven-point scale, ranging from “rural” (1) to “urban” (7). Percentages in the table represent response categories 1, 2 and 3 for rural; 5, 6 and 7 for urban; and response category 4 for neutral. 34 Research design and data sources 4.6. Study 6: Immunocastration versus surgical castration with anaesthesia/analgesia A large-scale consumer survey was performed in France, Germany and the Netherlands between January 21 and February 4, 2008, and further in Belgium between November 19 and 24, 2009. The study aimed to gain insights in the market acceptance of immunocastration and surgical castration with anaesthesia or analgesia, as an alternative practice for the routine practice of surgical piglet castration without anaesthesia. A total of 4,031 online interviews were conducted among pork consumers (individuals who consume at least 2 or 3 times pork a month). Participants were randomly recruited from online consumer panels in each country. In order to maximise the representativeness of the data for the overall population, data were weighted per country to reflect the demographics of pork consumers in terms of age, gender and education. The survey took approximately 15 minutes in duration. The fieldwork was performed by Lieberman Research Worldwide (LRW Inc.), an international market research company, and was financed by Pfizer Inc.. Sample characteristics are provided in Table 1.9. Table 1.9. Sample characteristics of the survey respondents within Study 6 Total sample Belgium France Germany The Netherlands (25.6%) (24.6%) (25.0%) (24.9%) Gender (%) Male 49.6 52.7 45.7 49.6 50.4 Female 50.4 47.3 54.3 50.4 49.6 18-24 years 9.3 9.2 11.2 9.8 7.1 25-34 years 16.6 14.5 16.5 13.7 21.7 35-44 years 21.2 19.7 18.7 22.5 23.7 45-54 years 20.9 22.7 21.7 19.2 19.8 55-64 years 15.1 21.2 14.3 12.7 12.0 > 65 years 17.0 12.7 17.6 22.1 15.7 2.51 (1.17) 2.54 (1.17) 2.60 (1.19) 2.47 (1.11) 2.45 (1.18) Age category (%) Family size Mean (SD) 35 Part I 5. Thesis outline The dissertation consists of a compilation of papers that have been published, accepted or submitted as contributions to international peer-reviewed journals or as conference proceedings, cross-covering the scientific disciplines of agricultural and food marketing, animal, veterinary and ethical sciences, sociology and consumer behaviour. The thesis includes five parts in total. Part II to Part IV consists of ten chapters, corresponding with single research papers. Figure 1.7 presents the positioning of the different parts relative to the conceptual framework. Each part covers relevant literature and focuses on analysing specific parts of the framework, following the rationale presented below. PART I. General introduction PART II. Public conception of farm animal welfare THEORY OF REASONED ACTION (Fishbein and Ajzen, 1975) (CITIZEN) (CONSUMER) ATTITUDE BEHAVIOURAL INTENTION PART III. Farm animal welfare: attitudes, perceptions and concerns Attitude modifiers Social structural variables place-based variables other social structural variables Individual animal-related experiential variables PART IV. Ambivalence: Influencing factors Perceived consumer effectiveness Cost Information Availability Dissociation Product attribute importance Animal welfare versus others Altruistic versus Egoistic Part V. Discussion and conclusion Figure 1.7. Outline of the thesis in relation to the conceptual framework Important to notice, the different chapters have a scope that stretches beyond the research objectives and research hypotheses formulated within this dissertation. The papers have been left largely untouched in the result section, though in order to frame each chapter within the dissertation and its research objectives and hypotheses, a short abstract will be provided at the beginning of each chapter. 36 Thesis outline Part II is an umbrella chapter in which a public conception of farm animal welfare is developed. The development of the conception has been effectuated in two steps, corresponding to the two chapters within this chapter. The first chapter combines the findings from secondary data sources and primary data from qualitative focus group discussions (Chapter 1), and has been used as preliminary work for the second chapter. There, a quantitative public survey has been used, followed by in depth interviews with stakeholders (Chapter 2) to construct a model that represents an integrated conception of farm animal welfare in a format that is comprehensible and understandable for the public. Both chapters relate the importance of insights in the public conception to the current information asymmetry that is a contributory cause of the low relative market share of higher welfare animal products. Additionally this part has some points of contacts with other parts. The findings from the first chapter also detail on different modifiers of public attitudes and concerns in relation to farm animal welfare (Part III) and on the influencing factors of the attitude-behaviour ambivalence (Part IV). Part III compiles seven research papers in which societal differences in attitudes, perceptions and concerns with regard to farm animal welfare are discussed in correspondence with the framework of Kendall and co-authors (2006). The first three chapters detail on the discordance between both chain’s end users, i.e. producers and citizens/consumers. This fits in with the framework of Kendall and co-authors under place-based factors. The discordance is discussed applying the conception for farm animal welfare developed within this dissertation (Chapter 3) and applying the Welfare Quality® conception (Chapter 4). The latter also details on the impact of being a vegetarian, which is considered an individual animal-related experiential variable in the conceptual framework. The next three chapters are related to the specific case of piglet castration. Two chapters present the public opinion, in which immunocastration is evaluated against the routine practice of surgical castration without anaesthesia (Chapter 5) and against surgical castration with anaesthesia or analgesia as alternative short term solution (Chapter 6). A third chapter details on the evaluation of Flemish pig producers of different alternative methods to the routine practice (Chapter 7). Comparing the findings from Chapters 5 and 6 with the findings from Chapter 7 further contributes to the discussion on the producer-consumer discordance. Next, Chapter 8 discusses societal differences in relation to concern for stocking density as a topical animal welfare theme. The final chapter of this part attempts to segment and profile citizens based on public attitude toward farm animal welfare and consumers’ relative importance attached to farm animal welfare as a product attribute during food purchasing decisions (Chapter 9). Profiling variables are in correspondence with the framework of Kendall and co-authors (2006) and are complemented with some other variables. This chapter covers many research hypotheses. Next to the discussion on attitude modifiers, it also contributes to a better understanding of different viewpoints within society, given that it combines a rather concrete consumer-related measure (relative importance attached to farm animal welfare as a product attribute) with a more abstract public opinion (attitude toward animal welfare). In addition, the chapter provides a segmentation analysis, which has been discussed as a necessary tool to understand how to make higher welfare products relevant to different consumers and how to 37 Part I position these products in a competitive marketing environment. Finally, the chapter documents on the balance between human- and animal-centred concerns in relation to animal welfare. Part IV elaborates on some issues that impact positively or negatively on the relationship between animal welfare concerns and stated pro-welfare behaviour. This part includes one research paper in which survey respondents are grouped in five groups based on self-reported pro-welfare behaviour regarding eggs and chicken meat and touches the different influencing factors on the attitudebehaviour ambivalence as described in the conceptual framework (Chapter 10). Finally, Part V provides the general discussion and conclusions. The most important findings of this doctoral research are discussed. Conclusions, implications and recommendations from the different research parts are tied together. 38 PART II Public conception of farm animal welfare Chapter 1 : Public conception – qualitative approach Chapter 1 Public conception of farm animal welfare – qualitative approach Adapted from Vanhonacker, F., E. Van Poucke, F. A. M. Tuyttens, and W. Verbeke. 2010. Citizens’ views on farm animal welfare and related information provision: exploratory insights from Flanders, Belgium. Journal of Agricultural & Environmental Ethics 23(6): 551-569. Abstract This chapter combines the results of two independent empirical studies with Flemish citizens and addresses the problem of a short fall of information provision about higher welfare products. The specific objectives of this chapter are (1) to improve our understanding of how citizens conceptualise farm animal welfare, (2) to analyse the variety in the claimed personal relevance of animal welfare in the food purchasing decision process and (3) to find out people’s expectations in relation to product information about animal welfare and the extent to which the current information caters to these expectations. The first study consisted of a survey conducted in three consecutive years (2000-2002, n=521) and was complemented with qualitative data from four focus group discussions (2006, n=29). The resulting citizens’ conceptualisation of farm animal welfare matched reasonably well with those in the scientific literature, although it was clearly influenced by a lower level of practical experience and a higher weight of empathy. In general, respondents indicated animal welfare as an important product attribute, although less important than primary product attributes such as quality, health and safety. Moral issues rather than a perception of higher quality were the main influence on preferences for higher welfare products. At present, higher standards of animal welfare are mostly guaranteed within more general quality assurance schemes. People’s decisions not to choose higher welfare products seemed to be related to the perceptual disconnection between eating animal food products and the living producing animals (dissociation). Respondents generally thought better information provision was required and the present level of provision was strongly criticised. In combination, the findings of both studies help inform the discussion about how citizens can be informed about animal welfare and the preferred content, source and medium of such information. In summary, this chapter provides insights into citizens’ semantic interpretation of the concept of animal welfare (what wordings they use) and the range of relevance that animal welfare has for different groups that, in turn is useful in identifying which segments can be targeted. This can contribute to a more effective valorisation of animal welfare as a product attribute. 41 Part II 1. Introduction Over the past century, Western animal production has changed dramatically due to a strong focus on maximising productivity and profitability. This has led to the development of intensive animal rearing conditions (e.g. indoor housing to maximise production within the minimum space) that rarely took animal welfare into account and frequently compromised it (e.g. Tuyttens et al., 2008). The increased recognition of these welfare problems (Brom et al., 2007; European Commission, 2006b) has coincided with the growing importance and increased complexity and sophistication of socio-cultural production aspects (Vermeir and Verbeke, 2006) and increased societal interest in, and concerns about farm animal welfare. Nevertheless, these societal interest and concerns are poorly reflected in the consumers’ market behaviour and high welfare products still have a relatively small market share. Harper and Henson (2001) argue that low market shares for high welfare products should not necessarily be interpreted as reflecting a lack of consumer concern. They identified several barriers that impede the translation of citizen concerns into consumer behaviour: these were lack of (appropriate) information, lack of available products, lack of belief in personal influence, pervasive ignorance of modern production methods, separation of the food product from the animal of origin (dissociation) and cost. This chapter elaborates on the information issue. The specific purpose of this chapter consists of three objectives that are adopted in order to identify ways of communicating animal welfare more effectively to consumers and touches a wide range of hypotheses postulated in this dissertation. The first objective was to understand the way in which the public (as citizens) conceptualises farm animal welfare and the wordings they use (semantic meanings). Several attempts have already been made to conceptualise farm animal welfare, although most of these have been based on points of view of animal welfare scientists, without taking into account the public opinion. Scientific efforts to include citizens’ perceptions and attitudes in the conceptualisation of farm animal welfare have thus far been limited to the integrated European Welfare Quality® project (www.welfarequality.net), in which 12 issues of concern (also termed “criteria”) were grouped into four “principles” (Botreau et al., 2008; Botreau et al., 2007; Miele and Evans, 2005; Tuyttens et al., 2010a; Veissier and Evans, 2007). We argue that attempts to effectively communicate information about animal welfare (as a product attribute) to consumers are hindered by the absence of input of public opinion into science-based definitions of welfare and by citizens’ limited awareness and knowledge about the issues. The second objective was to scrutinise the variety in the claimed personal relevance of animal welfare concerns in relation to food purchasing decisions. In today’s society, there is a growing demand for farm animals to have better living conditions (European Commission, 2005) and a distinct segment of consumers who are attentive to animal welfare issues (Boogaard et al., 2006; FAWC, 2006; Vanhonacker et al., 2007). The existence of a distinct market segment calls for a 42 Chapter 1 : Public conception – qualitative approach targeted strategy of information provision, since information is most likely to be effective and efficient when it meets the specific needs of a target audience (Verbeke, 2009). The third and final objective was to learn more about citizens’ preferences and expectations for information about animal welfare. At present, the general public is poorly informed about animal welfare conditions during rearing, transport and slaughter. This is at least partly due to poor communication. At the same time, consumers can easily get lost in the numerous, sometimes competing labels that often cause confusion (Passantino et al., 2008; Verbeke and Viaene, 1999b). Investigating societal informational expectations about animal welfare issues should provide insights into the public’s preferences about content (what), source (who) and medium (how) of information about animal welfare. 2. Material and methods This chapter compiles data from a quantitative survey (Study 1) and qualitative focus group discussions (Study 2) and is exploratory in nature. These two approaches are complementary. Qualitative research techniques are very suitable for relatively unexplored and abstract or moral themes and can illustrate underlying motivations and attitudes, which are not generally revealed through closed survey questions (Malhotra and Peterson, 2005). The lack of a priori information about how citizens conceptualise farm animal welfare means that qualitative research is highly appropriate for exploring this relatively unexplored concept. The qualitative methods are also used to deepen elements from the larger quantitative study. There are several advantages in combining these two approaches. Qualitative research is often seen as having the disadvantage of drawing on a limited number of participants and thus potentially being unrepresentative. Triangulating the results between the two approaches can strengthen the robustness of the conclusions and make them more acceptable in decision-making processes. Quantitative findings by contrast, can statistically substantiate differences between groups of people. Thus, in this analysis the first research objective was addressed solely through the qualitative results, while both research methods were applied to address the second and third objective. 2.1. Survey (Study 1) Methodological aspects in terms of period, research procedure (sample selection and data collection), number of respondents, sample description and the context in which the study is performed can be found in the methodology section of this dissertation (Part I.4). The non-random sampling procedure involves that the sample cannot be considered representative for the Flemish population. Hence, results are only indicative and should not be generalised to the whole population. The female majority reflects the selection of the main person responsible for food purchases in the household. Age ranged from 17 to 86 years. Mean age was slightly below the population mean of 40.2 years (source: FOD Economy) and different age categories were well represented. The sample varied in terms of respondents’ living environment, educational level and 43 Part II whether or not households had children. There was a slight overrepresentation of urban people, higher educated people and families with children. Two themes in the questionnaire were relevant and compatible with the aim of the present chapter. First, respondents were probed about their basic expectations of animal food products. They were asked about the importance they allocate in their choice of animal food products to different product attributes (health, safety, quality, trustworthiness and origin) in relation to animal welfare. The question reads “how important are the following product attributes for you, compared to animal welfare, when making your choice for buying animal food products?”. The answers were registered on five-point Likert scales that ranged from “much less important” (1) through “equally important” (3) to “much more important” (5). In addition, respondents were asked how they associated higher welfare products with other product attributes: “More attention for animal welfare will yield products that are … (more acceptable; more authentic; more available; more environmentally friendly; healthier; cheaper; more profitable for the producer; of better quality; safer; tastier; more traditional; more trustworthy).” This question was scored on a five-point Likert agreement scale that ranges from “disagree strongly” (1) through “neither disagree nor agree” (3) to “agree strongly” (5). Both questions contributed to exploring the second objective within this chapter. Second, respondents were provided with a list of statements about the relation between information and animal welfare and between animal welfare and claimed consumer behaviour. Statements were rated on five-point Likert agreement scales. Statements were: “There is enough information available about animal welfare,” “I believe the information provided about animal welfare,” “I am rather suspicious of information provided about animal welfare,” “Animal welfare should be monitored more strictly,” “Animal welfare should be guaranteed by means of a product label” and “I am willing to pay more for higher welfare products.” These statements provided information relevant to the third objective within this chapter. In addition to socio-demographical variables, respondents were asked to indicate their meat consumption frequency on a six-point scale that ranged from “never” (1) to “several times per day” (6). Data were analysed using SPSS 15.0. The mean scores with standard deviations and frequencies were shown in table format or bar diagrams. Bivariate analyses were done through correlations (in the case of ratio-scaled variables) and comparisons of mean scores by means of independent t-tests (in the case of categorical variables with two categories) and one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores (in the case of categorical variables with more than two categories) were used to detect differences in attitudes and perceptions between different socio-demographical groups. The critical p-value was fixed on 0.05. 44 Chapter 1 : Public conception – qualitative approach 2.2. Focus group discussions (part of Study 2) Focus group participants are often referred to as “prime witnesses,” people who are particularly interesting because of their specific socio-demographic, attitudinal or behavioural profile. Four focus group discussions were organised in March 2006. The participants (29 in total) were selected based on different meat consumption profiles (primary criterion) and on gender (secondary criterion) (Table 2.1). Group 1 (G1) was an all female group (n=7) who had recently lowered their meat consumption, Group 2 (G2) consisted of eight females whose meat consumption had not dropped recently, Group 3 (G3) were female vegetarians (n=7), and Group 4 (G4) were male vegetarians (n=7). All the focus group participants were aged between 18 and 58 and were responsible for the purchase of food for their household. The meat consumption profile was deemed more important than gender in explaining variations in attitude towards animal welfare and was thus used as the primary selection criterion. Given that a focus group should be homogenous in terms of demographics and socio-economic characteristics (Malhotra and Peterson, 2005), we chose to keep the groups separate according to these selection variables. The choice of two vegetarian groups was due to the expected higher awareness and involvement of vegetarians in animal welfare issues (Hopkins and Dacey, 2008; Lea and Worsley, 2002). Females are generally more concerned about animal welfare than men (Kendall et al., 2006) and women are more frequently responsible for household food purchases. These factors led to an imbalance in the gender representation (the secondary selection criterion) giving three female groups and one male group. The discussions were guided by a moderator who followed a structured topic list. Before starting the discussion a short general introduction was given, explaining that the meeting was about citizen perceptions towards farm animals and food (without specifying animal welfare) and the participants introduced themselves to each other. Following Krueger (1988), participants were always asked to recount their personal experiences in order to avoid stereotypes. The moderators tried to avoid drifting into stereotyping when directing the flow of the discussions. The topic list followed a funnel approach, starting with general questions and building towards more specific ones. The topic list covered all three chapter objectives. Respondents’ conceptualisation of animal welfare (chapter objective 1) was approached through direct questioning as well as hidden issue questioning. Examples of direct questions in the topic guide are “What do you understand by animal welfare? What influences good/bad animal welfare for farm animals? What influences your image of animal welfare (i.e. which animal species and/or production systems)? Are there differences between animal species? What are the welfare issues for each animal species? Which animal species have better welfare than others and why? What are the conditions for good welfare? What are the most important aspects to be fulfilled to guarantee good welfare and why? How can animal welfare be improved?” The time spent discussing each question varied according to how the discussion evolved. For the hidden issue questioning, participants were asked to work in small groups of two or three. Each group was asked to think 45 Part II about two imaginary countries, one where living conditions for farm animals were good and one with bad conditions. They were asked to describe the animal welfare conditions in each country, such as how the animals were kept and housed. Chapter objectives 2 and 3 were approached through the direct questioning method and questions were strongly related to the survey questions in the quantitative study (see above). Each group discussion lasted for approximately 2-2.5 hours. The participants received a monetary incentive to participate. All the discussions were attended by the authors and were audio-taped. Table 2.1. Characteristics of the focus group participants Number of Meat consumption participants profile 1 7 Meat consumer 2 8 3 7 4 7 Group Min. age Max. age (years) (years) Female 22 57 Female 21 58 Vegetarian Female 18 34 Vegetarian Male 23 32 Recently lowered meat consumption Gender 3. Results 3.1. Citizens’ conceptualisation of farm animal welfare For reasons discussed in the materials and methods section, this section is based solely on the outcome of the focus group discussions. The direct and hidden questioning of these focus groups led us to draw up a list of aspects that reflect how people think about or conceptualise farm animal welfare. A full list of these aspects can be found in Table 2.2. Our descriptions of these aspects attempt to stay loyal to the verbal formulations made by the participants. Where participants used sentences or phrases to formulate these concepts, we sought to capture the formulation in a single or a few words. Aspects with a very similar content that were differently formulated in different groups were grouped together. Aspects that are incongruous with animal production were reoriented (e.g. “…animals should die a natural death… (G3)” is captured in the aspect “lifespan”). We then grouped these aspects into six dimensions, according to the context in which they were raised. This was done to facilitate further discussion and comparisons with the main themes within the literature. These six dimensions are discussed below. The order in which they are presented does not reflect any relative importance between them. The first dimension involves aspects related to feed: here a distinction can be made between aspects about the composition of the feed and those relating to the way of feeding. A second dimension relates to the housing of farm animals and can be subdivided into aspects that relate to barn design and those that highlight other environmental factors. A third dimension relates to the health of the animal and consists of aspects describing physical as well as psychological health. Fourthly, a dimension about the animal’s ability 46 Chapter 1 : Public conception – qualitative approach to engage in natural behaviour can be distinguished. Fifth, aspects dealing with human-animal relationships (management issues) are grouped together. Finally, a sixth dimension clusters aspects related to the final stage of the animal’s life: “transport and slaughter.” Table 2.2. List of aspects that reflect how people think about or conceptualise farm animal welfare. The list is based on the focus group discussions, and aspects are arbitrarily assigned to a dimension (in capitals), and possible sub-dimensions (in italic) FEED Feed composition natural – varied (not monotonous) – nutritionally adequate – quality – healthy – tasty for the animal – adjusted to the species – free of synthetics – free of antibiotics Feeding regime not too little – not too much – constant access to water – fixed schedule HOUSING Barn design sufficient space (for locomotion) – group housing – not too many animals together – acceptable stocking density – separate areas (sleeping area, feeding area, …) – shelter – floor type – hygiene Environmental factors outdoor access – choice of being inside or outside – fresh air – natural light – natural sounds HEALTH Physical health no preventive treatments – individual medication for diseases – natural growth rate – avoiding injuries – absence of pain – mutilations (castration, beak trimming) – being active Psychological health feeling good – absence of fear – absence of stress – absence of frustration – toys (distraction materials) – not being bored ABILITY TO ENGAGE IN NATURAL BEHAVIOUR natural behaviour – social behaviour – maternal behaviour (leave the young with the mother) – natural birth HUMAN-ANIMAL RELATIONSHIP respectful treatment – personal farmer-animal bond – favourable working conditions – individual attention TRANSPORT AND SLAUGHTER minimal duration of transport – minimal distance of transport – loading procedure – unloading procedure – number of transports per animal – adapting transport to the animal – feed/water during transport – no early slaughter 47 Part II 3.2. Personal relevance attached to animal welfare 3.2.1. Survey On average, respondents in the survey sample considered animal welfare as a less important product attribute than health, trustworthiness, quality and safety. Only origin was attributed a lower importance and taste did not differ significantly from the importance of animal welfare (Table 2.3). For each of the product attributes there was a strong concentration of answers on the midpoint of the scale (from 47.3 percent for health to 59.3 percent for safety). However, a distinct group of respondents (about 10 percent in our sample, see Table 2.3) perceived animal welfare as more important than primary product attributes such as quality, health and safety. This response was influenced by place-based variables, given that rural inhabitants assigned a lower importance to animal welfare relative to quality (p=0.030), taste (p<0.001) and origin (p=0.002). Regarding the impact of other social structural variables, no relation with either gender or educational level was found (p>0.05). The relative importance of animal welfare generally decreased with increasing age. Households with children gave relatively less priority to animal welfare (p<0.05) than those without children. Finally, a negative and significant correlation was found between the frequency of meat consumption and the relative importance given to animal welfare (p<0.05). Respondents positively associated higher welfare products with health, safety, quality and trustworthiness. There were other positive associations for taste, acceptability, traditionality and the authenticity of the product, environmental friendliness and hygiene. The most positive association of higher welfare was with the product’s acceptability. By contrast, the sample considered higher welfare products to be more expensive and less available and believed that they were less profitable for the producer (Figure 2.1). Table 2.3. Comparing animal welfare to other product attributes. Mean values with standard deviations on five-point Likert scales are provided, together with the share of the sample that attributes a higher importance (scores 1 and 2; % less important), an equal importance (score 3), and a lower importance (scores 4 and 5; % more important) to animal welfare Origin Taste Safety Quality Health Trustworthy Mean 2.77 2.99 3.17 3.25 3.49 3.31 SD 0.99 0.94 0.88 0.92 0.96 0.89 % less important 32.2 23.3 12.7 12.0 8.5 9.4 % equally important 51.4 54.0 59.3 56.3 47.3 55.5 % more important 13.4 22.7 28.0 31.8 44.2 35.0 48 Chapter 1 : Public conception – qualitative approach acceptability quality health environmental friendliness safety trustworthy hygiene taste authenticity traditionality availability profitability price 1 1.5 2 2.5 3 3.5 4 4.5 5 Figure 2.1. Association of higher welfare standards with other product attributes. Mean scores above 3 correspond with positive associations, mean scores below 3 with negative associations 3.2.2. Focus group discussions All four focus groups spontaneously expressed similar basic expectations regarding food, though they differed slightly in their emphasis. They all attached a high importance to the quality of food, which was mainly associated with the product’s health and safety. Taste, origin, availability and price were mentioned as important decision criteria, with the latter mainly associated with pricequality ratio. In all the groups, the price aspect was not mentioned very quickly, though once cited, a consensus emerged about its importance. In other respects there were quite wide variations between groups, mainly between the vegetarian (G3, G4) and the non-vegetarian groups (G1, G2). Vegetarians attached a higher value to sustainability issues (production method, environmental impact) – although these issues were also mentioned by the non-vegetarian groups – and to freshness, local production and seasonality. These latter three issues were strongly associated with health. Vegetarians also highly valued socio-cultural product attributes such as environmental friendliness, animal welfare and fair trade, whereas the two non-vegetarian groups attributed more importance to the product’s appearance and trustworthiness. Basic expectations for meat products were expressed as being higher than for other food products. The reasons for this were the (recent) crises and scandals regarding meat production, the fact that meat is derived from living beings and its limited shelf life. There were differences between the vegetarian and non-vegetarian groups in the way that they regarded the relation between animal welfare and other product attributes. The vegetarian groups provided relatively homogenous responses, possibly due to shared values that influence their attitude towards animal welfare. In general, vegetarians were convinced about the positive relation 49 Part II of animal welfare with product quality and taste (“…you consume the stress of the animal, that can’t be healthy (G3)…stress induces inferior pork quality (G3)…organic eggs are better (G4)…”). They saw higher product quality as being subordinate to higher welfare conditions (“…I have no higher expectations of the quality of the product (eggs) but I do for the hen (G4)…”). Some of the meat consumers did not associate animal welfare with quality (“… animal welfare is unrelated to quality; it is a separate aspect (G1)…quality is related to the processing and the treatment of the meat (G1)...”). Frequent positive associations were made between animal welfare and health, taste and safety, with all groups making reference to the use of antibiotics and hormones and concerns about their residuals (“…less antibiotics and residuals (G2)…”). High welfare products were generally thought be less widely available (“…the problem for consuming more is the limited availability (G1)...”), carry a higher price and trustworthiness of the label/claim to be questionable. The importance of price as a product attribute differed considerably both between and within the two groups. Within groups this variation was mainly linked to age and correlated to socio-economic status, with younger participants having a lower budget and feeling more constrained by the higher product prices. Vegetarians gave a lower importance to price than the meat eating groups. 3.3. Information and animal welfare 3.3.1. Survey In general, the majority of the sample (52.9 percent) considered the current level of welfare-related information to be inadequate, while only one quarter was satisfied with it (Table 2.4). The responses concerning the credibility and reliability of information were also quite negative, although a substantial share of respondents (about 40 percent) neither agreed nor disagreed with the statements about these two issues. Respondents strongly agreed on the need for stricter welfare monitoring. A large majority of the sample (71 percent) thought labels have a positive role to play as a communication tool about the level of animal welfare. Finally, 60 percent indicated a willingness to pay more for higher welfare products. Opinions did not differ strongly between socio-demographic groups. Females were slightly more in favour of stricter welfare monitoring (p=0.026), the need for product label relating to welfare standards (p=0.023) and showed more the willingness to pay for higher welfare products (p=0.004). Households without children more strongly expressed the lack of available information (p=0.003) and were a little more in favour of the need for stricter welfare monitoring (p=0.036). Age only influenced respondents’ perceptions about the availability of information, with the youngest age group (age 17 to 23 years) having a more negative opinion than the oldest age category (age above 54 years) (p=0.017). Educational level had no influence on any of the issue statements, while urban people only differed from rural inhabitants in being more in favour of stricter monitoring of animal welfare (p=0.001). 50 Chapter 1 : Public conception – qualitative approach Table 2.4. Citizens’ view on information provision and monitoring in relation to animal welfare. Mean value and standard deviation on five-point Likert agreement scales. Sample share disagreeing (scores 1 and 2), answering neutral (score 3), and agreeing (scores 4 and 5) There is sufficient information available about animal welfare I believe the information about animal welfare I am rather suspicious about information about animal welfare Animal welfare should be monitored more strictly Animal welfare should be guaranteed by means of a product label I am willing to pay more for higher welfare products 3.3.2. Mean SD Disagree Neutral Agree 2.66 1.08 52.9 20.1 27.0 2.84 0.96 35.7 39.0 25.4 3.15 0.89 24.7 38.5 36.8 4.04 0.89 4.8 16.2 78.9 3.89 0.96 7.9 20.9 71.2 3.55 1.01 15.9 23.6 60.5 Focus group discussions Participants in the focus groups generally thought that there was insufficient available information about animal welfare when purchasing animal food products. Only a few respondents were content with the present levels of information provided and saw no need for additional information (“… that is none of my business (G2)…”). Even so, these respondents recognised the relevance of information provision, given the substantial societal interest in animal welfare. The majority were receptive to animal welfare related information and for some people it was a necessity: “…I would like to know more about it (animal welfare), then I would deal even more consciously with my food (G2) …people should be confronted with reality, now they are misled by advertising (G4)…” Considering the source of the information, participants stressed the importance of objectivity and expertise, with information being scientifically sound (i.e. credible and reliable) and straightforward (i.e. simple and understandable). They preferred a neutral source (e.g. the government instead of animal welfare organisations), recognising the potential influence of vested interests and their tendency to highlight only one side of the story. Other potential information sources mentioned were butchers, farmers, schools, advertising, educational and documentary television programs, leaflets (although these were thought to have a limited impact) and labelling. Labels were seen as an opportunity, though not in the way they are applied now (“…there is a proliferation of labels (G3)… labels are often to comfort the consumers, who do not know what the label is actually about (G3)…the meaning of labels should be more clear (G2)…”). 51 Part II Participants were often confused about the actual meaning of animal welfare labels and thought that they needed to be clear and reliable if they were to be effective or trustworthy. Participants stressed that the content of a label should be understandable by everyone. Several ideas were suggested about how a good label might be developed. These included an integrated approach (“…a welfare label should be established based on the consultation of different interest groups like producers, retailers, consumers, organisations related to vegetarianism and animal rights, the government (G4)…”) and a demand for simplicity (“…is it not possible to work with classes like A to E where A corresponds with the best welfare and E with the worst? (G3) solely two possibilities namely good or bad animal welfare…. (G2)”). A simple and readily recognisable label was deemed important as the population has a very different level of interest and origin of interest in higher welfare products. Some participants suggested that some shelves should only contain products from animals raised under good welfare conditions, so that they would not have to search for welfare related information. When asked who should be responsible for improving animal welfare, respondents focused on those who ran current intensive production systems, agreeing that animal suffering is mainly caused by mass production (“…the agricultural system is on the rocks (G2)…animal production has evolved at the service of provenance (G2)…the more people that want to consume meat the more meat that has to be produced (G3)…”). However, the majority of the participants thought that responsibility could not be pinned entirely on farmers (“…taking care of animal welfare is only possible on small farms (G4)…we have created a farmer that is industrial and no longer a farmer (G1)…they have to fight to survive (G1)…it is not a matter of not being willing but of not being able (G1)…”). Nonetheless, some participants – mainly vegetarians – thought that farmers were overly focused on profits instead of their animals. Focus groups also thought that consumers were partly responsible. Vegetarians in particular emphasised the effects of contemporary lifestyles (“…mankind has a feeling of superiority, he wants more and more (G2)…only buy what is necessary for a healthy life (G4)…eating no or fewer animal products has an impact on mass production (G4)…”). Participants varied in their opinions over this issue: some thought that consumers can influence the system by buying more animal friendly products, while others did not see how they could have an impact given the lack of direct involvement. Government was seen as having an important responsibility, especially in formulating labels, monitoring welfare and disbursing subsidies. The participants agreed that the responsibility did not lie solely with one actor, but that there is a need for more general awareness and change in mentality. 52 Chapter 1 : Public conception – qualitative approach 4. Discussion and conclusion This chapter has explored aspects of the increasing societal interest in farm animal welfare and has focused on the short fall in information provision. Three specific objectives could be distinguished in this chapter. First, it has analysed the way that Flemish citizens think about farm animal welfare: how they conceptualise it and the terminology they use to describe it. This relates to the first research objective of this dissertation. Second, it has investigated the claimed relevance of animal welfare as a product attribute versus other attributes. It has also explored how this varies between different social groups. In this perspective, it links with the second research objective of this dissertation. Finally, it has provided insights into how citizens think welfare information should be framed: in terms of content, source and medium. This ties up with this dissertation’s third research objective. Combined, these study findings can help to support further development of communication and marketing strategies. Regardless of the lay status (in terms of practical knowledge and awareness about animal production) of participants in the focus groups, the discussions provided a valuable (although relatively informal) list of aspects that the participants associated with farm animal welfare. The list showed much overlap with existing definitions provided in the literature. In fact, the four principles from Welfare Quality® (good housing, good feeding, good health, appropriate behaviour) were quite evident in the results, although the focus groups added two further dimensions, relating to transport and slaughter and to human-animal relationships. Both these additional dimensions received considerable attention during the focus group discussions and were discussed in a context that was distinct from the other dimensions. In comparison, the four principles of Welfare Quality® are designed to be applied both on farm and during transport and slaughter, and the human-animal relationship is contained within the principle of “appropriate behaviour”. On the aspect level, animal welfare measurements that require good knowledge of animal production and that are often used in welfare assessments (e.g. lameness, mastitis...) were not incorporated in the citizens’ conceptualisation of farm animal welfare. In general, the list of aspects provided by the focus groups seemed to some extent influenced by an idyllic image of the countryside and the position of farm animals (e.g. Korthals, 2002; Te Velde et al., 2002) and by an anthropomorphic perspective (McInerney, 2004; Rushen, 2003), views that are both fundamentally antithetical to current patterns of intensive animal production. Finally, the impact of recent food crises and scandals that undermined the public’s trust in animal production is still strong and was reflected in the presence of aspects such as “free of synthetics/antibiotics” and “no preventive medication”. These issues notwithstanding, the insights into the concerns of focus group members and the language they use make a useful contribution in developing understanding about how to effectively communicate about animal welfare. Respondents to the survey gave a much higher priority to primary product attributes, such as quality, health and safety than to socio-cultural aspects, such as animal welfare, environmental 53 Part II friendliness, and fair trade. Theoretically, this corresponds with Maslow’s hierarchy of needs (1943) and empirically it corresponds with earlier findings (Heleski et al., 2004; Ingenbleek et al., 2006a; Kanis et al., 2003; Verbeke and Viaene, 1999a). However, with the exception of a group that ignores animal welfare as a product attribute, both the survey and focus group discussions revealed the existence of a specific market segment that values animal welfare and considers it when choosing animal food products. The present study shows that urban people, younger people and people without children gave a relative higher importance to farm animal welfare as a product attribute. People’s preference (or rejection) of higher welfare products was more closely associated with moral issues than with quality or sensory ones. The main reasons for choosing animal welfare products were the higher acceptability of higher welfare production methods and respect for animals. By contrast those who did not have a preference for higher welfare products tended to ignore the animal origin in order to avoid feelings of guilt (Hopkins and Dacey, 2008; Niesten et al., 2003). In current marketing strategies, animal welfare is incorporated in broader quality assurance schemes (Verbeke, 2009) rather than having a specific and dedicated ethical claim or label. Respondents’ negative associations between higher welfare products with price and availability matched findings in the literature (e.g. Harper and Henson, 2001; Harper and Makatouni, 2002). Yet, the extent to which these negative associations inhibited claims of pro-welfare behaviour was inversely related to the personal priority that individuals gave to animal welfare. Efforts to communicate animal welfare issues with the public need to acknowledge the variety in people’s interest in animal welfare and their engagement in pro-welfare behaviour (Wilson, 1981). While people’s behaviour towards animal welfare varies, most citizens agree about the inadequacy of animal welfare-related information on food products, both in terms of availability and the way in which it is currently provided. They see the need for information that is credible and reliable and that can be interpreted at glance. The public’s need for credible and reliable information led to them expressing a preference for a source that is perceived as free from vested interests (such as the government) to communicate welfare-related information. Ideally, such information should be the result of an integrated approach, reflecting the general opinion that farm animal welfare is a joint responsibility. Respondents favoured a product label that is focused specifically and solely on animal welfare in a transparent format as the medium for communicating this information. Although this study sheds some light on the public’s expectations on, and interests in information related to animal welfare, practical implementation remains particularly challenging with respect to unambiguously determining what “good animal welfare” is and what amount of information should be communicated to the public and by whom. In conclusion, the study findings provide valuable insights that might be applied to the development of future communication and information provision about farm animal welfare. The results also positively contribute to bridging the perceptual discordance between the views of the general public and of welfare scientists and producers about what constitutes animal welfare and can thus be valuable to support a constructive dialogue. 54 Chapter 2 : Public conception – quantitative approach Chapter 2 Public conception of farm animal welfare – quantitative approach Adapted from Vanhonacker, F., W. Verbeke, E. Van Poucke, Z. Pieniak, G. Nijs, and F. A. M. Tuyttens. 2010. The concept of farm animal welfare: citizen perceptions and stakeholder opinion in Flanders, Belgium. Journal of Agricultural & Environmental Ethics doi 10.1007/s10806-010-9299-6. Abstract The objective of this chapter is to develop a conception of farm animal welfare that starts from the public’s perception and integrates the opinion of different stakeholder representatives, thus following a fork-to-farm approach. It constitutes a continuation of the previous chapter. The findings from the qualitative citizen focus group discussions were used to develop a quantitative questionnaire. The questionnaire has been completed by a sample of 459 respondents that have been selected to be representative of the Flemish (adult) citizen population in terms of gender, age, living environment (rural versus urban) and region (province distribution). Exploratory and confirmatory factor analyses are applied to develop a conception of farm animal welfare starting from an extended list of aspects that relate to animal production and associate with farm animal welfare in the public perception. In depth interviews with stakeholder representatives were used to match and adapt the structure of the animal welfare conception model. The resulting conception revealed seven dimensions grouped in two different levels. Three dimensions were animal-based: “Suffering and Stress”, “Ability to Engage in Natural Behaviour” and “Animal Health”. Four dimensions were resource-based: “Housing and Barn Climate”, “Transport and Slaughter”, “Feed and Water” and “Human-Animal Relationship”. This conception is distinct from earlier attempts since it is based on public perceptions; it addresses the opinion of different stakeholders and it distinguishes empirically between animal-based and resource-based dimensions in the conception of farm animal welfare. The relevancy of a definition that incorporates the public perception is supported by the present demand oriented economy, in which animal welfare is a non-trade concern and mainly left to the market where consumers still mainly act as individuals who calculate and weigh pros and cons. 55 Part II 1. Introduction “Animal welfare” is increasingly referred to as an important goal by different stakeholders along the food production chain (Verbeke, 2009), yet its meaning and conception differs depending on who is using the concept, which hampers effective policy implementations. Many of the conceptions, especially the ones developed from an animal science perspective, are criticised for being too limited and not sufficiently addressing the issues of actual public concern (Fraser, 2008; Lund et al., 2006; Rushen, 2003; Rushen and Depassille, 1992), whereas animal welfare is by nature a social concept that reflects societal values (Fisher, 2009; McInerney, 1991). Fisher (2009) suggests that the term should be considered as a judgment to be undertaken by combining varied understandings and to acknowledge underlying values and assumptions, both moral and scientific, rather than risking to exclude valid viewpoints through pursuing a prescriptive definition. In addition, Rushen (2003) stressed the relevancy of research into societal views on animal welfare, as well as citizen values and norms to shed light on the factors that steer the behaviour of people. Fisher (2009) further indicates that the essential reason for defining a term is to communicate. Considering the fact that animal welfare is to a large extent left to the market (MacArthur Clark, 2008; Sørensen and Fraser, 2010), the target audience for communication should be the public. In this perspective, a definition for farm animal welfare should appeal to the public and should take into account the public’s low practical knowledge base and awareness in relation to animal production practices. An accepted societal definition could act as a focus for debate and a target for improving and communicating about animal welfare. However, a societal conception of farm animal welfare is, compared to conceptions from other perspectives and despite its accredited relevancy (Fraser, 2001; Rushen, 2003), not yet extensively covered in scientific literature. The present chapter acknowledges the importance that the chain end user can play in the debate about animal welfare. Animal welfare is still considered a non-trade concern at WTO-level and the responsibility is by many stakeholders delegated to this end user who can vote for or against a product through purchasing it or not. As such, insights in how the public perceives the concept of farm animal welfare are important and relevant, more so in the present information society, where inadequate information gets lost in the abundance of information available. This study’s research question is therefore formulated as how the scientific definitions for animal welfare translate in a “popular” definition, necessary to support a constructive dialogue and an effective communication to the public, both in their role as consumers and citizens. Consumers can contribute through the purchase of higher welfare products. Therefore, animal welfare characteristics need to be simple and definable, recognised and explicit and certifiable as product characteristics (McInerney, 2004). With respect to the individual in his/her role as citizen, it is acknowledged that the concern of the public dictates the need for animal welfare standards and animal welfare legislations (Caporale et al., 2005; Edwards and Schneider, 2005; Garnier et al., 2003). In both cases, a good understanding of the public’s conception of farm animal welfare is determinant. 56 Chapter 2 : Public conception – quantitative approach The aim of this chapter is to translate the opinion of citizens with regard to farm animal welfare into a structured concept that is supported by the different stakeholders and interests groups along the food production chain. The study uses a more formal, transparent and quantitative methodological approach as compared to previous studies. The consistency with existing conceptions will be discussed, together with the applicability of our findings with respect to communication and how to put into practice. 2. Material and methods This study is a continuation of the citizen focus group discussions, considered in the previous chapter. Together with a profound literature search the focus group discussions served as the basis for the construction of a survey questionnaire. Data from this survey allowed conceptualising the public view on farm animal welfare in a quantitative way. Finally the outcome of the quantitative study was presented to, and discussed with representatives from different stakeholders and interest groups who are active in the meat production chain in order to incorporate a wide vision into our conception, following Fisher (2009). The finality of each part of the study is shown in Figure 2.2, and further methodological details on each part are provided in the next sections. All studies have been performed in Flanders. Focus group discussions (Chapter 1) 56 aspects in 6 dimensions Literature search and internal discussion 73 aspects Input for confirmatory factor analysis Exploratory factor analysis Construct validation Confirmatory factor analysis Quantitative public conception of farm animal welfare In depth interviews Figure 2.2. Methodological summary 57 Part II 2.1. Qualitative research –focus group discussions & literature search (Study 2) Full details about the methodology and results of the focus group discussions have been reported in Chapter 1. In brief, four focus group discussions with a total of 29 participants were performed in March 2006. Analysing the focus group discussions involved transcribing, structuring and content analysing the data. The main goal of the focus groups was to obtain a workable list of aspects that are related to animal production and that were believed to associate with farm animal welfare in the public perception, thus factors or production practices that people perceive as important for animal welfare. In total 56 aspects were obtained and were arbitrarily classified in the following six dimensions: feed and feeding / housing / animal health / engagement in natural behaviour / human-animal relationship / transport and slaughter. To account for the limited number of respondents involved in the focus group discussions, the list was modified and further extended to 73 aspects based on additional literature review and internal discussion among the scientists involved in this study (Table 2.5). Retaining or rejection of aspects in the final citizens’ conception of farm animal welfare was based on quantitative data analysing techniques. Table 2.5. Full list of aspects used for framing the public conception of farm animal welfare. Between brackets the aspect coding is given that will be used in the syntax of the confirmatory factor analysis and in further reporting Functional areas in barn (i30) Disease (i31) Lairage time (i32) Feed on fixed moments (i33) Foraging behaviour (i34) Flooring type (i35) Body care (i36) Boredom (i37) Freshness of feed (i38) Stunning (i39) Static groups (i40) Transport of living animals (i41) Outdoor access (i42) Duration of transport (i43) Stress (i44) Mortality (i45) Barn temperature (i46) Hygiene in the barn (i47) Play behaviour (i48) Social behaviour (i49) Taste of feed (i50) Sexual behaviour (i51) Shelter (i52) Shockproof & calm transport (i53) Space during transport (i54) 58 Respect for animals (i55) Preventive medication (i56) Having fun (i57) Slaughter without pain&stress (i58) Pain by human intervention (i59) Pain by conspecifics (i60) Farmer-animal bond (i61) Handling of animals (i62) Natural environment (i63) Natural growth rate (i64) Natural feed (i65) Natural birth (i66) Natural behaviour (i67) Music in the barn (i68) Maternal behaviour (i69) Procedure of (un)loading (i70) Mixing of groups during t&s (i71) Curative medication (i72) Technical noise (i73) Air quality (i74) Light regime (i75) Life span (i76) Ability to rest (i77) Climate during transport (i78) Design of slaughterhouse (i79) Group housing (i80) Hunger during transport (i81) Size of livestock herd on farm (i82) Group size (i83) Growth hormones (i84) Variation in feed (i85) Genetic selection (i86) Frustration (i87) Frequency visual inspection (i88) Explorative behaviour (i89) Balanced feed (i90) Thirst during transport (i91) Daylight (i92) Comfort (i93) Pen size (i94) Availability of water (i95) Availability of feed (i96) Skilled animal handlers (i97) Fear (i98) Distraction material (i99) Number of transports (i100) Stocking density (i101) Attention for animals (i102) Chapter 2 : Public conception – quantitative approach 2.2. Quantitative research – survey (Study 2) 2.2.1. Study design and subjects Cross-sectional survey data were collected during April 2006. A quota sampling procedure was followed with gender, age, living environment and province as quota control variables. Questionnaires were distributed via web-links, mail and personal contact procedures and has resulted in a sample of 459 respondents that closely reflects the Flemish adult (> 18 years) population in terms of the quota control variables (see Table 1.6). With respect to gender, a representative distribution relative to the population is obtained. The age of the respondents ranged from 18 to 75 years, with a mean age of 37.8 (SD=14.1), which is slightly below the population mean (40.2 years) (NIS 2002). Concerning urban versus rural living environment, a 35/65 ratio was aimed for. In comparison with census data, we obtained a small over-representation of urban respondents. Finally, a small over-sampling of the provinces West- and East-Flanders (resp. +6.2 and +4.7 percent) resulted at the expense of the provinces of Antwerp and Limburg (resp. -5.9 and -4.6 percent). With regard to the non-quota control variables, education and farming background, a small over-sampling of higher education occurred, while a realistic percentage of respondents with a farming background (around 6 percent of the population) were obtained. 2.2.2. Survey The analysis to conceptualise farm animal welfare from a public point of view was based on a single question that had to be answered for each of the 73 aspects. Responses on the question “According to your personal opinion, how important is this aspect in obtaining an acceptable level of farm animal welfare” were registered on a five-point Likert scale with the following response categories: “totally unimportant” (1), “unimportant” (2), “moderately important” (3), “important” (4), and “very important” (5). In order to avoid item order bias, five different questionnaire versions were used that differed in the presentation order of the aspects. 2.2.3. Exploratory factor analysis In order to prepare the data for a confirmatory factor analysis, the full list of aspects were factor analysed using the principal components extraction method with varimax rotation. The purpose of the initial exploratory factor analysis was to discover structures in the pattern of relationships among the aspects. In particular, it aimed to discover if the observed variables (the 73 aspects) can be explained largely or entirely by a smaller number of unobserved factors or components (referred to as dimensions in this study). Thus the exploratory factor analysis indicates how responses to individual animal welfare factors are correlated with one another. The selection of dimensions was based on the Kaiser Criterion (Eigenvalues >1), while factor loadings were used to interpret the meaning of the resulting dimensions and nominate them accordingly. 59 Part II Assigning an aspect to a particular dimension was informed by a high factor loading on its respective dimension. Given that the exploratory factor analysis was considered in this study to prepare the data for a further confirmatory factor analysis, no aspects were excluded at this stage from further analysis. Rather they were assigned by the authors to one of the resulting dimensions if the factor loadings’ criterion failed. Exploratory factor analysis was performed using the statistical software package SPSS 15.0. 2.2.4. Confirmatory factor analysis and construct validity Confirmatory factor analysis was used to investigate the discriminant and convergent validity of the measurement model, which consists of a priori defined relationships between observed variables (aspects) and their underlying latent constructs (dimensions). The postulated relationships between the aspects were based on the outcome of the exploratory factor analysis and in accordance with theoretical insights. Analyses were performed using the robust maximum likelihood procedure in LISREL 8.72. With the use of structural equation modelling the examination of all the relationships between items (aspects) and constructs (dimensions) was performed simultaneously, which is a substantial advantage compared with single equation modelling (Bollen, 1989). To evaluate the fit of the model, the χ²-value together with the degrees of freedom is reported, as well as three other indices: the Root Mean Square Error of Approximation (RMSEA), the Non-Normed Fit Index (NNFI) and the Comparative Fit Index (CFI). Values below 0.08 for RMSEA (Browne and Cudeck, 1993) and above 0.90 for NNFI and CFI (Bollen, 1989) indicate an acceptable fit of the measurement model. Confirmatory factor analysis was also performed to determine whether measures of a construct actually converge the intended latent variable or share a high proportion of variance in common (convergent validity) and whether the constructs were distinct from each other (discriminant validity). In order to assess convergent validity examination of unidimensionality and reliability estimates was performed (Hair et al., 2006). 2.3. In depth interviews In the final stage of this chapter, seven in depth interviews with stakeholder representatives and interest groups were carried out. Three persons were recruited to represent the producer side, one person represented the retail side, one represented animal welfare organisations, and finally two persons with a scientific background in animal welfare or veterinary sciences were interviewed. In the first part of the interview the interviewees were asked about their personal conception of farm animal welfare, before confronting them with the outcome of the confirmatory factor analysis. Thereafter, they were asked to comment on the public conception, both in terms of the resulting dimensions and their aspects. This should allow incorporating also their vision into the conception. 60 Chapter 2 : Public conception – quantitative approach 3. Results 3.1. Exploratory factor analysis The exploratory factor analysis yielded a 12 factor solution (Table 2.6) explaining a total of 62.5 percent of the variance in the original data. On the first factor high loadings were related to aspects that were associated with housing as well as with natural behaviour and the factor was termed accordingly (dimension: housing and natural behaviour (H&NB)). The second factor included aspects that were related to the process of transport and slaughter of animals (dimension: transport and slaughter (T&S)). Three aspects (number of transports per animal; thirst during transport; mixing of groups during transport and slaughter) that had a slightly higher factor loading on the dimension H&NB, were deemed more appropriate under the dimension T&S and were accordingly attributed to the latter dimension. Aspects that loaded high on the third factor were related to animal suffering and stress (dimension: suffering and stress (S&S)). Aspects loading high on the fourth factor could not be unambiguously categorised under a common nominative. As such we verified whether the aspects had also high loadings on other factors. Accordingly, flooring type, static groups, functional areas in the barn and air quality were allocated to the dimension H&NB, while design of the slaughterhouse was allocated to the dimension T&S. The fifth factor mainly contained aspects which have an affinity with the animals’ diet (dimension: feed). The sixth factor included aspects that were related to the relationship between the persons taking care of the animals and the animals themselves (dimension: human-animal relation (H-A)). Also the aspect frequency of visual inspection, with an original highest loading on the undetermined fourth factor, was allocated to this dimension. No further factors with a logical composition of aspects were found. Aspects that were not attributed to one of the aforementioned dimensions based on the exploratory factor analysis were: availability of water, barn temperature, music in the barn, genetic selection, feeding on fixed moments, barn hygiene, growth hormones, preventive medication and curative medication. In order not to exclude them from the confirmatory factor analysis these aspects were arbitrarily assigned to one of the five aforementioned dimensions. Availability of water, feeding on fixed moments and growth hormones were assigned to the dimension feed; barn temperature, genetic selection, barn hygiene and music in the barn to the dimension H&NB; and preventive and curative medication to the dimension S&S. 3.2. Confirmatory factor analysis and construct validity The conception of farm animal welfare in which the 73 aspects are categorised in five dimensions (Model 1 (M1)) based on the exploratory factor analysis gave acceptable goodness of fit statistics in confirmatory factor analysis (Table 2.7), though could be further optimised. Before optimising the measurement model through elimination and/or reallocation of aspects based on the statistical output of the confirmatory factor analysis, we verified the impact of separating the aspects related to housing from the aspects related to natural behaviour. This separation matched better with 61 Part II theoretical based conceptions from animal scientists where natural behaviour has been referred to as a singular dimension (e.g. Fraser, 2008). Also this separation allowed to better distinguishing between an animal-based and a resource-based dimension. Goodness of fit statistics further improved, thus yielding a conception in six dimensions (M2) (Table 2.7). Table 2.6. Output of the exploratory factor analysis on the full list of aspects. Only the factor loadings above 0.30 are presented for the twelve factors with Eigenvalue above 1 Body care (i36) Play behaviour (i48) Natural behaviour (i67) Outdoor access (i42) Social behaviour (i49) Explorative behaviour (i89) Natural environment (i63) Sexual behaviour (i51) Having fun (i57) Daylight (i92) Group housing (i80) Natural birth (i66) Maternal behaviour (i69) Distraction material (i99) Foraging behaviour (i34) Stocking density (i101) Shelter (i52) Pen size (i94) Natural growth rate (i64) Boredom (i37) Technical noises (i73) Light regime (i75) Group size (i83) Size of livestock herd (i82) Ability to rest (i77) Number of transports (i100) Thirst during transport (i91) Mixing of groups during t&s (i71) Life span (i76) Comfort (i93) Handling of animals during t&s (i62) Space during transport (i54) Procedure of (un)loading (i70) Stunning (i39) Slaughter without pain or stress (i58) Duration of transport (i43) Shockproof and calm transport (i53) Climate during transport (i78) Lairage time (i32) Transport of living animals (i41) Hunger during transport (i81) Disease (i31) Stress (i44) Fear (i98) Frustration (i87) Pain by human intervention (i59) Pain by conspecifics (i60) Mortality (i45) Flooring type (i35) Static groups (i40) Design slaughterhouse (i79) Functional areas in barn (i30) 62 1 .76 .76 .74 .72 .72 .70 .69 .68 .65 .63 .63 .62 .60 .59 .58 .54 .51 .51 .50 .50 .50 .49 .48 .45 .44 .44 .41 .40 .38 .38 .35 .40 .33 .30 .38 .45 .32 .37 2 4 5 6 7 8 9 10 11 12 .33 .33 .34 .30 .38 .40 .31 .36 .43 .35 .34 .44 .44 .35 .41 .37 .36 .40 .30 .32 .36 .30 .32 .32 .67 .59 .59 .56 .52 .50 .47 .43 .41 .40 .39 .36 .42 .32 .41 .42 .47 3 .31 .37 .34 .31 .36 .40 .32 .30 .66 .65 .63 .59 .58 .52 .48 .35 .35 .38 .31 .55 .49 .48 .48 Chapter 2 : Public conception – quantitative approach Frequency of visual inspection (i88) Air quality (i74) Natural feed (i65) Variation in feed (i85) Freshness of feed (i38) Balanced feed (i90) Attention for animals (i102) Respect for animals (i55) Farmer-animal bond (i61) Skilled animal handlers (i97) Availability of water (i95) Barn temperature (i46) Availability of feed (i96) Music in the barn (i68) Taste of feed (i50) Genetic selection (i86) Feed on fixed moments (i33) Hygiene in the barn (i47) Growth hormones (i84) Preventive medication (i56) Curative medication (i72) .35 .49 .34 .45 .44 .39 .62 .61 .59 .55 .31 .66 .65 .53 .53 .40 .32 .63 .50 .46 .32 .64 .58 .46 .39 .32 .37 .39 .56 .35 .35 .78 .34 .57 .47 Table 2.7. Results of the confirmatory factor analysis: measurement model and goodness of fit statistics for the consecutive models Measurement model M1 Housing & Natural Behaviour: i30 i34 i35 i36 i40 i42 i46 i47 i48 i49 i51 i52 i57 i63 i64 i66 i67 i68 i69 i73 i74 i75 i76 i77 i80 i82 i83 i86 i89 i92 i93 i94 i101 * Goodness of fit statistics χ²/df = 3.26 RMSEA = 0.070 Transport & Slaughter : i32 i39 i41 i43 i53 i54 i58 i62 i70 i71 i78 i79 i81 i91 i100 CFI = 0.98 Feed &Water: i33 i38 i50 i65 i84 i85 i90 i95 i96 NNFI = 0.98 Suffering & Stress: i31 i37 i44 i45 i56 i59 i60 i72 i87 i98 i99 Human-Animal relationship: i55 i61 i88 i97 i102 M2 Housing: i30 i35 i40 i42 i46 i47 i52 i68 i73 i74 i75 i77 i80 i82 i83 i93 i94 i101 χ²/df = 3.04 Natural Behaviour: i34 i36 i48 i49 i51 i57 i63 i64 i66 i67 i69 i76 i86 i89 i92 RMSEA = 0.067 Transport & Slaughter: i32 i39 i41 i43 i53 i54 i58 i62 i70 i71 i78 i79 i81 i91 i100 CFI = 0.98 Feed & Water: i33 i38 i50 i65 i84 i85 i90 i95 i96 NNFI = 0.98 Suffering & Stress: i31 i37 i44 i45 i56 i59 i60 i72 i87 i98 i99 Human-Animal relationship: i55 i61 i88 i97 i102 M3 Housing: i30 i35 i40 i42 i52 i73 i74 i75 i77 i80 i83 i93 i94 i101 χ²/df = 2.51 Natural Behaviour: i34 i36 i48 i49 i51 i57 i63 i64 i66 i67 i69 i89 i92 RMSEA = 0.057 Transport & Slaughter: i39 i41 i43 i53 i54 i58 i62 i70 i78 i79 i81 i91 i100 CFI = 0.98 Feed & Water: i38 i50 i85 i90 i95 i96 NNFI = 0.98 Suffering & Stress: i44 i45 i59 i60 i87 i98 Human-Animal relationship: i55 i61 i88 i97 i102 M4 Housing: i30 i35 i40 i42 i52 i73 i74 i75 i77 i80 i83 i93 i94 i101 χ²/df = 2.24 Natural Behaviour: i34 i36 i48 i49 i51 i57 i63 i64 i66 i67 i69 i89 i92 RMSEA = 0.052 Transport & Slaughter: i39 i41 i43 i53 i54 i58 i62 i70 i78 i79 i81 i91 i100 CFI = 0.98 Feed & Water: i38 i50 i85 i90 i95 i96 NNFI = 0.97 Suffering & Stress: i44 i45 i59 i60 i87 i98 Human-Animal relationship: i55 i61 i88 i97 i102 Animal Health: i31 i45 i47 i56 i72 * Model fit improves if χ²/df and RMSEA decreases and CFI and NNFI increase 63 Part II In the next step we sought to further improve the measurement model through elimination or reallocation of aspects (M3). Considering the eliminated aspects in more detail we found several aspects to have some affinity with each other, possibly due to a dimension that is not covered in the previous models. In this perspective, disease, barn hygiene, preventive medication and curative medication were grouped into a seventh dimension (dimension: animal health (AH)). Also the aspect mortality that was originally allocated to the dimension S&S was placed in the animal health dimension. The fit statistics for this model (M4) further improved in terms of a lower RMSEA and χ²/df-value, thus supporting the introduction of a seventh dimension. No further adjustments were made to the model (M4) based on the confirmatory factor analysis outcome. Standardised factor loadings and reliability estimates are presented in Table 2.8. Almost all individual aspect loadings on the dimensions were highly significant with values ranging from 0.57 to 0.83. Two aspect loadings were relatively low, particularly 0.48 (availability of water) and 0.37 (preventive medication). Nevertheless we decided to keep the aspects in the further analysis. Availability of water matches with “absence of prolonged thirst”, which is specifically mentioned as one of the 12 criteria under the Welfare Quality® conception (Botreau et al., 2008) and the use of preventive medication associates with the negative perception citizens’ hold with respect to the use of antibiotics in animal production. Further analysis has shown that including those aspects still revealed good constructs. No cross-loadings of 0.4 or more appeared. Hence, all other aspects were considered in the interpretation of the dimensions (Hair et al., 2006). Additionally, Cronbach’s alpha reliability coefficients were equal to or above the threshold value of 0.7 for satisfactory scales. Our results herewith fulfil the criteria for convergent validity for internal constructs (Bagozzi et al., 1991; Hair et al., 2006). Correlations between the dimensions are presented in Table 2.9. All correlations were significant. However, the correlation matrix revealed two (bivariate) correlations above 0.80, namely between housing & barn climate and ability to engage in natural behaviour; and between housing & barn climate and transport and slaughter, suggesting that those dimensions are closely related. 64 Chapter 2 : Public conception – quantitative approach Table 2.8. Standardised factor loadings of the retained aspects to the final model (M4) with 4 environment-based and 3 animal-based dimensions Ability to rest (i77) Comfort ( i93) Outdoor access ( i42) Group housing (i80) Light regime (i75) Air quality (i74) Group size (i83) Shelter (i52) Available space (i94) Technical noise ( i73) Stocking density (i101) Functional areas in barn (i30) Stable groups (i40) Flooring type (i35) Space during transport (i54) Shockproof and calm transport (i53) Duration of transport (i43) Number of transports (i100) Transport of living animals (i41) Handling of animals (i62) Procedure of (un)loading (i70) Climate during transport (i78) Thirst during transport (i91) Slaughter without pain or stress (i58) Hunger during transport (i81) Design of slaughterhouse (i79) Stunning (i39) Attention for animals (i102) Respect for animals (i55) Farmer-animal bond (i61) Frequency of visual inspection (i88) Skilled animal handlers (i97) Balanced feed (i90) Freshness of feed (i38) Availability of feed (i96) Variation in feed (i85) Taste of feed (i50) Availability of water (i95) Reliability estimates Inter-item correlations Housing & Barn Climate 0.73 0.71 0.74 0.71 0.68 0.64 0.68 0.68 0.66 0.64 0.66 0.64 0.60 0.63 Environmental based dimensions Transport & Human-Animal Slaughter Relationship Feed & Water Play behaviour (i48) Body care (i36) Natural behaviour (i67) Explorative behaviour (i89) Social behaviour (i49) Having fun (i57) Natural environment (i63) Maternal behaviour (i69) Foraging behaviour (i34) Sexual behaviour (i51) Natural birth (i66) Daylight (i92) Natural growth rate (i64) Disease (i31) Mortality (i45) Hygiene in the barn (i47) Curative medication (i72) Preventive medication (i56) Frustration (i87) Stress (i44) Fear (i98) Pain by conspecifics (i60) Pain by human intervention (i59) 0.81 0.80 0.76 0.75 0.71 0.71 0.71 0.70 0.69 0.68 0.66 0.66 0.59 Animal based dimensions Ability to Engage in Animal Suffering Animal Health Natural Behaviour and Stress 0.83 0.81 0.80 0.81 0.74 0.75 0.79 0.71 0.70 0.68 0.67 0.68 0.65 0.63 0.63 0.61 0.57 0.37 0.73 0.71 0.71 0.70 0.66 0.72 0.69 0.67 0.63 0.60 0.72 0.63 0.58 0.72 0.60 0.48 0.92 0.45 0.93 0.51 0.78 0.44 0.79 0.39 0.94 0.55 0.68 0.32 0.83 0.50 65 Part II Table 2.9. Correlation matrix of the animal welfare dimensions 1 2 3 4 5 6 1. Animal Health 1.00 2. Animal Suffering & Stress 0.57 1.00 3. Ability to Engage in Natural Behaviour 0.39 0.64 1.00 4. Housing & Barn Climate 0.51 0.70 0.88 1.00 5. Feed & Water 0.59 0.50 0.59 0.65 1.00 6. Transport & Slaughter 0.57 0.73 0.78 0.83 0.63 1.00 7. Human-Animal Relationship 0.59 0.51 0.58 0.65 0.65 0.63 3.3. 7 1.00 In depth interviews Expert opinions were gathered in order to fine tune the model and frame it in a way that it is appropriate for the different stakeholders along the chain. The background of the interviewees was well reflected in the way they approached animal welfare, and perceptions varied from “animal welfare is economics” to “animal welfare should be determined starting from the animal”. The following selected quotes more clearly reflect the opinions of the different stakeholders: Producers: “…the motivation to keep animals is not important, animals do not realise this anyway…what is important is the way animals are reared…this should happen as good as possible…but you have to be careful with anthropomorphism…” “…for animal welfare in farm animals one should take into account the fact that it is related to economic consequences, many people do not see this…animal welfare means that animals should be treated in a good way…” “…animal welfare is meeting the five freedoms in my opinion…” Retail: “…animal welfare is a general right of animals…it involves having respect for the animals…quality demands respect…what is animal welfare?...the problem is that you cannot ask it to the animals…” Animal welfare organisations: “…a definition for animal welfare should start from the perception and preferences of the animal, otherwise the issue of animal welfare does not really make sense…the ethical point of view should not be neglected…” 66 Chapter 2 : Public conception – quantitative approach Animal scientists: “…animal welfare is a situation in which the animals feel well...a situation in which all primary necessities of life are fulfilled…no hunger or thirst…as much natural behaviour as possible…no pain…” “…animal welfare means the absence of something negative…the welfare is damaged only if the organism is aware of something negative…” In general the different stakeholders reasonably well agreed with the structuring of the aspects into the seven dimensions of Model 4. Some comments were raised in relation to the formulation of the names of some dimensions. Based on these comments three dimensions were renamed. Housing was changed to “Housing and Barn Climate”. The aspects indeed relate, next to housing characteristics in general, also to the barn climate in particular. Feed was changed to “Feed and Water” and natural behaviour was modified to “Ability to Engage in Natural Behaviour”. Suffering and stress was also criticised by some interviewees, who suggested alternatives like “Absence of Suffering and Stress” and “Suffering and Chronic Stress”. However we aimed to avoid suggestive terminology like “absence of”. Next to chronic stress, acute stress can be baleful too (e.g. acute fear, castration of piglets, tail docking, and beak trimming). Therefore, the original formulation was kept. The other dimensions were debated less by stakeholders. Next to the formulation of the dimension names, some comments were raised about the structuring of the dimensions. It was argued that the seven dimensions could not be placed on the same level. It was suggested to distinguish between resource- or environment-based dimensions (housing and barn climate, transport and slaughter, feed and water, human-animal relationship) and animalbased dimensions (ability to engage in natural behaviour, animal suffering and stress, animal health) (Figure 2.3). Animal-based dimensions are directly related to the animal’s welfare and are determined by the resource-based dimensions. 67 Part II Resource-based dimensions Animal-based dimensions Housing & Barn climate Suffering & Stress Transport & Slaughter Animal Health FARM ANIMAL WELFARE Feed & Water Ability to engage in natural behaviour Human-Animal relationship Figure 2.3. Final structure of farm animal welfare conception The stakeholders were also asked to discuss the relevancy of the aspects in the public conception. Within all dimensions, next to aspects that were commonly accepted, opinions differed for some aspects depending on personal background and interpretation of the concept rather than on scientific facts. Interviewees suggested aspects in some cases to be removed from a dimension or suggested to incorporate new aspects in a dimension (see Table 2.10 for an overview). Suggested new aspects were mainly specificities that could be understood by existing variables (e.g. “possibility to separate at delivery” and “display nesting behaviour” resort under “maternal behaviour” or “natural birth”; “design of transport vehicle” falls under “calm and shockproof transport”; “escape possibilities” falls under “shelter”; etc.). Related to the personal interpretation, most comments were raised concerning the dimension “ability to engage in natural behaviour”. Still, only one aspect was debated by all stakeholder interviewees, namely “mechanical noise”. The stakeholders argued that it should be replaced by “sudden, loud and strong noise”. However, this term was included in Bartussek’s (1999) Animal Needs Index, where the importance of a stable climate was stressed, among which is absence of “mechanical noise”. Research further showed that continuous noise influences the acoustic behaviour of fowl (Huber and Fölsch, 1978) and that the continuous noise of the ventilation system negatively impacts on the nursing of piglets (Algers and Jensen, 1991). 68 Chapter 2 : Public conception – quantitative approach Table 2.10. Aspects that were suggested to be removed or added by experts Dimension Housing & Barn Climate Aspect suggested to be removed (number of stakeholders) stocking density (1), comfort (3), group size (1), group housing (1), outdoor access (1), functional areas in the barn (2) Aspects suggested to be added dry, control on harmful components, temperature Ability to Engage in Natural Behaviour day light (4), exploratory behaviour (1), natural behaviour (4), natural birth (3), natural growth rate (4), natural environment (3), having fun (4), sexual behaviour (4), social behaviour (1), play behaviour (1), foraging behaviour (2) stable environment, possibility to separate at delivery, nesting behaviour, control over own life, relational autonomy Feed & Water variation in feed (3) shape of the feed Transport & Slaughter number of transports (4), thirst/hunger during transport (1), duration of transport (1), transport of living animals (3), stunning (1) ability to rest during transport, way of stunning, design transport vehicle, age at slaughter, market in function of slaughter, comfort, protection, copying natural environment during transport Animal Health mortality (1) Suffering & Stress fear (1), frustration (3), stress (1) killing, separation of clinical ill animals, euthanasia of terminal clinical ill animals pain through disease, escaping possibilities, behavioural abnormalities, stunning before killing, pain by housing and environment Human-Animal Relation frequency of visual inspection (2), farmeranimal bond (2), skilled animal handlers (1) calm handling of animals 4. Discussion and conclusion Anticipating on the increasing relevancy of animal welfare as a concept on itself and as a part of the broader sustainability concept (Boogaard et al., 2008) in terms of public concern (European Commission, 2005) as well as in relation to consumer food purchasing behaviour (Grunert, 2006), this study was conducted with the aim to develop and validate a public or citizen-driven conception of farm animal welfare. Understandings of the public conception are necessary to actively involve end users in the ongoing debate with the aim to optimise the balance of farm animal welfare and economic output (Van Tichelen, 2009), to effectively communicate to the unaware society (Fisher, 2009) and to determine the need for new animal welfare legislation (Caporale et al., 2005; Garnier et al., 2003). A proper conception of farm animal welfare from the public point of view can be used as a structure and tool to inform the public in a transparent and understandable manner about efforts, actions and policies undertaken to improve farm animals’ welfare (MacArthur Clark, 2008). Similar to animal science based conceptions, this study has resulted in a complex and multidimensional conception of farm animal welfare (Fraser, 1995; Mason and Mendl, 1993). Seven dimensions of farm animal welfare were distinguished that could be grouped in three animal-based 69 Part II dimensions and four environment or resource-based dimensions. The three animal based dimensions – termed “Suffering and Stress”, “Animal Health” and “Ability to Engage in Natural Behaviour” – are affiliated with the three common perspectives on animal welfare described in literature (“mind”, “body”, and “nature”, respectively) (Hewson, 2003b), though are interpreted and given meaning by citizens against a background of lower familiarity, lower practical knowledge and lower awareness about animal production (Buller, 2009). In addition, four antecedent environment- or resource-based dimensions have been distinguished. These are termed “Housing and Barn Climate”, “Transport and Slaughter”, “Feed and Water” and “Human-Animal Relationship”. For each of these dimensions and for many of the aspects within the dimensions, numerous studies are available that acknowledge the significant impact each of them has on the welfare of the farm animals. These dimensions have also been picked up and translated in several legislations that abandon practices proven to compromise good animal welfare. These aspects and dimensions typically correspond with the resource-based welfare indicators or parameters that are currently used in animal welfare assurance schemes. This approach however is debated and different sources plead for bringing both animal-based and resource-based indicators into account (Butterworth, 2009; Fraser, 2004; Hewson, 2003a; Van Tichelen, 2009; Webster, 2005b). Animal-based indicators are a direct reflection of the animal’s welfare. Although Welfare Quality® gave preference to animal-based indicators of welfare, the Welfare Quality® list of 12 criteria and 4 principles contains both animal- and resource-based items. Our conception in two levels, to the contrary, seems a more logic and correct structure. Further comparison between this study’s conception with the Welfare Quality® conception, as an alternative conception based on citizen consultation, shows that they match reasonably well and differ mainly in a nuanced structure and categorisation. Taking into account only the terminology of the four dimensions of the Welfare Quality® conception (see Table 1.4), they could be related to four of our dimensions (“Feed and Water”, “Housing and Barn Climate”, “Animal Health” and “Ability to Engage in Natural Behaviour”) and thus mixing animal- and environment-based dimensions. Having a closer look into the 12 criteria constituting the four WQ dimensions, two additional dimensions of this study’s conception could be retrieved. “Good human-animal relationship” is a criterion under the principle “Appropriate Behaviour” and matches our dimension “Human-animal relationship”. Furthermore, “Absence of pain induced by management procedures” and “Absence of general fear”, criteria that resort under “Good health” and “Appropriate Behaviour” in WQ, are in this study’s conception included in the dimension “Suffering and Stress”. Finally, the absence of “Transport and Slaughter” in the WQ conception can also be explained, since the WQ conception is designed to score the animal’s welfare either on farm or during transport and slaughter. Our conception’s first goal was to map and validate the way in which the public perceives and conceptualises farm animal welfare. In general, the public conception as obtained from this study has much communalities with existing conceptions described in literature, though is couched in a format that is more transparent and understandable for the public. Yet, it is supported by different 70 Chapter 2 : Public conception – quantitative approach stakeholders and interest groups active in the meat production chain and thus answers to the need for a conception that unites varied understandings (Fisher, 2009). These insights make it an appropriate construct and tool to be applied in the context of communication and providing information about animal welfare and animal production to the wider public, and to stimulate dialogue and debate on how to improve farm animal welfare from a market-based point of departure. Correlations between the different dimensions were significant and correspond with the previously reported interrelationships between the dimensions (Appleby, 1999b; Fraser et al., 1997; Hewson, 2003b). Very high correlations were found between “Housing and Barn Climate” on the one hand and “Ability to Engage in Natural Behaviour” and “Transport and Slaughter” on the other hand. The high dependence of the ability of farm animals to engage in natural behaviour on the housing as projected by citizen opinions in this study is obvious. Nevertheless we favoured a split in order to make abstraction between animal- and environment-based dimensions. This decision was supported by a better model fit. The high correlation with transport and slaughter is also reflected in the WQ conception, where the identified principles apply both on farm and during transport and slaughter. However, citizen focus group discussions have shown the crucial importance of this final stage in the farm animal’s life from the citizen perspective (Chapter 1), which justifies treating transport and slaughter as a separate dimension. Discussions in the focus groups and debate in literature are often related to whether or not humans have the right to end an animal’s life, to the way the animals are transported, and to whether or not and how to stun animals before slaughter. 71 Part III Public attitudes, perceptions and concerns towards farm animal welfare & attitude modifiers Chapter 3 : Concept of farm animal welfare – citizens versus consumers Chapter 3 Concept of farm animal welfare : citizens versus farmers Adapted from Vanhonacker, F., W. Verbeke, E. Van Poucke, and F. A. M. Tuyttens. 2008. Do citizens and farmers interpret the concept of farm animal welfare differently? Livestock Science 116: 126136. Abstract Animal welfare is of increasing importance in animal production and consumption debates. However, it is hypothesised that attitudes towards farm animal welfare differ between citizens and farmers, which can impose problems to an effective implementation of welfare related policies. Since the search for, and the realisation of improvements towards farm animal welfare is strongly driven by citizen expectations, it is of utmost importance to better understand the perceptual discordance that exists between citizens and farmers. A quantitative study was done in Flanders, Belgium during 2006, including citizens as well as farmers, to obtain a detailed insight in the way the multi-dimensional concept of farm animal welfare is valued. This allowed discriminating between issues of agreement and disagreement. In general, a similar interpretation of farm animal welfare in terms of animal welfare related aspects’ ranking was found. Differences were mainly related to aspects dealing with the ability to engage in natural behaviour on the one hand and with production process-related aspects on the other hand. Citizens evaluate the current state of animal welfare as rather problematic, while farmers report a more satisfactory evaluation of the present condition of farm animal welfare. Especially differing opinions regarding the ability to engage in natural behaviour, together with aspects related to pain, stress and the availability of space seem to contribute to the discordance between farmers and citizens in terms of evaluative beliefs. 75 Part III 1. Introduction There is a general tendency in European society for an increasing influence of post-materialistic values, in particular on product attribute evaluation and food choice decision-making. For animal production, this entails that ecological and socio-cultural aspects occupy an increasingly important role in the buying decision process of food and animal products (Boogaard et al., 2006). One of the issues of growing citizen concern is farm animal welfare. Even though there is still a tendency to buy the cheapest meat, animal welfare is considered a priority by an increasing number of European citizens (European Commission, 2007). Hence, the buying behaviour does not simply reflect the attitude towards animal welfare (Te Velde et al., 2002). In addition, an effective incorporation of farm animal welfare as a product attribute in decision-making is hampered, partly because of diverging perceptions and the lack of consensus among the different stakeholders with respect to farm animal welfare (Lassen et al., 2006; Marie, 2006). This chapter will focus on the perceptions and beliefs held by stakeholders at both ends of the chain, i.e. the farmer and the individual in his role as citizen. Perceptions are constructed according to their frames of reference, which are in turn influenced by convictions (“opinions about the way things are”), values (“opinions about the way things should be”), norms (“translations of these values into rules of conduct”), knowledge (“constructed from experiences, facts, stories, and impressions”) and interests (“economic, social and moral interests”) (Te Velde et al., 2002). Citizens have a rather negative overall perception of the current state of farm animal welfare and are not satisfied with the amount and content of the information they receive in relation to the welfare conditions and protection of farm animals. Farmers to the contrary have a much more positive perception of the current state of farm animal welfare. This has been demonstrated by Te Velde et al. (2002), who explained the opposite perceptions between citizens and farmers based on a different interpretation of the components by which perception is determined. Regarding values, Te Velde et al. (2002) indicated that both farmers and citizens associate animal welfare with issues such as physical health, an adequate amount of food and drinking water and sufficient heating and protection. Citizens, however, tend to include two additional values: freedom to move and freedom to fulfil natural desires (Lassen et al., 2006; Marie, 2006; Te Velde et al., 2002). Furthermore, farmers’ norms are clearly related to factors that are important for optimising production such as fast and efficient growth (Dockes and Kling-Eveillard, 2006; Lassen et al., 2006; Te Velde et al., 2002). Citizens on the other hand remain vague in their formulation of norms and just want that the translation of values into rules of conduct is arranged (Harper and Henson, 2001). In addition, both stakeholder groups have different interests. Whereas individuals in their role as consumers would be mainly interested in healthy, tasty and cheap food that is easily available, individuals in their role as citizen also assign much importance to acceptable production methods, thus including moral and ethical interests (Bennett et al., 2002; Brom, 2000). The farmers’ interests 76 Chapter 3 : Concept of farm animal welfare – citizens versus consumers are more economically driven and centred on financial concerns and the need to make a living. Other interests of farmers include the wish to supply high quality products, to have a satisfying job and to establish a more positive image of agricultural and animal production. Also differences in knowledge contribute to the discordant perception. Farmers rely on their daily experience and practical knowledge of how their animals will optimally produce, which is something they positively relate to animal welfare. However, they are rather indifferent towards possible other aspects of animal welfare, for instance the ability to engage in natural behaviour (Marie, 2006; Milne et al., 2007; Morgan-Davies et al., 2006; Te Velde et al., 2002). Due to an increasing dissociation of citizens from farming practices as a consequence of agricultural intensification and growing urbanisation, citizens’ knowledge of the circumstances in which meat livestock is produced is much more limited (Frewer et al., 2005; Harper and Henson, 2001; Harper and Makatouni, 2002; Maria, 2006; Verbeke, 2005). Moreover, (negative) media coverage has a strong influence on citizens’ knowledge and perceptions (Swinnen et al., 2005; Verbeke and Ward, 2001). This limited knowledge could also be at the base of the vagueness of the norms formulated by citizens. Finally, the values and norms of citizens seem to be based on the same set of convictions as those of farmers: animals are meant to serve humans; meat is a substantial part of the human diet; keeping animals and killing them for meat is legitimate; and farmers are there to provide food for the rest of the population (Te Velde et al., 2002). In the present chapter, a cross-sectional survey was conducted among Flemish citizens and farmers in order to obtain a more detailed insight in the perceptual difference concerning the concept of farm animal welfare between both stakeholder groups. Whereas perceptions can be analysed based on a frame of reference including values, norms, convictions, interests and knowledge, we will focus in the analysis of this study on the quantification of the values only, i.e. opinions about the ways things should be. Differences in values will then be discussed based on the other components of the frame of reference, as they are described in literature. The values will be assessed based on two measures. For an extended list of farm animal welfare aspects (see Chapter 2) respondents were asked to indicate firstly how important they feel the aspect is for the welfare of farm animals in general, and secondly to what extent this aspect is believed to be problematic with respect to farm animal welfare in current animal production practices in Flanders. Responses to both questions are combined to identify the issues of highest and lowest potential conflict between citizens and farmers. 2. Material and methods 2.1. Study design and subjects Cross-sectional survey data were collected through questionnaires in Flanders during April 2006, after validation via a pre-test. Methodological details concerning the citizen sample are described in the previous chapter (Chapter 2). Farmers were selected through purposive sampling. Respondents 77 Part III were defined as a farmer if they or their parents had a farm. The choice for this definition is justified by a pre-analysis that placed respondents who indicated to have themselves or their parents a farm, but who are not a farmer themselves, very close to respondents indicating to be a farmer as compared to the difference with the citizen sample. As a result, a more elaborate sample is obtained, allowing analyses which are quantitatively more reliable. Since this population group is hard to define based on socio-demographic characteristics, the representativeness of the farmer sample can not be evaluated. Following Te Velde et al. (2002), who concluded that farmers’ perceptions of animal welfare are liable to a minor variation (since they are grounded on a collective tradition with shared convictions, values, norms and interests and on knowledge that is derived from comparable rearing, schooling and daily experience on the farm), the impact of this definition of the farmer group on the reliability of the results is limited. Questionnaires were distributed electronically through the use of websites frequently visited by farmers (www.vilt.be; www.groenekring.be), yielding a sample of 204 farmers. 2.2. Survey The key questions were related to the perception of farm animal welfare in terms of values. The values were pictured using an exhaustive list of 72 aspects concerning farm animal welfare. To facilitate the discussion, aspects were assigned to one of the seven key dimensions of the interpretation of farm animal welfare as described by Van Poucke et al. (2006) (Housing & Climate; Transport & Slaughter; Feed & Water; Human-Animal Relationship; Animal Suffering and Stress; Animal Health; Ability to Engage in Natural Behaviour) (Table 3.1)8. The decision to assign aspects to a particular dimension was taken based on preliminary insights from focus group discussion, as well as expert and scientist opinions involved in the field of animal welfare and behaviour. For each aspect, both citizens and farmers were asked to indicate its perceived importance for obtaining an acceptable level of farm animal welfare (“perceived importance”) and whether they believe the aspect poses a potential problem with respect to animal welfare in present Flemish animal farming (“evaluative belief”). The latter has been used frequently as a component of attitude, based on the Theory of Reasoned Action (Fishbein, 1967) and the Theory of Planned Behaviour (Ajzen, 1991), while the former can be considered as a weight for the relevance of each aspect in the interpretation of farm animal welfare. Moreover, combined these measures are a good representation of values within the frame of reference of perception. As both scores were probed for in adjacent positions (physically in the questionnaire) and scored on the same type of scale (fivepoint Likert scale), they could be combined to assess a “gap” corresponding with the difference between the perceived importance score and the evaluative belief score. As a result, a big positive gap corresponds with a high perceived importance and a low evaluative belief, i.e. a strong perceived potential improvement. Aspects with a small or negative gap represent either aspects perceived as important for overall farm animal welfare but with a low potential of contributing to 8 This paper has been written before the conception using structural equation modelling was fully developed, and explains why the categorisation differs slightly from the results discussed in the previous chapter. 78 Chapter 3 : Concept of farm animal welfare – citizens versus consumers improving farm animal welfare in Flemish animal production; or, aspects perceived as not important for welfare even if evaluated negatively; or, aspects scoring rather neutral on both questions. The latter aspects often correspond to issues that are not well understood or ambiguous. Finally, as it is the aim of the chapter to gain insight in the perceptual discordance between farmers and citizens, comparison is made for each individual aspect between the gap at citizen level and the perceived importance score at farmer level. This comparison is most suitable to reveal the aspects subject to the highest tension, since it allows to uncover the aspects which are associated with important problems at citizen level (big gap), but are only assigned limited importance by the farmer. 2.3. Statistical analyses Data were analysed using SPSS 12.0. Mean scores and standard deviations are presented in table format. Bivariate analyses through correlation and comparison of mean scores, i.e. independent samples t-tests were used to detect differences in perceived importance and evaluative belief between farmers and citizens. Linear regression analyses were performed to assess relationships between farmers’ and citizens’ responses. 3. Results 3.1. Perceived importance of animal welfare aspects For the citizen sample, mean perceived importance scores for the 72 aspects ranged from 3.14 (distraction material) to 4.56 (availability of water) (Table 3.1). At the top of the list, mainly aspects emerge relating to the dimensions Feed & Water, Human-Animal Relationship and Animal Health, together with some specific aspects from the dimensions Housing & Climate (stocking density, available space and air quality), Animal Suffering & Stress (pain through human intervention and stress), and Transport & Slaughter (slaughter without pain or stress). Among farmers, perceived importance scores ranged from 2.40 (size of livestock herd on farm) to 4.35 (availability of water) (Table 3.1). Farmers clearly considered aspects relating to the dimensions Feed & Water, Animal Health and Human-Animal Relationship as the most important in obtaining an acceptable level of farm animal welfare. Results from the citizen and farmer samples are compared in Figure 3.1. A positive linear relationship is found between perceived importance scores of farmers (Y) versus citizens (X) (Y=1.09x-0.84; R²=0.55). The negative intercept indicates that, in general, citizens gave higher perceived importance scores than farmers. This holds for most aspects, but particularly for aspects relating to the dimension Ability to Engage in Natural Behaviour. In addition, singular aspects from within Housing and Climate (stocking density (1), outdoor access (5), size of livestock herd on farm (15)) and Transport and Slaughter (thirst (2) and hunger (13) during transport, stunning (5)) also yield (relatively) large differences. Aspects where farmers’ scores are higher than citizens’ scores are genetic selection (14), taste of feed (8), preventive medication (6), frequency of visual inspection (4) 79 Part III and attention of farmer fro animals (3). Other aspects had rather similar mean importance scores, and hence are close to the bisector in the graph. These are aspects within Feed and Water (availability of water (1) and feed (3), freshness of feed (2), balanced feed (4)), Human-Animal Relationship, Animal Health (with the exception of lifespan animal (5)), plus some singular aspects within Housing and Climate (barn temperature (8), flooring type (11)) and Transport and Slaughter (lairage time (12)). The correspondence between citizens’ and farmers’ opinions regarding the first three dimensions (Feed and Water, Human-Animal Relationship, Animal Health) originates from the particular aspects being perceived as top priorities for farm animal welfare both at citizen and farmer level, while the similarity regarding the latter aspects is mainly due to an average perceived importance at farmer level and a rather moderate score at citizen level. The positive slope close to unity reflects a systematic and consistent pattern, indicating a quite similar view on most aspects’ relative importance in obtaining an acceptable level of farm animal welfare. 4.5 Perceived importance farmer 3 4 4 4 8 6 43 12 46 3 8 11 4 14 6 7 9 10 5 9 7 2 7 15 7 5 3 21 13 5 10 11 5 14 9 6 5 12 11 12 6 5 6 13 10 8 14 13 9 4 5 8 3.5 3 7 2.5 2 32 2 3 4 2 2 1 3 1 1 1 1 1 16 15 2 3 3.2 3.4 3.6 3.8 4 4.2 4.4 4.6 4.8 5 Perceived importance citizen Figure 3.1. Mean scores of perceived importance of 72 aspects in obtaining an acceptable level of farm animal welfare according to farmers (Y-axis) and citizens (X-axis). The line in the figure is the bisector, indicating equal mean scores. Colours correspond to aspects classified in the same dimensions. For the aspects corresponding with the numbers, please see Table 3.1. Mind that the scales of the X and Y axes are not equal 80 Chapter 3 : Concept of farm animal welfare – citizens versus consumers Table 3.1. Perceived importance* and evaluative belief** scores on 72 aspects related to animal welfare by citizens and farmers. Aspects are classified per dimension and ranked based on citizens’ mean perceived importance score (5-point Likert scales) PERCEIVED IMPORTANCE CITIZEN FARMER Mean (SD) Mean (SD) I. HOUSING & CLIMATE (red) Stocking density (1) 4.28 (0.74) 4.16 (0.73) Air quality (2) Available space (3) 4.16 (0.75) Ability to rest (4) 4.05 (0.75) 4.02 (0.96) Outdoor access (5) 3.99 (0.93) Comfort (6) Shelter (7) 3.94 (0.90) 3.84 (0.82) Barn temperature (8) Light regime (9) 3.74 (0.88) Group housing (10) 3.71 (0.85) Flooring type (11) 3.67 (0.89) 3.63 (0.94) Group size (12) Functional areas in barn (13) 3.53 (0.97) Static groups (14) 3.43 (0.90) Size of livestock herd on farm (15) 3.40 (1.11) Technical noise (16) 3.22 (0.97) II. ABILITY TO ENGAGE IN NATURAL BEHAVIOUR (green) Daylight (1) 4.08 (0.85) 4.04 (0.90) Natural growth rate (2) Natural behaviour (3) 4.00 (0.88) Body care (4) 3.89 (0.97) Natural environment (5) 3.88 (0.96) Explorative behaviour (6) 3.86 (0.92) Social behaviour (7) 3.84 (0.90) Natural birth (8) 3.81 (1.00) Maternal behaviour (9) 3.80 (0.96) 3.72 (0.94) Sexual behaviour (10) Having fun (11) 3.67 (1.04) Foraging behaviour (12) 3.64 (1.00) 3.58 (1.03) Play behaviour (13) Genetic selection (14) 3.30 (1.13) III. ANIMAL HEALTH (blue) Disease (1) 4.37 (0.71) Hygiene in the barn (2) 4.29 (0.68) Curative medication (3) 4.23 (0.66) 4.04 (0.91) Mortality (4) Life span (5) 3.64 (1.02) Preventive medication (6) 3.36 (1.08) EVALUATIVE BELIEF CITIZEN FARMER Mean (SD) Mean (SD) PERCEIVED IMPORTANCE CITIZEN FARMER Mean (SD) Mean (SD) IV. TRANSPORT & SLAUGHTER (brown) Slaughter without pain or stress (1) 4.30 (0.86) 3.75 (1.05) Thirst during transport (2) 4.10 (0.89) 3.33 (1.05) Handling of animals (3) 4.07 (0.84) 3.62 (0.95) Duration of transport (4) 4.03 (0.86) 3.46 (1.02) Stunning (5) 4.01 (1.08) 3.05 (1.36) Procedure of (un)loading (6) 3.97 (0.80) 3.54 (0.92) Space during transport (7) 3.97 (0.86) 3.30 (0.93) Transport of living animals (8) 3.93 (0.89) 3.48 (0.90) Shockproof and calm transport (9) 3.90 (0.86) 3.32 (0.90) Climate during transport (10) 3.81 (0.88) 3.33 (0.95) Number of transports (11) 3.79 (0.93) 3.07 (1.11) Lairage time (12) 3.72 (0.97) 3.53 (1.06) Hunger during transport (13) 3.66 (0.99) 2.88 (1.05) Mixture of groups during t/s (14) 3.38 (1.09) 2.89 (1.16) Design of slaughterhouse (15) 3.61 (0.96) 3.27 (1.01) V. FEED & WATER (pink) 4.56 (0.62) 4.35 (0.81) Availability of water (1) 3.22 (1.12) 2.54 (1.01) 3.47 (1.05) Freshness of feed (2) 4.24 (0.73) 4.08 (0.79) 3.19 (1.14) 2.29 (1.09) 3.54 (1.07) Availability of feed (3) 4.23 (0.68) 4.02 (0.86) 3.18 (1.04) 2.45 (1.05) 3.50 (0.98) Balanced feed (4) 4.17 (0.71) 3.99 (0.86) 2.80 (1.11) 2.43 (1.06) 3.39 (1.04) Variation in feed (5) 2.92 (1.12) 2.38 (1.08) 3.48 (1.07) Natural feed (6) 3.88 (1.01) 3.04 (1.12) 2.93 (1.09) 2.43 (1.05) 3.46 (1.07) Feed on fixed moments (7) 3.69 (0.89) 3.37 (1.03) 3.21 (0.95) 2.62 (1.02) 3.62 (0.90) Taste of feed (8) 3.59 (0.90) 3.61 (0.98) 2.91 (1.18) 2.66 (1.08) 3.47 (1.10) Growth hormones (9) 3.26 (1.55) 2.83 (1.55) 3.01 (1.07) 2.66 (1.03) 3.54 (0.96) VI. ANIMAL SUFFERING & STRESS (purple) 2.91 (1.12) 2.72 (0.97) 3.58 (0.94) Pain by human intervention (1) 4.23 (0.87) 3.59 (1.07) 2.94 (1.14) 2.66 (1.04) 3.61 (1.04) Stress (2) 4.15 (0.86) 3.85 (0.90) 2.94 (1.06) 2.66 (1.03) 3.62 (0.95) Fear (3) 4.04 (0.85) 3.57 (0.99) 2.91 (0.99) 2.66 (1.06) 3.62 (0.92) Pain by conspecifics (4) 3.96 (0.79) 3.54 (0.91) 3.38 (1.20) 2.92 (1.04) 3.74 (1.01) Frustration (5) 3.66 (0.96) 3.10 (1.03) Boredom (6) 3.50 (1.02) 2.91 (1.03) 4.13 (0.79) 2.82 (1.03) 3.56 (1.03) Distraction material (7) 3.14 (1.06) 2.64 (1.07) 4.17 (0.79) 2.91 (0.98) 3.74 (1.00) VII. HUMAN-ANIMAL RELATIONSHIP (black) 4.10 (0.83) 2.98 (1.00) 3.66 (1.09) Respect for animals (1) 4.39 (0.71) 4.25 (0.83) 3.83 (0.99) 2.86 (0.94) 3.56 (0.94) Skilled animal handlers (2) 4.29 (0.70) 4.15 (0.80) 2.99 (1.26) 2.83 (1.05) 3.73 (0.99) Attention for animals (3) 4.25 (0.74) 4.31 (0.82) 3.43 (1.09) 2.85 (1.04) 3.52 (1.05) Frequency of visual inspection (4) 3.68 (0.86) 3.75 (0.95) Farmer-animal bond (5) 3.59 (1.05) 3.34 (1.21) * Response to: “How important is this aspect, according to your personal opinion, in obtaining an acceptable level of farm animal welfare?” ** Response to: “What is your personal opinion about this aspect, i.e. do you think that this aspect poses a potential problem for the animal welfare in present Flemish animal farming?” 3.53 (0.96) 3.81 (0.82) 3.69 (0.83) 3.62 (0.90) 2.79 (1.26) 3.62 (0.89) 3.30 (1.15) 3.75 (0.87) 3.34 (1.02) 3.11 (1.04) 3.49 (1.02) 2.96 (1.16) 3.12 (1.11) 3.03 (1.07) 2.40 (1.30) 2.56 (0.97) 2.11 (1.03) 2.52 (0.94) 2.29 (1.07) 2.77 (1.00) 2.46 (1.08) 2.63 (1.01) 2.82 (0.99) 3.10 (0.93) 2.77 (0.97) 2.77 (0.99) 2.91 (0.98) 2.72 (1.01) 2.67 (1.01) 3.00 (0.88) 2.80 (1.09) 2.91 (0.96) 3.44 (1.10) 3.45 (0.98) 3.52 (1.05) 3.70 (0.90) 3.44 (1.14) 3.70 (0.92) 3.40 (1.01) 3.74 (0.92) 3.69 (0.90) 3.62 (0.94) 3.53 (0.98) 3.71 (1.00) 3.65 (1.01) 3.71 (0.95) 3.92 (1.07) 3.76 (0.92) EVALUATIVE BELIEF CITIZEN FARMER Mean (SD) Mean (SD) 2.36 (1.09) 2.37 (0.96) 2.28 (0.99) 2.46 (1.03) 2.66 (1.10) 2.51 (0.99) 2.24 (1.01) 2.47 (1.03) 2.44 (1.00) 2.58 (1.02) 2.71 (1.04) 2.68 (0.98) 2.67 (0.94) 2.77 (1.00) 2.70 (0.98) 3.47 (1.11) 3.27 (1.08) 3.35 (1.03) 3.46 (1.05) 3.55 (1.17) 3.51 (0.99) 3.38 (1.01) 3.54 (0.97) 3.40 (1.02) 3.50 (0.96) 3.61 (0.95) 3.47 (1.04) 3.59 (0.96) 3.29 (1.05) 3.41 (1.01) 3.27 (1.02) 3.03 (1.01) 3.28 (0.97) 2.92 (1.06) 3.95 (0.95) 3.91 (0.91) 4.02 (0.87) 3.93 (0.92) 2.61 (1.07) 3.34 (0.94) 3.11 (0.93) 2.18 (1.16) 3.68 (0.99) 3.90 (0.85) 3.99 (0.80) 3.29 (1.22) 2.43 (1.04) 2.22 (0.99) 2.55 (0.97) 2.69 (0.94) 2.57 (1.02) 2.47 (1.05) 2.82 (1.08) 3.56 (0.96) 3.31 (1.00) 3.47 (0.93) 3.46 (0.93) 3.39 (1.04) 3.40 (0.96) 3.60 (0.97) 2.92 (1.09) 2.79 (1.03) 3.08 (1.02) 3.07 (0.92) 2.97 (1.01) 4.00 (0.96) 3.55 (1.09) 3.98 (0.96) 3.91 (0.91) 3.84 (0.94) 81 Part III 3.2. Evaluative beliefs of the aspects’ practice in Flemish animal production Among citizens, evaluative belief scores for the 72 aspects ranged from 2.11 (stocking density) to 3.34 (feed on fixed moments), with only eight aspects receiving a mean score above the mid-point score of 3 (Table 3.1). The aspects evaluated most positively mainly relate to the dimensions Feed & Water (with the exception of growth hormones) and Human-Animal Relationship. The lowest evaluative belief scores were found for aspects related to the availability of space (stocking density, available space, space during transport and outdoor access), to the ability to engage in natural behaviour and also to a high number of the aspects within Transport & Slaughter and Animal Suffering & Stress. For farmers, evaluative beliefs are scored above the mid-point score for all aspects and ranged from 3.27 (thirst during transport) to 4.02 (availability of feed) (Table 3.1). Aspects related to Feed & Water (with the exception of growth hormones) and Human-Animal Relationship are perceived as the least problematic by farmers. Evaluative beliefs are compared between both stakeholder groups in Figure 3.2. A positive linear relationship is found between evaluative belief scores of farmers (Y) and citizens (X) (Y= 0.56x + 2.09; R²=0.63). The positive intercept indicates that the current level of animal welfare is generally evaluated much more positively by farmers as compared to citizens. The positive slope on the other hand involves that, in general, a comparable ranking of the aspects is found for both samples, with a broader range though in the scores attributed by the citizens (slope < 1). 4.2 4 Evaluative belief farmer 15 1 4 3.6 2 3 1 3.4 7 3 2 9 7 5 5 3.8 3 1 38 2 4 14 14 216 6 12 49 5 6 13 3 12 11 10 7 13 11 7 13 10 1 8 2 1 4 11 95 6 3 6 1 10 812 5 1 6 45 2 3 4 15 7 5 49 6 8 14 2 3.2 3 2 2.2 2.4 2.6 2.8 3 3.2 3.4 Evaluative belief citizen Figure 3.2. Mean scores of evaluative belief of 72 aspects according to both farmers (Y-axis) and citizens (X-axis). The line in the figure is the bisector, indicating equal scores. Colours correspond to aspects classified in the same dimensions. For the aspects corresponding with the numbers, please see Table 3.1. Mind that the scales of the X and Y axes are not equal 82 Chapter 3 : Concept of farm animal welfare – citizens versus consumers 3.3. Discordance between citizens and farmers Aspects with the highest potential discordance are characterised by a large positive gap at citizen level (i.e. the perceived importance score largely exceeds the evaluative belief score) combined with a relatively low perceived importance at farmer level. This combination is depicted for each of the dimensions in Figure 3.3. The aspects with the lowest discordance are situated in the lower right quadrant of the figures, showing a small gap at citizen level and a high perceived importance at farmer level. The upper right quadrant corresponds with aspects subject to a strong positive gap at citizen level but at the same time to a high perceived importance at farmer level. The lower left quadrant is composed of aspects causing no discordance at present. Finally, the aspects situated in the upper left quadrant contribute in the highest degree to the perceptual discordance between citizens and farmers. Hence, the degree of discordance diminishes from the upper left corner to the lower right corner of the graph. With regard to Housing & Climate, outdoor access (5) emerged as the aspect most susceptible to conflict, whereas the wide positive gap corresponding with stocking density (1), amount of space (3) and air quality (2) was at the same time subject to a strong perceived potential improvement in farmers’ opinion. All other aspects show up as less susceptible to discordance and conflict. The aspects within Engagement in Natural Behaviour were clustered around the centre of the map, with the exception of genetic selection (14), which corresponds to a relatively high farmer perceived importance and a small gap at citizen level. The least discordance was perceived for the dimensions Animal Health and Human-Animal Relationship, where the aspects were subject to a moderate gap at citizen level and a high perceived importance at farmer level. Also Feed & Water are only to a relatively small degree susceptible to a perceptual conflict, with only growth hormones (9) located more to the left part of the graph. Natural feed (6) on the other hand is located close to the centre of the graph, similar to the aspects within Ability to Engage in Natural Behaviour. Regarding Transport & Slaughter, big positive gaps are found for almost all aspects, with a polarisation of the aspects in the upper right quadrant of the graph and slaughter without pain and stress (1) and space (7), thirst (2) and handling of animals during transport (3) as the most susceptible to conflict. Finally, with regard to Animal Suffering & Stress related aspects, pain caused by human intervention (1) and stress (2) were located in the upper right quadrant, while the others aspects were less sensitive to conflict. 83 Part III 2. 5 2. 5 1 2 2 3 2 5 1. 5 1 7 9 12 10 13 15 16 0. 5 1 11 2 31 5 4 6 8 9 10 11 1312 1. 5 6 4 8 7 0. 5 14 14 0 0 1. 5 2 2. 5 3 3. 5 4 1.5 4. 5 2 2. 5 3 3. 5 4 2. 5 2. 5 2 2 1 3 72 1 2 3 1. 5 4 1. 5 1 5 6 4 9 86 10 12 15 14 0. 5 0 5 11 13 1 0. 5 4. 5 A b il i t y t o Eng ag e i n N at ur al B ehavi o r Ho usi ng & C l i mat e 0 1. 5 2 2. 5 3 3. 5 4 4. 5 1. 5 2 2. 5 A ni mal Heal t h 3 3. 5 2. 5 2.5 2 2 1. 5 1.5 1 6 9 1 4 4. 5 4 4. 5 T r ansp o rt & Sl aug ht er 0. 5 5 6 1 0.5 8 7 3 4 1 4 2 3 5 2 7 0 0 1. 5 2 2. 5 3 3. 5 4 1. 5 4. 5 2 2. 5 3 3.5 A ni mal Suf f er ing & St r ess F eed & W at er 2. 5 2 2 1 1. 5 3 1 5 0. 5 4 0 1. 5 2 2.5 3 3.5 4 4. 5 Human- A ni mal R el at io nshi p Figure 3.3. Discordance in perception between citizen and farmer, depicted for each dimension separately. The X-axis corresponds with the perceived importance at farmer level; the Y-axis corresponds with the gap (perceived importance – evaluative belief) at citizen level. For the aspects corresponding with the numbers, see Table 3.1. 4. Discussion In this chapter, the perception of the concept of farm animal welfare by both citizens and farmers is quantified and compared in terms of values, i.e. opinions about the way things should be. These values are based on the respondents’ perceived importance and evaluative beliefs of 72 aspects that are considered relevant for the welfare of farm animals. These results and gathered insights respond to a contemporary issue under concern and are in compliance with the goals of The Community Action Plan on the Protection and Welfare of Animals 2006-2010 (European 84 Chapter 3 : Concept of farm animal welfare – citizens versus consumers Commission, 2006a), where citizens’ perception of animal welfare and scientific research have been identified as key issues for the improvement of farm animal welfare. The results will be discussed and explained within the scope of the components of the frame of reference of perception, namely values, norms, knowledge, interests and convictions. In general, a rather analogous interpretation of the concept of farm animal welfare by citizens and farmers was found, given the roughly similar ranking of the perceived importance scores of these aspects. However, citizens generally attributed higher absolute importance scores to the aspects as compared to farmers. Possibly, acquiescence bias is at the base of this. This type of response bias corresponds with a tendency to report a positive connotation and especially occurs in the case of opinion or attitudinal questions, rather than factual or knowledge-based questions (Watson, 1992). It could be assumed that citizens mainly answered to this particular question from a perceptual perspective, while the farmer grounded his answer possibly more on expertise and knowledge. However, one should be careful to take only knowledge into account as an explanatory factor for the observed differences, as has been criticised by Hansen et al. (2003) and referred to as the knowledge deficit model. Probably, also the need for a clean conscience, expressed as a citizen’s interest, contributes to the systematic higher perceived importance scores among the citizen sample. Aspects with low differences in mean value between citizens and farmers pertain either to the aspects perceived as the most important by both groups or to aspects which require a practical knowledge base of farm animal rearing conditions. The former may be due to a ceiling effect, since mean scores are close to the upper value of the scale. Examples are the majority of the aspects within the dimensions of Feed & Water, Animal Health and Human-Animal Relationship. The latter category confirms the more fine-grained perception of the farmer, which is in accordance with the findings from Beekman et al. (2002), Te Velde et al. (2002) and Verbeke (2002). The exception on the overall similarity in interpretation is found in aspects related to the ability to engage in natural behaviour. These aspects received a very high perceived importance score by citizens and yielded a relatively strong difference with the respective farmers’ importance scores, even though also at farmer level a relatively high score was found. The difference regarding the ability to engage in natural behaviour is in correspondence with the two additional values that citizens include within their frame of reference as described in previous studies: freedom to move and freedom to fulfil natural desires (Lassen et al., 2006; Marie, 2006; Milne et al., 2007; Morgan-Davies et al., 2006; Te Velde et al., 2002). Also, from a farmers’ perspective, animal welfare is judged positive, since a fast growth and a satisfactory food conversion – two issues farmers relate to good animal welfare – are strived for and are usually also realised. Moreover, a switch towards a production system that gives more consideration to the ability to engage in natural behaviour involves huge investments and may jeopardise the economic performance and continuity of their farm. Te Velde et al. (2002) also mentioned that farmers consider this as returning to more traditional ways of farming, with worse working conditions as a consequence. Concerning the evaluative beliefs, several discordant beliefs between citizens and farmers were detected. As a result of the norms expressed by the farmers, no problems were perceived for any of 85 Part III the aspects, since they relate animal welfare primarily to health. Hence, if an animal eats well, it will grow fast, which means that it is healthy and that the welfare is good. A much more negative picture emerged from the citizen sample. Only the evaluation of issues relating to feed and water, together with the relationship between the farmer and his animals were judged positively by the citizens. Aspects related to availability of space, ability to engage in natural behaviour, transport & slaughter and suffering & stress were evaluated most negatively at citizen level. Summing up on the discordance between farmers’ and citizens’ perception of the concept of farm animal welfare, five categories of aspects covering a similar issue have been distinguished based on a qualitative interpretation of the findings. The first two categories include aspects that occupy a similar location in the upper right quadrant of the graph, and pertain to aspects dealing with space (category 1) and pain and stress (category 2). These are typically aspects whose perception towards animal welfare has become more negative with an increasing intensification of animal production, at least at the citizens’ level. Farmers do attach importance to these aspects, but see little opportunity for alleviating problems relating to these aspects without serious economic drawbacks, hence causing a conflict between interests and values at the farmers’ level. The third category is composed of aspects related to the ability to engage in natural behaviour, indicating citizen concern and limited farmers’ perceived importance. Consequently, these are issues that face a high susceptibility to debate. The fourth category is composed of the three dimensions Feed & Water, Animal Health and Human-Animal Relationship and does not yield perceptual discordance between citizens and farmers. Finally, the fifth category relates to aspects that are of relatively high importance to farmers, and subject to a small gap at citizen level. Seen the aspects that belong to this category (e.g. flooring type, barn temperature, lairage time, static groups, mixture of groups during transport & slaughter, …) together with citizens’ dissociation from agriculture, this position is most likely resulting from the difficulties citizens perceive in estimating the true impact of these specific aspects on farm animal welfare. It is important to stress that the composition of these categories and their interpretation may change over time. The components of the frame of reference explaining the perception of the concept of farm animal welfare can be assumed to be more stable over time among farmers (Te Velde et al., 2002) and even though a change at farmer level will be harder to realise e.g. through training and extension services, once established, it is likely to be more permanent. The gap at citizen level on the other hand can be expected to be more susceptible to fluctuation, dependent on for example the amount and tonality of news coverage or information provision; on the content of the information (positive versus negative); on a shift in citizen interests; and on an increasing citizen consciousness and knowledge about the issue. It is also important to notice that the results only focus on differences between farmers and citizens as an aggregate group. Also within each group, perceptions are likely to differ due to differences in values, interests, knowledge, norms and convictions. Examples are vegetarian individuals within the citizen group (see Chapter 4) and organic farmers within the farmer group (Van Huik and Bock, 2006). 86 Chapter 3 : Concept of farm animal welfare – citizens versus consumers 5. Conclusion This chapter provides a quantification of the similarities and differences between aspects relating to farm animal welfare as valued by farmers and citizens and has tried to explain the findings based on differences in values, interests, knowledge, norms and convictions, constituting the frame of reference of the perception towards farm animal welfare. The interpretation of the complex and multi-dimensional concept of farm animal welfare has proven to be quite compatible from a citizen and farmer perspective. The main differences are found in the importance attached to animals’ ability to engage in natural behaviour and in aspects which require some basic understanding about production conditions and the way farm animals are reared. Furthermore, citizens evaluated the current state of farm animal welfare rather negative with mean evaluative belief scores for almost all aspects below the mid-point of the scale, while a much more positive image was present among farmers. Discordance between citizen and farmer perception appeared to be the highest for aspects related to natural behaviour, pain, stress and availability of space. The findings from the present study can be considered useful to anticipating societal debates relating to farm animal welfare and can be valorised for diminishing the perceptual gap between livestock producers and citizens. Furthermore, these insights could offer business and market opportunities for stakeholders by means of for instance product differentiation based on prominent aspects in the perception of animal welfare at citizen level, or by paying particular attention to aspects scoring high in terms of citizen concern, both in adapting animal production practices and in developing comprehensive marketing strategies for animal food products. 87 Chapter 4 : WQ® definition – opinion of citizens, farmers and vegetarians Chapter 4 Welfare Quality® definition : opinion of citizens, farmers and vegetarians Adapted from Tuyttens, F. A. M., F. Vanhonacker, E. Van Poucke, and W. Verbeke. 2010. Quantitative verification of the correspondence between the Welfare Quality® operational definition of farm animal welfare and the opinion of Flemish farmers, citizens and vegetarians. Livestock Science 131: 108-114. Abstract This chapter constitutes an extension on the previous chapter. It investigates the impact of placebased variables on the interpretation of the operational definition of farm animal welfare as developed in the Welfare Quality® project, more specific with regard to the importance allocated to the twelve criteria and the four principles within the definition. Additional to place-based variables, the possible impact from other attitude modifiers incorporated in the dissertation’s conceptual framework are discussed. Vegetarians and females gave consistently higher importance scores to all 12 criteria as compared to farmers, citizens and males, respectively, particularly for the criteria belonging to the principle “appropriate behaviour”. Farmers allocated more weight to “good feeding” than did vegetarians or citizens. These differences confirm that farmers view animal welfare more in terms of biological functioning (instead of affective states or natural living) as compared to other citizens and to vegetarians in particular. 89 Part III 1. Introduction The Welfare Quality® project (an EU-funded integrated project in which the development of a standardised system for assessing farm animal welfare and to convey this information into food products has been among the main aims, www.welfarequality.net) has identified 12 areas of concern in relation to farm animal welfare (also termed “criteria”), grouped into four “principles”, by combining analyses of citizen perceptions and attitudes with existing knowledge from animal welfare science (Botreau et al., 2007; Veissier and Evans, 2007) (see Table 1.4). If the public is to have confidence in these animal welfare monitoring protocols, they ought to concord with the public conceptions of animal welfare and fully address public concerns over animal welfare (Fraser, 2006). The operational definition (and accompanying measures) used by scientists ought to reflect the socially constructed meaning of the term farm animal welfare (Fraser, 2003; Stafleu et al., 1996; Tannenbaum, 1991). The objective of this chapter, therefore, was to verify whether importance allocated to the 12 criteria and the relative weights for the four principles differs between citizens, farmers and vegetarians and between other social structural variables (age, gender, rural versus urban residence, being a parent, level of education). 2. Material and methods 2.1. Survey Details about the sampling procedure and the general questionnaire have been reported in previous chapters and correspond with Study 2 discussed in the methodology section of this dissertation (Part I.4). In brief, a total of 1,081 adult respondents filled in a questionnaire distributed via weblinks, email and a more targeted approach through a personal contact procedure during 2006 in Flanders, the Dutch-speaking region of Belgium. A total of 247 returned but incomplete questionnaires were discarded. Of the remaining sample, 72 respondents were vegetarian, and 204 respondents were farmers or farmer’s offspring (termed “farmers” from now onwards). As 30 percent of these “farmers” stated farming as profession, it seems that the remaining 70 percent of this stakeholder group were farmer’s offspring. A quota sampling procedure was used to select a sample of 459 respondents (termed “citizens” from now onwards) that is representative of the Flemish adult population with regard to gender, age, living environment (rural versus urban) and province. These three stakeholder groups were not mutually exclusive. For example, the citizen sample also included some farmers and vegetarians, in accordance with these groups’ share in the overall population. The present chapter reports on the responses given to two specific questions of the survey which allow verification whether these three different stakeholder groups allocate similar relative importance to the 12 criteria and four principles as derived by Welfare Quality® to form an overall welfare assessment. In the first question, respondents were asked to give a score between 1 (“not 90 Chapter 4 : WQ® definition – opinion of citizens, farmers and vegetarians at all important to the welfare of farm animals”) and 10 (“extremely important to the welfare of farm animals”) to each of the 12 criteria defined by Welfare Quality® (Table 1.4). In the second question, respondents were asked to allocate a total of 100 points between the four principles for the welfare of farm animals as defined by Welfare Quality®. These four principles are: “good feeding”, “good housing”, “good health” and “appropriate behaviour”. As allocating 100 points across a large number of aspects might overtax the respondents, we opted not to use this method for determining the relative importance of the 12 criteria. In the questionnaire the 12 criteria were not organised into the four Welfare Quality® principles. The criteria and principles were listed separately for the first and second question, respectively. Questions and items were presented in a standardised format and a fixed order to all participants. 2.2. Statistical analyses Data were analysed using SPSS 15.0. Differences between types of respondents (citizens, farmers and vegetarians) for criteria importance scores and principle relative weights were assessed using one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores. Pairwise comparison of mean scores allocated to the 12 Welfare Quality® criteria (66 comparisons in total) and the four Welfare Quality® principles (6 comparisons in total) was done using pairedsamples t-tests for each type of respondent separately. Bivariate analyses through comparison of mean scores, including independent samples t-tests and one-way ANOVA F-tests, were used to assess the effect of various social structural variables on the importance allocated to the different components of farm animal welfare. 3. Results 3.1. Importance of the 12 Welfare Quality® criteria With all average scores well above the neutral point of the scale (µ=5.5), all 12 criteria were judged rather important to extremely important for farm animal welfare by all three types of respondents (citizen, farmer or vegetarian) (Table 3.2). Nevertheless, there were many differences in the scores given to the various welfare criteria between these types of respondents. Vegetarians gave consistently higher scores than farmers, and the citizens’ scores were between the other two groups. The difference between vegetarians and farmers was most pronounced for the four criteria (criteria 9-12) of the welfare principle “appropriate behaviour”, followed by those belonging to the “good housing” principle (criteria 6-8). Citizens allocated significantly more importance to three criteria (“absence of disease”, “absence of pain induced by management procedures” and “absence of prolonged thirst”) compared to any other criteria. Farmers allocated greatest importance to “absence of disease” and “absence of prolonged thirst” and the least importance to “expression of other behaviours”. Vegetarians 91 Part III allocated greatest importance to “absence of pain induced by management procedures”, “absence of prolonged thirst”, “absence of injuries” and “absence of general fear”. They allocated the least importance (although still very high) to “thermal comfort” and “good human-animal relation”. Intriguingly, all types of respondents seemed to allocate a lot of importance to criteria that were phrased as “Absence of...” as compared to the other criteria that are phrased positively. Contrary to differences found between citizens and farmers, the scores allocated by rural respondents did not differ significantly from those of urban residents. Regarding other social structural variables, female respondents allocated more importance to all 12 criteria (and those belonging to the “appropriate behaviour” principle in particular) than males (largest p=0.022). Respondents with children tended to give higher scores to most criteria as compared to respondents without children, although the difference was significant only for criteria 1, 3-5 and 810 (p<0.05). Education level did not associate with the scores, except for criterion 11: respondents with a lower level of education allocated more importance to “human-animal relations” than respondents with a higher level of education (|t|= 3.79, p<0.001). 3.2. Relative weights of the four Welfare Quality® principles The relative weights allocated to the four welfare principles differed between types of respondents (citizen, farmer or vegetarian). Vegetarians gave relatively more weight to “good housing” and “appropriate behaviour” but relatively less weight to “good feeding” and “good health” as compared to farmers and citizens (Table 3.3). Farmers allocated more weight to “good feeding” than did citizens or vegetarians. Citizens and farmers allocated most weight to “good health”, followed by “good feeding”, “good housing” and “appropriate behaviour”. Vegetarians allocated more weight to “good housing” and “good health” than to “good feeding” and “appropriate behaviour”. Young respondents (18-23 years old) allocated relatively more weight to “good housing” than respondents aged between 38-53 years (|t|=3.57; p=0.004) and >54 years (|t|=3.86, p=0.005), but less weight to “good feeding” than respondents aged between 38-53 years (|t|=3.95, p=0.001) and >54 years (|t|=3.16, p=0.002). Weights for “good feeding” were also lower for respondents between 24-37 years old compared with respondents between 38-53 years (|t|=3.53, p=0.006) and >54 years of age (|t|=3.40, p=0.016). Rural respondents tended to allocate more weight to “good feeding” than urban respondents (|t|=1.75, p=0.081). Respondents with children allocated more weight to “good feeding” (|t|=5.81, p<0.001) and less weight to “appropriate behaviour” (|t|=2.33, p=0.020) and “good housing” (|t|=4.19, p<0.001) than respondents without children. Level of education was not significantly associated with the relative weights allocated to the four principles. 92 Chapter 4 : WQ® definition – opinion of citizens, farmers and vegetarians Table 3.2. Structure of the Welfare Quality® operational definition of farm animal welfare with 12 criteria aggregated into four principles. For each of the 12 criteria the mean weights (for the different animal types) derived by Welfare Quality® are presented (WQ weight) as well as the mean (+SD) importance score (on a scale from 1 to 10) allocated by adult Flemish citizens, farmers and vegetarians. Significant differences between these types of respondents (rows) are indicated by a different letter (x, y, z) in superscript before the score. Within type of respondent (columns), importance scores without a common superscript (a-n) after the number differ significantly (p<0.05) Welfare Quality® Principles Criteria Survey WQ weight* Citizen 8.5 (1.7)a Farmer Good 1. Absence of prolonged hunger 0.41 x x y feeding 2. Absence of prolonged thirst 0.59 x x y Good 3. Comfort around resting 0.34 x x y housing 4. Thermal comfort 0.31 x y z 8.7 (1.5)dg 5. Ease of movement 0.35 x y z 9.4 (0.9)ak Good 6. Absence of injuries 0.32 x y z 9.7 (0.7)b health 7. Absence of disease 0.39 x y x 8. Absence of pain induced by management procedures 0.29 x y z 9.8 (0.5)e 9. Expression of social behaviours 0.26 x y z 9.1 (1.1)fc Appropriate 10. Expression of other behaviours 0.22 x y 6.3 (2.0)h z 8.9 (1.2)gj behaviour 11. Good human-animal relationship 0.23 x x 7.1 (2.2)ln y 12. Absence of general fear 0.29 x y 9.0 (1.5)b 7.8 (1.6)c 7.3 (1.8)d 8.1 (1.6)e 8.8 (1.4)f 9.0 (1.4)b 9.0 (1.5)b 7.7 (1.8)c 7.2 (1.8)d 7.7 (2.0)c 8.4 (1.8)a 8.0 (2.1)a Vegetarian 8.6 (1.9)bjk 7.4 (1.7)clm 6.8 (1.8)d 7.2 (1.8)en 8.1 (1.8)a 8.6 (1.6)fj 8.3 (1.7)gk 6.7 (2.0)d 7.5 (2.1)im 9.4 (1.1)a 9.7 (0.6)be 9.1 (1.3)cj 9.3 (1.2)ack 8.7 (1.6)hg z 9.6 (0.7)ib * based on the Shapley values reported in Botreau et al. (2008) 93 Part III Table 3.3. Average (+SD) relative weights (scored as %) allocated by adult Flemish citizens, farmers and vegetarians to the four principles of animal welfare as defined by Welfare Quality®. Significant differences between types of respondents (rows) are indicated by a different letter (xz) in superscript before the figure. Within type of respondent (columns), significant differences are indicated by a different superscript (a-d) after the figure Principles Good feeding Good housing Good health Appropriate behaviour Citizen x 27.6 (10.2)a x 24.4 (9.5)b x 31.0 (11.3)c x 17.7 (10.9)d Farmer y 30.7 (13.6)a x 23.6 (11.3)b x 30.8 (15.1)a x 15.5 (12.0)c Vegetarian z 22.8 (5.8)a y 28.3 (7.5)b y 26.2 (6.9)b y 22.7 (8.7)a 4. Discussion and conclusion The importance scores and relative weights allocated to the 12 welfare criteria and the four welfare principles defined by Welfare Quality® differed between Flemish citizens, farmers and vegetarians. Overall, farmers gave lower importance scores to the 12 welfare criteria as compared to citizens, and much lower scores than vegetarians. The magnitude of the differences, however, varied depending on the criterion. The absolute difference between farmers versus citizens/vegetarians was highest for criteria such as “expression of social behaviour”, “expression of other behaviours”, “ease of movement” and “absence of general fear” and lowest for criteria such as “absence of disease”, “absence of prolonged thirst” and “absence of prolonged hunger”. Similar differences in perception of farm animal welfare between farmers and consumers/citizens have been documented previously (Lassen et al., 2006; Marie, 2006; Te Velde et al., 2002). People involved in animal production are known to have a much more positive perception of the current state of farm animal welfare and to view the welfare of farm animals more in terms of biological functioning (instead of affective states or natural living) as compared to other citizens (European Commission, 2005; Kendall et al., 2006). Farmers attribute more importance to criteria that have a direct economic impact, such as animal health and feeding. In the present study, vegetarians allocated significantly higher scores to all criteria compared with farmers, and to all but one criterion (“absence of disease”) compared to citizens. The allocation of weights to the four principles revealed that vegetarians allocated relatively more weight to “good housing” and “appropriate behaviour” and less weight to “good feeding” and “good health” than the other respondent groups. Vegetarians have previously been reported to attribute a lot of importance to animal welfare relative to other aspects of food and food production and to have a more negative view on the current state of animal welfare in Flemish animal production. Objections to the suffering and killing of animals kept for meat production are among the main motivations for being vegetarian (Hopkins and Dacey, 2008). Based on the focus group discussions presented in Chapter 1, it was reported that 94 Chapter 4 : WQ® definition – opinion of citizens, farmers and vegetarians vegetarians in general highly value socio-cultural product attributes like environmental friendliness, animal welfare and fair trade. A better understanding of these differences in opinion about what constitutes the concept of animal welfare may be beneficial for facilitating a constructive dialogue and improving the communication between farmers, vegetarians and other citizens. This study further illustrates that the concept of animal welfare has a different meaning to different people, depending on placebased variables such as stakeholder type and rural versus urban residence, and other social structural variables such as age, gender, being a parent, and – to a lesser degree – level of education. This variation underlines the relevance of a segmented and targeted market approach for higher welfare products. 95 Chapter 5 : Piglet castration – Flemish citizen sample Chapter 5 Piglet castration – Flemish citizen sample Adapted from Vanhonacker, F., W. Verbeke, and F. Tuyttens. 2009. Belgian consumers’ attitudes towards surgical castration and immunocastration of piglets. Animal Welfare, 18, 371-380. Abstract In the next three chapters, the thesis scope will be narrowed to piglet castration, as an animal welfare related topic. In order to fit in with the discussion about the impact of place-based variables on public attitudes, perceptions and concerns, the topic is approached from both the public (Chapters 5 and 6) and the producers’ view (Chapter 7). Male piglets are surgically castrated in the large majority of European countries in order to avoid the development of boar taint and a corresponding degraded meat quality. This practice is strongly debated for its negative impact on the piglet’s welfare, integrity and health, and has induced the search for more humane alternatives. Immunocastration is one such alternative and involves the injection of vaccine that inhibits the production of the hormones responsible for boar taint. This practice shows satisfactory results in terms of meat quality and production parameters, though uncertainty about consumer acceptance is often put forward as one of the key factors for a successful market adoption. This chapter focuses on the consumer awareness of piglet castration and their attitude towards immunocastration relative to the routine practice of surgical castration without anaesthesia by means of a web-based questionnaire among 225 Flemish citizens. About 40 percent indicates to be aware of the routine practice of piglet castration. This limited awareness is accompanied by a modest level of concern about castration, especially as compared to food safety issues and other pork production systemrelated animal welfare issues. Sixty percent of the sample had a general appreciation in favour of immunocastration as compared to surgical castration without anaesthesia. Informing consumers about possible benefits and/or risks from immunocastration did not have a major impact in terms of shifting their attitudes. Immunocastration did not emerge as a problem in terms of consumer acceptance, although special attention should be addressed to consumers’ perception of price, food safety and taste of the meat from immunocastrated pigs. 97 Part III 1. Introduction Background information about the problem of boar taint and the range of available and/or inquired practices to avoid the development of boar taint has already been discussed in the general introduction of this dissertation and will not be repeated here. The focus of this chapter will be on immunocastration and will be approached from the side of the chain end user. Presently, immunocastration is practiced in several countries including among others Australia, New Zealand, Brazil, Russia, Switzerland and several Latin-American and Asian countries, while (at the time of this specific study) also a positive opinion on an initial marketing authorisation application for the vaccine ImprovacTM on the EU market has been granted beginning 2009 by the EU Committee for Medicinal Products for Veterinary use9. Next to its advantages in terms of animal welfare, immunocastrated pigs have been shown to have a more favourable feed conversion (Cronin et al., 2003; Dunshea et al., 2001; Jaros et al., 2005; Turkstra et al., 2002) and a higher percentage of lean meat (Jaros et al., 2005) as compared to surgically castrated pigs and reduced sexual and aggressive behaviour as compared to entire males (Cronin et al., 2003; Velarde et al., 2008). Drawbacks include the costs of vaccination (vaccine and labour), the chance of self-injection and uncertainties regarding consumer/market acceptance of the method (Font i Furnols et al., 2008; Prunier and Bonneau, 2006; Prunier et al., 2006). Nonetheless, estimations indicate that the vaccination costs can be compensated by the benefits from an increased feed efficiency (de Roest et al., 2009). Also, specific injection devices have been developed that reduce the chance on self-injection in addition to the development of anti-vaccines. Although the consequences of immunocastration on production performance and animal behaviour have been investigated in several studies (for a review see Prunier et al., 2006), little research has been conducted on consumer acceptance. In a study by Font I Furnols et al. (2008) the focus was on the sensory evaluation and acceptability of meat from immunocastrated pigs in Spain, concluding that Spanish consumers were not able to distinguish between cooked pork from immunocastrated pigs, surgically castrated pigs and female pigs. Despite this favourable result in terms of sensory evaluation, these findings do neither provide any insights in consumers’ buying intention and behaviour nor do they reveal any insights in terms of consumers’ beliefs and attitudes, which can be expected to be antecedents of behavioural intentions. To the authors’ knowledge, the only studies that have dealt with consumer issues, apart from sensory studies, are a study in Australia (Hennessy and Newbold, 2004), a study in Sweden (Lagerkvist et al., 2006), a study in Norway (Fredriksen et al., 2010), and two studies in Switzerland (Giffin et al., 2008; Huber-Eicher and Spring, 2008). Considering the relatively high animal welfare standards and divergence in public interest in farm animal welfare and ethical issues across countries, generalisation of findings from these studies to other European countries remains quite speculative. 9 Presently, the vaccine has already been registered, and is thus available on the market (see next chapter) 98 Chapter 5 : Piglet castration – Flemish citizen sample Hennessy and Newbold (2004) found through qualitative focus group discussion with Australian female responsibles for food purchases that immunocastration was strongly accepted, despite a limited awareness about the practice. The practice was associated with a natural process and was perceived as a favourable alternative for surgical castration without anaesthesia. Lagerkvist et al. (2006) also found immunocastration to be a socially acceptable alternative, based on willingness to pay estimates from a choice experiment with Swedish consumers. They found that animal welfare concerns dominated aversion against biotechnology or perceived food safety risks, while animal welfare concerns were subordinate to sensory quality concerns. Regarding the two Swiss studies, conflicting results appeared. The study of Giffin et al. (2008), which was based on 971 on-line interviews about pork consumers’ acceptance, revealed that two out of three consumers considered immunocastration more acceptable than surgical castration. Huber-Eicher and Spring’s (2008) study in contrast revealed a low acceptance of meat from immunocastrated animals among 800 Swiss consumers, with 56 percent indicating a negative willingness to buy this meat. Surgical castration under anaesthesia/analgesia was preferred since it was most transparent to the consumer and it eliminated reliably the risk of boar taint. Within the PIGCAS project, where the consumer voice is represented by consumer organisations, a similar preference for castration under anaesthesia was found (von Borell et al., 2008). The objective of the present chapter was to investigate the awareness, concern, attitude and selfreported willingness to pay of Flemish consumers concerning immunocastration in pig production and pork from immunocastrated pigs. Given that this research topic is largely unexplored and that no primary data as such is yet available in Flanders, this research is organised and designed as a preliminary qualitative study. It provides insights into the degree to which it is justified to consider consumer acceptance as a key issue for the implementation of immunocastration. More specifically, issues on which this practice scores favourable or disadvantageous, as perceived by consumers, are reported. Additionally, the impact of communication messages informing consumers about the benefits and/or risks from immunocastration is explored. 2. Material and methods Methodological details in relation to the research approach and the sampling procedure are discussed in Part I.4, and apply to Study 5. 2.1. Survey The questionnaire started with some general questions about animal welfare and pig production. First respondents were provided with an extended and diversified list of issues and handlings (27 in total) that are related to the pig production chain. A large number of the issues and handlings dealt with characteristics of the pig production system that have an impact on the animal’s welfare, among them piglet castration. These issues and handlings were selected from an exhaustive list of issues/handlings that play a role in the citizen’s interpretation of the concept of farm animal welfare 99 Part III (see Chapter 2). The selection of issues was organised in such a way that each of the seven dimensions constituting the citizens’ (perceptual) interpretation of the concept of farm animal welfare as described was represented by at least one item. These dimensions were: Animal Health; Engagement in Natural Behaviour; Feed and Water; Housing and Climate; Human-Animal Relationship; Suffering and Stress; and Transport and Slaughter. Another group of issues involved meat product characteristics like taste and price of the pork as well as food safety related matters such as hormones, residuals and food safety itself. Further issues pertained to singular issues and handlings (e.g. quality of the information, environmental impact …). For each of the issues and handlings respondents were asked to express their level of concern on a seven-point Likert scale that ranged from “not at all concerned” (1) to “very concerned” (7). Second, respondents were asked if they were aware of piglet castration in Flemish pig production and if “yes”, to formulate the reason for this practice using an open-ended question. In this introductory part, nothing explicitly was asked about immunocastration. Following on the general part, the respondents were provided with printed information about immunocastration. The information message included three components. The first component included a general description of why castration was performed, how it is currently practised (surgical castration without anaesthesia) and a short explanation of immunocastration. The second component consisted of a description of the advantages of immunocastration. Benefits mentioned in the message referred to a reduction of pain and stress as compared to surgical castration, a reduction of aggression and sexual behaviour as compared to entire males, a better feed conversion ratio as compared to barrows and sows and the reassurance that such meat is safe for the consumer. In the third component the major downsides/risks associated with immunocastration were given, in terms of the danger of self-injection, the costs associated with injecting the pigs, the fact that the vaccine has not been registered yet by the EU, and the uncertainty about consumer acceptance (Table 3.4). Whereas the first component of the message was provided to all respondents, the second and third message component was only presented to subsamples of the overall sample. In total, four different types of questionnaires were distributed. In a first message condition, only the general part was provided. The subsample that received this first version functioned as the control group in the study (below this subsample is referred to as “control”). In the second condition, benefits were shown together with the control message (“benefits”). In the third condition, the downsides were given together with the control message (“risks”). Finally, the fourth message condition provided the full picture, i.e. it contained the general part and both the possible benefits and risks from immunocastration relative to surgical castration without anaesthesia (“full info”). Message conditions were assigned randomly to the survey participants. After exposure to the previously described message conditions questions were asked specifically about attitudes and self-reported willingness to pay related to immunocastration and meat from immunocastrated pigs. First, respondents were probed for their general attitude towards immunocastration as an alternative for the present practice of surgical castration. The measurement scale was a seven-point Likert scale, ranging from “much worse (than surgical 100 Chapter 5 : Piglet castration – Flemish citizen sample castration)” (1) to “much better” (7). Second, immunocastration was evaluated against surgical castration on a list of process characteristics and pork attributes, applying the same seven-point scale. Third, consumers were asked to report their willingness to pay for meat from immunocastrated pigs for different levels of price premiums on a seven-point probability scale, ranging from “very unlikely” (1) to “very likely” (7). Table 3.4. Message conditions tested in the survey: message content and number of respondents (n) exposed to each message Message condition Control Message content n Why are male piglets castrated? – To avoid boar taint Current method of castration – Surgical castration 58 Potential alternative – Immunocastration Benefits from Control + immunocastration 53 Pain and stress reduction Reduction of level of aggression and sexual behaviour Better growth performance No risks for food safety Risks from Control + immunocastration Danger of self injection Extra costs (labour and vaccine) 57 Vaccine not yet authorised Uncertainty about consumer acceptance Full content message 2.2. Control + Benefits + Risks 57 Statistical analyses Data were analysed using SPSS 15.0. Frequencies, means and standard deviations are presented in table format. Factor analysis using principal components and varimax rotation is applied to analyse the association between consumers’ reported concerns about issues and handlings related to the pig production chain. A factor analysis is useful to identify common underlying dimensions (factors) that consist of items (in this case concerns) that are strongly interrelated (Hair et al., 2006). The selection of factors was based on Eigenvalues (>1 as threshold), while factor loadings were used to interpret the meaning of the resulting factors. Cronbach’s alpha was used to decide upon internal reliability consistency. Threshold value for a satisfactory construct is 0.6, which denotes that the different items measure one single construct and therefore may be aggregated. Aggregation was done through averaging the scores across issues assigned to a specific factor. Bivariate analyses through comparison of mean scores, including independent samples t-tests and one-way ANOVA Ftests with Bonferonni and Dunnett’s T3 post hoc comparison of mean scores, were used to assess the impact of the different message conditions on interval scaled variables related to consumer attitudes and self-reported willingness to pay. 101 Part III 3. Results 3.1. Consumer awareness of piglet castration Half of the respondents reported to be aware of the fact that male piglets are castrated. This group is further termed as the “aware group”, the other half as the “unaware group”. From the aware group, 78 percent was able to report that the reason for this practice was related to the quality of meat, with the large majority referring to the odour of the meat and 27 respondents (12 percent) explicitly mentioning boar taint. Taking together the 50 percent indicating to be aware of piglet castration and the 78 percent denoting the correct reason, it can be concluded that about 40 percent of our sample was well-informed on the topic. Incomplete and/or wrong answers most frequently pertained to the idea that castration was done primarily in order to reduce the level of aggression, to control reproduction and to avoid energy use from reproduction and leave more energy for growth and other production parameters. 3.2. Consumer concerns about pig production practices Factor analysing the concern scores related to issues and handlings in pig production yielded a four factor solution based on Eigenvalues > 1. Concerns that either did not load high (no loadings above 0.5) on any of the retained factors or that had loadings on multiple factors were excluded from the analysis (Table 3.5). Total variance explained was 74.0 percent. The first factor comprised production system-related animal welfare concerns and explained the major part of the variance in the original data. At least one issue from each of the seven dimensions that constitute the citizen’s interpretation of farm animal welfare (see Chapter 2) was strongly correlated to this factor. The second factor corresponded with food safety concerns. Together with the item food safety itself, concern related to hormones and residues of medication was categorised in this factor. Interestingly also concern about genetic modification correlated significantly with the food safety factor. Next, concern for some specific mutilations – among them piglet castration – was ranked as a third factor, separate from the factor dealing with other production system-related animal welfare concerns. Yet it is noteworthy that concern about castration also correlated with the factor of animal welfare concerns (factor loading of 0.416), thus not perceived as completely unrelated to the other production system-related animal welfare concerns. The fourth factor (Eigenvalue just above the threshold) consisted of concern related to the end product’s price and taste, two important product attributes, which reflect pork meat concerns. Mean values for the factors indicated the highest concern for food safety issues, followed by animal welfare and pork meat concerns. Concern about animal mutilations was attributed a mean score around the scale’s mid-point. Differences in the factors’ mean values between the aware and unaware group were not significant (p > 0.05). 102 Chapter 5 : Piglet castration – Flemish citizen sample Regarding the concern scores for the individual issues and handlings, it was noticed that Flemish consumers were suspicious about the use of hormones and residuals and about food safety in general (Table 3.5). Also the pork meat’s taste was attributed a high concern score. The lowest mean concern score – though still on the positive side of the scale – was reported for piglet castration, followed by the two other animal mutilations incorporated in the study. Table 3.5. Factor analysis: Flemish consumer concerns related to pig production. Factor loadings from principal component analysis. The right column reports the mean concern score and standard deviation (SD) for the total sample (n=225) Factor 1 0.828 0.776 0.776 0.775 0.756 0.744 0.730 0.719 0.698 0.695 0.682 0.669 0.639 Factor 2 Factor 3 Factor 4 Mean (SD) 5.12 (1.48) 5.27 (1.56) 5.04 (1.50) 4.96 (1.65) 5.06 (1.54) 5.11 (1.54) 4.94 (1.69) 5.00 (1.72) 5.38 (1.55) 4.61 (1.65) 5.56 (1.44) 4.77 (1.59) 5.32 (1.60) 5.48 (1.60) 5.56 (1.64) 5.55 (1.61) 4.56 (1.71) 4.17 (1.77) 4.15 (1.76) 4.14 (1.73) 4.74 (1.61) 5.37 (1.55) Space availability Slaughter without pain and stress Barn climate Conditions transport Stress Animal welfare Contact farmer-animal Skills of farmer Mortality Way of housing Disease Natural Behaviour Feed Residuals of medication 0.871 Use of hormones 0.862 Food safety 0.741 Genetic modification 0.726 Tail docking 0.839 Tooth resection 0.834 Castration 0.777 Price of the meat 0.817 Taste of the meat 0.681 Eigenvalue* 11.2 2.7 1.3 1.1 Explained variance (%) 34.9 16.3 14.1 8.7 Cronbach’s alpha 0.96 0.83 0.92 0.63 Mean (SD)_Total sample 5.10 (1.28) 5.31 (1.40) 4.17 (1.63) 5.05 (1.35) Mean (SD)_Aware group 5.05 (1.29) 5.21 (1.54) 4.28 (1.73) 4.95 (1.45) Mean (SD)_Unaware group 5.14 (1.23) 5.40 (1.24) 4.03 (1.50) 5.15 (1.24) * Only factors with Eigenvalue above 1 are reported; Items not included because of low or dual factor loadings: Import of foreign meat (mean=4.63; SD=1.69); Environmental impact (mean=4.88; SD=1.62); Quality of information (mean=4.83. SD=1.59); Intensive character (mean=4.54; SD=1.54); Growth rate (mean=4.35; SD=1.53) 103 Part III 3.3. Impact of the information message related to immunocastration Message condition did not have a significant impact on the respondents’ general attitude towards immunocastration (Table 3.6) nor on their self-reported willingness to pay for meat from immunocastrated pigs (p > 0.1). Conversely some significant effects of message condition were found regarding more specific aspects of immunocastration versus surgical castration (Table 3.6). However, these differences did not provide a consistent and coherent picture. For instance, the message condition where only risks were communicated yielded a better comparative evaluation for immunocastration in terms of farm profitability and farmer’s profit, despite the fact that immunocastration was identified with possible extra costs. Thus, it is doubtful whether these differences were due to the message condition or due to differences in the characteristics of the respondents exposed to a particular message condition. Table 3.6. Consumer evaluation of immunocastration relative to surgical castration. Figures are mean values on a seven-point scale, that ranges from “much worse (than surgical castration)” (1) to “much better” (7). Mean values are presented for the full sample and for the different message conditions Sample mean General attitude tw IC ANOVA Message condition (SD) Control Benefits Risks Full info F-test p- n=225 n=58 n=53 n=57 n=57 value 4.77 (1.44) 4.95 4.58 4.98 4.55 0.240 a 3.85 4.57 b 4.30 0.019 a,b Avoiding boar taint 4.23 (1.18) Meat quality (sensory) 4.06 (1.11) 3.91 3.81 4.27 4.21 0.090 4.08 (1.23) a a 4.65 b 4.32 <0.001 4.64 b 4.05 0.063 Profit per pig Type of labour for farmer Farm profitability 4.33 (1.20) 4.10 (1.17) 4.15 3.64 a 4.40 a 3.84 a,b a 4.21 a 4.65 b 3.98 0.001 a 3.29 4.31 b 3.82 <0.001 3.92 Food safety 3.82 (1.22) Animal welfare 5.44 (1.31) 5.45 5.53 5.33 5.45 0.889 Consumer acceptance 4.83 (1.35) 5.06 4.48 5.11 4.64 0.041* Work load for farmer 4.17 (1.32) 4.31 3.88 4.55 3.93 0.025* Competitivity of sector 4.16 (1.16) 4.11 3.98 4.51 4.02 0.069 Affordability for consumer 3.69 (1.02) 3.51 3.58 3.95 3.71 0.128 * 3.80 3.65 Despite significant one-way ANOVA F-tests, Bonferroni and/or Dunnett’s T3 post hoc tests did not indicate significant differences between the mean scores 104 Chapter 5 : Piglet castration – Flemish citizen sample 3.4. Consumer attitude towards immunocastration Given that message condition did not affect general attitude or stated willingness to pay towards immunocastration, the data were merged for further analyses. In general respondents evaluated immunocastration slightly better than surgical castration, with 60 percent of the sample indicating to prefer immunocastration over surgical castration (i.e. score > 4) (Figure 3.4). The respondents believed that immunocastration was superior to surgical castration for what concerns animal welfare (Table 3.6). Also respondents indicated higher consumer acceptability for immunocastration. Further evaluations significantly in favour of immunocastration in consumer’s perception pertained to the type of labour for the farmer and the avoidance of boar taint. In contrast, food safety and price were evaluated more negatively for immunocastration compared with surgical castration. Work load for the farmer, competitivity of the sector, farm profitability, profit per pig and meat quality were not pronounced in favour of either one of the practices (i.e. mean value did not significantly differ from the scale’s mid-point). The respondents’ answers were mainly characterised by a high number of respondents positioning themselves in the middle of the scale. Such answering behaviour reflects either that people judge both methods truly equivalent or that they express uncertainty and/or unawareness. In particular, more than half of the sample (52.8 percent) positioned themselves at the scale’s mid-point for the issue attribute of sensory meat quality. 30 Percentage of sample 25 25 24.1 19.5 20 13.6 15 10.9 10 5.5 5 1.4 0 IC much worse worse somewhat worse neither better nor worse somewhat better better IC much better Figure 3.4. General attitude of consumers towards immunocastration (IC) relative to attitude towards the traditional practice of surgical castration. Answers are registered on a seven-point scale ranging from immunocastration perceived as “much worse” to “much better” 3.5. Self-reported willingness to pay for pork from immunocastrated pigs Self-reported willingness to pay for meat from immunocastrated pigs was assessed for different levels of price premiums as compared to conventional pork, ranging from an equal price to the double. Respondents stated to prefer meat from immunocastrated pigs if the meat is offered at the same price (positive reported likelihood of purchase) (Figure 3.5). On average, also a price premium 105 Part III of five percent still corroborates with a positive probability of buying meat from immunocastrated pigs (i.e. μ > 4). Further price premiums result in a negative purchase probability (i.e. μ < 4). 0% 5% 10% Level of price premium 20% 50% 100% 1 2 3 4 5 6 7 Figure 3.5. Self-reported willingness to pay (WTP) a price premium (expressed as percentage over the price conventional pork) for meat from immunocastrated pigs. Self-reported WTP is measured on a seven-point scale ranging from “very unlikely to buy” (1) to “very likely to buy” (7) 4. Discussion and conclusion Research related to alternatives for surgical castration of male piglets without anaesthesia is presently very topical in various research fields. This interest originates from the debate that surgical castration currently faces, a debate that is also kept alive by animal welfare organisations (e.g. “Pigs in Pain” campaign by GAIA in Flanders). From the different alternatives currently available and/or under development, this chapter has focused on immunocastration. Motivations underlying this choice pertain to the fact that a vaccine is already available and used in several countries, and authorisation of the vaccine on the EU market was expected in 2009 (at the time of the study). Thus, immunocastration can be seen as a possible solution in the short term. Also, no surgery is needed, yielding positive implications for the animal’s welfare, integrity and health. In addition, this method gives promising results in terms of animal growth and feed efficiency (e.g. Cronin et al., 2003; Dunshea et al., 2001; Jaros et al., 2005; Prunier et al., 2006; Turkstra et al., 2002). Successful implementation of novel methods and practices, such as immunocastration is often believed to depend upon consumer acceptance (Frewer et al., 1997a). So far consumer studies about immunocastration have focussed more on sensory issues (e.g. Font i Furnols et al., 2008) instead of attitudes, perceptions and purchasing behaviour. In this context, an exploratory research design was constructed with consumers’ awareness, attitudes, perceptions and beliefs as central 106 Chapter 5 : Piglet castration – Flemish citizen sample themes. The data collection method used allowed to collect a substantial amount of data in a limited amount of time, but also implied that the sample does not statistically represent the total Flemish population. A first important finding was the low awareness about piglet castration among the survey participants. The low awareness corresponds with the alienation of consumers and citizens from animal production and animal production practices (Harper and Henson, 2001). Due to an (ever) increasing degree of urbanisation and industrialisation of animal production, consumers and citizen become the less and less aware of how animals are actually reared (De Tavernier et al., 2010). As such, opinions are strongly shaped by perceptions and (often distorted) external information rather than on facts and real experiences. Increasing awareness – thus decreasing the distance between perception and reality - however is not simply an issue of providing additional information. A considerable amount of people for instance consumes meat, ignoring the fact that the meat originates from an animal raised for human. Also, buying food is most often a routine process in which people do not extensively search for information. This is amplified by the present information era in which consumers are overloaded with information (Verbeke, 2005). In addition, large standard deviations for the reported concerns reflect a substantial degree of heterogeneity in the population regarding the impact of animal welfare, which is also impacting on attitude formation and food purchasing behaviour. As such communication efforts should be well-considered and passed on to the right target audiences. In addition to the limited awareness, the survey participants expressed a moderate concern about piglet castration, especially as compared to other production system-related animal welfare concerns. Moreover castration, together with the other animal mutilations incorporated in the survey, emerged as a separate factor in the factor analysis, apart from other welfare related concerns. This suggests that people perceive these issues as different from other common welfarerelated practices in pig production. The highest concerns in our sample pertained to food safety issues. The use of hormones and residues of medication worries consumers and reflects that the different scandals and crises that have hit Flemish animal production are still prominent in consumers’ minds. Also the categorisation of genetic modification as part of the food safety factor reflects the wariness about potential safety implications of biotech in animal production (Frewer et al., 1997a). These findings suggest that consumers are extremely vigilant about animal production practices that may impact (real or perceived) end product safety. The second main finding was that in the present study consumer acceptance of immunocastration does not seem to be a problem. Immunocastration was significantly better accepted than surgical castration. This degree of acceptability seemed to be largely related to the improvement the method has in terms of animal welfare. As a consequence, the uncertainty expressed in some literature towards the acceptance of immunocastration at the consumer level cannot be confirmed in this study. Nonetheless, one should bear in mind that consumer acceptance is not equal to actual consumer behaviour. Reasons that explain the possible discordance between acceptance and actual 107 Part III purchasing behaviour relate to animal welfare not being a priority product attribute in the consumer decision-making process of pork, since this attribute is traded off against other quality characteristics like price, taste, safety and healthiness (Verbeke and Viaene, 2000). This will be further detailed in Part III of this dissertation. In general, animal welfare is of secondary importance and influences purchasing choices only when other criteria are fulfilled (Harper and Henson, 2001; Ingenbleek et al., 2006a). In this sample meat from immunocastrated pigs was expected to be more expensive and less safe than pork from surgically castrated pigs, while the respondents’ expectations about taste did not differ between meat from immunocastrated or surgically castrated pigs. These findings probably explain why the favourable attitude expressed by our respondents towards immunocastration is not translated into a strong self-reported willingness to pay. It is important to emphasise that these results reflect consumer perceptions at the time the survey took place, meaning that they can shift and fluctuate over time. As a consequence, if meat from immunocastrated pigs is about to be introduced, it will be important to pay attention and clearly communicate about the impact on price, food safety and taste. It will be especially important to avoid negative and/or incorrect information about these issues, given disproportionally large impacts of possible negative publicity related to food safety issues (Verbeke and Vackier, 2004; Verbeke and Ward, 2001). Finally, the communication experiment in which the impact of an information message was tested on consumers’ attitude and behaviour towards immunocastration resulted in non-significant effects. There could be several reasons for this outcome. First, the message was a merely rational message, which requires a substantial degree of active and rational information processing from the recipient audience. Possibly an emotional message or a message providing images or a video would have triggered a higher impact. Second, as the questionnaires were completed on-line outside the control of an interviewer, we do not have any information on the extent to which the message was effectively read by the respondents. Given the rather low awareness and concern about piglet castration in our sample, one can expect a low degree of issue involvement. Typically, low issue involvement results in a low level of active reasoning and a low level of conscious information processing (Mittal and Lee, 1989; Verbeke and Vackier, 2004). Consumer involvement is also important for the formation of beliefs, attitudes and intentions (Verbeke and Vackier, 2004). The reported low involvement implies that attitudes and beliefs are not strongly shaped and can shift in either direction depending on the individual, the context and the information received. This illustrates the importance of avoiding incorrect and negative publicity and of encouraging communication that can shift consumers’ beliefs and attitudes in a favourable direction, especially on the product attributes that strongly impact on consumer purchase behaviour and where uncertainty is present at consumer level. 108 Chapter 6 : Piglet castration – European citizens Chapter 6 Piglet castration – European citizens Adapted from Vanhonacker, F. , and W. Verbeke. 2011. Consumer response to the possible use of a vaccine method to control boar taint versus physical piglet castration with anaesthesia: a quantitative study in four EU countries. Animal, in press. Abstract This chapter further investigates the topic of piglet castration from the consumer side. Opposite to the previous chapter, this chapter is based on a European sample and measures consumer attitudes towards immunocastration relative to physical castration with anaesthesia or analgesia. This chapter is further complemented with a segmentation analysis. This large-scale quantitative crosscountry study (n=4,031) involving representative samples of consumers in France, Germany, the Netherlands and Belgium does not support the reserved attitude of stakeholders who fear potential low market acceptance for immunocastration as alternative practice. The vaccine method was actually preferred by the majority of consumers surveyed (69.6 percent of the participants) and it was perceived as equally effective in terms of avoiding boar taint. 43.8 percent of the consumers reported an intention to seek out pork from pigs where the vaccine had been used to control boar taint, whereas 33.7 percent reported an intention to avoid pork from pigs physically castrated with anaesthesia. Consumers’ favourable dispositions towards the vaccine method were independent of dominant ethics-, health- or price-orientations when purchasing pork. 109 Part III 1. Introduction This chapter concentrates on the vaccine method (synonym of immunocastration or immunovaccination), more specifically on the consumer acceptance of the method as opposed to physical castration with anaesthesia or analgesia. Both are possible solutions for the contested current practice of physical castration without anaesthesia. The vaccine (ImprovacTM, Pfizer GmbH) has recently been registered and is thus available on the market. The vaccine method has been associated with higher daily weight gain, better feed conversion and a higher percentage of lean meat (Prunier et al., 2006). In addition, von Borell et al. (2009) reported a positive effect on welfare in relation to pain during and after castration for immunocastration, and Baumgartner et al. (2010) concluded that the behaviour of vaccinated male pigs does not create additional problems during the fattening period, when compared with physically castrated pigs. With regard to economic implications, de Roest et al. (2009) concluded that the improvement in feed efficiency may compensate almost entirely for the cost of the vaccination and it is noted that the vaccine is sold with the claim of providing a positive economic return to pig farmers. The benefits in terms of the animal’s welfare, though poorly investigated (von Borell et al., 2009), are widely recognised. Recently, Fàbrega et al. (2010) confirmed improved welfare aspects such as reduced aggression and reduced mounting behaviour. The pig industry however is rather reserved towards the vaccine method, because of uncertainties regarding marketplace and consumer acceptance of the method. Alleged reasons for concern relate to the possible perception of a (false) link with a hormone treatment (de Roest et al., 2009; Heinritzi et al., 2006; Prunier and Bonneau, 2006), which is a very sensitive issue among European meat consumers (Verbeke et al., 2007a; Verbeke et al., 1999). Clinical research has shown that the consumption of meat from vaccinated pigs does not involve any risk to human health, which would also be anticipated from the protein composition and immunological mode of action (Clarke et al., 2008). A positive consumer perception is a key determinant for the applicability of a new method (Frewer, 1999; Huber-Eicher and Spring, 2008), especially in a demand-driven market and economy. As discussed before, consumer acceptance is put forward as the key for market success of the vaccine method. Several studies already concentrated on investigating consumer acceptance and perception of boar meat, mainly through sensory experiments with experts, lay consumers or a combination of both (see Allison et al. (2009) and Heid and Hamm (2009) for a review). The studies that also investigated consumer perceptions of pork from vaccinated pigs (e.g. Font I Furnols et al. (2008)) indicated that pork meat from vaccinated pigs was experienced at least equally good as pork meat from castrated and/or female pigs. In contrast, pork meat from entire boars was most often evaluated worse than meat from vaccinated or castrated pigs. Hence, there appears to be no increased likelihood of dissatisfying consumers in terms of sensory properties of the meat compared to the current practice. The present challenge that will be addressed in this chapter is to 110 Chapter 6 : Piglet castration – European citizens investigate the likelihood that EU consumers would not be willing to accept or purchase pork from immunised pigs for reasons of a perceived food safety risk (e.g. fear for residuals among Norwegian consumers as reported by Fredriksen et al. (2010)). Few studies have concentrated on such prepurchase attitudes and decision processes, and the available studies pertain to survey approaches on a single country level, e.g. Hennessy and Newbold (2004) (Australia), Lagerkvist et al. (2006) and Liljenstolpe (2008) (Sweden), Giffin et al. (2008) and Huber-Eicher and Spring (2008) (Switzerland), Fredriksen et al. (2010) (Norway), Vanhonacker et al. (2009) (Belgium, see Chapter 5), and Fredriksen et al. (2010) (Norway). The lack of cross-national comparisons and analyses in previous studies is a major limitation in an international economic context. The PIGCAS project followed an international approach, though results were based on a limited number of representatives of the different stakeholders, a shortcoming that has been acknowledged by the project consortium (Fredriksen et al., 2009). The consumers’ opinion was investigated through interviewing consumer organisations, which does not necessarily coincide with individual consumers’ views. Except for the Swiss study by Huber-Eicher and Spring (2008), Liljenstolpe (2008) (who did not specifically address the vaccine method) and Fredriksen et al. (2010) in Norway (who used the term “medical castration” and presented the use of anaesthesia as the standard procedure), previous studies mostly provided reassurance to the pig industry, denoting a favourable consumer attitude with regard to the vaccine method. In Chapter 5 this favourable attitude was linked mainly to a better perception score on animal welfare and a higher consumer acceptance, in spite of some uncertainty among consumers related to price. Several studies have pointed towards the importance that communication and information provision about the method will play on the consumer acceptance of the method (Heid and Hamm, 2009; Spring et al., 2009). Given that consumers are only scarcely aware of piglet castration, and only few consumers ever heard of the vaccine method, large-scale surveys need to inform participants before attitudes can be measured. As such, differences between the findings of different studies are probably partly related to the design of the study including the ex-ante information provision to participants. In this perspective, there is a need for cross-country studies that evaluate the consumers’ acceptance of the vaccine method, their (non-)preference for the method over physical castration with anaesthesia (as the most feasible alternative short-term solution at the present moment), and their likelihood to actively seek out or avoid pork produced with this method. The specific objective of the present chapter was to investigate consumer response, preference and purchasing intention in four member states of the European Union: Belgium (BE), France (FR), Germany (DE) and the Netherlands (NL). The study herewith focuses on a cluster of four neighbouring West-European which are important players in terms of pig production, pork consumption and intra EU or international pork trade. Possible differences between and within countries will be presented and discussed. This chapter tackles the key issue that dominates the present (anno 2010) debate about the vaccine method, namely the uncertainty towards the prepurchase consumer acceptance or rejection of meat from vaccinated pigs. Covering different 111 Part III countries matches with international business goals, market organisations and trade economies and rules out a possible framing effect from comparing methodologically different national studies. 2. Material and methods Methodological details in terms of research approach and sampling procedure are provided under Part I section 4, and apply to Study 6. 2.1. Measurement and scaling Participants were asked to complete a self-administered, structured, electronic questionnaire, which consisted of two parts, separated from each other by information provision on boar taint, on physical castration with anaesthesia and on the vaccine method as alternative methods to eliminate boar taint. In the first part of the questionnaire, next to some general questions related to the socio-demographic profile of the respondents, pork consumption frequency and product attribute importance in the purchase decision process of pork was assessed. Response categories for consumption frequency were “more than once a week”, “once a week” and “two to three times a month”. Perceived product attribute importance was measured for price, taste, health, animal welfare and environmental friendly production. The selection of these five attributes was informed by previous studies (Grunert et al., 2004; Vanhonacker and Verbeke, 2009; Verbeke, 2009; Verbeke and Viaene, 1999a) and by the possible perceived relation between the vaccine method and each of these attributes in consumers’ perception. The importance of “reasonably priced”, “tastes good”, “is healthy for your diet”, “animal was treated as humanely as possible” and “production of meat is done in an environmental friendly way” was measured on a seven-point Likert interval scale that ranged from “not at all important” (1) to “extremely important” (7). In order to account for country-specific and individual-specific scale use, the raw data were centred by individual. As such, the respondent’s original score was replaced by a relative individual score that ranges from -1 to +1. These relative product importance scores will be used to identify market segments in our sample and to investigate how possible market segments with different purchase motives differ in terms of acceptance of, preference for, and purchasing intention in relation to pork from vaccinated versus physically castrated pigs. Before the second part of the survey, participants were informed by means of a textual message about boar taint, and the two alternative methods (of interest in this study) to avoid the development of boar taint, i.e. physical castration with anaesthesia and the vaccine method (Table 3.7). After exposure to the information, participants were asked about their (aided) awareness of boar taint and the two alternative methods. Response categories were “I have never heard of it”, “I have heard of it but do not know much about it” and “I have heard of it and know about this issue”. Next, acceptance of the two methods was measured on a seven-point scale that ranged from “completely 112 Chapter 6 : Piglet castration – European citizens unacceptable” (1) to “completely acceptable” (7). Respondents’ confidence in the efficiency of the method with regard to eliminating boar taint was measured on a five point scale, ranging from “not at all confident” (1) to “very confident” (5). A five point scale was also used to measure the respondents’ likelihood of seeking out or avoiding pork produced using either method, where 1 corresponded with “definitely avoid” and 5 with “definitely seek out”. The scale’s midpoint was anchored with “neither seek out nor avoid”. This variable will be analysed as a categorical variable. Willingness to pay a moderate price premium for pork raised with either method was registered on a seven-point scale that ranged from “definitely would not pay a slight premium” (1) to “definitely would pay a slight premium” (7). Willingness to pay for pork from the vaccine method (physical castration method) was only registered for people who had expressed a preference for the vaccine method (physical castration method). Finally, two questions directly opposed the two alternative methods one to another. First, respondents were asked to indicate their preference on a seven-point scale that ranged from “strongly prefer the physical castration method” (1) to “strongly prefer the vaccine method” (7), with “neutral” being the scale’s midpoint (4). Second, they had to indicate their choice between the following four response categories relating to purchasing intention: “I will try to only eat pork produced using the vaccine method”, “I am happy to eat pork produced using either method”, “I will try to only eat pork produced using the physical castration method”, and “I will not eat pork using either method (i.e. I will no longer eat pork)”. 2.2. Statistical analyses Data were analysed using SPSS 15.0. Descriptive analyses were used to report and discuss the responses on the different questions on country-level. Hierarchical clustering with Ward’s method and squared Euclidean distance was performed using the product attribute importance scores on taste, price, health and ethics (animal welfare and environmental friendly) as segmentation variables, followed by a K-Means cluster analysis with initial cluster centres that resulted from the hierarchical procedure. The optimal number of clusters (so-called segments) was based on an increase in distance indices, together with a split-run procedure. Cross-tabulations with chi-square statistics are used to profile the segments in terms of socio-demographics and pork consumption frequency. Possible differences in terms of preference and acceptance measures between market segments and different socio-demographic consumer groups were analysed through bivariate analyses (one-way ANOVA and independent samples t-tests). 113 Part III Table 3.7. Textual information provided to participants Boar taint In sexually maturing male pigs, the developing testes can give rise to the accumulation of certain substances in the meat. These substances are released during cooking and can give meat an unpleasant flavour and odour described as sweat or urine. This characteristic is referred to as boar taint. Unfortunately, as much as 75% of the population is sensitive to boar taint, making tainted meat inedible. Currently, pork producers follow different procedures to ensure boar taint is not present in the pork before reaching the stores for purchase. Physical castration with anaesthesia Using this method, male piglets are physically castrated by the farmer within the first week of birth. Once their scrotum is cut with a scalpel, their testes are pulled out and cut off. To improve animal welfare, the technique of using anaesthesia during castration is under evaluation such as the use of a gas for inhalation anaesthesia or the injection of a local anaesthetic into the testes. Throughout this survey, please assume that anaesthesia is always used during physical castration. • Castrated male pigs eliminates 99%+ of boar taint (unpleasant favour/odour) in the meat • Castrated pigs are less aggressive and easier to manage on the farm than non-castrated pigs • Nevertheless, the open wound could become a source of infection or disease, which may lead to death • Pain and stress from physical castration may be reduced by using anaesthesia during castration, but may return once the anaesthesia wears off • The anaesthesia does not leave any detectable residue in the meat • Castration early in life reduces the efficiency of the male pig’s metabolism, which means that castrated pigs will eat more food and produce more environmental waste • Meat produced from castrated pigs tends to have more fat in the meat Vaccine against boar taint As an alternative to physical castration, male pigs can be given a vaccine to prevent boar taint. The vaccine works by stimulating the pig’s own immune system to create antibodies which temporarily limit the function of testes, preventing the accumulation of the substances that cause boar taint. This product is an injected vaccine, and is not a hormone or a drug. • This vaccine eliminates 99%+ of boar taint (unpleasant flavour/odor) in the meat • At the time of use, this vaccine will be approved by the [country] authorities • In eight years’ experience with the product in Australia and New Zealand, the vaccine has performed well and there have been no safety concerns • The vaccine does not leave any detectable residue in the meat • This method has no pain, stress or health consequences associated with it • Since the pigs are vaccinated late in life, this method allows the male pig to spend most of its life growing and maturing naturally, eating less food and producing less environmental waste than pigs that use other methods for the removal of boar taint • The meat from vaccinated pigs tends to be leaner than the meat from pigs that use other methods for the removal of boar taint • To use this method, farmers will need to change their current operating procedures and follow strict quality controls to ensure every male pig is properly vaccinated • In addition, farmers will need to use special safety vaccinators to minimise the risk of accidental self-injection 114 Chapter 6 : Piglet castration – European citizens 3. Results 3.1. Descriptive analyses Consumers’ aided awareness of boar taint and the methods to eliminate its development was very limited (Table 3.8). More than half of the sample indicated to have never heard of boar taint. The low level of awareness with regards to boar taint was most evident in France and Belgium. The vaccine method was completely unknown to a large majority of the sample and less known than the physical castration method. French and German consumers were the least aware of both methods. Overall, once consumers were made aware of the two alternative methods, the acceptance score for the vaccine method surpassed the acceptability of the physical castration method. Differences in absolute acceptance scores between both methods were most pronounced in Belgium, France and the Netherlands. German responses were characterised by a higher number of neutral answers on both questions. For both methods a similar degree of confidence was expressed with regard to eliminating boar taint. Absolute confidence scores were slightly in favour of the physical method in Germany, while the opposite was found in all other countries. Mean confidence scores were in the range of 3.2 to 3.5 on the five-point scale, thus on the positive side but with a high share of neutral responses (39.4% and 39.2% for physical castration and the vaccine method, respectively), which is in line with the reported unawareness about boar taint and both presented methods to prevent it. Similar results were found for the degree of seeking out or avoiding pork from either method. While consumers in all countries were strongly in favour of the vaccine method, this was the least pronounced in Germany. Responses to the question about preference for one method over the other were characterised by a relatively high share of the German sample preferring physical castration (18.2%) or taking a neutral position (20.7%). Except by German consumers, a negative willingness to pay a price premium for pork from physically castrated pigs was expressed (absolute score below 4; significantly in Belgium). Small positive willingness to pay figures were found with respect to the vaccine method (absolute scores above 4; significantly in France, Germany and The Netherlands). Directly probing for the respondent’s preference for one method over another resulted in a clear preference for the vaccine method. About 70 percent of the total sample expressed a preference for the vaccine method, while about 12 percent favoured the physical castration method. Finally, about half of the sample reported the intention to eat only pork from vaccinated pigs. A minority of between 11.6 percent in Germany and 3.8 percent in the Netherlands reported an intention to eat only pork from physical castrated pigs. 115 Part III 3.2. Segmentation analysis Consumer preferences do not stop at the country’s borders. Therefore, a cross-country segmentation analysis on the total sample has been performed in order to provide a more realistic and economically relevant picture of the market structure. Following the clustering procedure as described in the analysis section, a three-cluster solution came out as most optimal. The respective size of the clusters and the scores on the segmentation variables are reported in Table 3.9. Segment 1 (59.1 percent of the sample) can be typified as ethics-oriented consumers or citizens. The relative importance of the ethical attributes was the highest in this segment as compared to the other segments and these attributes were rated equally important as price and health. Taste received the highest relative importance score within this segment as compared to other attributes, but was the lowest as compared to the other segments. Given the high reported importance of ethical issues relative to other segments, and the contrast with market shares of meat with enhanced ethical characteristics, it can realistically be assumed that individuals belonging to this segment have reflected a citizen rather than consumer opinion (Krystallis et al., 2009; Vanhonacker et al., 2008). The second segment (21.7 percent of the sample) is strongly taste and price oriented, and reports a very low relative importance attached to ethical issues. Compared to both other segments a healthy diet is deemed relatively more important. Consumers belonging to the third segment (19.2 percent of the sample) focus solely on taste and price. Especially the minor importance of health is remarkable in this segment. The relative importance attached to ethical issues by segment 3 corresponded with the sample’s average. When the absolute scores of the segmentation variables were considered, mean scores for perceived importance of taste did not differ between the three segments (F=0.99, p=0.371); taste was considered the most important product attribute across all consumer segments. In segment 1, each of the four attributes was deemed very important, with reasonably prices receiving the lowest score in absolute terms. Segment 2 attributed very high importance to both taste and a reasonable price. A healthy diet, though perceived as important, was somewhat subordinate to both taste and price. Ethical issues were unimportant to this segment. Segment 3 differs from segment 2 because of its low rating on perceived importance of health. In this segment, taste and price appeared to be the most relevant attributes taken into account. Based on these profiles, the segments will further be referred as (1) “average, ethics-oriented”, (2) “low-ethics, health-oriented” and (3) “priceoriented”. Differences in country-distribution were mainly related to segment 2 and 3 (χ²= 99.94, p<0.001) (Table 3.10). Segment 2 was characterised by a higher share of Dutch and a lower share of Germans and French. The opposite was found for segment 3. Males were slightly overrepresented in segment 2 and 3; females in segment 1 (χ²=32.02, p<0.001). With regard to age, differences were mainly related to the youngest (< 35 years) versus the oldest age categories (> 55 years). The youngest belonged more to segment 2 and less to segment 1. The oldest on the other hand were found more 116 Chapter 6 : Piglet castration – European citizens in segment 1 and less in the two other segments. In terms of pork consumption frequency, the highest consumption frequency was found in segments 2 and 3 (χ²=27.04, p<0.001). Nevertheless, all respondents consumed pork regularly given that this variable was an inclusion criterion. Table 3.9. Profile of consumer segments on segmentation variables; relative (sum across attributes equals 4) and absolute importance scores Segmentation variables/ Segment 1 Segment 2 Segment 3 Sample mean* Reasonably priced 0.96 1.15 1.24 1.05 Good taste 1.08 1.22 1.27 1.15 Healthful 0.98 1.01 0.63 0.92 Ethical 0.98 0.62 0.86 0.88 Reasonably priced 5.73a (1.19) 6.10b (0.97) 6.22c (1.00) 5.91 (1.13) Good taste 6.40 (0.84) 6.42 (0.80) 6.37 (0.85) 6.40 (0.83) Attributes Relative scores Absolute scores** c Healthful 5.85 (1.03) c Ethical Segment size Name * ** b 5.38 (1.07) a a 5.25 (1.44) b 3.25 (1.08) 5.86 (0.95) 3.31 (1.03) 4.42 (1.42) 5.03 (1.50) n 2,386 873 772 4,031 % 59.1 21.7 19.2 Average, Low ethics, Price-oriented Ethics-oriented Health-oriented a-c Reference value; Scale: 1= not at all important, 7= extremely important; Scores in a row with different superscript are significantly different at p<0.05 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 117 Part III Table 3.8. Descriptive analyses for the different countries Pooled sample (n=4,031) BE (n=1,031) FR (n=993) DE (n=1,006) NL (n=1,001) 53.7 35.0 11.3 58.2 30.8 11.1 64.9 26.2 8.9 44.4 45.2 10.3 48.5 33.1 18.4 44.5 34.6 20.9 59.4 19.6 21.0 86.8 11.2 2.0 85.3 11.6 3.1 3.74 (1.74) 5.38 (1.61) Test statistic p-value 47.1 38.0 14.9 χ²=63.01 <0.001 52.7 33.5 13.8 37.7 44.6 17.7 χ²=169.56 <0.001 90.0 8.5 1.4 90.8 7.8 1.4 81.0 17.1 1.9 χ²=65.02 <0.001 3.84b (1.78) 5.62c (1.51) 3.48a (1.72) 5.39b (1.66) 3.91b (1.71) 4.83a (1.70) 3.71b (1.70) 5.69c (1.40) F=11.78 F=62.09 <0.001 <0.001 3.32 (0.96) 3.45 (0.90) 3.25a (0.99) 3.40a (0.98) 3.17a (0.97) 3.40a (0.92) 3.60b (0.97) 3.53b (0.92) 3.24a (0.90) 3.47ab (0.78) F=42.67 F=4.69 <0.001 0.003 15.4 50.9 33.7 16.6 52.7 30.7 13.8 50.8 37.0 20.5 47.0 32.5 10.5 55.0 34.5 χ²=48.49 <0.001 43.8 44.4 11.8 42.6 48.6 8.8 44.9 41.0 14.1 34.5 55.0 10.5 42.0 51.6 6.4 χ²=103.58 <0.001 3.85 (1.92) 4.47 (1.87) 3.25a (2.03) 4.03a (1.94) 3.80ab (2.00) 4.39b (1.93) 4.27b (1.75) 4.88c (1.67) 3.76ab (1.80) 4.62bc (1.80) F=6.44 F=25.61 <0.001 <0.001 5.50 (1.71) 11.9 18.5 69.6 5.55 (1.72) 11.1 17.4 71.5 5.59 (1.72) 11.7 16.8 71.5 5.10 (1.81) 18.2 20.7 61.1 5.77 (1.50) 6.5 19.0 74.5 F=28.33 χ²=78.22 <0.001 <0.001 50.1 36.3 7.5 6.0 51.9 37.6 5.7 4.7 52.4 31.7 9.0 7.0 46.1 35.7 11.6 6.5 49.8 40.3 3.8 6.0 χ²=69.05 <0.001 Awareness boar taint (%) Never heard of it Heard of, but don’t know much about Heard of and know a lot about Awareness physical castration (%) Never heard of it Heard of, but don’t know much about Heard of and know a lot about Awareness vaccine method (%) Never heard of it Heard of, but don’t know much about Heard of and know a lot about Acceptance (7-pt scale; Mean (SD)) Physical castration Vaccine method Confidence in elimination boar taint (5-pt scale; Mean (SD)) Physical castration Vaccine method Seek out vs. Avoid (%) Physical castration Would seek out Neutral Would avoid Vaccine method Would seek out Neutral Would avoid Willingness to pay (7-pt scale; Mean (SD)) Physical castration Vaccine method Preference (7-pt scale) Mean (SD) Preference for physical castration (%) Neutral (%) Preference for vaccine method (%) Pork consumption intention (%) Would eat only pork from vaccine method Would eat pork from either method Would eat only pork from physical castration Would stop eating pork a-c Scores in a row with different superscript are significantly different at p<0.05 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 118 Chapter 6 : Piglet castration – European citizens Table 3.10. Profile of the different segments in terms of socio-demographics, pork consumption frequency and absolute product importance scores Total * sample Segment 1 Segment 2 Segment 3 Average, Ethicsoriented Low ethics, Healthoriented Priceoriented Belgium France Germany The Netherlands 25.6 24.6 25.0 24.9 25.4 25.7 24.8 24.1 27.1 18.0 20.2 34.7 24.4 28.6 30.8 16.2 Male Female 49.6 50.4 46.0 54.0 56.3 43.7 53.2 46.8 18-24 years 25-34 years 35-44 years 45-54 years 55-64 years > 65 years 9.3 16.6 21.2 20.9 15.1 17.0 7.1 14.8 20.2 21.1 16.7 20.0 13.3 20.4 21.2 20.8 12.3 12.0 11.5 17.7 24.1 20.3 13.2 13.2 Pork consumption frequency 2 to 3 times a month Once a week More than once a week 17.6 36.8 45.6 18.6 39.2 42.2 16.6 32.9 50.5 15.7 33.9 50.5 Country (%) Gender (%) Age category (%) * Test statistics p-value χ²=99.94 <0.001 χ²=32.02 <0.001 χ²=91.21 <0.001 χ²=27.04 <0.001 Reference value 3.3. Between-group comparisons Central questions in the analyses were the respondents’ self-reported acceptance of both methods, the preference for one method over the other, the extent to which respondents would seek out or avoid pork from either method and their claimed pork purchasing intention. These questions will be related to the three market segments, the respondent’s demographic profile, awareness and confidence in the methods, in order to detect the underlying motivations for choosing one method over another. In the first part, the focus will be on acceptance and preference ratings; the second part will deal with the other two central questions that relate more to claimed purchasing intention. 3.4. Acceptance of both alternative methods Acceptance ratings for the physical castration method were not significantly correlated with acceptance ratings for the vaccine method (r=0.005, p=0.753). Thus, a high acceptance for one method did not necessarily involve a rejection of the other method. After a recoding of the original seven-point scale into a categorical three point scale (1=unacceptable, 2=neutral, 3=acceptable), it resulted that 73.4 percent of the respondents who accept the physical castration method also accept the vaccine method (n=942, 23.4 percent of the total sample). A minority of 3.8 percent of 119 Part III the sample (n=154) accepts the physical castration method, while rejecting the vaccine method. In contrast, 31.3 percent (n=1,261) accept the vaccine method and reject physical castration. Finally, 4.7 percent (n=188) rejects both methods. The different market segments differed in their acceptance of the physical castration (p<0.001), with the lowest acceptance score in segment 1 (“average, ethics-oriented”; μ=3.59) and the highest acceptance in segment 2 (“low-ethics, health-oriented”; μ=4.05). For all three segments, significantly higher acceptance scores appeared for the vaccine method as compared to physical castration, but no differences were found between segments for the acceptance of the vaccine method (p=0.110). Combining the two acceptance measures (in their categorical format) indicates that the “low-ethics, health-oriented” segment does not have a strong preference for one method over the other, as long as the elimination of boar taint is guaranteed. This segment accounts for 26.2 percent and 26.0 percent of the consumers who are neutral to or accepting both methods, respectively (Table 3.11). These shares are significantly different from the segment’s share of 21.6 percent in the total sample. Segment 1 (“average, ethics-oriented”) was more pronounced in its opinion, with a lower presence in the double neutral group (50.4 percent as compared to the segment’s share of 59.2 percent in the total sample). They appeared to be most rejecting physical castration, given a higher presence in response categories where physical castration was reported as unacceptable (around 65 percent as compared to the segment’s share of 59.2 percent in the total sample). Segment 3 was rather equally present across the different combinations. The acceptance scores were strongly linked with the respondents’ confidence in an effective elimination of boar taint. Further, acceptance scores for the physical castration method were higher among men (p<0.001) and did not differ between gender for the vaccine method (p=0.155). Age did not impact on the acceptance of physical castration (p=0.418), while the youngest respondents (1824 years) indicated a higher acceptance of the vaccine method than respondents aged above 54 years (p<0.001). Table 3.11. Combination of the responses on acceptance of physical castration method and acceptability of vaccine method Does not accept physical castration Does not accept vaccine method Neutral on vaccine method Accept vaccine method Neutral on physical castration Does not accept vaccine method Neutral on vaccine method Accept vaccine method Accept physical castration Does not accept vaccine method Neutral on vaccine method Accept vaccine method *Share in the total sample 120 Segment 1 (59.2%)* Segment 2 (21.6%)* Segment 3 (19.1%)* 65.1 65.1 63.8 15.9 15.1 18.6 19.0 19.7 17.6 67.3 50.4 59.5 16.3 26.2 22.9 16.3 23.3 17.6 60.8 55.9 53.2 21.6 23.4 26.0 17.6 20.7 20.8 Chapter 6 : Piglet castration – European citizens 3.5. Preference for the vaccine method over physical castration The preference score (the higher the stronger the preference is for the vaccine method) differed significantly between segments (p=0.004), with a lower mean score for segment 2 (“low-ethics, health-oriented”; μ=5.38) as compared to segment 1 (“average, ethics-oriented”; μ=5.58). In line with gender differences with respect to acceptance of both methods, a higher preference for the vaccine method was found among females (p<0.001). In a similar vein, a higher preference score was found among younger (18-24 years) consumers as compared to consumers aged above 45 years (p<0.001). No significant differences in acceptance and preference scores were found for different levels of pork consumption frequency (p>0.05). 3.6. Associations with claimed pork purchasing intentions On the question whether the respondent would seek out or avoid pork from either vaccinated or physically castrated pigs, about one third of the sample (30.3 percent) answered neutral for both methods. This reflects that the method of eliminating boar taint in rearing male pigs is not top of mind among a substantial share of consumers when purchasing pork. Nevertheless, 17 percent would avoid pork from physically castrated pigs while seeking out pork from vaccinated pigs. In contrast, only 2.6 percent would actively seek pork from physically castrated pigs and avoid pork from vaccinated pigs. The “average, ethics-oriented” segment 1 was underrepresented in the neutral groups, while more strongly represented mainly in the group that would avoid pork from physically castrated pigs and the group that would seek out pork from vaccinated pigs (Table 3.12). The “low-ethics, health-oriented” segment 2 was characterised by a higher presence in the neutral groups and an underrepresentation in the groups that would avoid or seek out either of both methods. Segment 3 could not be clearly profiled on these behavioural intention items. Table 3.12. Intention to seeking out and avoidance of pork from physically castrated and vaccinated male pigs, % of consumers Physical castration method Share in total Would sample seek out Segment 1 59.2 63.7 Segment 2 21.6 Segment 3 19.1 Vaccine method Would Would avoid seek out 52.0 68.0 66.2 51.2 62.9 15.8 27.6 15.3 16.3 28.3 16.6 20.5 20.4 16.7 17.5 20.5 20.4 Neutral Neutral Would avoid Similar results were found when the market segments were compared in relation to the question that asked about pork purchasing intentions (results not shown). The “average, ethics-oriented” segment 1 appeared more strongly in the groups with a pronounced opinion and thus less in the group that does not care about the method of castration. This indifferent response option was strongly reported by respondents from the “low-ethics, health-oriented” segment 2. Segment 2 was also most absent in the group that indicated to stop eating pork. 121 Part III Respondents who claimed to try to eat only pork from vaccinated pigs had the highest preference score (i.e. highest preference for the vaccine method), opposed to the lowest preference score for respondents who claimed to eat only pork from physical castrated pigs (Table 3.13). Positive acceptance scores (> 4) were found for the group that does not care about the method used. Within the group that indicates to stop eating pork, the lower acceptance score for pork from physically castrated pigs (μ=2.75) as compared to pork from vaccinated pigs (μ=3.62) reveals that rejection of physical castration is a much stronger potential motivation to stop eating pork than possible rejection of the vaccine method. Confidence in the effectiveness of the method appeared to be related to the preferred method and a low confidence in a proper elimination of boar taint seemed to be a motive for a refusal for eating pork. With respect to age, there was a tendency that more elderly would seek out pork from physically castrated pigs and more youngsters would seek out pork from vaccinated pigs. Men were more likely to seek out pork from physically castrated pigs and females to seek out pork from vaccinated pigs. Females were also more present in the group that claims to try to eat only pork from vaccinated pigs. Males were overrepresented in the group that would buy pork produced using either method. Table 3.13. Relation between claimed behaviour and attitude measures related to piglet castration Preference score Acceptance of physical castration Acceptance of the vaccine method Confidence about physical castration Confidence about the vaccine method a-d Would eat only Would eat pork Would eat only pork from the from either pork from physical vaccine method method castration c 4.68 b 4.50 d 5.25 b 3.51 c 3.35 6.55 3.07 6.00 3.20 3.75 b 2.15 c 5.30 c 3.34 c 3.77 b 2.63 Would stop eating pork a 4.88 b d 2.75 a 3.62 d 2.60 a 2.58 a b a a Scores in a row with different superscript are significantly different at p<0.05 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 comparison of mean scores) 122 Chapter 6 : Piglet castration – European citizens 4. Discussion This chapter contributes to the contemporary (anno 2010) debate on the castration of male piglets and has placed particular focus on the vaccine method (the alternative term used in this study for immunocastration) relative to physical castration with anaesthesia. This study allows drawing crossnational conclusions and herewith is an international comparison complement to the existing national studies that often differ from each other in their design and information provision, and are therefore hard to compare. In the present survey particular focus has been placed on the vaccine method relative to physical castration with anaesthesia, given that both methods are considered as possible and economically feasible solutions in the short-run to the reduction of boar taint (von Borell et al., 2009). In this study, pre-purchase consumer attitudes and claimed purchasing behaviour with regard to (pork from) the vaccine method and physical castration with anaesthesia were investigated. A study of this scale with cross-country comparison and segmentation analysis is unprecedented. Crosscountry differences observed in terms of awareness, acceptance and preference most likely reflect, first, (dis)similarities in terms of consumer familiarity with pig production and consumer interests when purchasing pork. The issue has been covered explicitly by means of the identification and profiling of cross-national consumer segments. Second, differences in the national pork markets and pork markets environments, including policies in relation to piglet castration and related communications, may explain another part of the observed cross-cultural differences. Detailed analysis and discussion of the latter market-environmental factors are outside the scope of this chapter. Descriptive analyses showed a low consumer awareness of boar taint, probably related to the effective elimination and the very rare prevalence of tainted meat, and corroborates earlier findings (e.g. Hennessy and Newbold, 2004) and findings from the previous chapter (Chapter 5). A corresponding low awareness was found for physical castration as a method to eliminate boar taint. The vaccine method was even more unfamiliar to the large majority of the sample. A slightly larger percentage of consumers who claimed to be well aware of the vaccine method was found in Belgium (though this percentage was still very low), which is most likely due to the later data collection as compared to the other countries. Low awareness corresponds with consumers and citizens dissociation from animal production and animal production practices (Harper and Henson, 2001). An (ever) increasing degree of urbanisation and industrialisation of animal production has led to a situation whereby consumers and citizens become less and less aware of how animals are actually reared and managed (De Tavernier et al., 2010). In this perspective, any publicity and information that consumers will receive about the current practice of physical castration of piglets will be important and determinant in shaping their opinion and will be possibly also influential on their purchasing behaviour. The sector is mainly concerned about false communications that relate the vaccine to a hormone treatment. Previous research already indicated the strong negative 123 Part III impact such communication can have on consumers’ attitudes and meat purchasing behaviour (Verbeke and Ward, 2001). No such indications of adverse consumer reactions to the vaccine method were found in the present study. This is in correspondence with a previous study in Sweden. Lagerkvist and co-authors (2006) concluded from a consumer choice experiment that food safety risks were subordinate to animal welfare concerns, which were, in turn, dominated by food quality (i.e. taste) concerns. In contrast, Fredriksen et al. (2010) attributed Norwegian consumers’ scepticism towards the vaccine method to fear for residuals in meat from vaccinated pigs and to their apparent contentment with the current local practice of physical castration with anaesthesia. Taste was the dominant product attribute in the pork purchasing process in our study. A segmentation analysis based on product attribute importance resulted in three distinct segments, which did not differ in their perceived importance of taste. This dominance was reflected throughout the survey findings, also in relation to the acceptance of, and preference for the vaccine method. Respondents were in the first place interested in meat with a high sensory quality (i.e. without boar taint). These findings herewith underscore the relevance of adequate monitoring of taste and sensory quality, including the prevention of off-flavours. Only a very small number of respondents (4.7 percent) evaluated both methods (physical castration and the vaccine method) as unacceptable after being informed about the methods. The stronger taste was dominating the pork purchasing process and the less important animal welfare was, the more neutral respondents were with regard to the method used to eliminate boar taint, as long as it is eliminated. Respondents preferring the physical castration method were not willing to pay a price premium for this meat. Respondents that preferred the vaccine method were only moderately willing to pay a price premium. Hence it seems that the acceptance of the vaccine method will rely more on an effective elimination of boar taint and no impact on product price, rather than on animal welfare considerations during purchase. Although increasingly important, animal welfare is not top of mind for the individual during the food purchasing process (Verbeke, 2009). The likelihood of consumers searching actively for pork that is produced using the vaccine method is thus rather small. This chapter’s conclusion of the impact of pork price determining method acceptance is different to the conclusion of de Roest et al. (2009), who researched the economic implications of the alternatives to physical castration without anaesthesia. They concluded that the improvement in feed efficiency will probably compensate the costs associated with the vaccine and therefore defined consumer acceptance of the vaccine method as a dependent factor for the economic feasibility of the method. Our findings indicate that acceptance and preference with respect to the vaccine method is not associated with consumers’ price-orientation. Combining the results of both studies argues in favour of the vaccine method. The exact role of the different product attributes in shaping the preference for the vaccine method can be further researched in specifically adapted (e.g. conjoint) research designs. 124 Chapter 6 : Piglet castration – European citizens 5. Conclusions This cross-country study demonstrates that the sector’s uncertainties with respect to consumer acceptance of the vaccine method in France, Germany, the Netherlands and Belgium are void. The vast majority of pork consumers in these four EU countries accept the vaccine method and strongly prefer this method over physical castration with anaesthesia, after being informed about both alternative methods for piglet castration. In addition, confidence that the vaccine method allows for a proper elimination of boar taint, and acceptance and preference for the vaccine method did not differ significantly between pork market segments that are ethics-, health- or price-oriented. These findings indicate that the adoption of the vaccine method is unlikely to result in marketplace loss of any particular target market. In contrast, physical castration with anaesthesia is significantly less preferred in general and even rejected by a substantial share of the pork consumers. The observed differences between the identified market segments suggest that in particular the majority segment of ethics-oriented consumers/citizens reacted negatively towards physical castration with anaesthesia. Furthermore, the study shows no evidence that providing consumers with information about the vaccine or immunisation method would evoke risk perceptions and adverse effects in terms of product acceptance. The conclusion of this cross-national European consumer study is that the vaccine method is the most preferred by consumers, irrespective of their main motivation for purchasing pork. Therefore, from a market differentiation perspective, the vaccine methods also presents itself as a more neutral alternative than physical castration with anaesthesia, in the sense that the application of the vaccine method might not differently affect different consumer segments in the pork market. 125 Chapter 7 : Piglet castration – Flemish producers Chapter 7 Piglet castration – Flemish producers Adapted from Tuyttens, F.A.M., F. Vanhonacker, B. Verhille, D. De Brabander, and W. Verbeke. 2011. Pig producer attitude towards surgical castration of piglets without anaesthesia versus alternative strategies. Research in Veterinary Science resubmitted 6 December 2010. Abstract In addition to the two previous chapters, that approached the theme of piglet castration from a consumer perspective, this chapter focuses on the opinion of the producer. This allows extracting possible differences in line with the impact of place-based variables. Also, pig producers are a key stakeholder in the discussion about more humane alternatives to the surgical castration of male piglets for avoiding the problem of boar taint, though their attitude has rarely been documented. This chapter reports the findings from a questionnaire among 160 Flemish pig farmers about their preference ranking of, and attitude towards surgical castration with (SCA) and without anaesthesia (SCN), immunocastration (IC), sperm sexing (SS) and raising entire males (EM). The farmers’ general preference was recorded as SCN>SS>SCA=IC>EM. The farmers’ self-reported knowledge of the strategies was associated with famers’ preference for SCA (negatively) and SS (positively). Herd size was correlated with farmers’ preference for SCA (negatively) and SS (positively). Farmers perceived surgical castration without anaesthesia as the most favourable strategy in terms of farm profitability, animal performance and effectiveness against boar taint, but they expected the lowest consumer acceptance for this strategy. Surgical castration with anaesthesia was ranked the least favourable in terms of labour conditions. Production of entire males was viewed as the least profitable and the least effective strategy. Sperm sexing was positively perceived, particularly in terms of labour conditions, animal welfare, effectiveness and expected consumer acceptance. The farmers’ opinion was quite homogeneous, especially regarding surgical castration without anaesthesia and production of entire males. Increasing our understanding of farmers’ perception will hopefully benefit communication about this intricate issue. 127 Part III 1. Introduction Annually, approximately 100 million pigs or about 83 percent of the total population of male piglets, are surgically castrated in the EU, predominantly without anaesthesia or analgesia (Fredriksen et al., 2009). De Roest et al. (2009) stated that within the EU this practice is very unlikely to be tenable in the future. Indeed, there are alternatives to this practice of which some are already implemented in certain countries. For example, the use of a local or general anaesthetic is mandatory in Norway, Switzerland and the Netherlands (for the domestic market only). With the possible exception of Lithuania, Hungary, Poland, Slovenia and Slovakia, however, anaesthesia is very rarely used in other European countries (Fredriksen et al., 2009). In some other countries, though, nearly all (Ireland and the UK) or the majority (Cyprus, Spain, Portugal) of the male piglets are not castrated but raised as entire males (Fredriksen et al., 2009). Immunocastration is used in various countries such as Australia and New Zealand and has recently been licensed for use in the EU. Other alternatives require more research and development before they can be implemented in practice. For example, artificially inseminating sows with sexed sperm so that they produce female offspring only has been suggested as a possible solution in the long term. Similarly, automated real-time sorting out of tainted carcasses at the slaughter line is theoretically possible but has not yet been proven to be ready for implementation in practice. There is a large body of research in which the pros and cons of these alternatives are compared with regard to animal welfare (reviewed by Prunier et al., 2006; von Borell et al., 2009), meat quality (reviewed by Lundstrom et al., 2009) and economic implications (reviewed by de Roest et al., 2009). Recently some studies have also been conducted on stakeholder attitudes towards the alternative strategies for dealing with the problem of boar taint. Most of these studies have focused on the attitude of consumers/citizens towards the vaccination method (Giffin et al., 2008; Hennessy and Newbold, 2004; Huber-Eicher and Spring, 2008; Lagerkvist et al., 2006). Studies that have focused on other alternatives or other stakeholder groups are rare. The attitudes of representatives of various stakeholder groups in the EU towards the different alternatives were investigated as part of the EU-PIGCAS project (EU FP6-2005-SSP-5A PIGCAS project nº 043969), but the results have not been published in a refereed scientific journal. Besides the general public, a key stakeholder is of course the pig producer who would have to implement the alternative strategy in practice. Fredriksen and Nafstad (2006) conducted a survey among Norwegian veterinarians and pig producers two years after the mandatory use of local anaesthesia for piglet castration was implemented. The effect of the anaesthesia was subjectively evaluated to be good by 54 percent of the veterinarians, but by only 19 percent of the pig producers. Two thirds of the veterinarians versus one third of the pig producers judged that the use of anaesthesia improved the level of pig welfare. On a much smaller scale Eijck et al. (2007) interviewed six organic farmers whose veterinarians had started using local anaesthesia. Four farmers agreed completely or partly that the use of anaesthesia was beneficial for the pig’s welfare, while the other two disagreed. A large proportion of the 174 European representatives of the mainstream pork production chain that were 128 Chapter 7 : Piglet castration – Flemish producers questioned as part of the PIGCAS-project were proponents to keep things as they are, at least until better solutions for the boar taint problem are available (Ouedraogo et al., 2009). They were, however, prepared to accept alternatives, provided that there is no extra cost or extra risk for them. Their mean order of preference for the different alternatives was SCN (surgical castration without anaesthesia) = SCA (surgical castration with anaesthesia) > SS (raising females only via sperm sorting) > IC (vaccination against boar taint) = EM (production of entire male pigs). Although valuable, this study had some important limitations. Responses were obtained via national contacts, from a limited and unbalanced number of organisations and from representatives rather than from individual farmers. They were not checked for accuracy and could not be considered representative at the level of the EU, nor the country itself. Moreover, the opinion of the representatives of the pork production chain may differ from the opinion of the pig producers themselves. The specific objective of the present chapter was, therefore, to use a similar questionnaire as the one used in the PIGCAS-project in order to investigate the attitude of individual Flemish (Dutch speaking part of Belgium where 95 percent of the Belgian pig production is situated) pig producers towards surgical castration and the various alternatives. These survey data also allowed investigation of the effect of variables such as farmer’s age, farmer’s knowledge about the alternatives and pig herd size on their self-reported attitude and perceptions. Insights with respect to farmers’ attitudes can be useful for future policy decisions in relation to possible market bans, and the introduction or adoption of one (or more) of the possible strategies to avoid boar taint. 2. Materials and methods Methodological details in terms of research approach and sampling procedure are provided in Part I, section 4, and apply to Study 4. 2.1. Survey The questionnaire assessed the attitude of Flemish pig farmers with regard to the routine practice of surgical castration of male piglets without anaesthesia/analgesia (SCN) and various alternative strategies to deal with the problem of boar taint. These alternatives included the surgical castration with anaesthesia (SCA), the vaccination against boar taint (also commonly known as immunocastration or the vaccine method) (IC), the production of entire males with detection and removal of carcasses with boar taint at the slaughter line (EM) and raising females only via sperm sexing (SS). The questionnaire was accompanied by a text providing background information on each alternative. This text was a Dutch translation from the English version drafted by the PIGCAS project (http://w3.rennes.inra.fr/pigcas). For every alternative, the method, the advantages and the possible disadvantages were described in a concise and objective way. 129 Part III The questionnaire consisted of two parts. In the first part, some questions were asked about the farm characteristics and management (e.g. number of sows, boars, fattening pigs and piglets, weaning age, slaughter age and weight, breed and gestating sow housing system) and the farmer (e.g. date of birth, gender, self-reported knowledge level about piglet castration and its alternatives). Self-reported knowledge level was assessed as the extent to which the respondent believed to be informed about the routine practice of castration and its alternatives, measured on a five point Likert scale that ranged from “I have almost no knowledge” (1) to “I am well informed” (5) about each alternative. The second part of the questionnaire contained statements intended to measure the farmers’ attitude towards the current practice of surgical piglet castration without anaesthesia and the alternative strategies. This part included exactly the same statements that had been asked to stakeholder representatives as part of the PIGCAS-study, with the exception of questions related to the production of entire males, which were not included in the PIGCAS-study. The statements are provided in full in Table 3.14 to Table 3.18. The respondents were asked to express their level of agreement on a five-point Likert scale, that ranged from “totally disagree” (1) over “neither disagree, nor agree” (3) to “totally agree” (5), concerning three to eight issue statements for each of the alternative strategies. Respondents could also choose for a “don’t know” option. Additionally, one specific question balancing immunocastration against the routine practice (SCN) was included and scored on a seven-point Likert scale that ranged from “IC is much worse than the routine practice” (1) over “neither worse, nor better” (4) to “IC is much better” (7). Farmers were further asked to rank the five practices in general from “most suitable practice” (1) to “least suitable practice” (5), as well as on six specific aspects (i.e. labour conditions, animal welfare, effectiveness against boar taint, production performance, farm profitability and consumer acceptance) from “most favourable practice” (1) to “least favourable practice” (5). 2.2. Statistical analyses Data were analysed with SPSS 15.0. Frequencies (as percentages) are presented in table format or as histograms. Ranking data from the practices are statistically compared using the non-parametric Friedman test with chi-square statistic and p-values below 0.05 as decision criterion for statistical significance. 3. Results 3.1. Sample profile The geographical distribution of the farmers closely matched with the partitioning in the SANITEL record. The age of the participating farmers ranged from 21 to 73 years, with a mean of 46.4 years (SD=10.0). On the majority of the farms (63.9 percent), piglets were weaned between 25 and 30 days of age. Slaughter age of fattening pigs varied from 90 to 125 kg live weight and was predominantly done in the range of 110 to 115 kg (73.6 percent of the farms). On about three 130 Chapter 7 : Piglet castration – Flemish producers quarters of the farms, fattening pigs were slaughtered between 24 and 30 weeks. On 62.3 percent of the farms, all gestating sows were housed individually. The number of pigs on the farm ranged from 0 to 1,393 sows, from 0 to 55 boars, from 0 to 4,900 piglets and from 0 to 10,000 fattening pigs. About 20 percent of the farmers considered themselves to be poorly informed about piglet castration and its alternatives and 62.5 percent reported to be reasonably or well informed. Only 10 percent found it advisable for Belgium to do pioneering work in Europe with respect to implementing alternative practices for piglet castration. The majority (56.9 percent) was not in favour of actively seeking out alternative strategies. 3.2. Surgical castration without anaesthesia Pig farmers in this Flemish sample reported a rather homogeneous opinion with respect to surgical castration without anaesthesia. Responses were skewed to one side of the scale for almost all statements dealing with SCN and corresponded with a favourable attitude towards the current practice SCN (Table 3.14). The strongest agreement was found for the statements that positively link SCN with avoiding aggression and mounting behaviour (84.4 percent strongly agreeing) and with a proper elimination of boar taint (82.5 percent strongly agreeing). Sixty percent of the farmers answered negative on the statement that male pigs should not be castrated because of the expected higher production cost. This indicates that interest in avoiding aggression, mounting behaviour and boar taint through castration outweighs the expected lower production costs associated with the production of entire males. Further, 72.5 percent of the sample agreed that castration is well endured by the animals. Very few pig farmers (2.5 percent) thought that male pigs should not be castrated in order not to impact on the animal’s integrity. Opinions were the least homogenous for the statement that suggested to sell pork from entire males at a lower price. 131 Part III Table 3.14. Attitude of Flemish pig farmers with regard to surgical castration without anaesthesia. Percentage of responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option 1 2 3 4 5 DK I would rather have males castrated than have a 10% increase in the incidence of “off-flavour” when eating pork 2.5 1.3 2.5 6.3 82.5 5.0 I would rather have males castrated than a high level of aggression and mounting behaviour in pigs 0.6 1.9 0.0 8.1 84.4 5.0 Castration of male pigs is so painful that I think it should be avoided 52.5 16.3 13.1 7.5 0.6 10.0 I think that male pigs should not be castrated because castrates need more resources to be raised and thereby they cost more to be produced 40.6 20.0 14.4 8.8 2.5 13.8 It is fair to sell pork from entire males at a lower price than pork from castrates because of the possibility that it is tainted 26.9 11.3 12.5 15.6 24.4 9.4 It is fair to sell pork from entire males at a higher price than pork from castrates because avoiding castration is good for the animal’s welfare 45.0 10.6 20.0 5.0 10.0 9.4 Castration of pigs is a very old practice which is well endured by the animals 3.1 3.8 10.0 22.5 50.0 10.6 I think males should not be castrated, because I believe that we should not change the natural state of the animal 64.4 13.1 11.3 1.9 0.6 8.8 3.3. Castration with anaesthesia and/or analgesia In general Flemish farmers expressed a negative attitude towards SCA. They particularly disagreed with the statements that related SCA to the worthiness of extra production (85 percent) and labour costs (65.1 percent) and with the suggestion that it should be performed by a veterinarian (82.6 percent) (Table 3.15). In addition, 84.4 percent they did not believe that the animal’s welfare benefits from the use of an anaesthetic, because of the stress associated with its administration. Also, a majority (54.4 percent) deemed the risk of harmful residues from the anaesthetic more important than the ethical consideration associated with castration without anaesthesia. Consequently, the farmers were not willing to bear any extra costs or risks, given that the search for alternatives is driven by public concern in terms of animal welfare, something that is not considered a problem at farmer’s level. In the case consumers or society were willing to bear the extra costs, a more positive farmers’ attitude towards SCA was recorded (agreement by 68.2 percent). 132 Chapter 7 : Piglet castration – Flemish producers Table 3.15. Attitude of Flemish pig farmers with regard to surgical castration with anaesthesia and/or analgesia. Percentage of responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option 1 2 3 4 5 DK 51.3 13.8 10.0 15.0 3.8 6.3 81.3 1.3 4.4 3.1 3.8 6.3 75.6 9.4 5.0 1.9 3.8 4.4 11.9 4.4 10.0 14.4 53.8 5.6 40.6 13.8 21.9 6.9 4.4 12.5 1.3 3.8 6.3 18.1 66.3 4.4 68.8 10.6 9.4 1.3 1.9 8.1 I think the extra work needed to anesthetise pigs before castration is worth doing I think only veterinarians should be allowed to apply anaesthetics prior to castration Though pain prevention in castrating the animals entails extra production costs, I think this extra is worth paying by the farmer Pain prevention in castrating the animals will certainly increase the production costs; however, I think the extra cost is worth paying by the consumer I am less concerned with the risk of residues in pig meat when castrated with anaesthesia, than with the idea that the animal has been castrated without anaesthesia I think the animals’ welfare does not improve with castration with anaesthesia, given that the animals are more stressed preceding the castration Even with anaesthesia, I think boars should not be castrated, because I believe that we should not change the natural state of the animal 3.4. Immunocastration The farmers’ opinions related to IC were less strongly skewed to one scale’s end and a higher share of the participants chose the neutral response category or the ‘don’t know’ option as compared to the statements for SCN and SCA (Table 3.16). Farmers seemed uncertain especially in relation to an effective elimination of boar taint (38.8 percent answering neutral or don’t know), a possible safety risk for the farmer during vaccination (34.4 percent) and to a possible safety risk following on consumption (41.7 percent). 71.3 percent were unwilling to bear any extra costs from this method, while only 13.7 percent still disagreed on the adoption of IC when extra costs were to be covered by consumers. The statement that ‘the production of entire males should be preferred to IC because the latter is unnatural’ was rejected by the majority of the farmers and indicates that the importance of naturalness was perceived subordinate to the wish for a proper elimination of boar taint (and probably less aggressive male pigs). When asked directly about their general opinion 133 Part III about IC as compared to SCN, about half the respondents was in favour of SCN, while a minority of 13.2 percent favoured IC (Figure 3.6). Table 3.16. Attitude of Flemish pig farmers with regard to immunocastration (IC). Percentage of responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option 1 2 3 4 5 DK I doubt that IC of males can be done by the farmer without safety risk, so I think we should avoid IC as an alternative to surgical castration 10.0 7.5 18.1 10.0 38.1 16.3 IC may entail an increase in the production costs; however I think any extra price is worth paying by the farmer 61.3 10.0 13.8 3.8 1.9 9.4 IC may increase the production costs; however I think any extra is a price worth paying by the consumer 10.6 3.1 12.5 15.0 46.9 11.9 Because I think IC may affect consumer health, I would rather have all boars castrated without anaesthesia 3.8 4.4 17.5 12.5 50.0 11.9 Because I think IC may affect consumer health, I would rather have all boars castrated with anaesthesia 33.1 12.5 26.3 10.0 2.5 15.6 I doubt that all vaccinated males will be free from boar taint, so I think we should avoid IC 5.0 5.0 16.3 16.9 34.4 22.5 Because I think IC is unnatural, I would prefer farmers to rear entire males 40.0 12.5 16.9 7.5 5.6 17.5 better IC much better 40 Percentage of the sample (%) 35 30 25 20 15 10 5 0 IC much worse worse somewhat worse neither worse nor better somewhat better Figure 3.6. General attitude of farmers towards immunocastration (IC) relative to attitude towards the routine practice of surgical castration without anaesthesia. Answers were registered on a seven-point scale ranging from IC perceived as (1) “much worse” to (7) “much better” 134 Chapter 7 : Piglet castration – Flemish producers 3.5. Sperm sexing In general, farmers’ responses were rather positive towards the idea of SS as an alternative practice, although a high number of neutral answers suggested some degree of uncertainty (Table 3.17). Similar to other alternative methods, farmers were particularly open to sperm sexing if the consumers would be willing to pay (agreement by 47.6 percent, disagreement by 18.8 percent). About half of the sample agreed that SS would be the best way to avoid castration and were as such in favour of further research to optimise this technique. Opinions about the statement that boars should be castrated because sexing sperm is unnatural were more diverse. Table 3.17. Attitude of Flemish pig farmers with regard to sperm sexing (SS). Percentage of Responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option 1 2 3 4 5 DK 12.5 6.3 17.5 18.8 28.8 16.3 8.8 5.6 11.3 23.1 35.6 15.6 21.3 9.4 18.8 11.3 25.0 14.4 SS is not cost effective for the farmer, but I think the extra price is worth paying by the consumer Sexing semen adequately would be the best way to avoid castration, so I think that techniques to do it properly need to be developed Because SS to raise females only is unnatural, I would rather have all boars castrated 3.6. Production of entire males Farmers were not in favour of EM as an alternative strategy. This was mainly concluded from three quarters of the sample expressing negative opinions on the issue statements which associated the production of entire males with an increased level of aggression and mounting behaviour and with losses from an improper elimination of boar taint (Table 3.18). Similar to the former alternatives, the farmers rejected the idea of having to bear any extra cost themselves, while their opinions were less adverse if the consumer was willing to bear the extra costs. 135 Part III Table 3.18. Attitude of Flemish pig farmers with regard to the production of entire males (EM). Percentage of responses are registered on five-point Likert scales that range from “disagree strongly” (1), over “neither agree, nor disagree” (3), to “agree strongly” (5), and an additional “don’t know” (DK) option 1 2 3 4 5 DK 0.6 3.8 10.0 17.5 57.5 10.6 64.4 10.6 5.0 4.4 5.0 10.6 15.6 8.8 12.5 18.1 33.1 11.9 3.1 5.0 13.1 7.5 56.3 15.0 The animal’s welfare is harmed too much through an increased level of aggression and mounting behaviour of entire males This practice can cause increased costs at slaughter because of sorting out tainted carcasses, but I think the extra cost is worth paying by the farmer This practice can cause increased costs at slaughter because of sorting out tainted carcasses, but I think the extra cost is worth paying by the consumer I don’t fancy raising entire males 3.7. Ranking the alternatives Overall, the current practice of SCN was ranked as the most suitable practice by the pig farmers involved in this study (Table 3.19). Nearly two thirds (62.5 percent, result not shown) of the farmers gave the best rank to SCN. A large chi-square value indicated a pronounced difference in the general ranking of the practices (χ²=224.36). Further, a large chi-square value was associated with a proper elimination of boar taint and with farm profitability (respectively χ²=244.72 and χ²=122.05), two characteristics on which SCN received the most favourable mean ranking. As such, these two characteristics most probably contributed a lot to the general ranking of the alternatives by the farmers, possibly together with expected benefits in terms of reduction of aggression and mounting behaviour (favourable ranking on labour conditions, especially as compared to the ranking of the production of entire males). In contrast, SCN was ranked the least favourable alternative in terms of animal welfare and expected consumer acceptance. SS was perceived as the most suitable alternative. It ranked second best (after SCN) regarding effectiveness against boar taint, performances and profitability and first on the other characteristics. The general ranking for SCA and IC did not differ. Compared with SCN, SCA ranked particularly less favourable in the farmers opinion with respect to labour conditions (mean rank of 2.59 for SCN versus 4.13 for SCA). The farmers believed this latter practice would be better accepted among consumers (mean rank of 3.94 for SCN versus 2.67 for SCA), while only a minor difference in mean ranking appeared regarding animal welfare (3.51 for SCN versus 3.29 for SCA). IC ranked second for animal welfare, but was attributed a lower mean ranking for perceived effectiveness for the elimination of boar taint and expected consumer acceptance (respectively 3.82 and 3.24). Finally EM was considered the least suitable practice in general, mainly due to an unfavourable ranking for farm profitability, performance indicators, and a proper elimination of boar taint. 136 Chapter 7 : Piglet castration – Flemish producers Table 3.19. Mean rankings of the different practices to eliminate boar taint. A lower score corresponds with a more positive ranking (SCN: surgical castration without anaesthesia; SCA: surgical castration with anaesthesia; IC: immunocastration; SS: sperm sexing; EM: production of entire males) SCN SCA IC SS EM General Labour Animal welfare Effective against boar taint Performances Profitability Expected consumer acceptance 1.67 3.14 3.28 2.68 4.24 2.59 4.13 2.95 1.97 3.36 3.51 3.29 2.81 1.99 3.40 1.84 2.59 3.82 2.03 4.71 2.42 3.21 3.22 2.68 3.47 1.62 3.22 3.38 2.82 3.96 3.94 2.67 3.24 2.16 2.99 χ² 224.36 108.16 64.49 244.72 30.09 122.05 70.95 p <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 Non-parametric Friedman Chi-square tests are used to statistically compare differences in mean rankings within each column. 3.8. Influencing factors Age, farm size and self-reported knowledge significantly influenced the rank order of the different alternatives (p < 0.001) (Table 3.20). SCN was ranked as the most suitable practice in each category, EM as the least suitable practice. In general, a similar trend in mean rankings of the practices was seen between the categories within each of the farm or farmer characteristics. Regarding the age of the farmer, SCN was ranked slightly better (i.e. lower rank) with increasing age (1.75 for the youngest versus 1.58 for the oldest age category). Farmers younger than 41 years indicated a better ranking for IC as compared to older farmers (2.85 versus 3.44 and 3.39). The ranking of SS improved with increasing farm size (as approximated by the number of sows) (from 3.24 to 2.43), mainly at the expense of the rank order of SCA. Similarly, a greater self-reported level of knowledge was associated with an improved rank order of SS (from 3.03 to 2.54) and a worsened rank order of SCA (from 2.58 to 3.33). 137 Part III Table 3.20. Influencing factors on the rankings of the different practices to eliminate boar taint. Figures present mean rankings, where a lower score corresponds with a more favourable ranking (SCN: surgical castration without anaesthesia; SCA: surgical castration with anaesthesia; IC: immunocastration; SS: sperm sexing; EM: production of entire males) Age farm manager Farm size (# sows) Self-reported knowledge < 41 41-50 51+ 0 1-99 100-170 171+ Low Neutral High SCN 1.75 1.70 1.58 1.68 1.56 1.59 1.83 1.74 1.45 1.71 SCA 3.25 3.14 3.07 2.82 3.05 3.23 3.49 2.58 3.10 3.33 IC 2.85 3.44 3.39 3.11 3.27 3.45 3.24 3.48 3.14 3.25 SS 2.88 2.59 2.63 3.24 2.83 2.43 2.22 3.03 2.76 2.54 EM 4.28 4.13 4.33 4.16 4.29 4.30 4.19 4.16 4.55 4.17 χ² 52.62 84.48 93.18 48.23 63.06 74.71 53.47 41.45 56.88 136.64 p <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 Non-parametric Friedman Chi-square tests are used to statistically compare differences in mean rankings within each column 4. Discussion and conclusion Increased public concern about animal welfare issues has brought many livestock production practices into discussion, among them surgical piglet castration without anaesthesia. Research on alternative strategies to deal with the problem of boar taint has highlighted surgical castration with anaesthesia and/or analgesia and immunocastration as feasible solutions in the short run and no castration as the final target in the long run. The latter can be achieved through sperm sexing (female offspring only), genetic selection of breeds with no boar taint or sorting the tainted carcasses at the slaughter line (Prunier and Bonneau, 2006), although additional research is needed before these alternatives can be implemented on a large commercial scale. The pig production chain is composed of different stakeholders with different interests and different attitudes. As such the marketplace success of any alternative method is likely to depend on a positive acceptance by stakeholders throughout the chain and by pig producers in particular. This study has quantified the attitude of pig farmers in Flanders, Belgium, about current practice of surgical castration without anaesthesia and its alternative short and long term strategies. In general, the farmers preferred the current practice. This preference was mainly grounded on an expected reduced level of aggression and mounting behaviour among pigs as a positive consequence of castration, on delivering a high quality pork (i.e. without boar taint), and on perceived economic benefits (e.g. farm profitability). These issues were perceived negatively in relation to the production of entire males. As such, the latter practice was strongly rejected by a large proportion of the pig farmers. Sperm sexing was attributed the second best rank, followed by surgical castration with anaesthesia and immunocastration. The use of anaesthetics or analgesics was not preferred, mainly owing to its expected additional costs in terms of higher expenses and labour conditions, while producers did not consider it better for the welfare of the piglets. Compared to 138 Chapter 7 : Piglet castration – Flemish producers other practices, farmers chose more frequently for the neutral or ‘don’t know’ response category in case of sperm sexing and immunocastration, which can be an indication of a higher degree of unfamiliarity with or uncertainty about these practices. Sperm sexing is not yet ready to be put in practice and immunocastration was, at the time of the survey (end of 2007 – beginning of 2008), still a rather new concept on the Belgian market, which may have triggered a more reserved response behaviour. The rank order obtained from the present study deviates to some extent from the rank order obtained in the PIGCAS-study (SCN=SCA>SS>IC=EM, Ouedraogo et al., 2009), which was based on European pig producer representatives. The main differences are related to a better ranking of SCA and a worse ranking of SS and IC among the stakeholder representatives (PIGCAS-study) versus the actual producers (present study). The farmers reported a quite homogeneous opinion, especially with regard to surgical castration without anaesthesia (positive) and the production of entire males (negative). Indeed, attitudes were often skewed to one side of the scale. Such a homogeneous opinion corresponds with earlier surveys among farmers (e.g. Boogaard et al., 2006; Te Velde et al., 2002; Vanhonacker et al., 2008). Farmer’s perception is clearly distinct from the public’s perception. The public is mainly led by ethical and moral considerations such as animal welfare concerns and societal acceptability (Vanhonacker et al., 2009). The resulting farmers’ low importance for the animal’s naturalness corroborates with literature about conceptions of animal welfare, where animal welfare is mainly related to a good functioning of the animal in the view of farmers (e.g. Chapters 3 and 4). The farmers’ perception did not always match with scientific evidence. For instance, surgical castration without anaesthesia was ranked most favourable in terms of production performance, while growth parameters and feed efficiency have been shown to be superior in entire males (de Roest et al., 2009). Also the farmers’ expected lower consumer acceptance (Vanhonacker et al., 2009), reduced profitability (de Roest et al., 2009), and uncertainty with respect to consequences on human health (Clarke et al., 2008) and to the elimination of boar taint related to immunocastration (Jaros et al., 2005) have been contradicted by the referenced scientific studies. Interestingly, farmers did not associate the current practice of surgical castration without anaesthesia with a strong impairment of animal welfare. They did not judge the use of anaesthesia or the production of entire males as superior alternatives for the welfare of the pigs, and yet they believed that both alternatives would be distinctly more acceptable among consumers. This could explain their reported preference for the current practice of surgical castration without anaesthesia, especially given that the search for alternative practices stems from public concern related to animal welfare. This also supports their rejection to bear any costs associated with a possible adoption of the alternatives, while costs borne by consumers or society increased farmers’ willingness to adopt new strategies to avoid boar taint. Based on these findings we predict that resistance and dissatisfied farmers’ reactions can be expected if a ban on the current practice of surgical castration without anaesthesia were to be implemented. A general voluntary producerinitiated shift away from the current practice seems unlikely without additional scientific arguments, economic incentives or coercion. 139 Chapter 8 : Modifiers of societal concerns Chapter 8 Modifiers of societal concerns related to stocking density Adapted from Vanhonacker, F., W. Verbeke, E. Van Poucke, S. Buijs, and F. A. M. Tuyttens. 2009. Societal concern related to stocking density, pen size and group size in farm animal production. Livestock Science 123: 16-22. Abstract In the following two chapters, the discussion on attitude modifiers goes beyond place-based variables and also discusses the impact of other social structural variables and individual animalrelated experiential variables. In the present chapter, this discussion is embedded in the context of space allowance for farm animals, as one of the most topical and vividly debated sub-matters of farm animal welfare. As such, this chapter focuses on societal concerns related to space allowance in Flemish animal production. It reports a quantitative analysis of cross-sectional surveys conducted among citizens of Flanders during 2000-2002 (n=521) and during 2006 (n=459). Compared with other animal welfare aspects, the survey participants perceived stocking densities and pen sizes as very essential for attaining an acceptable level of farm animal welfare. In addition, they evaluated stocking densities and pen sizes in current Flemish animal production as inadequate. Combined these results indicate that societal concern, defined as the difference between perceived importance and evaluative belief, is very high for stocking density and pen size, though less for group size. Societal concern about stocking density was found to be higher among younger people and households without children, while societal concern was not significantly associated with living environment (rural versus urban). Gender (higher concern among women in dataset 2006) and educational (higher concern among higher educated people in dataset 2000-2002) differences were not consistently significant in all datasets. The findings from this study suggest that larger gains in terms of animal welfare image among the public can be realised from providing farm animals with more space relative to reducing animal group size. 141 Part III 1. Introduction The importance of farm animal welfare is recognised by all stakeholders in the farm animal production chain including scientists, government, retailers, producers, consumers and society as a whole (Bracke et al., 2005). This has led to a substantial body of animal welfare related research originating from varying scopes and resulting in many – often conflicting – definitions of and approaches to welfare. Within these various studies the aspect of space allowance in stables or confinements is an often recurring theme, irrespective of the animal species considered, the science research perspective or the stakeholder at word. At the institutional level, EU welfare legislation has already introduced some abolitions and/or obligations linked to minimal space allowance for farm animals. At retailers’ level, examples of efforts relate to the establishment of quality assurance schemes and labels claiming more attention for animal welfare. Almost invariably, issues relating to space are a cornerstone of such quality assurance schemes. Farmers also increasingly consider space allowance as an important issue in obtaining an acceptable level of farm animal welfare and end product quality (Bock and van Huik, 2007; Borgen and Skarstad, 2007; Hubbard et al., 2007; Kling-Eveillard et al., 2007; Menghi, 2007). However farmers face a difficult trade-off between economic interests and animal welfare-related concerns (Appleby, 2004; McInerney, 1991). Specifically, the optimal space allowance per animal from the perspective of animal welfare and health is usually larger than that for maximising farm economic profits. In addition, severe (inter)national competition on animal food product markets pushes farmers to produce at minimal costs. As a result, it is neither self-evident that farmers voluntarily house their animals at a stocking density that is sub-optimal in terms of economic profitability, nor that mandatory systems with lower densities are readily accepted. With regard to the citizen’s view related to farm animals’ available space, very little research is currently available (e.g. Hall and Sandilands, 2007), in contrast to the vast amount of studies concentrating on individuals’ (citizen or consumer) perceptions about farm animal welfare in general (Frewer et al., 2005; Harper and Henson, 2001; Heleski et al., 2006; Kanis et al., 2003; Maria, 2006; McGlone, 2001; Phan-Huy and Fawaz, 2003; Te Velde et al., 2002; Verbeke and Viaene, 2000) and to many studies evaluating the effect of stocking density on welfare from the animal’s point of view (Estevez, 2007; Postollec et al., 2008; Spinu et al., 2003; Spoolder et al., 1999; Zimmerman et al., 2006). Despite criticism towards focusing on citizen opinions due to the so-called “citizen-consumer duality”, the proposition of citizens being hypocrite is refuted (Dagevos and Sterrenberg, 2003). Consequently, citizen perceptions cannot simply be ignored despite their often loose relationship with marketplace behaviour, even more so taking into account the trend of reversal of the food chain, which has shifted from a supply orientation to a demand orientation. 142 Chapter 8 : Modifiers of societal concerns Focusing on space related issues is topical since the lack of space is one of the issues most frequently brought into charge by animal welfare organisations and subject to a proportionally high share of media coverage. It is also one of the most tangible and visible features of animal production systems, and therefore highly relevant in a society that is largely alienating from agriculture and farm animal production practices. In this chapter available survey data from Flemish citizens, collected during the period 2000-2006 in different waves and with different purposes, are explored in order to gain insights in the societal concern and perception of space allowance as an animal welfare issue. First, it was explored to what extent people associate space allowance with animal welfare and to what extent they consider it a problem or not in current Flemish animal production. Second, since citizens can not be considered one homogenous group regarding their concern towards animal welfare, significant differences between socio-demographic citizen groups will be assessed within each dataset. Finally, starting from the theoretical assumption that the perception related to pen size and group size contribute to the perception of stocking density, the extent to what pen size or group size perceptions weigh more heavily in people’s formation of beliefs relating to stocking density is examined. 2. Material and methods 2.1. Sample and procedure Cross-sectional data were collected through a series of surveys using self-administered questionnaires in Flanders in four different years (2000, 2001, 2002 and 2006). The data collected in 2000, 2001 and 2002 corresponds with Study 1, described in Part I, section 4. The data gathered in 2006 corresponds with Study 2 in the same section. Owing to the different study objectives, a significantly different between-sample gender composition was found (χ²=10.7; p=0.001). In dataset 1, respondents were responsible for food purchasing, which is clearly reflected in the gender distribution (female majority), while a representative fifty-fifty gender distribution was obtained in dataset 2. The mean age and the share of urban respondents corresponded reasonably well with the population census data (mean age: 40.2 years; 36.3 percent living in urban areas (FOD Economy)) and did not differ significantly between both datasets (t=0.78; p=0.438 and χ²=0.55; p=0.459, respectively). The proportion of households with children was somewhat higher in dataset 1 as compared to dataset 2 (χ²=4.46; p=0.035), though more closely matching the population composition of about 60 percent. Finally, education level was higher in dataset 2 than in dataset 1 (χ²=12.38; p<0.001) and biased towards higher education. 143 Part III 2.2. Dependent variables The surveys probed for the respondents’ perceived importance and evaluative belief concerning the aspect of stocking density (number of animals housed per m², i.e. the concept of space allowance) as well as of its two components - pen size (in both datasets) and group size (only in dataset 2) – relative to a list of other aspects related to the concept of farm animal welfare. Both measures reflect citizens’ perceptions or beliefs, without linking to potential marketplace behaviour as consumer. Perceived importance refers to individual’s perception of the relative importance or relevance of these aspects for the realisation of an acceptable level of farm animal welfare. In dataset 1, perceived importance was measured as: “According to your personal opinion, to what extent do the following acts or factors have to do with animal welfare”, and was probed on a five point Likertscale, ranging from “has nothing to do with animal welfare” (1) to “has a lot to do with animal welfare” (5). In dataset 2, the equivalent question was formulated as: “According to your personal opinion, how important is this factor or act in obtaining an acceptable level of farm animal welfare”, again probed on a five point Likert-scale but now ranging from “totally unimportant” (1) to “very important” (5). The difference in phrasing between both datasets possibly induces bias, which imposes limits on the comparability. However, the assumption is that the between-dataset response style for this measure is systematic, hence not preventing the comparison of the ranking of the aspects under consideration. Evaluative belief refers to a “good-bad judgment” (Zajonc, 1980) or to an evaluative opinion about what a specific issue is or ought to be, more specifically animal welfare related aspects in the present study. Evaluative belief was measured in dataset 1 as: “To what extent do you think the following acts or factors influence the welfare of the animals in current Flemish animal production?”, probed on a five point Likert-scale, ranging from “has a very negative impact on animal welfare” (1) to “has a very positive impact on animal welfare” (5). In dataset 2, this question was formulated as: “What is your personal evaluation about this factor or act: to what extent do you think this factor or act is a problem for the welfare of animals in current Flemish animal production?”, also probed on a five point Likert-scale, ranging from “is a big problem” (1) to “is no problem at all” (5). In this case, the different phrasings include a clear difference in connotation. In dataset 1, we asked the respondent to make an overall judgment about the actual state (hence relating rather to an overall image), while in dataset 2, we asked for the respondent’s personal evaluation or personal image. Hence, comparison of mean values and magnitude between both datasets should be handled very carefully. Since the perceived importance and the evaluative belief scores were probed for in adjacent positions (physically in the questionnaire) and scored on the same type of scale (five point Likertscale), the difference between both scores was calculated and used as a proxy of the degree to which the aspect is evoking concern among individuals. “Concern” refers on one hand to 144 Chapter 8 : Modifiers of societal concerns “something that is of interest or importance to a person” and on the other hand associates with “anxiety or worry” (source: http://www.thefreedictionary.com/concern). The former association is covered by a positive relationship with the perceived importance measure, the latter by an inverse relationship with the evaluative belief measurement. High perceived importance combined with low (or negative) evaluative belief (positive difference score) correspond with a high concern (Figure 3.7). Aspects with such a positive difference score are subject to a strong perceived potential improvement, which is perceived as badly needed. Aspects for which a small and/or negative difference score is obtained can have several meanings. First, the aspect can have a high score on both measurements, i.e. they are both perceived as very important and evaluated as good. Such aspects have the potential to contribute substantially to the positive image of farm animal welfare but should be preserved from negative news seen their high perceived importance (“positive attention point”). Second, if an aspect has a low perceived importance score and a good judgment, it evokes “no particular concern”. Third, aspects that receive a low perceived importance and a negative evaluation only have a low potential for contributing to the improvement of citizens’ perception of farm animal welfare (“not an issue”). Stocking density and pen size (dataset 1 and dataset 2) or group size (dataset 2) were included among a list of other relevant aspects with respect to farm animal welfare for which the respondents’ perceived importance and evaluative belief are assessed. In dataset 1, the list was composed of 23 aspects that were selected based on literature review (Table 3.21). In 2006, a more elaborated list of in total 72 aspects was used based on four focus group discussions with citizens and a profound literature review (see Part II). PerImp High Positive attention point High concern Neutral EvalBel Good / positive judgment Bad / negative judgment Not a major issue Low / no particular concern Low Figure 3.7. Conceptualisation of concern as a combination of perceived importance (PI) and evaluative belief (EB) 145 Part III 2.3. Statistical analyses Data were analysed using SPSS 15.0. Bivariate analyses through comparison of mean scores, i.e. independent samples t-tests and one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores, were used to assess association between the interval scaled space related aspects on one hand and the categorical scaled socio-structural variables on the other hand. Stepwise linear regression analysis was used to assess which of the components of stocking density (i.e. pen size or group size) most strongly influences citizen perceptions of stocking density. 3. Results 3.1. Relative importance of stocking density, pen size and group size In dataset 1, pen size and stocking density were attributed the second and third highest perceived importance score among the 23 aspects, only headed by hygiene in the barn (Table 3.21). The aspects pen size and stocking density were ranked 11th and 12th respectively, in terms of evaluative belief. When ranking all 23 aspects from highest to lowest concern (i.e. perceived importance score minus evaluative belief score), stocking density and pen size ranked third and fourth (Table 3.21). The highest concern was expressed for the transport of animals and the (un)loading of animals. Table 3.21. Perceived importance, evaluative belief, and concern scores for 23 animal welfare aspects in dataset 1 (n=521); mean scores with standard deviation on 5-point scale; aspects are ranked according to their (descending) concern score Transport Loading / Unloading Stocking density Pen size Growth promoters Stunning Slaughter Medication Feed ad libitum Breeding Outdoor access Hygiene in barn Barn Climate Feed composition Litter Group housing Lairage before slaughter Ventilation Flooring type Atmospheric humidity Temperature of surroundings Artificial insemination Selection 146 Perceived importance 4.15 (0.85) 4.16 (0.84) 4.44 (0.72) 4.47 (0.69) 3.48 (1.35) 4.13 (0.93) 4.12 (0.99) 3.79 (1.03) 3.49 (1.14) 3.70 (1.14) 4.38 (0.74) 4.47 (0.67) 4.20 (0.76) 4.08 (0.89) 4.11 (0.77) 4.00 (0.85) 3.61 (1.11) 4.16 (0.73) 3.93 (0.90) 4.01 (0.79) 4.06 (0.81) 2.98 (1.17) 2.99 (1.09) Evaluative belief 2.82 (1.19) 2.84 (1.17) 3.20 (1.34) 3.27 (1.34) 2.36 (1.11) 3.13 (1.14) 3.11 (1.15) 2.78 (1.11) 2.59 (1.05) 2.83 (1.04) 3.53 (1.29) 3.67 (1.14) 3.47 (1.03) 3.40 (1.13) 3.43 (1.02) 3.32 (1.07) 2.95 (0.99) 3.54 (1.00) 3.31 (0.99) 3.45 (0.94) 3.52 (0.94) 2.85 (0.86) 2.96 (0.84) Concern 1.33 (1.49) 1.32 (1.44) 1.24 (1.53) 1.19 (1.50) 1.11 (1.70) 1.00 (1.50) 1.00 (1.52) 1.00 (1.45) 0.90 (1.46) 0.88 (1.51) 0.86 (1.44) 0.81 (1.25) 0.73 (1.22) 0.69 (1.38) 0.67 (1.19) 0.67 (1.31) 0.64 (1.48) 0.62 (1.15) 0.61 (1.21) 0.56 (1.08) 0.55 (1.13) 0.14 (1.46) 0.05 (1.35) Chapter 8 : Modifiers of societal concerns In dataset 2, in particular stocking density’s perceived importance was very high (ranking 7th on 72 aspects) and also pen size ranked relatively high (14th). Group size, however, ranked only 58th. The top aspects were related to the availability of feed and water, the relationship between the farmer and his animals, animal health, and pain and stress related aspects (for more details, see Chapter 3). Most importantly, stocking density was attributed the lowest mean evaluative belief score (i.e. perceived as the most problematic aspect in relation to farm animal welfare) of all 72 aspects, while pen size also scored as highly “problematic” (6th). Group size was ranked in the middle (41th) (Table 3.22). Of all aspects, stocking density received the highest concern; pen size also ranked very high (4th), while group size ranked 55th. 3.2. Influence of socio-structural variables In order to account for differences in question phrasing and scale use between both datasets, the results of the two datasets for the aspect “stocking density” will be discussed separately (Table 3.23). Living environment, representing a place-based variable, did not associate significantly with levels of concern about stocking density. Regarding other social structural variables, no significant effect of gender on concern towards stocking density was found in dataset 1, despite a more negative evaluative belief among males. In contrast, in dataset 2, a highly significant effect of gender was found, with a higher concern among women. Women reported both a higher perceived importance and a more negative evaluative belief. Concerning age, a significant effect on concern was found in both datasets. In dataset 1, the concern decreased from the youngest to the oldest age category. In dataset 2, concern scores differed between the youngest two categories and the oldest two categories. This difference was based on a more negative evaluation among the younger group, while no significant difference in perceived importance score was found. This age-effect was also reflected in the ranking of stocking density relative to the other welfare related aspects in dataset 2 (results not shown). Stocking density was ascribed the highest level of concern among the first three age groups, while it only ranked seventh in the oldest group. Both datasets revealed a higher level of concern among households without children. In dataset 1, the association was resulting from a more negative evaluative belief, while in dataset 2, households without children both perceived stocking density more important as well as evaluated it as more problematic compared to households with children. Finally, dataset 1 revealed a significant higher concern among higher educated people, while no association with education was detected in dataset 2. 147 Part III Table 3.22. Perceived importance (PerImp), evaluative belief (EvalBel) and concern (Concern) scores of the 72 animal welfare aspects listed in dataset 2 (n=459); mean (standard deviation) on 5-point scale; aspect are ranked according to their (descending) concern score Stocking density Stress Slaughter without pain or stress Pen size Pain by human intervention Handling of animals Natural growth rate Thirst during transport Space during transport Air quality Duration of transport Outdoor access Natural behaviour Disease Daylight Natural environment Skilled animal handlers Fear Respect for animals Procedure of (un)loading Body care Transport of living animals Shockproof and calm transport Explorative behaviour Hygiene in the barn Comfort Stunning Availability of water Ability to rest Natural feed Pain by conspecifics Balanced feed Curative medication Climate during transport Social behaviour Freshness of feed 148 PerImp 4.28 (0.74) 4.15 (0.86) 4.30 (0.86) 4.16 (0.75) 4.23 (0.87) 4.07 (0.84) 4.04 (0.90) 4.10 (0.89) 3.97 (0.86) 4.16 (0.73) 4.03 (0.86) 4.02 (0.96) 4.00 (0.88) 4.37 (0.71) 4.08 (0.85) 3.88 (0.96) 4.29 (0.70) 4.04 (0.85) 4.39 (0.71) 3.97 (0.80) 3.89 (0.97) 3.93 (0.89) 3.90 (0.86) 3.86 (0.92) 4.29 (0.68) 3.99 (0.93) 4.01 (1.08) 4.56 (0.62) 4.05 (0.75) 3.88 (1.01) 3.96 (0.79) 4.17 (0.71) 4.23 (0.66) 3.81 (0.88) 3.84 (0.90) 4.24 (0.73) EvalBel 2.11 (1.03) 2.22 (0.99) 2.36 (1.09) 2.29 (1.07) 2.43 (1.04) 2.28 (0.99) 2.29 (1.09) 2.37 (0.96) 2.24 (1.01) 2.52 (0.94) 2.46 (1.03) 2.46 (1.08) 2.45 (1.05) 2.82 (1.03) 2.54 (1.01) 2.38 (1.08) 2.79 (1.03) 2.55 (0.97) 2.92 (1.09) 2.51 (0.99) 2.43 (1.06) 2.47 (1.03) 2.44 (1.00) 2.43 (1.05) 2.91 (0.98) 2.63 (1.01) 2.66 (1.10) 3.27 (1.02) 2.77 (1.00) 2.61 (1.07) 2.69 (0.94) 2.92 (1.06) 2.98 (1.00) 2.58 (1.02) 2.62 (1.02) 3.03 (1.01) Concern 2.17 (1.58) 1.93 (1.57) 1.94 (1.66) 1.88 (1.61) 1.81 (1.56) 1.78 (1.63) 1.75 (1.75) 1.73 (1.59) 1.73 (1.69) 1.64 (1.40) 1.57 (1.70) 1.56 (1.80) 1.55 (1.71) 1.55 (1.36) 1.54 (1.61) 1.50 (1.81) 1.50 (1.40) 1.49 (1.53) 1.48 (1.44) 1.46 (1.57) 1.46 (1.80) 1.46 (1.69) 1.46 (1.66) 1.44 (1.78) 1.38 (1.30) 1.36 (1.47) 1.35 (1.90) 1.29 (1.25) 1.28 (1.50) 1.27 (1.83) 1.27 (1.47) 1.26 (1.45) 1.25 (1.27) 1.24 (1.69) 1.22 (1.71) 1.21 (1.37) Mortality Attention for animals Natural birth Maternal behaviour Variation in feed Shelter Frustration Number of transports Growth hormones Lairage time Boredom Having fun Sexual behaviour Hunger during transport Light regime Foraging behaviour Availability of feed Group housing Play behaviour Design of slaughterhouse Group size Functional areas in barn Life span Flooring type Barn temperature Frequency of visual inspection Farmer-animal bond Mixture of groups during t/s Size of livestock herd on farm Preventive medication Taste of feed Static groups Genetic selection Feed on fixed moments Distraction material Technical noise PerImp 4.04 (0.91) 4.25 (0.74) 3.81 (1.00) 3.80 (0.96) 3.90 (0.89) 3.94 (0.90) 3.66 (0.96) 3.79 (0.93) 3.26 (1.55) 3.72 (0.97) 3.50 (1.02) 3.67 (1.04) 3.72 (0.94) 3.66 (0.99) 3.74 (0.88) 3.64 (1.00) 4.23 (0.68) 3.71 (0.85) 3.58 (1.03) 3.61 (0.96) 3.63 (0.94) 3.53 (0.97) 3.64 (1.02) 3.67 (0.89) 3.84 (0.82) 3.68 (0.86) 3.59 (1.05) 3.38 (1.09) 3.40 (1.11) 3.36 (1.08) 3.59 (0.90) 3.43 (0.90) 3.30 (1.13) 3.69 (0.89) 3.14 (1.06) 3.22 (0.97) EvalBel 2.86 (0.94) 3.08 (1.02) 2.66 (1.08) 2.66 (1.03) 2.75 (1.03) 2.82 (0.99) 2.57 (1.02) 2.71 (1.04) 2.18 (1.16) 2.68 (0.98) 2.47 (1.05) 2.66 (1.04) 2.72 (0.97) 2.67 (0.94) 2.77 (0.97) 2.66 (1.03) 3.28 (0.97) 2.77 (0.99) 2.66 (1.06) 2.70 (0.98) 2.72 (1.01) 2.67 (1.01) 2.83 (1.05) 2.91 (0.98) 3.10 (0.93) 3.07 (0.92) 2.97 (1.01) 2.77 (1.00) 2.80 (1.09) 2.85 (1.04) 3.11 (0.93) 3.00 (0.88) 2.92 (1.04) 3.34 (0.94) 2.82 (1.08) 2.91 (0.96) Concern 1.18 (1.50) 1.17 (1.35) 1.16 (1.77) 1.14 (1.74) 1.14 (1.63) 1.12 (1.64) 1.09 (1.73) 1.08 (1.73) 1.07 (2.08) 1.04 (1.68) 1.03 (1.85) 1.01 (1.88) 1.00 (1.65) 1.00 (1.74) 0.97 (1.57) 0.97 (1.74) 0.95 (1.31) 0.94 (1.58) 0.93 (1.88) 0.91 (1.70) 0.91 (1.72) 0.87 (1.68) 0.81 (1.79) 0.77 (1.63) 0.75 (1.41) 0.62 (1.45) 0.62 (1.67) 0.61 (1.84) 0.60 (1.97) 0.51 (1.55) 0.48 (1.56) 0.42 (1.49) 0.38 (1.71) 0.35 (1.47) 0.33 (1.94) 0.31 (1.72) Chapter 8 : Modifiers of societal concerns Table 3.23. Influence of socio-demographic characteristics on the perceived importance, the evaluative belief and the concern towards stocking density; mean scores on 5-point scales Dataset 1 (n=521) PerImp Gender Age Concern PerImp EvalBel *** *** 1.86*** ** 1.40 4.13 2.26 Concern Male 4.41 3.01 Female 4.46 3.32** 1.15 4.43*** 1.95*** 2.48*** 18-24 4.54b 2.86a 1.68c 4.40 2.00a,b 2.40a,b 25-39 4.35a,b 3.03a,b 1.32b,c 4.33 1.87a 2.46b 40-54 4.49b 3.48b 1.01a,b 4.16 2.31b,c 1.85a 55+ 4.19a 3.46b 0.73a 4.22 2.41c 1.81a 4.59*** 3.18 1.40* 4.25 2.04 2.22 3.20 1.13 * 4.30 2.15 2.14 Urban Living EvalBel Dataset 2 (n=459) environment Rural Presence of Yes 4.45 3.48*** 0.97*** 4.13*** 2.37*** 1.76*** children No 4.50 2.88*** 1.62*** 4.42*** 1.86*** 2.56*** Education Till 18 4.34 3.32** 1.03** 4.29 2.14 2.14 Beyond 18 4.53 3.09** 1.43** 4.28 2.09 2.19 ** 0.001<p<0.05; *** p<0.001; 4.33 *** a-c Scores in a row with different superscript are significantly different at p<0.05 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 3.3. Relation between perceptions of stocking density, pen size and group size By definition and mathematically, stocking density is determined by the number of animals per unit surface area. In the questionnaire, perceptions related to stocking density are measured directly, while perception related to group size (number of animals) and pen size (surface area) can be assumed indirect measurements. In this case, a relevant question to address is to what extent the citizens’ perception about stocking density is determined by their respective perceptions relating to pen size and group size. Using dataset 2, which includes the necessary information on each of the three aspects, three regression analyses were performed with perceived importance (PerImp), evaluative belief (EvalBel) and citizen concern (Concern) towards stocking density, respectively, as the dependent variable and perceived importance, evaluative belief and citizen concern of both pen size and group size, respectively as independent variables: PerImp (stocking density ) = a + b1 ⋅ PerImp ( pen size ) + b2 ⋅ PerImp( group size ) EvalBel(stocking density ) = a + b1 ⋅ EvalBel( pen size ) + b2 ⋅ EvalBel( group size ) Concern (stocking density ) = a + b1 ⋅ Concern ( pen size ) + b2 ⋅ Concern( group size ) 149 Part III For all three models, a consistent picture of effects and effect sizes was found (Table 3.24). Both pen size and group size were confirmed to be significant as explanatory variables of perceptions of stocking density (p<0.05) and a satisfactory goodness-of-fit was obtained for each of the regression equations. Higher standardised regression coefficients were found for pen size as compared to group size, indicating a stronger effect of perceived pen size relative to group size on public opinions about stocking density. Table 3.24. Regression analysis with stocking density as dependent variable and pen size and group size as explanatory variables; dataset 2 (n=459) Perceived importance Evaluative belief Concern b SE β R²adj Constant 1.717 0.160 Pen size 0.473 0.041 0.481 0.363 Group size 0.163 0.033 0.207 Constant 0.135 0.092 Pen size 0.553 0.037 0.575 Group size 0.260 0.039 0.257 Constant 0.856 0.073 Pen size 0.603 0.037 0.617 Group size 0.202 0.034 0.220 0.575 0.591 4. Discussion and conclusion The data used for this study was based on surveys that differed considerably in the construction of the questionnaire, in question wording and in the characteristics of the samples. As a consequence, the aim of the discussion is not primarily to compare results between the two datasets and the respective time periods. Instead, the aim was to find out whether certain trends and effects were robust in the sense that they were apparent in both datasets despite the differences in methodology, sample and timing. Trends that could be detected in one dataset only could indicate either that the trend is less robust, that a real shift in opinion took place during the first and last survey or that the trend could not be statistically proven in the other dataset due to inadequate setup or statistical power for that particular aspect. Indeed, the reasons for between dataset differences cannot be discerned with certainty as it cannot be verified whether the observed difference is a reflection of real evolutions in the societal perception and/or consciousness of the issue at stake, or rather due to differences in question wording or sample composition or due to social phenomena, like for instance media coverage related to animal welfare. In both datasets, the application of appropriate standards regarding stocking density and pen size were perceived as fundamental requirements for an acceptable level of farm animal welfare. Group size however, though attributed a positive importance score, was not perceived as a top priority in 150 Chapter 8 : Modifiers of societal concerns attaining positive animal welfare. Regarding the aspects’ evaluative beliefs, stocking density was identified as the most problematic aspect in relation with animal welfare in Flemish animal production in dataset 2. Also pen size was ranked among the most problematic ones, while fewer problems were perceived with respect to group size. In dataset 1 these aspects were not perceived as the ones having the most negative influence on the welfare of the animals in Flemish animal production. Combining the perceived importance and evaluative belief score yielded a measure for “concern”. As a common – and therefore robust - conclusion, we found that the societal concern of both stocking density and pen size is imperative within the full picture of farm animal welfare. In comparison with other animal welfare related aspects, both stocking density and pen size rank high in terms of perceived importance and receive relatively low evaluative belief scores. Both aspects are herewith clearly positioned in the upper left “high concern” quadrant of Figure 3.7. The perceived importance for group size is considerably lower as compared to stocking density and pen size, and citizen judgments are milder towards this aspect. Regarding the social structural attitude modifiers, some consistent results were found indicating that younger people and households without children are the higher concerned citizen groups regarding stocking density in present animal production. Another consistent picture pertains to the impact of living environment, which did not associate significantly with the concern of Flemish citizens regarding stocking density. Finally, a higher concern among women and higher educated people in one of the datasets was neither confirmed nor disconfirmed in the other dataset. In literature, no specific data is available concerning the association of social structural variables with citizen concerns about the issue of stocking density or space allowance in particular, although several studies have investigated associations between socio-demographics and concern towards animal welfare in general. Our findings corroborate with regard to a higher concern among younger people (Boogaard et al., 2006; Harper and Henson, 2001; Kendall et al., 2006; Maria, 2006) and households without children (Kendall et al., 2006). The higher concern among households without children could be attributed to families with children prioritising health and safety above production-related attributes like animal welfare. In contrast with Boogaard et al. (2006) and Kendall et al. (2006), our data did not confirm a significant effect of living environment. The role of gender (Boogaard et al., 2006; Harper and Makatouni, 2002; Maria, 2006) and education level (Boogaard et al., 2006; Harper and Makatouni, 2002; Kendall et al., 2006) were only partly confirmed. Given the very strong gender effect in the second dataset, the non significant impact in the first dataset is expected to be largely due to methodological issues. Specifically, samples were restricted to people responsible for food purchasing in dataset 1, whereas this selection criterion was not used for the sample in the second dataset. This suggests that concern related to stocking density is more apparent among females in their role as a citizen than in their role as food shopper. The non-significance of living environment can probably be ascribed to the self-administered and subjective interpretation of the measurement scale. It only allowed the respondent to indicate whether he/she lived in either a city or in the countryside without further specification. Our findings 151 Part III related to the role of education may be attributed to the composition of our samples, in which only a small share of respondents had a low level of education (less than secondary school). Indications of a different degree of concern between different social structural groups in society are very valuable since it gives insight into which citizen groups efforts should be addressed to alleviate public concern and improve perceptions regarding the welfare of farm animals, particularly regarding stocking density. Finally, it is also important to anticipate on the best way to communicate, in order to most effectively build favourable public perception and livestock farming image. The perception of stocking density appeared to be most strongly dependent on the perception of pen size. This indicates that especially the perception of spatial aspects (such as ideas, images or expectations about the available space, the pen size or cage size) is of importance in public perception and citizen concern regarding stocking density, rather than the perception of group size (i.e. the number of animals housed together). These findings corroborate with insights obtained during the focus groups, discussed in Chapter 1. When the discussion was concentrated on spatial aspects in relation to animal welfare, the focus group participants rarely related an excessive stocking density to a too high number of animals, but almost invariably to too limited space availability. In addition, group size seemed not to be perceived as a negative factor only, in the sense that both a too high and a too small number of animals were indicated as potentially detrimental for animal welfare. The focus group participants also remarked that for a good animal welfare, animals should not be housed individually (referring to sows; leaving calves with the cows; images from herds in nature). Consequently it could be expected that changing animal group sizes and associated communication will contribute less to building a more positive perception about stocking density compared to an increase of pen size. These findings also indicate that communication efforts that focus on enlarged pen size stand a better chance of being effective when the aim is to realise a more positive public perception about stocking density and higher societal acceptance of farm animal production practices. Seen the debated relationship between individuals’ attitudes or what they state as citizens and what they actually do as consumers, future research is recommended focusing on the extent to which the stated perceptions discussed in this study can be translated into purchasing intentions and actual buying behaviour of food products produced under better standards of stocking density and space allowance for farm animals. 152 Chapter 9 : Consumer variety – segmentation analysis Chapter 9 Acknowledging consumer variety: a segmentation analysis Adapted from Vanhonacker, F., W. Verbeke, E. Van Poucke, and F. A. M. Tuyttens. 2007. Segmentation based on consumers’ perceived importance and attitude toward farm animal welfare. International Journal of Sociology of Agriculture and Food 15: 91-107. Abstract In this chapter, further elaboration is made on determinants that modify the attitudes and concerns for farm animal welfare. In the introduction of the chapter, the attitude modifiers discussed in this dissertation’s conceptual framework are compared with study findings published in literature. This chapter applies the attitude modifiers as profiling variables of market segments that differ in public opinion toward animal welfare as well as in animal welfare related behaviour as consumers. In that perspective, a cluster analysis is performed using a cross-sectional dataset of 459 residents of Flanders, Belgium, gathered in 2006. The perceived importance attached to animal welfare as a product attribute in the food purchasing decision process relative to other product attributes (relative importance; RI) is investigated, as well as the subjective evaluation of the current state of farm animal welfare (evaluation; EV) as segmentation variables. Six clusters are obtained: cluster 1 with moderate RI and positive EV (21.1 percent); cluster 2 with very low RI and strong positive EV (12.9 percent); cluster 3 with low RI and moderate EV (18.7 percent); cluster 4 with moderate RI and low EV (12.6 percent); cluster 5 with high RI and moderate EV (23.5 percent); cluster 6 with very high RI and very negative EV (11.1 percent). Based on this segmentation exercise and the segments’ profiles, market opportunities for higher animal welfare products are identified. 153 Part III 1. Introduction 1.1. Background Animal welfare in general and more specifically in animal production has become a major issue of interest, not only among direct interest groups involved in the food production chain, such as producers, retailers, government and consumers, but also in a wide variety of scientific research disciplines. This tendency of a general increased interest in animal welfare can be explained to a large extent by the prosperity level in the Western society (Seamer, 1998). Food supply has largely exceeded food demand, which has turned markets into demand-driven economies where the goal of exchange and marketing is to better meet consumers’ expectations, demands and preferences. Together with the growing influence of post-materialistic values – of which interest in farm animal welfare is just one example – on product attribute evaluation and food choice decision-making, this has recently led to numerous studies about public and citizen concerns on the one hand (Boogaard et al., 2006; Kanis et al., 2003; Lassen et al., 2006; Maria, 2006; Verbeke, 2002), and about consumer attitudes and behaviour in relation with farm animal welfare on the other hand (European Commission, 2005; Frewer et al., 2005; Harper and Henson, 2001). This research focus was anticipated by Verbeke and Viaene (1999a), who concluded based on the analysis of a 1998 consumer sample in Flanders that animal welfare and acceptable production methods emerged as key attention points for the future of animal production and marketing, as well as public and consumer acceptance of animal-based food products. However, the interpretation of the concept of farm animal welfare differs considerably between different interest and stakeholder groups, and its conceptualisation is heavily influenced by convictions (opinions about the way things are), values (opinions about the way things should be), norms (translations of these values into rules of conduct), knowledge (constructed from experiences, facts, stories, and impressions) and interests (economic, social and moral interests) (Te Velde et al., 2002). This framework explains why the animal production and processing sector and the broader public tend to speak different languages when talking about animal welfare. Producers tend to position themselves as knowledgeable and rational actors, while they dismiss the concerns of the lay person as emotional and uninformed. The public, however, often associates the industry’s interest in animal welfare as strictly economic and profit oriented and considers their own citizen and consumer viewpoint as ethically motivated (Kendall et al., 2006). 1.2. Public attitude toward animal welfare Within the broader public, diverse opinions exist. Despite the occurrence of differing opinions and the relevance of the topic, little research has focussed thus far on a conceptual approach to determinants of public or consumer attitudes to animal welfare. In this dissertation, I make use of the theoretical framework developed by Kendall et al. (2006), based on an extension of existing 154 Chapter 9 : Consumer variety – segmentation analysis literature in sociology, mainly stemming from U.S. studies. Summarising, Kendall and co-authors distinguish between social structural variables to explain different attitudes between social groups, and individual animal-related experiential variables to explain differences within social groups (see Part I). Regarding social structural variables, distinction is made between place-based variables and other social structural variables. In this introduction, this framework is compared with findings from published studies. With regard to the place-based variables, utilitarian motives were used to hypothesise a lower concern about animal welfare among persons with a rural background and/or experience with farming. This was confirmed by Verhue and Verzeijden (2003) and Frewer et al. (2005), who found that people living in rural neighbourhoods evaluated the state of animal welfare more positively. Other social structural variables comprised gender, socio-economic class, age and family status. Women expressed a higher concern with animal welfare as compared to men. The task of women as primary family caretakers and as being more likely to engage in household tasks that put them in contact with animals, like caring for pets and preparing food were considered as possible explanations for this gender difference. Similar conclusions are found by Burrel and Vrieze (2003) and Verhue and Verzeijden (2003). Next, lower income categories and less educated persons were hypothesised to express a greater concern for animals, which has also been referred to as the underdog-hypothesis (Kendall et al., 2006). Opposite results were found by Burrel and Vrieze (2003) and Verhue and Verzeijden (2003), where especially higher educated people expressed a higher concern for animal welfare. Further, age was hypothesised to be inversely related to the concern for animal welfare. This was linked to one’s life-cycle stage, hence subject to a change over time depending on the evolution and change of factors in a person’s direct social environment, such as family relations (Kendall et al., 2006). Verhue and Verzeijden (2003) confirmed this hypothesis indicating a higher concern toward animal welfare among younger people. Finally, with regard to family status, expectations were that people with dependent children would express less concern about animal welfare, since they have to attribute time and energy toward their own offspring instead of to non-human others. With regard to the individuals’ experiences, a positive relationship with concern for animal welfare is hypothesised for people who do not hunt, have a pet, are vegetarian, are more involved in cooking and food shopping and have higher concerns about the environment and food in general. With regard to vegetarianism, the hypothesis was grounded on the worldview accompanying vegetarianism, which encompasses greater awareness of the origin of one’s food (McDonald, 2000). 1.3. Animal welfare related consumer behaviour At present, there is a lack of studies that combine the citizen and the consumer perspective on farm animal welfare, i.e. studies that consider both variations in citizen attitude toward animal welfare on the one hand and variations in the impact of animal welfare as a product attribute on consumers’ food choice decisions on the other hand. Such studies are relevant because the market 155 Part III for high welfare products is rapidly evolving, due to some contemporary changing food patterns. Increased disposable incomes have caused that food shoppers in many markets can afford to pay premium prices for differentiated quality products. As a result, food has begun to provide an emotional as well as a functional role in consumers’ lives. At the same time, consumer confidence in food production has dropped due to some consecutive food scares in Europe at the end of the nineties, in particular within the animal production chain. Furthermore, consumers are increasingly aware of the association between food intake or their dietary behaviour and their personal health and overall well-being. They believe that food produced in a more natural way will suit them better and provide them with more benefits (Grunert et al., 2000). Considering these tendencies, a higher willingness to pay for high welfare products can be assumed. However, many studies have criticised claimed consumer willingness to pay, referring to the ambivalent position of an individual in the role of consumer versus citizens (Bennett et al., 2002; Carlsson et al., 2007; Korthals, 2001a; Liljenstolpe, 2008). Moreover, animal welfare is an ethical issue and as a consequence highly sensitive to social desirable answering, for example in survey research and interview questionnaires. As a consequence, some deduced that consumers do not prioritise animal welfare considerations while shopping for food. Such conclusions, however, may be too general and based on sample average scores, without acknowledging for different segments that might exist. Only few studies noticed the existence of a specific segment taking animal welfare more into account when shopping (e.g. Grunert et al. (2004)). More specifically, little information is available with respect to the segmentation of individuals based on their perceived importance of animal welfare when purchasing food in general (thus in their role as a consumer) and to our knowledge the segmentation of individuals based on the relative importance attached to animal welfare when purchasing food has not yet been studied in depth. 1.4. Scope and objectives The aim of this chapter is to perform a segmentation analysis based on the two topics discussed above: i.e. public attitude toward farm animal welfare in relation to animal production in Flanders, Belgium, and consumers’ relative importance attached to farm animal welfare as a product attribute during food purchasing decisions. The contribution of this approach is two-fold. On the one hand, positioning different segments based on those two dimensions can help to better understand different viewpoints within society (citizens versus consumer), yielding a valuable basis to improve the societal (public and market-related) debate about the issue. On the other hand, segmentation is a necessary tool in order to understand how to make higher welfare foods relevant to different consumers and how to position these products in a competitive marketing environment. From this angle, distinct consumer profiles can be established which can provide insights as to how to target, communicate and convince these distinct groups to purchase higher welfare products. Attitude toward animal welfare is used as an indicator for the market opportunities of high welfare products, while the relative importance of animal welfare as a product 156 Chapter 9 : Consumer variety – segmentation analysis attribute will be considered as the leverage or selling proposition for how to promote and communicate these products. The strength of this segmentation exercise is that it combines a rather concrete consumer-related measure (relative importance attached to farm animal welfare as a product attribute) with a more abstract public opinion (attitude toward animal welfare). Profiling variables pertain to the attitude modifiers considered in the theoretical framework of Kendall et al. (2006), meat consumption, knowledge and interest in information. These results should provide a more balanced picture with respect to the existence of socially and ethically engaged segments, integrating both evaluations from a public and a consumer perspective. 2. Materials and methods Methodological details in terms of research approach and sampling procedure are provided in Part I, section 4, and apply to Study 2. 2.1. Measurement of constructs First, 13 product attributes were probed for their perceived importance (PI) in the food purchasing decision process of animal food products on a five-point interval scale ranging from “totally unimportant” (1) to “very important” (5). The product attributes were: safety, quality, reliability, taste, origin, health, price, appearance, freshness, environmental friendliness, availability, animal welfare and production method. PI reflects the individuals’ reaction from a consumer perspective, i.e. someone who has to weigh and evaluate different product attributes before coming to a purchase decision. Second, respondent’s attitude was measured in terms of evaluative belief with respect to the current state of animal welfare in Flemish animal production (EV). Therefore the statement: “Do you believe the current state of farm animal welfare in Flanders in general is…” was used. This item was measured on a seven-point interval scale anchored at the left pole by “very poor” (1) and at the right pole by “very good” (7), with “moderate” (4) as the mid-point of the scale. This second measure is much more a public opinion, which is presumed to be held rather independent of the consumption decisions (see previous chapters). Third, consumption of beef, pork, poultry, fish and meat substitutes was scored on a six-point selfreported consumption frequency scale, ranging from “daily” to “never”. Fourth, both subjective and objective knowledge about farm animal welfare were probed. Subjective knowledge was assessed using four relevant items of the five-item scale described by Flynn and Goldsmith (1999) and measured on a five-point Likert-scale. Items were: “Compared to an average person, I know a lot about animal welfare”; “I have a lot of knowledge about how to evaluate the welfare of farm animals”; “I have a lot of knowledge about how farm animals are kept”; “My friends consider me as an expert on farm animal welfare.” Objective knowledge was 157 Part III investigated using five statements, corresponding with the major five groups of farm animals in Flanders: “Barn eggs are from chicken that have outdoor access” (not correct); “Male pigs are castrated because otherwise the meat can have a bad smell” (correct); “Broilers are slaughtered at the age of 4 months” (not correct); “A dairy cow gives milk only after calving” (correct); “Doublemuscled beef cattle mostly calve via a Caesarean section” (correct). Respondents could either agree or disagree with each statement. To account for guessing, respondents were also asked to indicate how confident they were about their answer to each item on a scale from “very uncertain” (1) to “very certain” (5). Fifth, consumer opinions toward information related to farm animal welfare were assessed using six statements, each scored on a five-point Likert-scale. Statements were: “Labels should indicate more clearly the rearing conditions of the animals”; “Information about animal welfare is too little available”; “Animal welfare should be controlled more severely”; “Animal welfare should be guaranteed through a label on the product”; “I’m willing to pay more for food produced with more attention to animal welfare”; “More information about animal welfare would influence my meat consumption”. Sixth, variables representing the determinants for the attitude toward animal welfare as described by Kendall et al. (2006) were included. The place-based factors were represented by living environment and farm experiences. Respondents were asked in what type of place they live (rural or urban). The response is coded 1/0 and is based on individuals’ self-identification. To tap experience with farming, respondents were probed with the following statements on which they could answer “yes” or “no”: “I have / my parents have a farm”; “My grandparents / other relatives have a farm”; “Close neighbours / good acquaintances have a farm”. Other structural factors involve gender, age, education and presence of children. Finally, being a vegetarian (yes/no) was included as variable measuring individuals’ animal-related experience. 2.2. Statistical analyses Data were analysed using SPSS 12.0. First, hierarchical clustering with Ward’s Method as cluster method and K-means cluster analysis were performed to obtain segments. Bivariate analyses including cross-tabulation with χ²-statisitics, independent samples t-test and one-way ANOVA comparison of means were used to profile the clusters in terms of behaviour, knowledge, information opinions and structural determinants for attitude toward animal welfare. Given the high reliability coefficient of the four subjective knowledge items (Cronbach’s alpha = 0.93), we calculated a summated subjective knowledge score corresponding with a score ranging from 4 to 20. Also with respect to objective knowledge, we computed a cumulative score. For each objective knowledge item, a wrong answer was coded as zero, while a correct answer was coded with its corresponding reported certainty level, i.e. ranging from one to five. As a result of this coding procedure and after summation across the five items, an overall range from zero to 25 was obtained for objective knowledge. 158 Chapter 9 : Consumer variety – segmentation analysis 3. Results 3.1. Segmentation analysis 3.1.1. Segmentation variables The first segmentation variable pertains to the perceived importance attached to animal welfare in the food purchasing decision-process. Since this perceived importance has little meaning in absolute terms, though only relative as compared to the perceived importance attached to other product attributes, a relative score was computed for each of the 13 attributes assessed by the respondents. This score, corresponding with the relative importance, was computed using (1): RI i = 13 * PI i 13 ∑ PI i =1 RI = relative perceived importance PI = absolute perceived importance (1) i A RI-score below the value of 1 indicates that the specific product attribute ranks among the less important product attributes, while a score above 1 corresponds with a relatively important product attribute. As the focus will be on the relative perceived importance of animal welfare, we will use the abbreviation RIAW in further discussion as reference for the relative perceived importance score assigned to the attribute animal welfare. RIAW ranges from 0.27 to 2.60 within the sample, with a mean score of 0.98 (SD=0.23). The second segmentation variable is the respondent’s attitude in terms of evaluation (belief) of the current state of animal welfare in Flemish animal production (EV). EV ranges from 1 to 7 with a mean score of 4.13 (SD=1.49). For the clustering procedure, we have opted to work with the standardised score (z-score) of both variables rather than with the actual scores, in order to obtain a segmentation that better puts the relative position of the segments into perspective. In further discussion, absolute perceived importance score and evaluation of farm animal welfare in the current Flemish animal production refer to the mean scores of the non-standardised variables RIAW and EV. 3.1.2. Cluster analysis A hierarchical clustering followed by a K-means cluster analysis was used to determine the optimal number of clusters (so-called segments) yielding the highest degree of differentiation. This resulted in a six-cluster solution (Table 3.25 and Figure 3.8). 159 Part III Table 3.25. Profile of the segments on the segmentation variables (n=459) S1 S2 S3 S4 S5 S6 Segment size (% of sample) 21.1 12.9 18.7 12.6 23.5 11.1 Absolute perceived importance (PI) 4.14 2.51 3.15 3.95 4.55 4.98 Relative importance (RIAW) 1.03 0.66 0.81 0.95 1.12 1.30 RI z-score (segmentation variable) 0.22 -1.40 -0.75 -0.15 0.59 1.39 Evaluation (EV) 5.44 5.95 4.53 2.36 3.76 1.67 EV z-score (segmentation variable) 0.88 1.22 0.26 -1.18 -0.24 -1.65 EV 1.5 S2 1 S1 0.5 S3 RI 0 -1.5 -1 -0.5 0 0.5 S5 1 1.5 -0.5 S4 -1 -1.5 S6 -2 Figure 3.8. Mapping of the cluster centres according to relative importance (RI z-score) and evaluation (EV z-score) of farm animal welfare; the size of markers reflects cluster size Segment 1 (S1; 21.1 percent of the sample) corresponds to respondents who attached high absolute importance to animal welfare when purchasing animal food products. However, all product attributes received a high perceived importance score among these consumers. As a consequence, their RIAW is rather neutral, thus animal welfare is considered moderately important compared to other product attributes. Nonetheless, animal welfare received a higher importance than some other production system-related attributes, such as production method and environmental friendliness. With respect to their evaluation of farm animal welfare in the current Flemish animal production, an above average EV was found (“rather good” to “good”) (Table 3.25). Segment 2 (S2; 12.9 percent of the sample) corresponds to respondents who claim not to take animal welfare into account in their food purchasing decision-making, given that no other product attribute received a lower RI-score than the attribute animal welfare. Their most important product attributes when purchasing animal food products were freshness, quality and taste. Also, people belonging to S2 reported the most positive EV. 160 Chapter 9 : Consumer variety – segmentation analysis Respondents belonging to Segment 3 (S3; 18.7 percent of the sample) showed a lot of similarities with S2 regarding the absolute perceived importance of the product attributes. However, their opinion is not that sharply pronounced, meaning that a similar ranking occurred, with smaller differences between the relative scores. Moreover, in contrast to the very positive EV given by S2, S3 is somewhat less positive, evaluating animal welfare mainly as “moderate” to “rather good”. Next, Segment 4 (S4; 12.6 percent of the sample) mirrors S1 in terms of the standardised RIAW and EV (Figure 3.8). Similar as for S1, high perceived importance scores are attached to all product attributes. However, respondents belonging to S4 attribute less importance to animal welfare as compared to environmental friendliness and production method. Segment 5 (S5; 23.5 percent of the sample) is the largest segment and is composed of respondents who indicate to take animal welfare quite heavily into account, i.e. to the same extent as quality and health attributes and even to a higher extent than taste and safety. With respect to EV, a rather average score was found. This segment mirrors S3 with respect to the segmentation variables. Finally, Segment 6 (S6; 11.1 percent of the sample) is very concerned about animal welfare, as is expressed first in a very high PI of animal welfare and second in a very poor EV. Animal welfare emerged as the most important product attribute, followed by the two other production systemrelated attributes and outweighing the quality traits and taste. Moreover, product appearance, availability and price are relatively considered as much less important among S6-members as compared to the other segments. 3.2. Profiling of the clusters 3.2.1. Determinants of animal welfare perception Table 3.26 presents a set of variables within the determinants of public attitude toward animal welfare as described by Kendal et al. (2006). First, the place-related variables are considered. No significant differences between the segments related to living environment were found, although S1 and S2 were composed of the highest amount of rural inhabitants. Living environment is not associated with RIAW (t=0.86, p=0.39), while a marginal influence on EV resulted (t=1.79, p=0.074), with a more positive evaluation of current farm animal welfare given by rural inhabitants (μurban=3.97; μrural=4.22). Concerning farming experience, differences were more pronounced. The respondents who have themselves a farm or have parents with a farm are distributed over S1, S2 and S3, with the highest relative share corresponding to S2 (21.6 percent) (χ²=43.09, p<0.001). Segment 1 also includes a relatively large share (13.3 percent) of respondents whose grandparents or other relatives have a farm (χ²=14.98, p<0.001). The percentage of close neighbours or good acquaintances having a farm does not differ significantly between the segments (χ²=9.18, p=0.102). Ownership of a farm or parents having a farm has its consequences with regard to RIAW and EV: a significant lower RIAW (t=3.65, p<0.001) and a more positive EV (t=10.95, p<0.001) appeared among respondents with the highest farming experience. Conversely, a lower degree of farming experience 161 Part III did not longer impact RIAW (p>0.1), while it still showed a significantly positive relationship with EV (p<0.1). Second, gender, age, education level and the presence of children were included as other social structural variables. For gender, significantly more men belong to S2 (χ²=39.55, p<0.001), whereas women rather belong to the S5 and S6. In general, females (χ²=1.03) attached more importance to animal welfare relative to other product characteristics as compared to men (μ=0.93) (t=4.44, p<0.001) and they evaluate the current state of animal welfare as more negative (μmale=4.45, μfemale=3.82; t=4.59, p<0.001). Also for age, a significant distinction between the segments was present (χ²=36.59, p<0.001). Segment 1 was composed of a relatively low amount of people aged between 24 and 37, while we found an over-representation of the two oldest age categories. A similar age distribution appeared for S2, although deviations from the distribution in the total sample were smaller as compared to S1. Segment 3 is characterised by a rather low amount of youngsters, while S4 consists of a large group of respondents aged between 24 and 37. Finally, S5 does not show large deviation from the samples’ age distribution, whereas S6 has an overrepresentation of the two youngest age categories. No age-effect was found on RIAW (F=0.96, p=0.41), while EV was clearly age-dependent (F=8.13, p<0.001). The second age category expressed the lowest evaluation, significantly different from the two oldest groups. The remaining three categories do not differ significantly from each other. Furthermore, education level does not differ between the six segments (χ²=2.73, p=0.74), and did not associate with neither RIAW (t=0.239, p=0.811) nor EV (t=0.998, p=0.319). Finally, between-segment differences appeared for the presence of children in the household (χ²=36.31, p<0.01). The highest share of households with children was present in S1 and S2, while within S6, only one fifth of the households had children. No association of the presence of children was found with any of the segmentation variables (p>0.1). Third, vegetarianism was the single-item included within the set of individual experiences. Almost no vegetarians were present in S1, S2 and S3, whereas the majority of respondents within S6 (60 percent) indicated to be vegetarian. S4 and S5 were positioned in between, with respectively about 20 percent and 10 percent being vegetarian. Consequently, this difference in segment composition was significant (χ²=148.86, p<0.01). Also, a highly significant association of vegetarianism was found with both segmentation variables (p<0.1). 162 Chapter 9 : Consumer variety – segmentation analysis Table 3.26. Determinants of animal welfare for the different segments (n=429), frequency distributions (%); total sample characteristics are mentioned between brackets Sample S1 S2 S3 S4 S5 S6 Urban (38.9) 33.7 30.5 46.4 41.4 38.9 43.1 I / parents have a farm (6.1) 13.3 21.6 3.7 0 0 0 Grandparents / other family have a farm (29.1) 43.2 35.2 21.3 28.6 21.4 26.5 Neighbours / acquaintances have farm (40.1) 50 48.2 30.9 38.9 35.0 40.8 Male (48.5) 54.2 82.5 44.0 48.3 36.1 33.1 Age <24 (18.8) 17.7 18.6 12.9 20.7 21.3 23.5 Age 24-37 (32.4) 15.6 25.4 36.5 39.7 36.1 49.0 Age 38-53 (31.7) 38.5 32.2 35.3 27.6 31.5 17.6 Age 54+ (17.1) 28.1 23.7 15.3 12.1 11.1 9.8 Mean age (years) (37.8) 42.5 41.0 37.4 35.2 36.0 32.6 Higher education (67.7) 63.2 67.8 67.9 74.1 65.7 72.5 Presence of children (48.9) 63.2 66.1 52.4 38.6 43.5 19.6 (12.2) 0 0 2.4 21.1 9.4 62.0 Place-based variables (% yes) Social Structural Location (% yes) Individual Experience Vegetarian (% yes) 3.2.2. Meat and meat substitute consumption Differences in claimed consumption behaviour between the segments are clearly reflected in their reported consumption frequency of meat, fish and meat substitutes (Table 3.27). The segments S1, S2 and S3 appeared as the heaviest meat consumers. Segment 6 on the other hand reported a very low meat consumption frequency, while S4 and S5 indicated a meat consumption frequency in between these two extremes. Reported fish consumption frequency, was lower for S6 as compared to the other segments. Consumption frequency of meat substitutes was the inverse of the meat consumption frequency, with a very regular consumption of meat substitutes in S6 and a very low consumption frequency by S1, S2 and S3. Table 3.27. Profiling of the segments based on meat and meat substitute consumption frequency; mean scores S1 Consumption frequency Beef Pork Poultry Fish Other meat substitute S2 S3 S4 S5 S6 Sample * a 2.51 a 2.61 a 2.91 3.34 a 5.18 c a,b 2.61 a 2.59 a 2.72 a 3.39 c 5.35 b,c 3.01 a,b 2.92 a 2.89 a 3.23 c 5.07 c 3.64 c 3.81 b 3.63 a 3.42 b 3.96 c 3.29 b,c 3.40 b 3.40 a 3.24 b 4.37 d 3.22 d 3.25 c 3.32 b 3.42 a 4.55 4.98 4.90 5.00 4.31 2.69 * Six-point frequency scale: 1 = every day; 2 = several times a week; 3 = weekly; 4 = monthly; 5 = less than a-d monthly; 6 = never ; Scores in a row with different superscripts are significantly different at p<0.05 (oneway ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 163 Part III 3.2.3. Knowledge The level of both subjective and objective knowledge about animal production practices correlates quite well and is not equal among the six segments (F=10.89, p<0.001; F=5.21, p<0.001, respectively). S6 and S2 reported the highest subjective knowledge about farm animal welfare. S3 on the other hand indicated the lowest subjective knowledge and all other segments did not differ significantly from each other (Table 3.28). In line with the reported subjective knowledge, S6 and S2 turned out to be also effectively the most knowledgeable about animal welfare issues based on the objective knowledge score (Table 3.28). S3, S4 and S5 were the least knowledgeable. 3.2.4. Information variables The different segments show clear differences with regard to the evaluation of the current information about animal welfare, the expectation they have about animal welfare information and the stated impact of information on their behaviour. The lowest score for each of these issues was given by people belonging to S2. With regard to the first four information items as they are presented in Table 3.28, a neutral average segment score was found within S2. This indicates that they are not against a more intensive and more clear information provision, but that they are also not explicitly asking for more information. The difference with the other segments was the largest with respect to the attitude toward more severe controls on animal welfare. The low interest in information is also reflected in a very low expected impact of receiving more information on their meat consumption. Furthermore, S2 was the only segment which disagreed to pay more for food produced with more attention for animal welfare. S1 and S3 are somewhat more positive toward the information statements. They score neutral (mean value approaching the value of 4) on the statements related to more and clearer information provision and on more severe controls. Somewhat contradictory with this request is their neutral attitude toward the expected impact of more information on their meat consumption and the limited willingness to pay for food produced with specific attention for animal welfare. Probably, this could be driven by their daily consumption of meat. In contrast, the remaining segments (S4, S5 and S6) expressed high expectations for more and clearer information and for more severe controls on animal welfare. While these expectations are high for S4 and S5, it is extremely high for S6. Despite this strong request for information, only a moderate expected impact of receiving more information on the meat consumption is indicated by these segments. Explanations pertain to a segment being largely vegetarian (S6), hence not willing to eat meat whatever the amount of information provided, or to segments consuming already a large amount of meat (S4, S5), hence hardly leaving room for further increase. Regarding willingness to pay, we found a high score corresponding with S4 and S5 and a very high score for S6. Especially concerning S6, this high willingness to pay is in line with the low perceived importance of price in the food buying decision process. 164 Chapter 9 : Consumer variety – segmentation analysis Table 3.28. Profiling of the segments based on knowledge and information variables S1 S2 S3 S4 S5 S6 Sample Knowledge Subjective knowledge Objective knowledge 12.4 b b,c 13.1 a,b 14.7 b 3.17 b 3.08 b 2.81 b,c 3.19 b 2.32 b 2.08 12.7 a 10.2 b 10.7 a 3.78 a 3.92 a 3.99 a,b 11.2 a 10.2 b 4.43 b 4.40 b 4.57 a,b 11.3 a 10.6 c 4.34 c 4.38 c 4.55 c,d 4.51 c 4.32 b,c 3.78 c 12.02 b 11.97 d 4.05 c 4.02 d 4.12 e 4.13 d 4.73 b,c 3.19 15.5 a 14.2 c 4.82 c 4.44 c 4.94 d 4.86 c 4.78 Information statements Labels should indicate more clearly the rearing conditions Information is too less available 3.89 3.85 Animal welfare should be controlled more severely 3.84 Animal welfare should be guaranteed by a label 4.00 I am willing to pay for products with more animal welfare 3.56 More information about animal welfare would influence my choice a-e 2.93 a a,b 3.78 a 3.21 b a 2.95 b 4.47 4.19 3.60 c 3.63 Scores in a row with different superscript are significantly different at p<0.05 (one-way ANOVA F- tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 4. Discussion and conclusion Driven by several socio-economic evolutions, animal welfare has gradually come to the forefront in recent societal debates. Despite the fact that animal welfare is subject to an increasing amount of research, theoretical development and empirical evidence related to the topic within sociology and consumer science research is rather limited and often focused on the description of findings on a general population level. However, opinions related to animal welfare are much divided and often conflicting. Moreover, seen the ethical character of animal welfare, public opinions or reflections from a citizen perspective do not provide a straightforward picture that is relevant for actual consumer behaviour and food choice. A gap in literature exists with respect to combining public citizen-oriented measures as well as consumer-oriented approaches. This chapter applied the conceptual approach presented by Kendall et al. (2006), which aimed at fostering the sociological debate about attitudes toward animal welfare. With the performed segmentation analysis, insights in the existence of different population groups or segments were provided. It fuels the debate starting from a different and often conflicting perspective and identifies specific market opportunities for high welfare products. 165 Part III A six-cluster solution was achieved. Seen the resulting characterisation of the segments in terms of structural determinants of animal welfare, meat and meat substitute consumption frequencies, knowledge about animal welfare and attitudes toward information about animal welfare, the six segments could basically be interpreted as four groups with very distinct features. S2 and S6 appeared as two extreme groups that stand on their own. They have a completely opposite attitude and belief structure in relation to farm animal welfare and a very opposed food consumption behaviour pattern. The remaining two groups could be composed through a combination of two segments: S4 with S5 and S1 with S3, respectively. With regard to their features, they are positioned in between the two extreme groups, with S4/S5 rather tending to S6 and S1/S3 rather tending to S2. Along the axis of the attitudes toward animal welfare, the characteristics of the different groups largely corresponded with the set of determinants defined by Kendall et al. (2006) as affecting the attitude toward animal welfare. With regard to S2, especially the highest degree of farming experience and a high share of rural inhabitants, together with the predominantly male composition seemed to explain the very low concern about the current state of farm animal welfare. S6 on the other hand, which expressed the highest concern toward farm animal welfare, differed most strongly from the other segments in terms of the share of vegetarians (highest), their age profile (youngest) and the share of households with children (lowest). All these characteristics were described as determinants for a higher concern toward animal welfare. The perspectives of S2 and S6 toward farm animal welfare are strongly opposed, most likely because of their different type of involvement with animal welfare. S2 is mainly socio-economically involved with agriculture and animal production, in the sense that farming activities are a part of their daily lives and a source of livelihood in their living environment. S6 is mainly involved with animal welfare because of personal moral and ethical considerations. Both segments display a very consistent attitude-value profile (as individual in their role as a citizen or member of a particular societal group) and behavioural profile (as individual in their role as consumer, thus with respect to food choices). Most likely, the societal debate about farm animal welfare will continue to be fuelled mainly by those two societal groups with opposing interests. The position of the other segments is bridging between these two extremes with regard to determinants described by Kendall et al. (2006). Depending on the strength of arguments in the debate, through new personal experiences, changes in their social and living environment and exposure to information, these segments may evolve over time in either direction. Hence, from a communication point of view, these segments are particularly interesting as target audiences because of their rather moderate predisposition toward farm animal welfare. The segmentation exercise is especially valuable with regard to identifying market opportunities and formulating marketing strategies for high welfare products for each of the groups. An increased market opportunity for high welfare products is assumed with an increased concern for the current state of animal welfare. As a result, little or no animal welfare-related marketing possibilities are seen for people belonging to S2, who are very positive toward the current state of animal welfare. In addition, they indicate a very low importance for animal welfare as a product attribute in their 166 Chapter 9 : Consumer variety – segmentation analysis food purchasing process. Their low expressed information need and willingness to pay corroborate with this. The group composed of S1 and S3 are also considered as a group with rather low marketing opportunities for high welfare products. Seen their modest willingness to pay and information need, together with animal welfare not being ranked as a very important product attribute, high welfare products will need very strong tangible benefits (e.g. taste, tenderness …) without high price premiums for possible market success with these segments. Next, the group composed of S4 and S5 can be considered as a real marketing opportunity segment. This group reports concerns about animal welfare and does not neglect animal welfare as a product attribute in the food purchasing decision process. Moreover, they express high information needs and willingness to pay for higher welfare products. This group comprises 36.1 percent of the sample, hence constituting a considerable market. Within this group, animal welfare is important but not the top priority. Consequently, we do not expect a very strong commitment in terms of behaviour, i.e. people may not consistently buy high welfare products on every food purchase occasion. Seen the importance attached to animal welfare and the concern toward it, it will be important from a marketing point of view to do efforts in order to better match behaviour with attitude within this segment. Possible strategies pertain to stimulating awareness; a strong focus on associations of high welfare products with for instance a better taste or with benefits in terms of health and safety; and stimulation of trial purchases through free-samples and promotions. Finally, S6 also yields clear marketing opportunities. Seen their limited size (11.1 percent) and about 60 percent of them being vegetarian, this group constitutes only a small market for meat and other livestock products. Notwithstanding the small size, this group has a very high commitment and a very high willingness to pay. To most effectively reach this niche market, products should strongly focus on high animal welfare standards, for example through clear and credible labels backed up by trustworthy control, traceability mechanisms and personal reassurance. 167 Part IV Factors influencing the citizenconsumer ambivalence Chapter 10 : Citizen – consumer ambivalence Chapter 10 Citizen-consumer ambivalence: influencing factors Adapted from Vanhonacker, F., and W. Verbeke. 2009. Buying higher welfare poultry products? Profiling Flemish consumers who do and do not. Poultry Science 88: 2702-2711. Abstract In the previous chapter, discussion about the attitude-behaviour gap has already been introduced. This chapter further elaborates on this issue and groups consumers according to their reported buying frequency of higher welfare eggs and higher welfare chicken meat. Similarities and dissimilarities between these groups are mapped in terms of individual characteristics, product attribute importance, perceived consumer effectiveness, perception of higher welfare products and attitude toward a welfare label. These determinants relate directly or indirectly to the influencing factors included in this dissertation’s conceptual framework. The research methodology applied was a quantitative study with cross-sectional consumer survey data collected in Flanders in spring 2007 (n=469). Pro-welfare behaviour was unevenly distributed across different consumer segments, despite general interest and concerns for bird welfare. A consistent choice for standard (no welfare premium) poultry products was related to strong perceived price and availability barriers, to a low importance attached to ethical issues as product attributes and to a low perceived consumer effectiveness. A consistent choice for products with higher welfare standards to the contrast associated with a high importance attached to ethical issues, a low impact of price and availability perception, a strong association of higher welfare products with product attributes like health, taste and quality and high perceived consumer effectiveness. The identification of market segments with common characteristics is essential for positioning higher welfare products and developing effective communication strategies. Finally, a welfare label emerged as an appropriate communication vehicle for consumers who engage in pro-welfare behaviour and who experienced the label as a solution to lower the search costs for higher welfare products. Additionally and beyond the scope of this paper, a table is inserted at the end of the chapter, providing some data on ‘having a pet’ as attitude modifier, given that this variable was only incorporated in the survey used for this paper. 171 Part IV 1. Introduction 1.1. State of the art The present affluent Western society is increasingly interested and concerned about the welfare standards of food producing animals (Boogaard et al., 2006; European Commission, 2005; Harper and Henson, 2001). However, market shares for animal products with a welfare provenance remain rather marginal (FAWC, 2006). This inconsistency with Fishbein and Ajzen’s (1975) theory of reasoned action, where marketplace behaviour is assumed to follow an attitude, has already been explicitly discussed in the field of green consumerism (e.g. Laskova, 2007) and sustainable food consumption (e.g. Vermeir and Verbeke, 2006), but has never been the specific scope of research in relation to the consumption and/or purchase of higher welfare poultry products. Available data on determinants of the attitude-behaviour gap pertain to the specificity of the food buying process (IGD, 2007; Von Alvensleben, 1997), the general dominating importance of other product attributes (Verbeke and Viaene, 1999a), the inadequacy of information provision (Delezie et al., 2006; Harper and Henson, 2001; Korthals, 2001b) and to fact that consumers are a heterogeneous group of individuals (Verbeke, 2005). Besides a substantial consumer segment that hardly considers animal welfare when purchasing animal products, there is a segment that does buy higher welfare products very frequently (IGD, 2007; Ingenbleek et al., 2004). In literature, a smaller gap between attitude and behaviour (or behavioural intention) has been reported for consumers with a higher involvement, with a lower uncertainty in relation to information and knowledge, with a higher behavioural control in terms of availability and higher perceived consumer effectiveness (Vermeir and Verbeke, 2006) and with a higher importance attached to animal welfare (Verbeke and Vackier, 2004). 1.2. Rationale and scope This chapter concentrates on profiling consumer segments that differ in their buying frequency of higher welfare eggs and chicken meat (further termed as pro-welfare behaviour). The specific focus on the poultry sector is motivated by the significance of the sector in the study area, which is Flanders, by significant consumption rates of eggs and chicken meat among Flemish consumers and by the fact that welfare standards for laying-hens and broilers are perceived worse as compared to the other main production animals in Flanders (pigs and cattle). As a consequence eggs and chicken meat are considered as “welfare sensitive products” and improvements in bird welfare have been identified as outspoken strong expectation by European consumers (Castellini et al., 2008). As such the purchase of higher welfare animal products is thought to be concentrated in the poultry sector (FAWC, 2006; IGD, 2007). Whereas most studies have started from attitudes that relate to concerns and interest in animal welfare, this study will offset with distinguishing consumer segments based on their reported 172 Chapter 10 : Citizen – consumer ambivalence purchase frequency of eggs from hens that are kept under higher welfare standards (higher welfare eggs) and of meat from broilers that are raised under higher welfare standards (higher welfare chicken meat). The specific goal is to investigate the role that different personal characteristics play in guiding the level of pro-welfare buying behaviour of eggs and chicken meat. Given the scarcity of information about pro-welfare behaviour, we mainly formulate the study hypotheses based on findings from green and sustainable or ethical consumerism and seek to what extent they apply to pro-welfare behaviour. Vermeir and Verbeke (2006) defined the ethical consumer – a concept related to “higher welfare consumer” – as a middle-aged person with a higher income, who is above average educated, with a prestigious occupation. No gender differences were reported. We verify the interest and concern in animal welfare in general and in bird welfare more specific and the extent to what it translates in pro-welfare behaviour. Hypotheses are that a higher translation of positive interest and concern in pro-welfare behaviour is positively associated with a higher importance of animal welfare both as a social theme and as a product attribute, a higher perceived consumer effectiveness (PCE), a stronger association of higher welfare products with positive product attributes and lower perceived barriers (price and availability) for purchasing pro-welfare products. A final hypothesis relates to an expected higher need for a welfare label and higher willingness to pay for products with a welfare provenance among consumers who already engage more in pro-welfare behaviour. Establishing consumer profiles is relevant since it can provide a guide as to how and who to target and how to position higher welfare products or communicate welfare efforts effectively to consumers. As a secondary objective of this study, we are also interested to compare the results for higher welfare eggs and higher welfare chicken meat, given some different characteristics of both products: eggs do not really have a substitute, hence being more price inelastic as compared to chicken meat; eggs are a derivate of the animal whereas chicken meat concerns the consumption of the animal itself after slaughter; and eggs, more than chicken meat, have been subject to initiatives that are believed to associate with better animal welfare among consumers (e.g. free range, outdoor access, lower density, more space). 2. Material and methods Methodological detail in terms of research approach and sampling procedure are provided in Part I, section 4, and apply to Study 3. 2.1. Study design and subjects The age of the participants ranged from 18 to 79 years, with a mean age of 40.7 (SD = 13.5), which corresponds with the population mean (40.2 years). Compared to census data, the dataset consists of a higher share of females (+10 percent), a small over-representation of people from urban areas 173 Part IV (±5 percent), an over-sampling of the provinces West- and East-Flanders (resp. +18 and +4.7 percent) at the expense of Antwerp, Flemish Brabant and Limburg (-3.2; -8.7 and -10.8 percent, respectively) and a higher share of higher educated people. The study complied with standard ethics procedures in market research. All participants in the study were adult volunteers who were informed about the scope of the study. The study did not involve the collection of sensitive personal information. Data collection was fully anonymous and all data were stored in a non-identifiable format. Also, the number and percentage of vegetarians in the study is reported. Strictly speaking vegetarianism is the practice of following a diet that excludes meat (including game and slaughter by-products; fish, shellfish and other sea animals; and poultry). However vegetarianism has several variants, some of which are more relaxed and include fish (pesco-vegetarianism or pescetarianism) or stricter and excluding eggs and dairy products on top of the meat (veganism). “Vegetarians” are referred to in this study are mainly pesco-vegetarians, i.e. consumers who eat fish, eggs and/or dairy products, but no meat. Given the study purpose it will be important to consider this group separately, especially for analyses concerning the consumption of chicken meat. 2.2. Survey Attitude towards animal welfare and pro-welfare behaviour. Two survey questions were included to account for the interest in, and concern for animal welfare. First the extent to which the respondent agreed with the statement that he/she judges animal welfare an important issue was measured on a five-point Likert agreement scale with extreme values of “totally disagree” (1) and “totally agree” (5). Second the respondent was asked to evaluate the current condition of the welfare of laying hens and broilers in Flemish animal production. Evaluations were registered on a seven-point scale that ranged form “very bad” (1) to “very good” (7). The pro-welfare behaviour measure is assessed as the buying frequency of higher welfare eggs and higher welfare chicken and was formulated as: “Out of 10 times that you buy eggs (chicken meat), how often do you buy eggs (chicken meat) that have been produced with extra care for the animal’s welfare?” Answers were registered on a ratio scale, ranging from “0” to “10”. Since no normal distribution in the buying frequency is expected, the response scale will be divided in five sections for further analysis, corresponding with five groups with a different level of pro-welfare behaviour: consumers who never buy higher welfare eggs (chicken meat) (response grade 0; further referred to as No Welfare_eggs(chicken meat)); consumers who rarely buy higher welfare products (grades 1, 2 and 3; Little Welfare_eggs(chicken meat)); consumers who buy higher welfare products to some degree (grades 4, 5 and 6; Some Welfare_eggs(chicken meat)); consumers who regularly buy higher welfare products (grades 7, 8 and 9; Regular Welfare_eggs(chicken meat)); and consumers who always buy higher welfare eggs (chicken meat) (grade 10; All Welfare_eggs(chicken meat)). Vegetarians will be considered a separate group, irrespective of their response grade. 174 Chapter 10 : Citizen – consumer ambivalence Profiling variables – demographics and meat consumption. The different consumer groups will be described and compared in terms of socio-demographic characteristics (gender, age and education) and meat consumption. Self-reported meat consumption is measured on a six-point frequency scale, ranging from “never” (1) to “daily” (6) for different animal products (beef, pork, chicken meat, eggs, fish), as well as for meat substitutes. Role of attribute perceptions, PCE and consumption motives and/or barriers. First respondents were asked about the importance of a list of food product attributes in their decision to buy animal products (“importance score”). The importance was measured on a five-point scale, anchored by “totally unimportant” (1) and “very important” (5). Product attributes were “health”, “taste”, “price”, “safety”, “trustworthiness”, “quality”, “availability” and “ethical aspect”. For the latter, explicit reference to animal welfare and environment was given between brackets. Second, perceived consumer effectiveness (PCE) (i.e. the extent to which individuals believe that their actions make a difference in solving a problem (Ellen et al., 1991)) was probed through four statements focusing on the theme of animal welfare: “One person alone can do very little for the animal’s welfare”; “Efforts concerning animal welfare by one person are useless as long as other people do not want to do something”; “Refusing products that do harm to the animal’s welfare is a good way to change the production system and the production offer”; and “An individual person can make a difference for the animal’s welfare by carefully selecting the products”. Responses were measured on five-point Likert agreement scales. Finally, questions were provided specifically in relation to eggs and chicken meat and its production. Respondents were probed for their beliefs regarding higher welfare eggs and higher welfare chicken meat in terms of the product’s health, taste, price, acceptability, safety, trustworthiness, quality and availability. The discussion will make distinction between product attributes that can be considered a motive (health, taste, acceptability, safety, trustworthiness and quality) and product attributes that rather function as a barrier for higher welfare products’ purchase (price and availability). The eight product attributes considered are similar to the ones that were probed for their importance in the respondent’s general decision process of animal food products. As a result both scores can be combined into one measure that is a possible predictor value for the level of pro-welfare behaviour. To this end, the eight belief scores are rescaled from variables ranging from 1 to 5 into variables ranging from -2 to +2 (i.e. original variables minus 3). For each respondent a predictor value for the level of pro-welfare behaviour is calculated from the eight product attributes: 8 ∑ i =1 ( IMPORTANCE _ SCORE ∗ RESCALED BELIEF _ SCORE) The predictor value can range from -16 to +80. A higher score corresponds with a higher likelihood of engagement in pro-welfare behaviour. The interpretation of the value involves some assumptions and as a consequence it is a relative measure rather than absolute. Assumptions are that respondents answer consistently; that the eight product attributes cover the most important product attributes; and that the choice for higher welfare eggs and higher welfare chicken meat is determined by product attribute beliefs. Or conversely, a strong match between the predictor value 175 Part IV and the reported pro-welfare behaviour coincides with consistent answering behaviour, a high importance of the product attribute and a strong impact of product attribute beliefs on the choice for higher welfare eggs and chicken meat. Outcome variables – interest in welfare labelling. Respondent agreement on the statement “I use label info when I make a food choice” was registered on a five-point Likert agreement scale. Next respondents were asked more specific to indicate their need for a label that refers to the welfare provenance for both eggs and chicken meat as a food product category. Answers were registered on a five-point scale, ranging from “I have no need at all” (1) to “I have a strong need” (5). In addition, they were probed for the willingness of buying eggs or chicken meat with such a label for six different price premiums (equal price, +5 percent, +10 percent, +20 percent, +50 percent, +100 percent), using a five-point Likert scale that ranged from “very unlikely” (1) to “very likely” (5). 2.3. Statistical analyses Data were analysed using SPSS 15.0 (SPSS Inc., Chicago, 2006). Means with standard deviations are presented in table format. Frequencies are provided in table format or histogram presentation. A Kolmogorov-Smirnov test is performed to verify whether the buying frequency of both higher welfare eggs and higher welfare chicken is normally distributed. A construct reliability test with Cronbach alpha statistic is performed to check for the reliability of the PCE scales. Bivariate analyses including cross-tabulation with χ²-statisitics and one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores were used to profile the segments with different levels of pro-welfare behaviour in terms of the socio-demographic, attitudinal and behavioural variables that are included in the study. 3. Results and discussion 3.1. Attitude towards animal welfare and pro-welfare behaviour A pronounced interest and concern for animal welfare is shared among our sample. The importance of the concept of animal welfare corresponded with a mean score of 4.29 ± 0.75 on the five-point scale and both the welfare of laying-hens and broilers was negatively evaluated (2.91 ± 1.59 and 2.73 ± 1.58 on the seven-point scale, respectively). A non-normal frequency distribution is found for the sample’s consumption frequency of both eggs and chicken meat (Kolmogorov-Smirnov, p<0.001), allowing to convert the original scale into five groups with a different level of pro-welfare behaviour (see material & method section of this chapter). The distribution of the sample over the groups yielded a different pattern for the two food products (Figure 4.1). For the purchase of higher welfare eggs, a U-shaped distribution was found, meaning that the highest numbers of respondents were found in the two extreme positions, i.e. either never or always buying higher welfare eggs. Both categories together correspond with more 176 Chapter 10 : Citizen – consumer ambivalence than half of the sample. The mean value of the sample (without vegetarians) (mean=5.36) indicates that on ten purchasing occasions, consumers on average buy higher welfare eggs in somewhat more than half of the occasions. From the vegetarian subsample, more than three out of four individuals always buys higher welfare eggs and another twenty percent regularly buys them. For the purchase of higher welfare chicken meat, no distinct pattern appeared (Figure 4.1). About one third of the sample was categorised in the No_Welfare group and even half of the sample in the two categories with the lowest engagement in pro-welfare behaviour. The mean value coincides with higher welfare chicken meat being bought in four out of ten purchasing occasions (μ=4.12). The higher mean value for the frequency of purchasing higher welfare eggs as compared to higher welfare chicken meat is in correspondence with the higher familiarity or awareness of consumers with initiatives related with better animal welfare. Consequently consumers are able to make a better informed and more rational choice when it concerns higher welfare eggs, resulting in a Ushape distribution. The distribution of buying higher welfare chicken meat reflects the higher unawareness and higher uncertainty towards these products. 35 32.1 30 26.6 25.3 Percentage 25 21.3 20 15 17.3 12.4 19.1 17.6 14.4 14 10 5 0 eggs No_welfare Little_welfare chicken meat Some_welfare Regular_welfare All_welfare Figure 4.1. Buying behaviour of higher welfare eggs and higher welfare chicken meat (%, nonvegetarians: n=400) These findings confirm that reported concern and interest in the issue of animal welfare in general and the welfare for laying-hens and broilers more specific do not translate equally in consumer buying behaviour (Castellini et al., 2008; FAWC, 2006; IGD, 2007; Harper and Henson, 2001; Korthals, 2001b; Shrum, 1995; Vermeir and Verbeke, 2006). The attitude-behaviour inconsistency is mainly situated in the three groups with the lowest engagement in pro-welfare behaviour, suggesting a low significance of interest and concern in the buying decision process of higher welfare eggs and chicken meat (Table 4.1). Opposite, both the strong importance for animal welfare and the poor evaluation of the current poultry welfare status among the Regular_welfare and the All_welfare group (Table 4.1) seems to translate into consistent buying behaviour. 177 Part IV Table 4.1. Attitude towards animal welfare in terms of welfare importance and welfare evaluation for consumer groups with a different level of pro-welfare behaviour as a reflection of a possible attitude-behaviour gap Pro-welfare behaviour group Little Some Regular No Eggs Animal welfare is important to me Evaluation welfare laying-hens Chicken meat Animal welfare is important to me Evaluation welfare broilers b,c 3.95 c 3.55 b,c 4.37 a,b 2.70 a 4.37 b,c 3.36 b 4.32 a,b 2.59 3.73 c 3.63 A 4.11 b,c 3.19 a 3.97 c,d 3.11 3.82 d 3.88 Vegetarians All c,d 4.61 a 2.14 d 4.91 1.38 b 4.76 a 2.24 c N.A. N.A. a-d Scores in row with different superscript are significantly different at p<0.01 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 3.2. Socio-demographic characteristics and meat consumption For eggs as well as for chicken meat the All_welfare group was most strongly represented by the middle-aged groups and corroborates the study hypothesis (Vermeir and Verbeke, 2006). The oldest age category was strongly present in the Some_welfare groups. Engagement in pro-welfare behaviour is negatively associated with being male and higher education in our sample (Table 4.2), disconfirming our study hypothesis of no gender impact and a positive association with higher education (Vermeir and Verbeke, 2006). Possible explanations for this finding are, first, that ethical consumerism (from which the study hypothesis was derived) is a much broader concept than ‘prowelfare consumerism’, hence different characteristics can emerge. Second and more importantly the role of traditional demographic characteristics is strongly debated, especially seen that ethical (and by extension animal welfare) concerns and awareness have become a widespread issue (Diamantopoulos et al., 2003; Roberts, 1995). Table 4.2. Socio-demographic profile of segments with different levels of pro-welfare behaviour (n=400). The non-shaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat Gender (%) Age (years) Education (%) 178 Sample (%) 41.7 Male 41.7 Male No 53.5 54.3 Little 49.0 50.0 Some 51.8 31.6 Regular 42.2 37.1 All 21.2 23.6 41.7 41.7 14.3 17.3 35.7 32.7 14.3 17.3 32.7 35.7 Mean Mean 18-24 25-34 35-49 50+ 18-24 25-34 35-49 50+ 39.8 40.5 17.2 19.2 38.4 25.3 15.7 18.9 40.9 24.4 41.4 40.6 14.3 16.3 32.7 36.7 16.2 17.6 32.4 33.8 44.7 42.9 17.9 12.5 17.9 51.8 18.7 16.0 13.3 52.0 44.2 43.2 13.1 11.9 33.3 41.7 11.6 11.6 43.5 33.3 40.1 42.8 10.6 23.1 46.2 20.2 7.1 21.4 46.4 25.0 69.8 69.8 Higher Higher 69.7 74.0 85.1 87.7 53.6 62.7 77.4 67.1 65.7 56.6 χ² 27.18 21.40 p-value <0.01 <0.01 0.07 0.54 30.40 <0.01 32.75 <0.01 15.34 17.51 <0.01 <0.01 Chapter 10 : Citizen – consumer ambivalence Higher meat consumption frequency associates with lower levels of pro-welfare buying behaviour (Table 4.3). Despite the lowest (absolute) meat consumption frequency among the All_welfare group, they still do consume meat, which differentiates them from vegetarians. Lower meat consumption is compensated by a higher consumption frequency of meat substitutes, corresponding with literature findings, where the consumption of food products with a welfare provenance was associated with lower meat consumption, mainly as a compensation for the price premium they pay for higher welfare products (IGD, 2007; McEachern and Schröder, 2002). Total consumption frequency of eggs did not differ between groups (p>0.05), corresponding with the fact that there are no direct substitutes for eggs. Regarding the vegetarian group (results not shown) egg consumption frequency was slightly lower, which could be due to a share of the vegetarians being also vegan. Table 4.3. Meat consumption frequency of segments with different levels of pro-welfare behaviour (n=400). The non-shaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat Beef Pork Chicken meat Eggs Vegetarian Beef Pork Chicken meat Eggs Vegetarian Sample mean (2.89) (2.94) (2.88) (3.11) (4.68) (2.89) (2.94) (2.88) (3.11) (4.68) No 4.30c 4.33b 4.11a,b 3.84 1.75a 4.23b,c 4.29b 4.12a,b 3.89 1.92a Pro-welfare behaviour group Little Some Regular b,c b,c 4.46 4.35 3.93a,b b a,b 4.38 4.11 3.97a,b b a,b 4.37 4.27 4.09a,b 4.00 3.74 3.96 2.08a,b 1.94a,b 2.60b,c 4.42c 3.96a,b 4.13a,b,c b a 4.33 3.80 4.01a,b b b 4.28 4.30 4.14b 3.92 3.81 3.88 2.16a,b 2.28a,b,c 2.68b,c All 3.83a 3.72a 3.93a 3.90 2.89c 3.84a 3.80a 3.76a 4.07 2.88c p-value <0.01 <0.01 <0.01 0.36 <0.01 <0.01 <0.01 <0.01 0.39 <0.01 a-c Scores in a row with different superscript are significantly different at p<0.01 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 3.3. Role of product attribute perceptions In general, consumers want a healthy, tasteful, safe, trustworthy and high quality food product (Table 4.4). Product availability was commonly indicated as the least important attribute. Between group differences are mainly situated on the issue of price and the ethical aspect. The higher the engagement in pro-welfare behaviour, the higher the importance of the ethical aspect and the lower the relative importance of price (also availability for the vegetarians). Consumers generally agreed upon the higher price and lower availability of higher welfare products, two possible barriers for translating attitudes in consistent behaviour (Harper and Henson, 2001; Vermeir and Verbeke, 2006) (Table 4.5). Additionally, increasing perception scores related to possible motives of engagement in pro-welfare behaviour are found with increasing levels of pro-welfare behaviour (Table 4.5). Between group differences are stronger for eggs as compared to chicken meat. For eggs the highest between group differences (highest F-values) are found for “more acceptable” and further for “healthier”, “tastier” and “better quality”. Differences for chicken meat were only very clear for “more acceptable” and to a lesser extent “tastier”. Facing the two groups with the highest 179 Part IV levels of pro-welfare behaviour, the main perceived benefit seemed to correspond to the higher acceptability of the production method. Hence, whereas acceptability acts as a motive for consumers who engage in pro-welfare behaviour, the findings suggest that No_welfare consumers ignore the issue in order to avoid feelings of guilt (Harper and Henson, 2001). For the No_welfare_eggs group, none of the attributes was attributed a positive association (all mean scores below 3 on five-point scale), indicating that barriers clearly outweigh any possible benefit among this consumer segment. Table 4.4. Drivers in choosing animal products. Attributes are ranked with descending mean per column (n=400). The non-shaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat Health Taste Price Safety Trustworthiness Quality Availability Ethical aspect Health Taste Price Safety Trustworthiness Quality Availability Ethical aspect No 4.29 (2) 4.49 (1) 3.80 (6) 4.15 (5) 4.16 (4) 4.29 (3) 3.57 (7) 3.23 (8) 4.32 (3) 4.34 (2) 3.80 (6) 4.15 (5) 4.18 (4) 4.35 (1) 3.58 (7) 3.33 (8) Pro-welfare behaviour group Little Some Regular 4.22 (3) 4.50 (1) 4.51 (1) 4.23 (2) 4.15 (5) 4.24 (5) 3.57 (6) 3.66 (7) 3.74 (7) 3.85 (5) 4.31 (4) 4.36 (4) 3.93 (4) 4.37 (3) 4.40 (3) 4.28 (1) 4.39 (2) 4.47 (2) 3.41 (8) 3.50 (8) 3.56 (8) 3.46 (7) 3.70 (6) 4.22 (6) 4.35 (1) 4.55 (1) 4.27 (4) 4.35 (2) 4.36 (4) 4.27 (3) 3.55 (6) 3.69 (7) 3.43 (7) 4.02 (5) 4.40 (3) 4.21 (5) 4.07 (4) 4.34 (5) 4.38 (2) 4.34 (3) 4.47 (2) 4.47 (1) 3.35 (8) 3.68 (8) 3.39 (8) 3.50 (7) 4.12 (6) 4.14 (6) All 4.46 (2) 4.41 (3) 3.37 (8) 4.31 (6) 4.41 (5) 4.54 (1) 3.47 (7) 4.41 (4) 4.77 (1) 4.53 (6) 3.51 (8) 4.56 (5) 4.69 (2) 4.64 (5) 3.53 (7) 4.65 (3) Vegetarians 4.17 (2) 3.94 (5) 3.05 (8) 3.94 (5) 4.08 (3) 3.93 (6) 3.08 (7) 4.75 (1) N.A. Table 4.5. Possible barriers and motives for engagement in pro-welfare behaviour. The nonshaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat Cheaper Better available Healthier Tastier More acceptable Safer More trustworthy Better quality Cheaper Better available Healthier Tastier More acceptable Safer More trustworthy Better quality a-d No 1.88 2.00a 2.87a 3.00a 2.91a 2.86a 2.90a 3.11a 1.93 2.20a 3.04a 3.44a 3.28a 3.19a 3.11a 3.55a Pro-welfare behaviour group Little Some Regular 2.05 2.03 2.31 2.40a,b 2.40a,b 2.61b 3.21a,b 3.42b,c 3.88c,d a,b a,b 3.42 3.49 3.84b,c b b 3.55 3.61 4.26c a a,b 2.88 3.33 3.53b a,b b,c 3.15 3.44 3.60b,c a a,b 3.20 3.50 3.94b,c 2.09 2.25 2.28 2.42a,b 2.57a,b 2.40a,b 3.40a,b 3.45a,b 3.79b a,b b,c 3.64 3.98 4.20c a b 3.62 4.07 4.22b a a,b 3.24 3.67 3.75b a,b b,c 3.40 3.67 3.77b,c a,b a,b,c 3.78 3.97 4.15b,c All 2.01 2.65b 4.01d 4.15c 4.64c 3.64b 3.75c 4.15c 2.17 2.80b 3.58b 4.38c 4.65c 3.87b 3.89c 4.32c Vegetarians 2.46 3.04 4.01 4.06 4.71 3.81 4.08 4.32 N.A. Scores in a row with different superscript are significantly different at p<0.01 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 180 Chapter 10 : Citizen – consumer ambivalence Predictor values are in correspondence with the level of pro-welfare behaviour, for eggs as well as for chicken meat (Table 4.6). Especially in the case of eggs, between group differences were pronounced suggesting a high importance of product attribute perceptions and their association with higher welfare eggs in the purchase decision. In the case of chicken meat differentiation was less pronounced, where no significant differences in predictor value appeared between the three groups with the highest level of pro-welfare behaviour. The higher predictor scores for chicken meat as compared to eggs within the same group indicate that consumers are more convinced about the positive impact of higher welfare standards in chicken meat production on resulting product attributes. This probably relates to the fact that chicken meat consumption concerns the consumption of the animal itself whereas eggs are a derivate from the animal. Table 4.6. Predictor values for the level of pro-welfare behaviour Pro-welfare behaviour group No Little Some Regular All Vegetarians a a b b b Eggs -9.25 -0.49 0.17 15.53 22.20 26.57 a a,b b,c c c Chicken meat -3.54 4.72 13.04 20.42 25.37 N.A. a-c Scores in a row with different superscript are significantly different at p<0.01 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 3.4. Role of perceived consumer effectiveness A satisfactory Cronbach’s alpha of 0.72 allows considering the four different items as one construct representing PCE. The issue of PCE is frequently issued as a meaningful moderator within the literature of green and sustainable consumerism (Ellen et al., 1991; Laskova, 2007; Shrum, 1995; Vermeir and Verbeke, 2006). PCE is especially relevant for food choice, given that consumers often have limited knowledge and awareness of agriculture and its production processes and consequently lack insights into the implications of their food purchase decisions on the food supply chain (Dickson, 2001; Verbeke, 2005). Our study results confirm the moderating role of PCE in the translation of positive attitudes towards animal welfare in pro-welfare behaviour, with a higher PCE among groups with higher levels of pro-welfare behaviour (Table 4.7). In the case of eggs, the three groups with the lowest engagement in pro-welfare behaviour did not have a positive PCE (i.e. mean score lower or not significantly different from the scale’s midpoint). In the vegetarian subsample, a very strong belief in the positive impact of their own behaviour on animal welfare is seen. Regarding chicken meat, only the No_welfare and the Little_welfare group did not express a positive PCE. The All_welfare group to the contrast indicated a significant higher PCE as compared to all other groups. Table 4.7. Pro-welfare behaviour and perceived consumer effectiveness (PCE) Pro-welfare behaviour groups No Little Some Regular All Vegetarians a a a,b b,c c Eggs 2.95 2.99 3.12 3.46 3.81 4.51 a a a a b Chicken meat 3.04 3.07 3.33 3.49 3.96 N.A. a-c Scores in a row with different superscript are significantly different at p<0.01 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) 181 Part IV 3.5. Outcome variables – interest in welfare labelling A higher label use in the food buying decision process is reported by consumers with higher levels of pro-welfare behaviour (Table 4.8). A higher label use is often related to a higher consumer involvement (Verbeke and Vackier, 2004; Verbeke et al., 2007c). Table 4.8. Pro-welfare behaviour and perception on welfare label issues. The non-shaded rows represent the results for higher welfare eggs, the shaded rows correspond with the results for higher welfare chicken meat Pro-welfare behaviour groups No I take a label into account when buying food There is a need for an animal welfare label for eggs I take a label into account when buying food There is a need for an animal welfare label for chicken meat a-d Little a 2.99 a 3.17 a 3.11 a,b 3.56 b 3.75 a,b 3.63 Some b,c 3.71 b 3.67 b,c 4.10 Regular All Vegetarians b,c 4.27 c 4.58 b,c 4.37 4.07 4.20 4.01 c 4.74 c 4.89 c N.A. a 3.36 b 3.85 b,c 4.12 c 4.28 d 4.84 Scores in a row with different superscript are significantly different at p<0.01 (one-way ANOVA F-tests with Bonferroni and Dunnett’s T3 post hoc comparison of mean scores) Consumer segments with higher levels of pro-welfare behaviour indicate a higher willingness to buy eggs or chicken meat with a welfare provenance, irrespective of the price premium, confirming our research hypothesis (Figure 4.2 and 4.3). In addition to the higher purchase willingness we also witness a lower drop in willingness with increasing price premiums. For an equal price, a high purchase willingness of the labelled eggs and chicken meat was expressed in all groups. However a price premium of five percent already yields a strong drop in the purchase willingness of the No_welfare group. A further price increase even yields negative purchase willingness. Hence for this consumer group a price premium seem to be an insurmountable burden. Only among the groups with the two highest levels of engagement in pro-welfare behaviour, considerable monetary sacrifices are witnessed to attain the perceived benefits of higher welfare products. These results support the consistent translation of attitude into behaviour among the groups with higher prowelfare behaviour. 182 Chapter 10 : Citizen – consumer ambivalence 5 4.5 4 3.5 3 2.5 2 1.5 1 No Little 0% Some 5% 10% 20% Regular 50% All 100% Figure 4.2. Self-reported willingness to pay for higher welfare eggs, registered on a five-point Likert scale ranging from “very unlikely” (1) and “very likely” (5) 5 4.5 4 3.5 3 2.5 2 1.5 1 No Little 0% Some 5% 10% 20% Regular 50% All 100% Figure 4.3. Self-reported willingness to pay for higher welfare chicken meat, registered on a fivepoint Likert scale ranging from “very unlikely” (1) and “very likely” (5) It can be concluded that the gap between attitude – which is characterised by a high interest and concern for bird welfare in general – and actual engagement in pro-welfare behaviour are associated with some specific consumer characteristics. Consumers who engage strongly in prowelfare behaviour attach a high importance to ethical issues when buying animal food product, while price and availability were of minor importance. Further, they strongly associate higher welfare poultry products with important product attributes like health, taste and quality, and believe that they can contribute to the welfare problems through their choice for higher welfare products. A wide gap between attitude and pro-welfare behaviour to the contrast are related to 183 Part IV strong perceived price and availability barriers, to a low importance of ethical issues as product attributes and to low perceived consumer effectiveness. These low involved consumers do not invest extra efforts, e.g. in information search or label use and want good quality standards to be set and monitored by retailers, consumer organisations and governmental institutions (McEachern and Schröder, 2002). Segmenting consumers based on behavioural characteristics and gaining insights into common characteristics of consumers within a segment is essential for positioning higher welfare products and developing effective communication strategies. Consumer groups in between were less clear-cut to define. Especially for such less clearly profiled consumer groups, who are likely to be quite uncertain, marketing efforts could be highly influential. Vermeir and Verbeke (2006) already demonstrated that involvement, perceived consumer effectiveness and perceived product availability – proven to be key factors as motives or barriers for the purchase of pro-welfare behaviour in our study – can be successfully manipulated through communication efforts and information provision. In addition we want to stress the importance of the product attribute perception of higher welfare food products, especially in the case of eggs. It is important to create positive consumer expectations in terms of attribute evaluation, but it is even more important that these expectations are confirmed. A confirmation will yield satisfaction and stimulate re-purchase, whereas disconfirmation will make consumers to pull out (Oliver, 1980). This holds particularly for the case of higher welfare food. First of all food purchase is a process that is strongly influenced by habit, hence buying higher welfare food involves changing habits for the majority of consumers and second, uncertain consumers might face an additional barrier in terms of the price premium. Finally a welfare label seems an appropriate communication and marketing tool for consumers who engage in pro-welfare behaviour and who experience the label as a solution to lower the search costs for higher welfare products. 3.6. Impact of having a pet on attitudes towards farm animal welfare Additional to the objectives within this chapter, an analysis is included to investigate a possible impact of having a pet in the household on attitudes towards farm animal welfare (Table 4.9). The findings correspond in general with our research hypothesis. Especially of interest for this dissertation is the more utilitarian view of persons that do not have a pet in the household and a more humanitarian vision among persons with a pet in the household. This corresponds with Bradshaw and Paul (2010, p.110), who reported that anthropomorphism, defined as the projection of human thoughts, feelings and attributes on to non-human animals, is a defining component of pet-keeping. Noticeable, no difference was found in the relative importance of animal welfare as a product attribute, despite a higher importance attached to animal welfare, a poorer evaluation and higher perceived benefits from higher welfare products among persons with a pet. 184 Chapter 10 : Citizen – consumer ambivalence Table 4.9. Attitudes towards farm animal welfare. Impact of having a pet in the household Pet No pet p-value 3.76 3.43 <0.001 2.34 2.24 0.244 3.98 3.55 0.001 2.11 2.51 <0.001 1.02 0.97 0.180 3.18 2.45 <0.001 4.42 4.03 <0.001 3.87 4.35 0.002 Perceived benefits from higher welfare products Higher welfare products => better quality / taste / health / safety / trust 1 : not at all agree => 5 : strongly agree Perceived barriers from higher welfare products Higher welfare products => cheaper / more available 1 : not at all agree => 5 : strongly agree Animal treatment scale (Kendall et al. (2006)) higher score => greater concern with animal welfare 1 : not at all agree => 5 : strongly agree Animal utility scale (Kendall et al. (2006)) Evaluative concept of general utilitarianism 1 : not at all agree => 5 : strongly agree Relative importance of animal welfare as product attribute < 1 : less important than other product attributes 1 : equally important as other product attributes > 1 : more important than other product attributes “Humans and animals have the same value” 1 : not at all agree => 5 : strongly agree “Animal welfare is important to me” 1 : not at all agree => 5 : strongly agree “Evaluation farm animal welfare” 1 : very poor => 7 : very good 185 Part V Discussion and conclusion Discussion and conclusion The findings from the different studies have been profoundly discussed in the previous parts and chapters and have been linked to some study-specific conclusions. Part V has the aim to discuss and conclude on the findings in more general terms and is subdivided in three sections. The first section provides a recapitulation of the dissertation’s structure and indicates how this structure relates to the conceptual framework, to the four research objectives and to the ten research hypotheses. The second section is specifically dedicated to a general discussion of the four research objectives. The research hypotheses are verified and are related to policy recommendations. The third and final section is dedicated to the limitations of this dissertation, a critical discussion of the applied methodology and to suggestions for future research, based on the limitations or as a continuation of this dissertation’s findings. Annexed to this part and additional to the objectives and results of this dissertation, a brief review of solutions, that are proposed and discussed in literature, to realise better farm animal welfare is included (Annex 1). This Annex also reviews the role that different actors, such as stakeholders, interest groups, science or the media can play in improving farm animal welfare. 1. Recapitulation The overall objective of this PhD-dissertation was to contribute to a better understanding of the public conception of, and the public perception towards farm animal welfare. More specific, this dissertation has elaborated on the way people think about farm animal welfare. It has acknowledged that this way of thinking, together with people’s attitudes and concerns towards farm animal welfare differs between persons and it has raised some issues that contribute to a better understanding of the contested ambivalence between “citizen” attitudes and “consumer” market place behaviour. In addition, this dissertation has given special attention to piglet castration, as a welfare-related case study. Analyses were based on the conceptual framework that was developed based on literature findings and that has been outlined in Part I. Both qualitative exploratory and quantitative conclusive research methods have been carried out and statistical analysis techniques commonly used in social sciences have been applied to process the survey data. The relevance of addressing these objectives should be seen in the double role that is granted to the public in the adoption of more animal welfare production practices. On one hand, societal concerns related to farm animal welfare play a determinant role in the increased number of policies in the area – on national, European as well as global level – and stimulate sector initiatives (e.g. product differentiation). On the other hand, higher farm animal welfare is currently to a large extent left to the market. In this perspective, insights in public opinions, concerns and expectations are necessary in order to effectively address the expectations of this chain end user. Such insights also allow to better align public perceptions with scientific facts, for instance through communication about animal welfare criteria, whose importance is unknown by the lay public. 189 Part V The different parts in this dissertation correspond with the different research objectives. Each part consists of one or more chapters that are adapted from a published or submitted peer-reviewed scientific paper. Part II is devoted to the conceptualisation of farm animal welfare from a public point of view and is comprised of two chapters, corresponding with a qualitative and a quantitative research approach. The qualitative research was based on secondary data sources and focus group interviews with citizens. The quantitative study continued on the first chapter and was based on a Flemish survey, complemented with in depth interviews with stakeholder representatives and interest groups along the animal production chain. Multivariate analysis techniques and structural equation modelling were combined to obtain a multi-dimensional conception of farm animal welfare. The final concept is the resultant of an integrated approach that has started from the public opinion and has involved expert opinions to develop a structured model of farm animal welfare’s public conception. Its value as an advisory tool for the development of communication strategies and information provision, for reducing the gap between public perceptions and scientific facts through education and for supporting a constructive dialogue about the issue has been discussed. Part III is devoted to determinants of public attitudes, perceptions and concerns in relation with farm animal welfare. The first two chapters focused on the discordance between producers and citizens in their interpretation of farm animal welfare. Differences were discussed based on the model developed in this dissertation (Chapter 3), and based on the model developed by Welfare Quality® (Chapter 4). Additionally, the latter chapter elaborated on differences between demographical groups other than place-based variables and detailed on the interpretation of the concept by vegetarians. Three other chapters within this part concentrated on the issue of piglet castration. They were positioned in this part of the dissertation because of their correspondence with the place-based attitude modifiers. Namely, two chapters were based on consumer/citizen studies, one on a regional level (Chapter 5) and one on European level (Chapter 6). The third chapter was based on the opinion of Flemish pig producers (Chapter 7). Finally, two chapters were included to deal with the other attitude modifiers described in the conceptual framework. Chapter 8 discussed the impact of social structural variables on the concern for stocking density as a wellunderstood welfare-loaded aspect. Chapter 9 used these determinants as profiling variables in a segmentation analysis, in which respondents were grouped according to similarities on a typical citizen’s attitude measurement versus a more consumer-oriented measurement. In Part IV, different factors that are related to characteristics of higher welfare products as well as to consumer attitudes are discussed for their influence on the relationship between what individuals state in questionnaires and actually do in the shop. This part consists of one chapter only, in which self-reported pro-welfare behaviour for eggs and chicken is related to the different influencing factors included in the conceptual framework. 190 Discussion and conclusion 2. General discussion and policy recommendations The conceptual framework is developed in order to address the four research objectives and the ten research hypotheses. The next discussion is structured according to the four objectives of this dissertation and includes the verification of the respective research hypotheses. 2.1. Public conception of farm animal welfare An integrated study approach has resulted in a conception for farm animal welfare, that consists of seven dimensions on two levels: three animal-based dimensions with a direct impact on the animal’s welfare (Ability to Engage in Natural Behaviour; Animal Health; Suffering and Stress) and four resource- or environmental-based dimensions with an indirect impact on the animal’s welfare (Feed and Water; Housing and Barn Climate; Human-Animal Relationship; Transport and Slaughter). There are three general conclusions I want to bring out, corresponding with this outcome. First, the model, although based on measurements among lay persons, has many common grounds with the conceptions available in literature, be it in a format that betrays the limited awareness and low practical knowledge base of the public in relation to animal production practices (Beekman et al., 2002; Te Velde et al., 2002). Second, the subdivision in two levels contributes to the current discussion on the relative absence of animal-based parameters in welfare legislations, welfare assurance schemes or welfare indices10. These queries are for instance raised in publications by Bracke (2007), Evers et al. (2005), Gottardo et al. (2009), Knierim and Wincler (2009), Scipioni et al. (2009) and Sevi (2009). These parameters are more reliable for their direct impact on the animal’s welfare, though are more difficult (and costly) to obtain in a valid, reliable, repeatable and feasible way (Gottardo et al., 2009). At present, legislation and assurance schemes are almost exclusively based on resource-based indicators, an approach which can often cause (unintentional) adverse welfare effects. For example, ‘mortality rate’ in 2007 was the first animal-based parameter included in a European Council Directive (Council Directive 2007/43/EU). Third, the model incorporates the full life-span of the animal. This is important since it avoids potential negative side effects of an aspect or dimension that is not incorporated in the model, though that could strongly harm the societal value and relevance of the model. For instance, if slaughter conditions are negatively associated with the animal’s welfare, thought not part of the model, then the model would be of no value to represent ‘farm animal welfare’. 10 Examples of protocols that take animal-based parameters into account are: Bristol University Assessment System of Animal Welfare; protocol developed in Welfare Quality®; Wemelsfelder’s Quality Behaviour Assessment; Bracke’s Decision Support System; Bartussek’s Animal Needs Index 191 Part V The model and the findings of this part can be applied to address several imperfections of current welfare-oriented initiatives that originate from a poor understanding and/or a poor addressing of public expectations, concerns and perceptions in relation to farm animal welfare, and to positively stimulate market adoption of higher welfare products. Different sources have already plead to integrate the public opinion and concerns in decision making about regulatory purposes, the development of welfare assurance schemes and product differentiation (Boogaard et al., 2008; Fearon, 1998; Mirabito et al., 2008; Sørensen and Fraser, 2010). In this perspective, the popularity of free range systems could be framed, given the importance attached to natural living among the public. Garnier et al. (2003) also referenced the strength of consumer pressure in relation to a growth of the organic production in Switzerland, Austria and Germany. In addition, assurance on good animal welfare based on technical parameters should be communicated in a transparent, understandable and credible manner to consumers. Therefore, a good understanding of consumer perceptions is required. Education initiatives, communication efforts and information provision should be developed in relation to the ‘cognitive capacity’ of consumers, by which I allude to the limited consumer awareness and practical knowledge base in relation to animal production practices, in combination with the specificities of the food purchasing process. A better understanding of the public perception of the concept should also assist in reducing the discordance between citizens and other stakeholders along the chain. Additionally, it is important to stress that the model is based on public perception measures. Perceptions are based on human subjectivity and therefore do not necessarily match with reality. Perceptions differ between persons and can change or evolve over time. Given that the model is supported by different experts along the animal production chain, it can be assumed that the general structure of the model will be rather fixed, though the (perceived) importance of an aspect within a dimension and the (perceived) strength of the association between the different dimensions is likely to vary depending on the population segment considered or to be influenced by external environmental factors. In this perspective, the model allows to register possible evolutions and differences in public opinions. 2.2. Determinants of public attitudes towards farm animal welfare Public attitudes, perceptions and concerns towards farm animal welfare vary within a society. Already documented in literature, this is also confirmed throughout the different chapters of this dissertation. However, the findings from the different chapters were not all univocally consistent about the impact of the determinants and did not allow to fully confirming the first three research hypotheses. The first hypothesis, which assumed an impact of place-based variables, could only be partly confirmed (H1 partly confirmed). Farmers and/or first degree farmer’s offspring differed considerably from people with non-farming roots in both attitude towards, and conceptualisation of farm animal welfare. Animal welfare was perceived much less a problem among farmers, most strongly informed by a functional approach of the concept. Citizens from their side attribute more importance to the ‘natural’ aspect of the animal’s life and are more negative about the (present and 192 Discussion and conclusion perceived) level of farm animal welfare. Residency, i.e. rural or urban residency, was much less influential. Some of this dissertation’s study results confirmed the hypothesised higher concern among urban residents, while other studies could neither accept nor reject this hypothesis. Thus, life farm experiences appear to be more influential as compared to residency. Possible explanations pertain to the fact that animals are for the large majority housed indoors, such that contact with farm animals does not differ significantly between rural and urban inhabitants and to the relativity of the distinction between cities and the countryside in our study area. Flanders is a very densely populated area, where real distances and differences between cities and the countryside are rather small, especially in comparison to the United States, from where studies stemmed on which the research hypothesis was based. This has very recently been exemplified by a Eurostat study, in which the applied definition for ‘sparsely populated municipality’ resulted in only three rural villages in Flanders (Belga, 03/09/2010). Next to differences between farmers and citizens, it is also important to acknowledge variation within both population segments. Bock and van Huik (2007) and Hubbard et al. (2007) allocate variation within the farmer’s segment to a different conception, whereas Bruckmeier and Prutzer (2007) attribute variation to gender. Mann (2005) determined farm size, farmers’ education, age, less favoured areas and higher investments previously made as factors that influence the participation in agro-ethological programs. Differentiation within the citizen sample is discussed in the next two hypotheses (H2 & H3) and in the segmentation analyses based on animal welfare related issues (H10). With regard to social structural variables other than place-based variables, higher concerns were generally found among households without children and among younger respondents, two variables which are presumably interrelated with each other (H2b and H2c confirmed). The hypothesised higher concern among females only appeared in some of our studies, while other studies did not point on any gender impact (H2a partly confirmed). Socio-economic status through education level was non-significant in most studies and proportionate related to the level of concern in one study (H2d rejected). The impact of the individual animal-related experiential variables incorporated in this thesis followed the research hypothesis (H3 confirmed). Vegetarianism, having a pet in the household and a higher interest in animal welfare-related information all associated positively with a higher concern for the animal’s welfare. Interestingly, a higher concern for farm animal welfare was most often associated with a higher importance of natural living in the conception of farm animal welfare. These results provide some food for thought. The classical demographic variables, though significant in the study of Kendall et al. (2006) and in different other (European) studies, were not very explanatory in this dissertation. This was already reported by several authors (Dagevos, 2005; Diamantopoulos et al., 2003; Pouta et al., 2010), where the usefulness of these variables as explanatory variables for sustainable-related behaviour was explicitly questioned. Attitudes, perceptions and concerns were much stronger influenced by variables that relate to familiarity with the agricultural sector and to life-style values and beliefs with respect to animals, animal welfare and animal production. These variables could to some extent be reduced to the opposition between 193 Part V a utilitarian and a humanitarian vision, between utility versus affect, or ratio versus emotion. Different studies already abandoned the classical demographical variables in explaining differences in attitudes related to animal welfare or differences in sustainable behaviour. Vermeir and Verbeke (2006) made use of the Schwarz value scale to explain sustainable behaviour, similar to Aertsens et al. (2009) in relation to organic food consumption. Boogaard et al. (2006) applied value orientations to explain perceptual differences in relation to farm animal welfare; Te Velde et al. (2002) approached the discordance between farmers and consumers using a frame of reference consisting of values, norms, convictions, interests and knowledge to analyse perceptions; and Hoogland (2007, 2005) used the Portrait Values Questionnaire to explain societal differences in animal welfare concerns. Serpell (2004) defined, next to individual human characteristics, also animal characteristics and cultural factors (such as religion) as modifiers for human attitudes towards animal welfare. These findings involve some valuable information for policy makers and for the development of education initiatives, information provision and communication. Persons differ in their attitudes and concerns for the animal’s welfare, as well as in relation to their conception of farm animal welfare. This necessitates a segmented approach, in which profiling variables should go beyond the classical demographic variables, and information provisions, communication strategies and marketing strategies (f.e. advertising, assortment, packaging, display, product positioning, …) should address the specific expectations of the target segments (H10 confirmed). 2.3. Citizen attitudes versus consumer behaviour Citizens ask collectively for more animal friendly production practices. However, the average consumer does not act accordingly in the shop and most often chooses for the cheaper and more convenient alternative. Better insights in this apparent opposition are needed in the context of the increasing global dimension of trade in agricultural products, in which animal welfare is regulated through the market. Insights can contribute to narrowing the gap that exists between the producer, scientific evidence and the consumer, and can be supportive for future policy efforts and policy implementations. Perceived consumer effectiveness positively moderates the relation between attitudes and pro-welfare behaviour (H4 confirmed), analogous to green consumerism and sustainable behaviour. Important in this perspective, Vermeir and Verbeke (2006) have demonstrated that PCE can be successfully manipulated through communication efforts and information provision. This suggests a positive impact of programmes oriented on public education and raising consumer awareness and personal responsibility. “The more consumers understand about animal welfare and animal production, the more consumer power can make a difference” (Bill Reily, President of the British Veterinary Association, World Poultry.net, 21 June 2010). The hypothesised information asymmetry associated with farm animal welfare appeared throughout the different Parts and Chapters (H5 confirmed). The participants of the different 194 Discussion and conclusion studies share a negative evaluation of the current manner of information provision in relation to farm animal welfare, in terms of quantity, but the more so in terms of quality. Given the WTOagreements and the interests of retailers in satisfying as many consumers as possible, it is unlikely, in the short term, that animal food products produced with lower welfare levels, will entirely disappear from the supermarket shelves. In this perspective, labels are commonly seen as the most appropriate differentiation tool to market higher welfare food products. Labels, especially in the case of credence attributes, can function as a heuristic for consumers (Ingenbleek et al., 2004; Verbeke, 2005). Labels can be compulsory or voluntary, in which the latter is assumed most realistic in the present global food market (McInerney, 2004; Verbeke, 2009). The content of the schemes can vary depending on who initiates the voluntary labelling program and in this perspective McInerney (2004) distinguishes between three types: • • • Producer-driven schemes, where the aim is to create a good sector image; to improve competitiveness against substitute products like meat imports; and to maintain or recapture sales volumes; Retail-driven schemes, where the aim is to attract new consumer segments; and to secure competitive advantage in a particular market; Consumer-driven schemes, with the aim to deliver satisfaction through providing a combination of tangible benefits and intangible benefits. Specifically with regard to the content of the schemes, distinction can be made between schemes that focus on animal welfare only (e.g. Freedom Food), schemes that focus on various aspects including animal welfare (e.g. Certus, organic food, Marine Steward Council) and schemes that focus on aspects other than animal welfare but that have positive side-effects on animal welfare (e.g. traditional food products). In a similar vein, Ingenbleek et al. (2004) differentiates between four types of labelled products, depending on the strength of the requirements and the width of the domain that is covered with the requirements: “Idealists” are subject to high requirements and cover a wide domain (e.g. organic); “Case solvers” have high requirements, but more narrow (e.g. Freedom Food, Fair Trade); “Size-seekers” request lower requirements and cover a broad area (e.g. Unilever); and “Window seekers” demand low requirements and cover a small range of sustainability requirements. Thompson et al. (2007) from their side, define four types of standards: “Olympic”, with a single winner; “filters”, which separate good from unacceptable and which is the type of standard most often used; “ranks”, which categorise products from good to bad; and “divisions”, which make categories, but do not involve per se information about superior or inferior quality. This brief overview of variation in scheme content already indicates the multitude of options that are available for the consumer. This introduces the following discussion on the obstacles that the current application of welfare-related labels faces. I will present in brief some issues that should be taken into account for labels to be more effective. First and most importantly, the food purchasing process is a routine process that is largely driven by habit. Meat consumption is embedded in the gastronomic culture of the Western society and consumers buy animal food products very 195 Part V frequently in a routine way. This involves low information processing and a quick scan of the different alternative products on the supermarket shelves. Moreover, consumers are only limited aware of, and knowledgeable about animal production practices. Summarising from literature and own findings, labels should be visible at glance, simple, recognised, transparent, and credible. The requirement for easy visibility pertains to the information overload that consumers currently face on product packages. As postulated earlier, too much information can have an adverse effect and results in consumers ignoring the information (Verbeke, 2005). This further involves that the label should be present in a simple format (Harper and Henson, 2001; McInerney, 2004). With “recognised”, I refer in the first place to the content of the scheme behind the label. It should be based on a problem that is also recognised at consumer level, i.e. it should correspond with the consumers’ conception of farm animal welfare. McInerney (2004) exemplifies this with the success of free range (see also higher), and the problems that “lame-free milk” would face. Although lameness is a problem in intensive dairy production, consumers are not or poorly aware of this. Issues that currently appear on a large number of animal-based food products, linked to animal welfare (in consumers’ perception) pertain to “free range”, “grass fed”, “outdoor reared”, “absence of growth promoters” and “slower growth”. In addition, the success of a label will also depend on its notoriety. Enneking (2004) demonstrated a higher willingness to pay for well-known labelled quality products compared with less well-known ones. In a similar vein, harmonisation could be very important for future market success of a welfare label and is also recognised by the European Commission, though is challenging due to the global context of food trade and cultural and religious specificities (Doerfler and Peters, 2006). Finally, transparency and credibility both link to the issue of trust. Consumer trust in a label will be of paramount importance (Blokhuis et al., 2008; Nocella et al., 2010). Misleading, vague and confusing information should be banned, together with the use of labels as a pure marketing tool without the presence of an actual benefit (termed “valueengineering” by Veissier et al. (2008) and “greenwashing” by Lewis et al. (2008)). Further, the hypotheses related to perceived costs and availability were confirmed in this dissertation (H6 & H7 confirmed). Higher welfare products were in general perceived as more expensive and less available, which acts as barriers to purchasing intention. Nonetheless, this was less a barrier among respondents with a higher level of pro-welfare behaviour. The perception of low availability is linked to some extent to the information asymmetry discussed above, and could be improved accordingly. The price-issue has been covered in several studies, mostly related to willingness to pay experiments (Bennett and Blaney, 2002; Bennett and Larson, 1996; Bennett et al., 2002; Carlsson et al., 2007; Liljenstolpe, 2008; Nilsson et al., 2006; Olesen et al., 2010; Taylor and Signal, 2009). In addition, some interesting reflections have been raised, that merit some extra attention and that have some implications for market initiatives. Consumers are considered rather ‘calculating’ individuals than ‘price-conscious’ individuals. This involves that they are poorly aware of real food prices and are guided in their choice by price-differences between available alternatives. In this perspective, it is sometimes claimed that higher welfare products are not too expensive, but that the mainstream products are too cheap. In addition, mainstream meat products 196 Discussion and conclusion are very often used, through promotion actions, as a means to attract consumers to the shop, and further enforce the price-gap. Related, Ingenbleek et al. (2006a) point on the baleful impact of the current price-war between different retailers on the sales volume of higher welfare food products. The inverse relation between profit and higher welfare standards is commonly acknowledged, although Lawrence and Stott (2009) provide some examples where profits and welfare are aligned and can be simultaneously improved by addressing system components, e.g. dairy cow health, neonatal survival and animal temperature. The dissociation barrier was confirmed in this dissertation, given that chicken meat consumption was more strongly affected by animal welfare concerns than egg consumption, which is a derived animal product (H8 confirmed). In this perspective and in correspondence with the discussion of the fifth hypothesis, it will be important that the functional ignorance among consumers is taken into account, in information provision, communication and the development of marketing tools. “It is not because people want to eat a burger that they want to meet the cow” (quote from Steve Buscemi, as ‘James McCord’ in ‘The Island’) In this perspective, legislations requiring higher welfare standards would be most suitable, shifting away the responsibility from the consumer. However, as discussed higher, this option is little likely, certainly in the short run. As such, the adoption of a general, well-known and trusted label could be valuable. The incorporation of animal welfare as part of a wider quality label scheme or a label involving more sustainability issues could also address the dissociation barrier (and probably address a larger population), if this does not impair a proper recognition and adoption of the label. Altruistic concerns were more determinant for a lower ambivalence between animal welfare concerns and stated pro-welfare behaviour as compared to anthropocentric concerns (H9 confirmed). A positive association of higher welfare products with consumer acceptability and benefits for the animal, combined with a higher attribute importance of ethical issues, among which animal welfare, translated more strongly in pro-welfare behaviour as compared to a high association of higher welfare products with other product attributes such as quality, health, safety and sensory issues. Many studies however have coupled the success of higher welfare products to a premium value, thus relating the price premium to additional product quality benefits (‘Unique Selling Properties’) (e.g. Binnekamp and Ingenbleek, 2006). Also in recent studies in relation to organic food buying motives, both health and environmental traits were deemed important in the choice for organic products, with health traits dominating (Mondelaers et al., 2009; Shepherd et al., 2005). The dominance of health traits in the case of organic products however could be explained by the strong association of organic products with the non-use of pesticides, which has a direct impact on human health. Animal welfare on the other hand is mainly considered good for the animal and by extension, not bad for the quality of the meat. Thus, altruistic concerns inducing the purchase of higher welfare animal food products are more likely to stimulate repeat-purchase if the products do not compromise on experience attributes as compared to of lower welfare conventional alternatives. 197 Part V In relation to the findings of this PhD study, it is suggested that making consumers aware of the downsides of mainstream products in terms of health, animal welfare and environment compared to higher welfare products – within the limits imposed by legislation on comparative advertising – would be a more effective marketing strategy to justify the price premium than the, currently most often used, focus on the benefits of higher welfare products in terms of health, safety, animal welfare, environment and sensory characteristics. In other words, comparing higher welfare products with mainstream products should take the higher welfare products as a benchmark. Consumers assume that the products available for purchase meet the requirements in terms of food safety and health, if not indicated otherwise. As such, a price premium on these issues is difficult to convey to the consumer. Korthals (2003) compared this strikingly with the sales of an airline ticket. A passenger assumes that every airline company meets all criteria to guarantee a safe flight. As such, a marketing campaign in which an airline company would guarantee an even safer flight (against a price premium) would doubtfully be successful. On the other hand, a cheaper flight, marked as less safety, would probably not attract many passengers. I further want to raise a comment on the approach in which higher welfare products are related to better sensory properties. It has been shown that the consumer segment that most strongly values taste and related sensory attributes is a segment of meat lovers. This segment selects their meat based on taste as the main product attribute and wants to pay a price premium for it. However they do not take animal welfare issues into account. As such, this strategy might target the wrong segment and involves some marketing risks. Following on the expectancy-value or confirmationdisconfirmation theory (Oliver, 1980), a disconfirmation of higher expectations will result in dissatisfied consumers, and lowers the chance on repeat-purchase. If this strategy is applied, thus if the higher price is linked to tangible benefits such as taste and appearance, it will be crucial that these benefits are actually experienced by the consumer. Given that the experienced quality is dependent on many issues, apart from the production process (e.g. preparation) (Bernués et al., 2003a), this seems rather hard to accomplish. 2.4. Piglet castration The final research objective elaborated on the theme of piglet castration. The routine practice of surgical castration without anaesthesia and/or analgesia in Belgium as well as in many other European countries has become a very topical debate in recent years, with an intensive search for short and long-term feasible alternatives. In this dissertation I have mainly focused on immunocastration as a possible alternative. One of the major concerns that are often associated with this alternative method is whether or not consumers will readily accept and purchase pork from immunocastrated animals. This makes it relevant to research from a consumer perspective and adds to the existing knowledge available from sensory studies and data related to production parameters. 198 Discussion and conclusion Despite the debate on the issue on national and international level and the campaigns from animal welfare organisations, the general public appeared to be poorly aware of the practice and even less of possible alternative practices. This implies that people have to be provided with information, before attitudes can be measured, but equally allows researchers to measure the impact of different levels of information provisions. Throughout the different studies in this dissertation, consumer acceptance of immunocastration was not likely to be a problem. Consumers preferred immunocastration over surgical castration with or without anaesthesia or analgesia; a choice which was motivated most strongly in terms of better social acceptability and better animal welfare. Textual messages with different types of content did not have any impact on this preference. Other recent findings, not presented in the result part of this dissertation, further show that the introduction and the wordings used to present the practice as well as the use of audio-visual messages are likely to play a role (Tuyttens et al., 2010b). In a study done in January 2008 where immunocastration (which is the term used in scientific midst) was presented as “chemical castration with synthetic vaccine”, 44.6 percent of the respondents was in favour of this practice over the current practice, 28.2 percent was neutral, and 27.2 percent favoured the routine practice which was surgical castration without anaesthesia or analgesia. In the study presented in Chapter 5 (March 2008), where “immunocastration” was used, these percentages were 63.7 percent, 15.5 percent and 20.8 percent. Finally, the study presented in Chapter 6 (November 2009) resulted in 72.0 percent favouring the “vaccine method” versus 11 percent favouring surgical castration. All were thus in favour of immunocastration, though it is clear that also the word use will play a role. It is however important to notice that the percentages are obtained from different independent studies, such that the differences in percentages are not only due to this different word use. Next to the impact of word use, we also found that using audio-visual material was especially baleful for surgical castration. This actually corresponds with the conclusions in relation to the ninth hypothesis, i.e. that the preference for immunocastration (the higher welfare alternative) is related to a negative attitude towards surgical castration rather than to a beneficial attitude towards immunocastration per se. This message frame was much more an emotional message as compared to the more rational textual message (Tuyttens et al., 2010b). Interestingly and corresponding with previous parts of this dissertation, producers’ opinion differed from the public opinion. The producers did not associate the routine practice with a strong impairment of the animals’ welfare. They strongly preferred the routine practice to all other alternatives (and the production of entire males in particular), a choice that was strongly based on economic motives and claimed uncertainty about the effective elimination of boar taint (which can be considered an economic motive too for the producer). A successful adoption of an alternative practice will equally depend on the farmers’ support. This dissertation’s findings suggest that a legislative ban on the routine practice is likely to invoke farmers’ resistance and dissatisfaction, necessitating convincing arguments and/or incentives related to possible alternatives to encourage adoption by pig producers. 199 Part V Specifically for immunocastration, worries also exist among retailers (who question the proper elimination of boar taint, fearing to lose customers), slaughterhouses (who fear extra work and costs at the slaughter line associated with by products) and the farmers’ union (who have worries about the competitive position of their sector, related to a possible refusal of the meat in neighbouring countries, and about the difficulty to separate meat from immunised pigs and surgically castrated males and further question the farmers’ safety and the proper elimination of boar taint). As such, this theme remains interesting for further research and will need pioneering producers (e.g. pig farm ‘Wolkenhoeve’ in Geel, Belgium) and retailers (e.g. Colruyt) to make it successful. 3. Limitations and future research Every research is based on a specific research design and applies a certain methodology and analysing techniques. This choice imposes some limitations on this dissertation and provides, together with the findings from different other studies, material for future research. In this section, limitations related to the geographical scope, sampling method, analysing techniques, conceptual framework, scope of the research and the time frame are discussed. Finally some interesting reflections for future research are provided. Geographical scope. This research, with the exception of one study (Chapter 6), is limited to Flanders, Belgium. Due to an open market however, animal welfare is actually a global matter. Available studies in different European countries all indicate public concerns for animal welfare, though they do not all agree to the same extent with the introduction and adoption of higher welfare requirements (see for example Eurobarometer study on farm animal welfare (European Commission, 2005)). Further comparison between European countries can be valuable, especially for covering countries with importations from Belgium, but also for countries outside Europe, like Brazil and Thailand. These countries are not subject to EU-regulations on animal welfare and are able to market lower priced animal products on the European market due to scale economies. Also in many countries animal welfare is not (yet) an issue. Note in this respect that a Chinese expression for “animal welfare” still has to be invented. In addition, comparisons related to cultural and religious differences should be further researched, given the growing and market relevant segment of individuals with foreign origin. Also, farm animal welfare is a competence of the federal government in Belgium. As such, possible enactments following on citizen needs urge for an understanding of public attitudes in both parts of Belgium. Flanders and Wallonia differ considerably on some issues that could be expected to influence public attitudes, such as population density, ratio of extensive versus intensive animal production, types of animals and animal production systems and practices. Sampling method. In most of the studies presented in this dissertation, data were gathered through convenience sampling. Resulting, findings can not be extrapolated as such to the overall 200 Discussion and conclusion population, but only apply within the characteristics of the sample. In different studies, we have tried to compensate this limitation through for instance the use of weighing coefficients or to make a more representative selection of respondents within a larger sample. Still, ‘representativeness’ is always based on a selection of criteria or characteristics, thus excluding other possibly relevant characteristics (e.g. religion). Additionally, different studies involved a segmentation analysis, which also compensates partly for this limitation. Segmentation analyses reveal the presence of different market segments and allow profiling them. A non-representative sample does not influence the existence and the profile of these segments, but does not allow making general conclusions on segment sizes. Also, most of the data is gathered via web surveys. This excludes persons who do not have access to internet, but allows collecting a substantial amount of good quality data in a relatively short time notice against low costs. The resulting samples often display some bias towards higher education. Furthermore, the current internet penetration is much higher than a decade ago, when web surveys were much more criticised. Analysis techniques. Most of the analyses are performed in SPSS and pertain to the standard univariate and multivariate analysing methods, commonly used in social sciences for measuring and analysing attitudes and perceptions through questionnaires. Chapter 2 applied structural equation modelling using the LISREL software, which is more advanced and currently much applied and valued. Examples of studies that applied more advanced analysing methods to quantify attribute importance pertain to the use of choice experiments (e.g. Christensen et al., 2006; Pouta et al., 2010) and the application of Information Display Matrix (Zander and Hamm, 2010). Many studies also apply contingent valuation surveys to research willingness to pay for higher welfare products (e.g. Bennett and Larson, 1996; Carlsson et al., 2007; Liljenstolpe, 2008; Nocella et al., 2010; Olesen et al., 2010). Other variants pertain to mixed logit models (Ubilava et al., 2008) and Vickrey auctions (Napolitano et al., 2008). These more advanced techniques better allow taking social desirable answering behaviour into account and approaching more adequately insights in consumers’ behaviour or behavioural intention. Conceptual framework. Literature provides different models to conceptualise the role of perceptions and attitude modifiers in relation to farm animal welfare (e.g. Austin et al., 2005; Nelson, 2004; Te Velde et al., 2002). In this dissertation, I have made use of the framework presented by Kendall et al. (2006), which was mainly based on social structural variables. Apart from the level of familiarity with animal production and individual animal related experiential variables, this framework was not as explanatory in our local setting as it was in the United States. Further research could incorporate more attitudinal variables, beliefs, life-style variables, and value-orientations. In addition, the conceptual framework was mainly based on measuring attitudes and perceptions, and could further be elaborated with more research into behaviour and behavioural intention. 201 Part V Scope of the research. Most of this dissertation focused on farm animal welfare in general. Also the model with regard to the public conception of farm animal welfare is developed for farm animal welfare as a general concept and “animals” as an overall subject or category. It is obvious that the conception may differ dependent on the type of animal and housing system. As such further specification of the model will be necessary in order to make it a useful tool. Next to further research on attitudes depending on type of animal or housing system, it will also be relevant to research variation in consumer attitudes and citizen expectations depending on different meat types, like fresh meat, processed meat, meat mixtures, etc. Time frame. Societal perceptions are constantly evolving in relation to changing life styles, emerging innovations, demographic evolutions, media agenda, changes in knowledge, and so on. As such every cross-sectional study on attitudes is a snapshot in time. For this dissertation this limitation especially holds for the piglet castration theme and the respective attitudes of both consumers/citizens and producers towards the issue. For instance the producer data was gathered in 2007-2008. At that time immunocastration was not yet well known among producers, while at this moment immunocastration has been introduced on the market. Repeating the study would most probably give different results and would be interesting for comparison, especially given the debate on which the theme has been (and still is) subject to. Future research. Some ideas for future research have already been discussed in association with the limitations of this dissertation. Additionally the findings of this dissertation also provide a basis for future research, together with some recent developments that are related to some extent with animal welfare. Further research can be relevant to double check the resulting conception as proposed in this dissertation in order to increase its validity. For example, the resulting conceptualisation can be presented to and discussed with citizens in a qualitative study. Alternatively, the proposed conception can be quantitatively validated using additional consumer segments, e.g. from different geographical settings. Otherwise in a strict sense the result reflects a social-scientific conception rather than a citizen conception. Also, further research is necessary to evaluate different alternatives of communicating animal welfare to the consumer, either through wider quality label schemes, stand-alone animal welfare label schemes or the incorporation of animal welfare in sustainable oriented label schemes. Next, different possible solutions and strategies are presented in literature for animal welfare improvements, but are mainly discussed as a school of thought (see Annex 1). It would be interesting to verify their practicability and the opinion of consumers and citizens. Another issue is the price issue. Much will depend on the costs of welfare improvements at producer and consumer level. More advanced willingness to pay experiments and testing different price-strategies would be interesting and relevant in this perspective. Finally, future research could anticipate on recent innovations and evolutions and investigate the public’s opinion on for example genetic modification, cloning of animals, in vitro meat, the consumption of insects and the role of animal welfare in the acceptance and/or rejection of these issues. 202 Improving animal welfare – solutions and stakeholder roles ANNEX 1. Improving animal welfare – solutions and stakeholder roles In this Annex a brief and descriptive overview of some literature will be given that relates to solutions proposed to improve the level of animal welfare. First five specific studies or publications will be grasped in which different solutions are suggested and discussed. These solutions will not be discussed in depth here, but are incorporated because of their relevancy as a complement to the scope of this dissertation. Also, it can be of interest for further research to explore the citizen and consumer opinions on these propositions. In a second part, the role that different actors can play will be outlined, referencing studies that have focused specifically on this. As such, this section is informative and brings together existing knowledge, rather than adding knowledge. 1. Selected overview of solutions proposed to improve the level of animal welfare 1.1. Four ways to progress (Aerts and Lips, 2010) In a book, entitled “Animal and Welfare” [translated from “Dier en Welzijn”], four strategies were discussed that could make it possible to, within a certain region, actually contribute to improve the level of farm animal welfare, within a certain region. 1. Livestock farmers should show the present, intensive animal production as it is, and point on the existing welfare problems of the animals. This should motivate the consumer to pay a price premium for higher welfare animal food products. 2. Legislation should suppress dishonest communication about animal products. 3. Win-win situations should be searched for, such that the animal’s welfare improves, without an economic disadvantage for the farmer. 4. The government should pay the deficit to farmers that are willing to work with housing systems that are better for the animal’s welfare (and that are not yet competitive). 1.2. Defra report (McInerney, 2004) In a report for Defra, entitled “Animal Welfare, Economics and Policy”, McInerney (2004) distinguished between market based approaches and a focus on livestock farmers as policy measures for welfare improvements: 203 Annex 1 Market-based approaches Focus on livestock farmers • Pull strategy (market-led) • Welfare subsidies • Farm assurance schemes • Welfare taxes • Information and education programmes • Tradable permits • Consumer subsidies • Labelling 1.3. EU FAIR CT98-3678 (Harper and Henson, 2001) In the final report of a project entitled “Consumer Concern about Animal Welfare and the Impact on Food Choice”, and financed by the European Commission’s FAIR programme, the following five possible scenarios were developed: 1. Compulsory labelling. 2. Minimum standards. 3. Change in agriculture standards. 4. Education of consumers. 5. Voluntary code of practice. 1.4. LEI-report a (Ingenbleek et al., 2006a) This report is resulting from a study conducted in the Netherlands, on the authority of the Ministry of Agriculture, Nature and Food Quality, and is entitled “Paying for animal welfare; Barriers and solution directions in consumer markets and business-to-business markets” [translated from ‘Betalen voor dierenwelzijn: Barrières en oplossingsrichtingen in consumentenmarkten en businessto-business markten’]. The report is focused on how the additional costs, associated with improvements in animal welfare, can be managed, and is based on a literature review that has concentrated on consumers, retailers and chains. In the conclusions of the report, the following five future directions were suggested: 1. Removal of animal welfare unfriendly products from the shelves. 2. Linking animal welfare to parallel performed cost reductions. 3. Linking animal welfare to customer value. 4. Price premiums for welfare costs. 5. Separate payments for animal welfare. 1.5. LEI-report b (de Bakker and Dagevos, 2010) A second LEI-report, also performed on the authority of the Dutch Ministry of Agriculture, Nature and Food Quality, has explored some possibilities to realise more sustainable meat consumption from a consumer perspective. The report is entitled “Meat lovers, meat reducers, and meat avoiders”, with “sustainable protein consumption in a carnivorous consumption culture” as subtitle 204 Improving animal welfare – solutions and stakeholder roles [translated from: “Vleesminnaars, vleesminderaars en vleesmijders. Duurzame eiwitconsumptie in een carnivore eetcultuur”]. It presents eight marketing strategies for alternative protein products that can make the consumption pattern more sustainable. Though this is not directly related to higher welfare animal products, all of the suggested strategies associate with lowered meat consumption, and thus a less intensive animal production. The strategies are: 1. Hybrid meat products. 2. Animal species with a lower ecological foot print. 3. Meal concept with none or less meat. 4. Vegetarian meat substitutes. 5. Sustainable farmed fish. 6. Organic meat. 7. Proteins from insects. 8. Lower meat consumption. 2. Role of different actors In this part a literature overview is given in which the role that is allocated to different stakeholders, interest groups, or scientific disciplines along the animal production chain are presented and discussed. In total, I distinguished between nine groups of actors. However, it is commonly acknowledged that the way to welfare improvements should be an integrated approach, with constructive dialogues, and an effective and transparent transfer of information to the consumer. 2.1. European Union Only some decades ago, the European Union (at that time still called the European Economic Community (EEC)) has initiated initiatives to protect animals. The main driver was related to economics and trade, notably to avoid unfair competition within the common market due to different national standards of farm animal welfare (Veissier et al., 2008). The first Community legislation on animal welfare was adopted in 1974 and concerned the stunning of animals before slaughter (Council Directive 74/577/EEC) (Blokhuis et al., 2008; Horgan and Gavinelli, 2006). Later initiatives were increasingly informed by the animals’ interest, the more so with the recognition of animal sentience in the “Protocol on Protection and Welfare of Animals” annexed to the EC Treaty of Amsterdam in 1999, and reconfirmed in the Treaty of Lisbon in 2007. The European Union can influence the welfare level mainly through legislation. Veissier et al. (2008) give a nice overview of how the legislative texts are finally achieved. In brief, animal protection falls under the responsibility of DG SANCO, which is the General Directorate for the Health and Consumer Protection. To create a new legislation, DG SANCO consults a scientific committee as part of EFSA (Panel on Animal Health and Welfare). This committee, in turn, appoints a working group of scientific experts to make a scientific report with recommendations. This report is then used by DG SANCO to draft a directive, and becomes a Council Directive when it is approved by the Council of 205 Annex 1 Ministers of the EU. These directives are then translated into national regulations. The legislation in the different EU Member States must be at least conforming to the European regulations. Auditing is done by the European Commission’s Food and Veterinary Office. A summary of Council Directives is given in Veissier et al. (2008). In general, EU Directives are mainly steered by the following trends (Veissier et al., 2008, p.282): • To increase space allowance per animal; • To permit interactions between animals, and hence to encourage group housing; • To give more freedom of movement; • To provide animals with an enriched environment; • To feed animals a regimen consistent with their physiological and behavioural needs; • To limit painful intervention. In brief (and without the aim of being exhaustive), the following initiatives further underline the response of the European Union to the public demand for more animal friendly production practices: • Cooperation with organisations such as the OIE to promote the development and implementation of internationally recognised animal welfare standards (Blokhuis et al., 2008); • Negotiate the status of animal welfare as a non-trade concern with the WTO (Horgan and Gavinelli, 2006); • Negotiate animal welfare standards to be incorporated into bilateral agreements between the EU and Third Country suppliers of animals and animal products (Horgan and Gavinelli, 2006); • Financing European projects such as Welfare Quality® (www.welfarequality.net); • Community Action Plan on the Protection and Welfare of Animals (http://ec.europa.eu/food/animal/welfare/com_action_plan230106_en.pdf); • Feasibility study on animal welfare labelling. Given the relevance of labelling in the context of this dissertation, I will include the summary table of this feasibility study, discussing on different options for a European label on animal welfare (Table A1.1) (http://ec.europa.eu/food/animal/welfare/farm/aw_labelling_report_part1.pdf); • Growing importance of animal welfare in recent CAP reforms (Blokhuis et al., 2008; Horgan and Gavinelli, 2006). Rather striking, original CAP reforms were at the base of the current intensive animal production and its related welfare problems (Winter et al., 1998). 206 Improving animal welfare – solutions and stakeholder roles Table A1.1. Policy options for indicating animal welfare related information on products of animal origin Option Description Baseline option 0. No change Continuation of the current situation (status quo option) Mandatory labelling 1. Mandatory labelling of the welfare standards under which products of animal origin are produced Entailing a requirement for all relevant products of animal origin to include a label of the standard/measure of animal welfare achieved for farm animals 2. Mandatory labelling of the farming system under which products of animal origin are produced Entailing a requirement for all relevant products of animal origin to include a label of declaration of the system of production of farm animals 3. Mandatory labelling of compliance with EU minimum standards or equivalence with those Entailing a requirement for all relevant products of animal origin to include a label indicating compliance with EU minimum regulated standards (or equivalent) Requirements for the voluntary use of claims 4. Harmonised requirements for the voluntary use of claims in relation to animal welfare EU law will regulate mandatory standards that must be achieved when suppliers voluntarily label products indicating a certain standard/measure of animal welfare achieved for farm animals 5. Harmonised requirements for the voluntary use of claims in relation to farming systems EU law will regulate standards that must be achieved when suppliers voluntarily label products declaring the system of production of farm animals Other options 6. A Community Animal Welfare Label open for voluntary participation A harmonised EU-wide label would be established, organised, and/or managed in a harmonised way, providing for voluntary participation 7. Guidelines for the establishment of animal welfare labelling and quality schemes Guidelines could be established at an EU level to harmonise the establishment of animal welfare labelling and quality schemes 2.2. National governments In common with the role of the European Union, the significance of national governments is mainly related to legislation. EU Member States have to follow the minimum standards set by the EU, and have to evaluate the impact of possible stricter regulations. I will briefly discuss some differences between European countries, mainly based on a presentation on the second Welfare Quality® stakeholder conference (Kjaernes et al., 2007), on a Dutch report of Ingenbleek et al. (2004), and on an article from Berg and Hammarström (2006). Kjaernes et al. (2007) distinguish between a welfare state, a supermarket, and a producer approach (Figure A1.1), while Berg and Hammarström (2006) detail on the Swedish case. In general, higher concerns for animal welfare are associated with the Northern countries, as opposed to Southern and Eastern European countries (although the assumed difference between Northern and Southern countries has been contradicted by Blandford and Fulponi (1999)). 207 Annex 1 The welfare state approach Non-differentiated market Little worry Welfare for all via state regulation NO SE HU GB The producer approach Provenance and local networks Encompassing quality Welfare via “traditional” production NL FR, IT The supermarket approach Segmented market Integration in the chain Politicisation of AW Welfare via modernisation Figure A1.1. Different country approaches (source: Kjaernes et al. (2007)) Norway exemplifies a welfare state approach. Norwegian standards are regulated by law at a level that stretches beyond the requirements of the EU directives, and involves that animal welfare is a fairly invisible product attribute. This approach works very well in Norway for some reasons. Norwegian consumers have high trust in their government, and have a superior quality perception of Norwegian products, which facilitates the justification of a price premium. In addition, Norway is not an EU member, and applies taxation on imported products, which decreases the price gap with, and the success of, cheaper animal food products. Sweden also applies a welfare state approach, with high regulatory standards and few labels and schemes. The importance of animal welfare is also reflected in the formation of an independent Animal Welfare Agency, where before, issues related to housing, management and welfare were part of the wider Swedish Board of Agriculture. Sweden, opposite to Norway, is part of an open market, and thus vulnerable for cheap imports and discount practices among retailers. France and Italy, two countries where gastronomic traditions prevail, are positioned alike, and match strongly with the producer approach. Legislation does not stretch beyond EU minimum requirements, and initiatives that are beneficial for the animal’s welfare mainly stem from producers and producer-initiated initiatives. Animal welfare is mostly embedded in wider quality assurance schemes or in label schemes that reflect a traditional production method (e.g. Label rouge, origin labels). The Netherlands (e.g. Albert Heijn), Great Britain (e.g. Marks & Spencer, Waitrose), and Switzerland (e.g. COOP) are examples of countries where a supermarket approach is applied. Retailers take initiatives with respect to animal welfare improvement and very often require producers to follow assurance schemes, such as Freedom Food in Great-Britain, and IKB in 208 Improving animal welfare – solutions and stakeholder roles the Netherlands. Great Britain is positioned somewhat higher in the triangle, given that national legislations are more often applying higher standards of animal welfare compared with the rest of the world, including the EU. Similar to Sweden, the UK does not put restrictions on the import of meat from countries that do not impose such standards, and where costs are as such often lower. This approach has some important implications. Taking pork and pig production into account, half of the pork market went overseas when sow pens and tethering were banned in the UK, while about one third of all meat sales in the UK is currently imported (from countries with lower animal welfare standards) (Pig progress newsletter, 19/02/2010). Finally, Hungary exemplifies the more Eastern Europe located countries, where lower animal welfare concerns are present. This is translated in low state interventions, and fewer producer- and retailer-led initiatives. The challenge there is to transfer the skills and knowledge available in Western European countries, necessary to adopt more animal friendly housing systems and production practices. Specifically for Belgium, as the study area of this dissertation, a position in between the Netherlands on one hand and France and Italy on the other hand can be concluded. National legislation follows the EU minimum requirements, with some exceptions (e.g. the recent decree on the use of anaesthesia during piglet castration), in order not to harm the competitive position. Initiatives mainly stem from the retail sector, which enforce their delivering producers to meet quality assurance schemes. Examples pertain to Meritus for beef, Certus for pork, and Quality Control as a scheme specifically adopted by the Colruyt group, or to the ban of cage-eggs in different retailers. Finally Ingenbleek et al. (2006b) reported that the role of the government could be fulfilled either through financial measures at the producer side (e.g. investment allowance, sustainable investing, fund as security, pilot projects, cross compliance); financial measures at the demand side (e.g. consumer taxes, indirect taxes), or through investments in arrangements between companies, knowledge institutes, and NGO’s. From these options, investment allowance and arrangements were discussed as most suitable for direct application. 2.3. Non-governmental organisations Non-governmental organisations are a very broad category. Organisations that play a role in the quest for more welfare friendly production practices are animal welfare organisations (e.g. GAIA in Belgium), consumer organisations (e.g. OIVO, Test-Aankoop), independent animal-welfare related advisory councils (e.g. Farm Animal Welfare Council), and global organisations that protect the interests of the animals (e.g. WWF, OIE). These organisations can assert their influence on legislative institutions, through governance, or through codes of practice. 209 Annex 1 2.4. Retail sector In a report of Ingenbleek et al. (2004), the important role that retailers can fulfil is related to five specific points: • Retailers are faster in the development and implementation of new standards; • Their power is not limited to the country’s borders, because of their purchasing network; • They are not restrained to the same extent to international conventions as national governments; • They have a lot of power, given that they act as gatekeepers, i.e. they can refuse suppliers who do not meet their requirements; • They have control mechanisms for safety, quality and hygiene standards. In addition, the structure of the distribution chain is shaped according to a sandglass (Figure A1.2), a position that further strengthens the power of the retail sector (Ingenbleek et al., 2004; Thompson et al., 2007). Many suppliers have to sell their products to a limited number of distribution points, which again is the place where many consumers can buy their products. The supermarkets decide upon the standards that products have to meet, control shelf space, influence through advertising, and therefore have a big influence on production practices, and consumer purchasing behaviour. Consumers (160,000,000) Shoppers (89,000,000) Stores (170,000) Formulas (600) Buying desks (110) Producers (8,600) Semi-producers (80,000) Suppliers (160,000) Farmers/fishers etc (3,200,000) Figure A2.2. Supply chain funnel on European level (Ingenbleek et al., 2004, p.58) 2.5. Veterinarians Veterinarians work very closely together with livestock producers and animals, and have as such a direct impact on the animal’s welfare. Veterinarians themselves are well aware of this. However, the focus from veterinarians is often criticised for being focused too strong on animal health (e.g. Koch, 2009). In this perspective, incorporating training in animal welfare for the veterinary profession, taking into account the physiological state of the animal next to it’s physical state is 210 Improving animal welfare – solutions and stakeholder roles deemed important (e.g. Algers and Berg, 2001; de Lange, 2009; Ellis et al., 2009). Veterinarians could fulfil a supervising role, work closely together with producers to share their knowledge, and could act as an independent, impartial expert information source about farm animal welfare in order to educate the public. “To use a human analogy, we can consider the example of child welfare. Although we have the choice to raise our children isolated in closets to protect them from illness and injury, we choose instead to grant them the abilities to move around more freely and to associate with other people. We recognise that they risk illness, injury or even death by doing so, but we feel that such freedom and interaction are crucial to their welfare. The same principles should be appropriate for sow welfare.” (Koch, 2009, p.199) 2.6. Scientists Science is a very broad concept, and covers a large variety of research disciplines that have something in common with farm animal welfare. From literature, a role in improving the level of farm animal welfare is appropriated to journal editors (Marusic, 2009; Osborne et al., 2009), ethologists and animal scientists (Bertoni, 2009a, b; Fraser, 1999; Grandin, 2000; Mejdell, 2006; Sørensen and Fraser, 2010), philosophers and ethicists (Fraser, 1999; Garner, 2010; Korthals, 2001a). Journal editors can play a role in the selection of papers accepted for publication, and the (ethical) standards they apply for these papers. Ethology and animal science are important in the determination of what good welfare is, how it can be improved, and how it can be objectively obtained and measured. Finally, the philosophical and ethical reflection should not be neglected in this whole story. Animal welfare is an ethical theme, and should not be decoupled from the animals’ interests. 2.7. Industry / Animal production chain The industry itself, among them the farmers, are the most closely related to the animals. The impact of the relationship between the farmer and his animals (e.g. Hemsworth and Coleman, 1998; Windschnurer et al., 2008), and the management as a whole (e.g. Abeni and Bertoni, 2009; Anonymous, 2001; Mench, 2004), has received a lot of attention, and is commonly recognised as important for the animals’ welfare and for inclusion in welfare assessment schemes. These issues are also captured in the final model of farm animal welfare developed in this dissertation. In a Belgian project where the role for the different actors was discussed among different stakeholders, the following points come affront for the producers and the industry (Goris and Baeten, 2005): • Open house; • Work on the existing negative image; • Learn to communicate: what, how, why; • Transparency; • Allowing more guidance by veterinarians or company management. 211 Annex 1 Higher costs associated with more welfare friendly production systems and low producers margins however, often are a burden for producers to apply more welfare friendly production systems. In addition, producers have little confidence in the consumers’ willingness to pay, and have little freedom because of their dependency on the supplier of feed, slaughterhouses, and banks. 2.8. Media The media plays a determinant role in the current (rather negative and often distorted) public perception of animal production, as the sector is mainly brought into the attention of the public and the consumers in times of food scares or scandals (Niesten et al., 2003; Verbeke and Viaene, 1999b). Nonetheless, given the large impact of media in the current society, a more nuanced representation could strongly contribute to educating the public and raising their awareness of agricultural life and animal production practices. 2.9. Citizens and consumers The role and importance of citizens and consumers has been intensively investigated and discussed in this dissertation. Citizens have an important role in pronouncing the existing public concerns, more from a political perspective, while the consumers are the actual actors, who can strongly stimulate the market of higher welfare products through their purchasing behaviour. With respect to improving animal welfare through a pull-strategy, the distinction from Beekman et al. (2002), between the short- and long term was interesting. A short term improvement was deemed fairly difficult, given that values, perceptions and consumer behaviour is not likely to change from one day to another. They were however positive for long term improvements in case an intelligent mix of communicative, economic and social instruments were to be applied. 212 Summary Summary Western society, in a food secure environment, is the more and more interested in, and concerned about the way their food is produced, and how this impacts on the environment and the animals’ welfare. In order to actively and effectively involve the chain end user in strategies that have the goal to improve the animals’ welfare, it is important to have thorough insights in how the public thinks about farm animal welfare, and in their attitudes, perceptions, and concerns in relation to farm animal welfare. Urbanisation and public alienation from animal production and animal production practices have created a burden between producers and consumers/citizens, resulting in a different (and often non-compatible) view on the concept of farm animal welfare. The public conception of farm animal welfare, along with the variety in societal attitudes towards animal welfare and the ambivalence between citizen sayings and consumer doings have been the key points of interest in this doctoral research. The overall objective of this PhD-dissertation has as such been defined as to contribute to a better understanding of the public conception of, and the public perception towards farm animal welfare. In addition this doctoral research has focused on the issue of piglet castration, and the role of the public in the search for alternatives to surgical castration without anaesthesia and analgesia. The research is based on both primary and secondary data, which have been collected in the scope of six different studies that are executed independently from each other, and include different sets of respondents. Both qualitative exploratory (focus group discussions and in depth interviews) and quantitative conclusive (survey) research methods have been carried out. Five studies were limited to Flanders, a sixth study involved data gathered in Belgium, France, Germany and The Netherlands. The main findings and conclusions are highlighted hereunder, following the four research objectives of this doctoral research. The first objective was related to unravelling the public conception of farm animal welfare. An integrated study approach has resulted in a conception for farm animal welfare, consisting of seven dimensions on two levels: three animal-based dimensions with a direct impact on the animal’s welfare (Ability to Engage in Natural Behaviour; Animal Health; Suffering and Stress), and four resource- or environmental-based dimensions with an indirect impact on the animal’s welfare (Feed and Water; Housing and Barn Climate; Human-Animal Relationship; Transport and Slaughter). This model, although based on measurements among lay persons, has many common grounds with the conceptions available in literature, be it in a format that betrays the limited awareness and low practical knowledge base of the public in relation to animal production practices. Also, the subdivision in two levels contributes to the current discussion on the relative absence of animalbased parameters in welfare legislations, welfare assurance schemes and welfare indices. The model and the findings of this part can be applied to address several imperfections of current welfare-oriented initiatives that originate from a poor understanding and/or a poor addressing of public expectations, concerns and perceptions in relation to farm animal welfare, and to better align public perceptions with scientific evidence. 213 Summary The second objective consisted of the aim to reveal determinants of differing attitudes and concerns in relation to farm animal welfare within a society. Classical demographic variables were not very explanatory in this dissertation. Attitudes, perceptions and concerns were more strongly influenced by variables that relate to familiarity with the agricultural sector, and to life-style values and beliefs with respect to animals, animal welfare and animal production. These findings involve some valuable information for policy makers and for the development of education initiatives, information provision and communication. Persons differ in their attitudes and concern for the animal’s welfare, as well as in relation to their conception of farm animal welfare. This necessitates a segmented approach, in which profiling variables should go beyond the classical demographic variables, and information provisions, communication strategies and marketing strategies (e.g. advertising, assortment, packaging, display, product positioning, …) should address the specific expectations of the target segments. The third objective pertained to getting a better understanding of some influencing factors that contribute to the poor reflection of the public’s interest in, and concerns for animal welfare in actual food purchasing behaviour. The findings confirmed the hypothesised impact of the different influencing factors, and were linked to some suggestions to better align citizen sayings with consumer doings. Opportunities from initiatives that raise public awareness and consumer involvement in relation to farm animal welfare issues and that narrow the distance with the producer were concluded from the positive influence of perceived consumer effectiveness; suggestions in terms of price strategies followed on the revealed differences in price perception of higher welfare products among consumers with a high versus low engagement in pro-welfare behaviour; recommendations for marketing strategies, mainly with regard to labelling and communication, were derived from findings in relation to the public’s opinion and perception towards animal welfare-related information, product availability, dissociation, and product attribute importance. The fourth objective was related to the issue of piglet castration, and on immunocastration as a possible alternative practice. The practice of male piglet castration without anaesthesia or analgesia – a routine practice to avoid boar taint, and applied in many European countries, among which Belgium (anno 2010) – became widely contested for its compromise to the piglet’s welfare and integrity during the course of this dissertation. Consumers were scarcely aware of the practice and its possible alternatives. The reserved attitude of the different stakeholders along the pig production chain towards immunocastration, because of uncertainty about consumer acceptance could not be confirmed. To the contrast, throughout the different studies and independent from the information provided, consumers preferred immunocastration over surgical castration with or without anaesthesia or analgesia; a choice which was motivated most strongly in terms of better social acceptability and better animal welfare. Farmers were more reticent to alternative practices, mainly related to uncertainty about economic implications in terms of cost and work load. It could be concluded that the adoption of immunocastration is unlikely to result in marketplace loss of any particular target segment, but will need pioneering producers and retailers to make it successful. 214 Samenvatting Samenvatting In de Westerse samenleving is een steeds meer uitgesproken interesse en bezorgdheid vast te stellen met betrekking tot de manier waarop voedsel wordt geproduceerd en wat de impact van deze productie is op het milieu en het dierenwelzijn. Om de eindgebruiker van de voedselketen op een meer actieve en efficiënte manier te betrekken in strategieën om het dierenwelzijn te verbeteren is een goed inzicht nodig in de manier waarop mensen denken over dierenwelzijn en in de maatschappelijke attitudes en bezorgdheden betreffende dierenwelzijn. Een sterke verstedelijking en de daaraan gekoppelde vervreemding van dierlijke productie heeft een kloof gecreëerd tussen producent en consument. Dit heeft er mede toe geleid dat beiden een andere visie hebben op dierenwelzijn. De publieke opvatting over het welzijn van landbouwhuisdieren, verschillen in maatschappelijke standpunten en de ambivalentie tussen wat een individu zegt als burger en doet als consument vormen de centrale onderzoeksthema’s. De overkoepelende doelstelling van dit doctoraat is in die zin gedefinieerd als het bijdragen tot een beter begrip van de publieke opvatting over, en de publieke perceptie ten aanzien van het welzijn van landbouwhuisdieren. Bijkomend is in dit doctoraatsonderzoek aandacht besteed aan het thema biggencastratie en de rol van de maatschappij in de zoektocht naar mogelijke alternatieven voor het onverdoofd chirurgische castrateren (huidige praktijk). De bevindingen binnen dit doctoraatsonderzoek zijn gebaseerd op zowel primaire als secundaire gegevens, die verzameld zijn in het kader van zes onafhankelijke studies. Zowel kwalitatief verkennende (via focus groep discussies en diepte-interviews) als kwantitatief besluitende (via enquêtes) onderzoeksmethodes zijn toegepast. Vijf studies waren beperkt tot Vlaanderen, terwijl bij een zesde studie data zijn verzameld in België (beide landsdelen), Frankrijk, Duitsland en Nederland. De hoofdbevindingen en conclusies worden hieronder weergegeven en sluiten aan bij de vier onderzoeksdoelstellingen die in deze thesis zijn opgesteld. In de eerste doelstelling werd de publieke opvatting over het welzijn van landbouwhuisdieren onderzocht. Een geïntegreerde studieopzet resulteerde in een model met zeven dimensies op twee niveaus: drie diergerelateerde dimensies met een directe impact op het dierenwelzijn (Vermogen om natuurlijk gedrag te vertonen; Diergezondheid; Lijden en stress) en vier omgevingsgerelateerde dimensies met een indirecte impact op het dierenwelzijn (Voeder en water; Huisvesting en stalklimaat; Mens-dier relatie; Transport en slacht). Dit model toont sterke gelijkenissen met bestaande experten definities, maar is gekenmerkt door een beperkte familiariteit met de dierlijke productie. Daarnaast draagt de opdeling in twee niveaus bij tot de huidige discussie in verband met de relatieve afwezigheid van diergerelateerde parameters in welzijnswetgevingen, maatstaven en indicatoren. Het model en de bevindingen uit dit deel kunnen aangewend worden om in te spelen op verschillende tekortkomingen van huidige welzijnsgeoriënteerde initiatieven. Vaak zijn deze het resultaat van een gebrekkig begrip en adressering van publieke verwachtingen, bezorgdheden en 215 Samenvatting percepties in relatie met dierenwelzijn. Ook kunnen deze resultaten dienen om publieke percepties en (wetenschappelijke) feiten dichter bij elkaar te brengen. In de tweede doelstelling werden verschillen in maatschappelijke houding ten aanzien van dierenwelzijn onderzocht. Klassieke demografische variabelen bleken niet erg verklarend in tegenstelling tot familiariteit met de landbouwsector, waarden in het leven en overtuigingen met betrekking tot dieren, dierenwelzijn en dierlijke productie. Deze bevindingen zijn waardevol voor beleidsmakers en marketeers en voor een efficiënte communicatie naar de eindgebruiker. Personen verschillen in hun attitudes, bezorgdheden en opvattingen ten aanzien van dierenwelzijn. Dit noodzaakt een gesegmenteerde benadering, waarbij men rekening dient te houden met de specifieke verwachtingen van de verschillende doelsegmenten bij het opstellen van informatie-, communicatie- en marketingstrategieën (bv. reclame, productassortiment, verpakking, labelling, ...). In de derde doelstelling werd de ambivalentie tussen burgers en consumenten bestudeerd. De bevindingen bevestigden de veronderstelde impact van verschillende factoren, en konden gelinkt worden aan enkele suggesties om hetgeen een individu als consument doet in de winkel beter af te stemmen op wat datzelfde individu zegt als burger. De resultaten toonden een positieve invloed van de “perceived consumer effectiveness” op de ambivalentie tussen burger en consument. Zo kan verwacht worden dat (consumenten)gedrag beter zal aansluiten bij (burger)houding onder impuls van initiatieven om het publieke bewustzijn betreffende dierenwelzijn te verhogen. Het verschil in prijsperceptie van welzijnsvriendelijke voedingsproducten tussen consumenten die veel versus weinig diervriendelijke producten kopen werd gekoppeld aan mogelijk te volgen prijsstrategieën. Aanbevelingen voor communicatie en labels volgden uit resultaten omtrent de publieke perceptie ten aanzien van welzijnsgerelateerde informatie, productbeschikbaarheid, het loskoppelen van het stukje vlees van het dier en het belang van productattributen. De vierde doelstelling was gerelateerd aan biggencastratie. Deze praktijk is sterk ter discussie komen te staan tijdens de loop van dit doctoraat omwille van de negatieve impact op het biggenwelzijn. Consumenten bleken slechts beperkt op de hoogte te zijn van deze praktijk en nog minder van mogelijke alternatieven. De eerder gereserveerde houding van de sector ten aanzien van immunocastratie (een alternatief waar binnen dit onderzoek dieper op is ingegaan) omwille van onzekerheid omtrent consumentenacceptatie kon niet bevestigd worden. Integendeel, resultaten wezen op een consumentenvoorkeur voor immunocastratie boven chirurgische castratie. Deze voorkeur was het sterkst gemotiveerd in termen van een betere maatschappelijke acceptatie en beter dierenwelzijn. Producenten waren eerder terughoudend ten aanzien van verschillende alternatieve praktijken, een houding die kon gelinkt worden aan onzekerheid inzake economische parameters. Er kon besloten worden dat het onwaarschijnlijk is dat de invoering van immunocastratie zal leiden tot een marktverlies bij om het even welk doelsegment, maar dat er producenten en retailers zullen moeten zijn die het voortouw nemen om deze praktijk succesvol te maken. 216 References References Abeni, F., and G. Bertoni. 2009. Main causes of poor welfare in intensively reared dairy cows. Italian Journal of Animal Science 8: 45-66. Adams, C. J., and J. Donovan. 1995. Animals and women. Feminist theoretical explorations. Durham: Duke University Press, 396 pages. Adams, T. E. 2005. Using gonadotropin-releasing hormone (GnRH) and GnRH analogs to modulate testis function and enhance the productivity of domestic animals. Animal Reproduction Science 88: 127-139. Aerts, S., and D. Lips. 2010. Dierenwelzijn en consumptie. In: J. De Tavernier, D. Lips and S. Aerts (eds.) Dier en welzijn. p 85-100. Tielt: Uitgeverij Terra-Lannoo nv. Aertsens, J., W. Verbeke, K. Mondelaers, and G. Van Huylenbroeck. 2009. Personal determinants of organic food consumption: A review. British Food Journal 111: 1140-1167. Ajzen, I. 1991. The theory of planned behavior. Organizational Behavior and Human Decision Processes 50: 179-211. Ajzen, I. 2001. Nature and operation of attitudes. Annual Review of Psychology 52: 27-58. Algers, B., and C. Berg. 2001. Monitoring animal welfare on commercial broiler farms in sweden. Acta Agriculturae Scandinavica Section a-Animal Science: 88-92. Algers, B., and P. Jensen. 1991. Teat stimulation and milk-production during early lactation in sows effects of continuous noise. Canadian Journal of Animal Science 71: 51-60. Allison, J. et al. 2009. Impact of using vaccination with Improvac® on the sensory perception of meat from finishing male pigs. Proceedings of the 55th International Congress of Meat Science and Technology (ICOMST), 15-21 August, Copenhagen, Denmark. Anonymous. 2001. Scientist' assessment of the impact of housing and management on animal welfare. Journal of Applied Animal Welfare Science 4: 3-52. Appleby, M. C. 1999a. Definitions of welfare. Applied Animal Behaviour Science 65: 159-160. Appleby, M. C. 1999b. Tower of babel: variation in ethical approaches, concepts of welfare and attitudes to genetic manipulation. Animal Welfare 8: 381-390. Appleby, M. C. 2004. Considerations of the relationship between food prices and animal welfare. Journal of Dairy Science 87: 238-238. Austin, E. J., I. J. Deary, G. Edwards-Jones, and D. Arey. 2005. Attitudes to farm animal welfare: factor structure and personality correlates in farmers and agriculture students. Journal of Individual Differences 26: 107-120. Bagozzi, R. P., Y. J. Li, and L. W. Phillips. 1991. Assessing construct-validity in organizational research. Administrative Science Quarterly 36: 421-458. Banks, E. M. 1982. Behavioral research to answer questions about animal welfare. Journal of Animal Science 54: 434-446. 217 References Bartussek, H. 1999. A review of the Animal Needs Index (ANI) for the assessment of animals' wellbeing in the housing systems for austrian proprietary products and legislation. Livestock Production Science 61: 179-192. Baumgartner, J. et al. 2010. The behaviour of male fattening pigs following either surgical castration or vaccination with a GnRF vaccine. Applied Animal Behaviour Science 124: 28-34. Bayvel, A. C. D. 2005. The use of animals in agriculture and science: historical context, international considerations and future direction. Revue Scientifique Et Technique-Office International Des Epizooties 24: 791-797. Beardsworth, A. et al. 2002. Women, men and food: the significance of gender for nutritional attitudes and choices. British Food Journal 104: 470 – 491. Beekman, V., M. B. M. Bracke, T. Van Gaasbeek, and S. Van der Kroon. 2002. Begint een beter dierenwelzijn bij onszelf? LEI, Den Haag, Rapport 7.02.02, 94 pages. Belga. 2010. Belgische gezinnen kopen vaker bio. Het Nieuwsblad (01/04/2010). Belga. 2010. Vlaanderen telt geen platteland meer. VILT nieuwsbrief (03/09/2010). Bennett, R., and R. Blaney. 2002. Social consensus, moral intensity and willingness to pay to address a farm animal welfare issue. Journal of Economic Psychology 23: 501-520. Bennett, R., and D. Larson. 1996. Contingent valuation of the perceived benefits of farm animal welfare legislation: an exploratory survey. Journal of Agricultural Economics 47: 224-235. Bennett, R., J. Anderson, and R. J. P. Blaney. 2002. Moral intensity and willingness to pay concerning farm animal welfare issues and the implications for agricultural policy. Journal of Agricultural & Environmental Ethics 15: 187-202. Berg, C., and A. Hammarström. 2006. The process of building a new governmental authority based on public demands for improved animal welfare. Livestock Science 103: 297-302. Berger, I. E., and R. M. Corbin. 1992. Perceived consumer effectiveness and faith in others as moderators of environmentally responsible behaviors. Journal of Public Policy & Marketing 11: 79-89. Bernués, A., A. Olaizola, and K. Corcoran. 2003a. Extrinsic attributes of red meat as indicators of quality in Europe: an application for market segmentation. Food Quality and Preference 14: 265-276. Bernués, A., A. Olaizola, and K. Corcoran. 2003b. Labelling information demanded by european consumers and relationships with purchasing motives, quality and safety of meat. Meat Science 65: 1095-1106. Bertoni, G. 2009a. Ethical motivations (consumers) of animal welfare and the interest for farmers. Italian Journal of Animal Science 8: 897-899. Bertoni, G. 2009b. Special issue criteria and methods for the assessment of animal welfare preface. Italian Journal of Animal Science 8: 5-8. Binnekamp, M., and P. Ingenbleek. 2006. Market barriers for welfare product innovations. NjasWageningen Journal of Life Sciences 54: 169-178. 218 References Binnekamp, M., P. Ingenbleek, H. van Trijp, and K. de Vlieger. 2005. Dierenwelzijn in de markt. Agrimonitor, Februari 2005, p 1-2. Birke, L. 1993. Review essay. Society & Animals 1: 191-206. Blandford, D., and L. Fulponi. 1999. Emerging public concerns in agriculture: domestic policies and international trade commitments. European Review of Agricultural Economics 26: 409-424. Blokhuis, H. J., L. J. Keeling, A. Gavinelli, and J. Serratosa. 2008. Animal welfare's impact on the food chain. Trends in Food Science & Technology 19: S79-S87. Bock, B. B., and M. M. van Huik. 2007. Animal welfare: The attitudes and behaviour of European pig farmers. British Food Journal 109: 931-944. Bock, B. B., M. M. van Huik, M. Prutzer, F. Kling Eveillard, and A. C. Dockes. 2007. Farmers relationship with different animals: the importance of getting close to animals - case studies of French, Swedish and Dutch cattle, pig and poultry farmers. International Journal of Sociology of Agriculture and Food 15: 108-125. Bollen, K. A. 1989. Structural equations with latent variables. New York: Wiley, 528 pages. Boogaard, B. K., B. B. Bock, S. J. Oosting, and E. Krogh. 2010. Visiting a farm: an exploratory study of the social construction of animal farming in norway and the netherlands based on sensory perception. International Journal of Sociology of Agriculture and Food 17: 24-50. Boogaard, B. K., S. J. Oosting, and B. B. Bock. 2006. Elements of societal perception of farm animal welfare: A quantitative study in the netherlands. Livestock Science 104: 13-22. Boogaard, B. K., S. J. Oosting, and B. B. Bock. 2008. Defining sustainability as a socio-cultural concept: citizen panels visiting dairy farms in the netherlands. Livestock Science 117: 24-33. Borgen, S. O., and G. A. Skarstad. 2007. Norwegian pig farmers' motivations for improving animal welfare. British Food Journal 109: 891-905. Botreau, R., P. Perny, and I. Veissier. 2008. Reports on the construction of welfare criteria to different livestock species, part 3: Criteria construction for all animal types on farm. Deliverable 2.8b, subtask 2.3.1.2, Welfare Quality® (EU FOOD-CT-2004-506508). Botreau, R., I. Veissier, A. Butterworth, M. B. M. Bracke, and L. J. Keeling. 2007. Definition of criteria for overall assessment of animal welfare. Animal Welfare 16: 225-228. Bourdieu, P. 1984. Distinction: a social critique of the judgment of taste. Cambridge, MA: Harvard University Press, 613 pages. Bourdieu, P. 1990. The logic of practice. Stanford, CA: Stanford University Press, 333 pages. Bracke, M. B. M., K. Greef, and H. Hopster. 2005. Qualitative stakeholder analysis for the development of sustainable monitoring systems for farm animal welfare. Journal of Agricultural & Environmental Ethics 18: 27-56. Bracke, M. B. M. 2007. Animal-based parameters are no panacea for on-farm monitoring of animal welfare. Animal Welfare 16: 229-231. Bracke, M. B. M., B. M. Spruijt, and J. H. M. Metz. 1999. Overall welfare reviewed. Part 3: Welfare assessment based on needs and supported by expert opinion. Njas-Wageningen Journal of Life Sciences 47: 307-322. 219 References Bradshaw, J. W. S., and E. S. Paul. 2010. Could empathy for animals have been an adaptation in the evolution of homo sapiens? Animal Welfare 19: 107-112. Brambell, F. R. 1965. The report of the technical committee to enquire into the welfare of animals kept under intensive livestock husbandry systems. London: H.M.S.O. Command paper no. 2836. Brom, F. W., T. Visak, and F. Meijboom. 2007. Food, citizens and the market: The quest for responsible consuming. In: L. Frewer and H. Van Trijp (eds.) Understanding consumers of food products. p 610-623. Cambridge, MA: Woodhead Publishing, CRC Press. Brom, F. W. 2000. Food, consumer concerns, and trust: food ethics for a globalizing market. Journal of Agricultural & Environmental Ethics 12: 127-139. Broom, D. M. 1986. Indicators of poor welfare. British Veterinary Journal 142: 524-526. Broom, D. M. 1991. Animal-welfare - concepts and measurement. Journal of Animal Science 69: 4167-4175. Broom, D. M. 2001. Coping, stress and welfare. In: D. M. Broom (ed.) Coping with challenge: Welfare in animals including humans. p 1-9. Berlin: Dahlem University Press. Browne, M., and R. Cudeck. 1993. Alternative ways of assessing model fit. In: K. A. Bollen and J. Long (eds.) Testing structural equation models. p 136-162. California, CA: Sage Publications. Bruckmeier, K., and M. Prutzer. 2007. Swedish pig producers and their perspectives on animal welfare: A case study. British Food Journal 109: 906-918. Buller, H. 2009. What can we tell consumers and retailers? Proceedings of the Final Stakeholder Conference ‘Delivering animal welfare and quality: transparancy in the food production chain’, Welfare Quality®, 8-9 October, Uppsala, Sweden, p 43-46. Buller, H., and C. Cesar. 2007. Eating well, eating fare: farm animal welfare in France. International Journal of Sociology of Agriculture and Food 15: 45-58. Burrell, A., and B. Vrieze. 2003. Ethical motivation of Dutch egg consumers. Tijdschrift voor Sociaal Wetenschappelijk Onderzoek voor de Landbouw 18: 30-42. Butterworth, A. 2009. Feeding support information back to management. Proceedings of the Final Stakeholder Conference ‘Delivering animal welfare and quality: transparancy in the food production chain’, Welfare Quality®, 8-9 October, Uppsala, Sweden, p 33-37. Caporale, V., B. Alessandrini, P. Dalla Villa, and S. Del Papa. 2005. Global perspectives on animal welfare: Europe. Revue Scientifique Et Technique-Office International Des Epizooties 24: 567-577. Carenzi, C., and M. Verga. 2009. Animal welfare: review of the scientific concept and definition. Italian Journal of Animal Science 8: 21-30. Carlsson, F., P. Frykblom, and C. J. Lagerkvist. 2007. Consumer willingness to pay for farm animal welfare: Mobile abattoirs versus transportation to slaughter. European Review of Agricultural Economics 34: 321-344. Casimir, G., and C. Dutilh. 2003. Consument en burger vanuit een genderperspectief. In: H. Dagevos and L. Sterrenberg (eds.) Burgers en consumenten. p 99-110. Wageningen, The Netherlands: Wageningen Academic Publishers. 220 References Castellini, C., C. Berri, E. Le Bihan-Duval, and G. Martino. 2008. Qualitative attributes and consumer perception of organic and free-range poultry meat. Worlds Poultry Science Journal 64: 500512. Christensen, T., M. Mørbak, S. Denver, and B. Hasler. 2006. Preferences for food safety and animal welfare - a choice experiment study comparing organic and conventional consumers. Proceedings of the 1st Joint European Organic Congress, Odense, Denmark, 2 pages. Clarke, L. et al. 2008. Inherent food safety of a synthetic gonadotropin-releasing factor (GnRF) vaccine for the control of boar taint in entire male pigs. International Journal of Applied Research in Veterinary Medicine 6: 7-14. Codron, J. M., L. Siriex, and T. Reardon. 2006. Social and environmental attributes of food products in an emerging mass market: challenges of signaling and consumer perception, with European illustrations. Agriculture and Human Values 23: 283-297. Cohen, N. E., M. A. P. M. van Asseldonk, and E. N. Stassen. 2007. Social-ethical issues concerning the control strategy of animal diseases in the European Union: a survey. Agriculture and Human Values 24: 499-510. Coleman, G. 2007. Public perceptions of animal pain and animal welfare. Proceedings of the Australian Animal Welfare Strategy Science Summit on Pain and Pain Management, 27 May, Melbourne, Australia, 8 pages. Colman, E., A. Dhondt, C. De Moor, S. Tanghe, and E. Vandamme. 2007. Dierenwelzijn en consumentengedrag. Student project within the Master’s Programme to obtain the degree of Bio-Engineer, Faculty of Bioscience Engineering, Ghent University. Cronin, G. M. et al. 2003. The effects of immuno- and surgical-castration on the behaviour and consequently growth of group-housed, male finisher pigs. Applied Animal Behaviour Science 81: 111-126. Dagevos, H. 2005. Consumers as four-faced creatures. Looking at food consumption from the perspective of contemporary consumers. Appetite 45: 32-39. Dagevos, H., and L. Sterrenberg. 2003. Burgers en consumenten: tussen tweedeling en tweeeenheid. Wageningen, The Netherlands: Wageningen Academic Publishers, 160 pages. Dawkins, M. S. 2006. A user's guide to animal welfare science. Trends in Ecology & Evolution 21: 7782. Dawkins, M. S. 2008. The science of animal suffering. Ethology 114: 937-945. de Bakker, E., and H. Dagevos. 2010. Vleesminnaars, vleesminderaars en vleesmijders. Duurzame eiwitconsumptie in een carnivore eetcultuur. LEI, Den Haag, LEI-rapport 2010-003, 207 pages. De Jonge, F., and E. Goewie. 2000. In het belang van het dier: over het welzijn van dieren in de veehouderij. Rathenau Instituut, Den Haag, 130 pages. de Lange, K. 2009. Animal welfare high on the agenda at fve assembly. Veterinary Record 165: 734735. de Roest, K., C. Montanari, T. Fowler, and W. Baltussen. 2009. Resource efficiency and economic implications of alternatives to surgical castration without anaesthesia. Animal 3: 1522-1531. 221 References De Tavernier, J., D. Lips, and S. Aerts. 2010. Dier en welzijn. Tielt: Uitgeverij Terra-Lannoo nv., 215 pages. Degrazia, D. 1998. Animal ethics around the turn of the twenty-first century. Journal of Agricultural & Environmental Ethics 11: 111-129. Delezie, E., W. Verbeke, J. De Tavernier, and E. Decuypere. 2006. Consumers' preferences toward techniques for improving manual catching of poultry. Poultry Science 85: 2019-2027. Demeritt, L. 2005. Understanding consumers goes beyond demographics and lifestyles. Available on http://www.djc.com/news/en/11170267.html (last consulted: 15/09/2010). Diamantopoulos, A., B. B. Schlegelmilch, R. R. Sinkovics, and G. M. Bohlen. 2003. Can sociodemographics still play a role in profiling green consumers? A review of the evidence and an empirical investigation. Journal of Business Research 56: 465-480. Dickson, M. A. 2001. Utility of no sweat labels for apparel consumers: profiling label users and predicting their purchases. The Journal of Consumer Affairs 35: 96-119. Diederen, P. 2003. Burger, laat die consument met rust! In: H. Dagevos and L. Sterrenberg (eds.) Burgers en consumenten: Tussen tweedeling en twee-eenheid. p 19-30. Wageningen, The Netherlands: Wageningen Academic Publishers. Dockes, A. C., and F. Kling-Eveillard. 2006. Farmers' and advisers' representations of animals and animal welfare. Livestock Science 103: 243-249. Doerfler, R. L., and K. J. Peters. 2006. The relativity of ethical issues in animal agriculture related to different cultures and production conditions. Livestock Science 103: 257-262. Duncan, I. J. H. 1996. Animal welfare defined in terms of feelings. Acta Agriculturae Scandinavica Section a-Animal Science: 29-35. Duncan, I. J. H. 2005. Science-based assessment of animal welfare: farm animals. Revue Scientifique Et Technique-Office International Des Epizooties 24: 483-492. Duncan, I. J. H., and D. Fraser. 1997. Understanding animal welfare. In: M. Appleby and B. O. Hughes (eds.) Animal welfare. p 19-31. Wallingford, UK: CABI Publishing. Dunshea, F. R. et al. 2001. Vaccination of boars with a GnRH vaccine (Improvac) eliminates boar taint and increases growth performance. Journal of Animal Science 79: 2524-2535. Eastwood, P. J. 1995. Farm animal welfare, Europe and the meat manufacturer. British Food Journal 97: 4-11. Edwards, J. D., and H. P. Schneider. 2005. The World Veterinary Association and animal welfare. Revue Scientifique Et Technique-Office International Des Epizooties 24: 639-646. EFSA. 2004. Welfare aspects of the castration of piglets. Scientific report of the scientific panel for animal health and welfare on a request from the commission related to welfare aspects of the castration of piglest. EFSA - AHAW/04-087, Parma, Italy. Available on http://www.efsa.europa.eu/en/scdocs/scdoc/91.htm (last consulted: 18/09/2010). Eijck, I., C. van der Peet-Schwering, M. Kiezebrink, and A. Vink. 2007. Effect of castration of organic swine with anesthesia on the veterinary costs and physical work load of pig farmers. Tijdschrift voor Diergeneeskunde 12: 476-479. 222 References Elder, G., J. Wolch, and J. Emel. 1998. Le pratique sauvage: Race, place, and the human-animal divide. In: J. Wolch and J. Emel (eds.) Animal geographies: Place, politics, and identity in the nature-culture borderlands. p 72-90. New York, NY: Verso. Ellen, P. S., J. L. Weiner, and C. Cobb-Walgren. 1991. The role of perceived consumer effectiveness in motivating environmentally conscious behaviors. Journal of Public Policy & Marketing 10: 102-117. Ellis, K. A., K. Billington, B. McNeil, and D. E. F. McKeegan. 2009. Consumer perceptions of dairy welfare and production: A role for vets in public education? Cattle Practice 17: 148-152. Engel, J. F., P. W. Miniard, and R. D. Blackwell. 1990. Consumer behavior. Chicago: The Dryden Press, 790 pages. Enneking, U. 2004. Willingness-to-pay for safety improvements in the German meat sector: The case of the Q&S label. European Review of Agricultural Economics 31: 205-223. Estevez, I. 2007. Density allowances for broilers: where to set the limits? Poultry Science 86: 12651272. European Commission. 2005. Attitudes of consumers towards the welfare of farmed animals. Special Eurobarometer 229. Avalaible on http://ec.europa.eu/food/animal/welfare/euro_barometer25_en.pdf (last consulted: 15/09/2010). European Commission. 2006a. Commission working document on a Community Action Plan on the Protection and Welfare of Animals 2006-2010. Strategic basis for the proposed actions. Available on http://ec.europa.eu/food/animal/welfare/work_doc_strategic_basis230106_en.pdf (last consulted: 18/09/2010). European Commission. 2006b. Risk issues. Special Eurobarometer 238. Available on http://ec.europa.eu/food/food/resources/specialeurobarometer_riskissues20060206_en.pdf (last consulted: 18/09/2010). European Commission. 2007. EU consumers willing to pay for better animal welfare. Press release 22/03/2007. Available on http://europa.eu/rapid/pressReleasesAction.do?reference=IP/07/398&type=HTML&aged=0 &language=EN&guiLanguage=en (last consulted: 18/09/2010). Evers, J., D. Lips, J. De Tavernier, J. Buyse, and E. Decuypere. 2005. Extension of the Animal Needs Index for laying hens with animal-related parameters. Animal Science Papers and Reports Quarterly 23: 271-272. Fàbrega, E., A. Velarde, J. Cros, M. Gispert, P. Suárez, J. Tibau, and J. Soler. 2010. Effect of vaccination against gonadotropin-releasing hormone, using ImprovacTM, on growth performance, body composition, behaviour and acute phase proteins. Livestock Science, doi: 10.1016/j.livsci.2010.04.021. FAWC (Farm Animal Welfare Council). 1993. Report on priorities for animal welfare research and development. Surbiton, UK. FAWC (Farm Animal Welfare Council). 2006. Report on welfare labelling, 40 pages. 223 References Fearon, J. D. 1998. Deliberation as discussion. In: J. Elster (ed.) Deliberative democracy. p 44-68. Cambridge, MA: Cambridge University Press. Festinger, L. 1964. Conflict, decision and dissonance. Stanford, CA: Stanford University Press, 163 pages. Fishbein, M. 1967. Attitude and predicting behaviour. In: M. Fishbein (ed.) Reading in attitude theory and measurement. p 389-400. New York, NY: Wiley. Fishbein, M., and I. Ajzen. 1975. Belief, attitude, intention and behavior: an introduction to theory and research. Reading, MA: Addison-Wesley Publishing Company, 578 pages. Fisher, M. W. 2009. Defining animal welfare - does consistency matter? New Zealand Veterinary Journal 57: 71-73. Flynn, L. R., and R. E. Goldsmith. 1999. A short, reliable measure of subjective knowledge. Journal of Business Research 46: 57-66. Font i Furnols, M. et al. 2008. Consumers' sensory acceptability of pork from immunocastrated male pigs. Meat Science 80: 1013-1018. Fraser, D. 1995. Science, values and animal welfare: exploring the 'inextricable connection'. Animal Welfare 4: 103-117. Fraser, D. 1999. Animal ethics and animal welfare science: bridging the two cultures. Applied Animal Behaviour Science 65: 171-189. Fraser, D. 2001. The "new perception" of animal agriculture: legless cows, featherless chickens, and a need for genuine analysis. Journal of Animal Science 79: 634-641. Fraser, D. 2003. Assessing animal welfare at the farm and group level: the interplay of science and values. Animal Welfare 12: 433-443. Fraser, D. 2004. Applying science to animal welfare standards. Proceedings of the Global Conference on Animal Welfare: an OIE initiative, 23-25 February, Paris, France, p 121-127. Fraser, D. 2006. Animal welfare assurance programs in food production: A framework for assessing the options. Animal Welfare 15: 93-104. Fraser, D. 2007. Animal welfare and intensive animal production: Are they compatible? In: W. Zollitsch, C. Winckler, S. Waiblinger and A. Haslberger (eds.) Sustainable food production and ethics: Preprints of the 7th Congress of the European Society for Agricultural and Food Ethics (EURSAFE). p 31-36. Wageningen, The Netherlands: Wageningen Academic Publishers. Fraser, D. 2008. Understanding animal welfare. Acta Veterinaria Scandinavica 50: S1. Fraser, D., and I. J. H. Duncan. 1998. 'Pleasures','pains' and animal welfare: toward a natural history of affect. Animal Welfare 7: 383-396. Fraser, D., D. M. Weary, E. A. Pajor, and B. N. Milligan. 1997. A scientific conception of animal welfare that reflects ethical concerns. Animal Welfare 6: 187-205. Fredriksen, B. et al. 2009. Practice on castration of piglets in Europe. Animal 3: 1480-1487. 224 References Fredriksen, B., and O. Nafstad. 2006. Surveyed attitudes, perceptions and practices in Norway regarding the use of local anaesthesia in piglet castration. Research in Veterinary Science 81: 293-295. Fredriksen, B., A. M. Sibeko Johnsen, and E. Skuterud. 2010. Consumer attitudes towards castration of piglets and alternatives to surgical castration. Research in Veterinary Science doi: 10.1016/j.rvsc.2010.06.018. Frewer, L. 1999. Risk perception, social trust, and public participation in strategic decision making: Implications for emerging technologies. Ambio 28: 569-574. Frewer, L. J., C. Howard, D. Hedderley, and R. Shepherd. 1997a. Consumer attitudes towards different food-processing technologies used in cheese production - the influence of consumer benefit. Food Quality and Preference 8: 271-280. Frewer, L. J., C. Howard, D. Hedderley, and R. Shepherd. 1997b. The elaboration likelihood model and communication about food risks. Risk Analysis 17: 759-770. Frewer, L. J., A. Kole, S. M. A. Van De Kroon, and C. De Lauwere. 2005. Consumer attitudes towards the development of animal-friendly husbandry systems. Journal of Agricultural & Environmental Ethics 18: 345-367. Furniere, D. 2008. Consumentengedrag ten aanzien van immunocastratie als alternatief voor chirurgische castratie bij biggen. Thesis to obtain the degree of Bio-Engineer, master Agriculture, Ghent University. Garner, R. 2010. Animals, ethics and public policy. The Political Quarterly 81: 123-130. Garnier, J. P., R. Klont, and G. Plastow. 2003. The potential impact of current animal research on the meat industry and consumer attitudes towards meat. Meat Science 63: 79-88. Giddens, A. 1986. The constitution of society: outline of the theory of structuration. Berkely, CA: University of California Press, 274 pages. Giffin, B. J., J. R. D. Allison, S. Martin, P. Ward, and A. Tschopp. 2008. Consumer acceptance of the use of vaccination to control boar taint. Proceedings of the 20th International Pig Veterinarian Science (IPVS) Congress, 22-25 June, Durban, South Africa. Goris, S., and J. Baeten. 2005. Maatschappelijke indicatoren voor dierenwelzijn. DP21 (Dierlijke Productie in de 21ste eeuw), projectbrochure. Gottardo, F., M. Brscic, B. Contiero, G. Cozzi, and I. Andrighetto. 2009. Towards the creation of a welfare assessment system in intensive beef cattle farms. Italian Journal of Animal Science 8: 325-342. Grandin, T. 2000. Livestock handling and transport. Wallingford, UK: CABI Publishing, 464 pages. Griffith, M., J. Wolch, and U. Lassiter. 2002. Animal practices and the racialization of Filipinas in Los Angeles. Society and Animals 10: 221-248. Grunert, K. G. 2006. Future trends and consumer lifestyles with regard to meat consumption. Meat Science 74: 149-160. Grunert, K. G., T. Bech-Larsen, and L. Bredahl. 2000. Three issues in consumer quality perception and acceptance of dairy products. International Dairy Journal 10: 575-584. 225 References Grunert, K. G., L. Bredahl, and K. Brunsø. 2004. Consumer perception of meat quality and implications for product development in the meat sector - a review. Meat Science 66: 259272. Grunert, K. G., and C. Valli. 2001. Designer-made meat and dairy products: consumer-led product development. Livestock Production Science 72: 83-98. Hair, J., W. Black, B. Babin, R. Anderson, and R. Tatham. 2006. Multivariate data analysis. New Jersey: Prentice Hall, 928 pages. Hall, C., and V. Sandilands. 2007. Public attitudes to the welfare of broiler chickens. Animal Welfare 16: 499-512. Hamilton, M. 2006. Disgust reactions to meat among ethically and health motivated vegetarians. Ecology of Food and Nutrition 45: 125-158. Hansen, J., L. Holm, L. Frewer, P. Robinson, and P. Sandøe. 2003. Beyond the knowledge deficit: recent research into lay and expert attitudes to food risks. Appetite 41: 111-121. Hansman, H. 1999. De consument gevangen in cijfers: Zoektocht naar het bestaan van consumentenbeelden. LEI, Den Haag, Rapport 3.99.07, 77 pages. Harper, G., and S. Henson. 2001. Consumer concerns about animal welfare and the impact on food choice. Final report (EU FAIR CT98-3678), 38 pages. Available on http://ec.europa.eu/food/animal/welfare/eu_fair_project_en.pdf (last consulted: 26/05/2010). Harper, G., and A. Makatouni. 2002. Consumer perception of organic food production and farm animal welfare. British Food Journal 104: 287-299. Harrison, R. A. 1964. Animal machines: the new factory farming industry. London, UK: Vincent Stuart Ltd, 186 pages. Haynes, R. P. 2008. Animal welfare: competing conceptions and their ethical implications. Dordrecht: Springer Science + Business Media B.V., 162 pages. Heid, A., and U. Hamm. 2009. Consumer acceptance of alternatives to piglet castration without anaesthesia. Fleischwirtschaft 89: 93-98. Heinritzi, K., M. Ritzmann, and W. Otten. 2006. Alternatives of castration of suckling piglets, determination of catecholamines and woundhealing after castration of suckling piglets at different points of time. Deutsche Tierarztliche Wochenschrift 113: 94-97. Heleski, C. R., A. G. Mertig, and A. J. Zanella. 2004. Assessing attitudes toward farm animal welfare: a national survey of animal science faculty members. Journal of Animal Science 82: 28062814. Heleski, C. R., A. G. Mertig, and A. J. Zanella. 2006. Stakeholder attitudes toward farm animal welfare. Anthrozoos 19: 290-307. Hemsworth, P., and G. Coleman. 1998. Human-livestock interactions: the stockperson and the productivity and welfare of intensively farmed animals. Wallingford, UK: CABI Publishing, 176 pages. 226 References Hennessy, D., and R. Newbold. 2004. Consumer attitudes to a boar taint vaccine, Improvac® - a qualitative study. Proceedings of the 18th International Pig Veterinarian Society (IPVS) Congress, 27 June – 1 July, Hamburg, Germany, p 612. Hewson, C. J. 2003a. Can we assess welfare? Canadian Veterinary Journal-Revue Veterinaire Canadienne 44: 749-753. Hewson, C. J. 2003b. What is animal welfare? Common definitions and their practical consequences - introduction. Canadian Veterinary Journal-Revue Veterinaire Canadienne 44: 496-499. Hewson, C. J. 2004. What's animal welfare science all about? Canadian Veterinary Journal-Revue Veterinaire Canadienne 45: 254-258. Hills, A. M. 1993. The motivational bases of attitudes toward animals. Society and Animals 1: 111128. Honkanen, P., and S. O. Olsen. 2009. Environmental and animal welfare issues in food choice: the case of farmed fish. British Food Journal 111: 293-309. Hoogland, C. T., J. de Boer, and J. J. Boersema. 2005. Transparency of the meat chain in the light of food culture and history. Appetite 45: 15-23. Hoogland, C. T., J. de Boer, and J. J. Boersema. 2007. Food and sustainability: do consumers-package information mers recognize, understand and value on production standards? Appetite 49: 47-57. Hopkins, P. D., and A. Dacey. 2008. Vegetarian meat: could technology save animals and satisfy meat eaters? Journal of Agricultural & Environmental Ethics 21: 579-596. Horgan, R., and A. Gavinelli. 2006. The expanding role of animal welfare within EU legislation and beyond. Livestock Science 103: 303-307. Hubbard, C., M. Bourlakis, and G. Garrod. 2007. Pig in the middle: Farmers and the delivery of farm animal welfare standards. British Food Journal 109: 919-930. Huber-Eicher, B., and P. Spring. 2008. Attitudes of swiss consumers towards meat from entire or immunocastrated boars: a representative survey. Research in Veterinary Science 85: 625627. Huber, A., and D. W. Fölsch. 1978. Akustische Ethogramme von Hühnern. Die Auswirkungen verschiedener Haltungssysteme. Basel und Stuttgart: Birkhäuser Verlag, Basel und Stuttgart, 50 pages. Hughes, D. 1995. Animal welfare: the consumer and the food industry. British Food Journal 97: 3-7. IGD (The Institute of Grocery Distribution). 2007. Consumer attitudes to animal welfare. A report for Freedom Food by IGD. Available on http://www.igd.com/index.asp?id=1&fid=1&sid=8&tid=33&cid=311 (last consulted: 15/09/2010). Ingenbleek, P., M. Binnekamp, and H. Van Trijp. 2006a. Betalen voor dierenwelzijn. Barrières en oplossingsrichtingen in consumenten- en business-to-business markten. LEI, Den Haag, Rapport 5.06.02, 96 pages. 227 References Ingenbleek, P., M. Binnekamp, J. Van Trijp, and J. de Vlieger. 2004. Dierenwelzijn in de markt. Een drieluik van consumenten, retailers en belangenorganisaties in Europa. LEI, Den Haag, Rapport 5.04.11, 129 pages. Ingenbleek, P. et al. 2006b. Dierenwelzijn in transitie; thema's rond de implementatie van de dierenwelzijnsindex. LEI, Den Haag, Rapport 5.06.03, 103 pages. Ingenbleek, P., and V. M. Immink. 2010. Managing conflicting stakeholder interests: an exploratory case analysis of the formulation of corporate social responsibility standards in The Netherlands. Journal of Public Policy & Marketing 29: 52-65. Jaros, P. et al. 2005. Effect of active immunization against GnRH on androstenone concentration, growth performance and carcass quality in intact male pigs. Livestock Production Science 92: 31-38. Jasper, J. M., and D. Nelkin. 1992. The animal rights crusade - the growth of a moral protest. New York, NY: The Free Press, 214 pages. Kahneman, D., J. L. Knetsch, and R. H. Thaler. 1990. Experimental tests of the endowment effect and the coase theorem. Journal of Political Economy 98: 1325-1348. Kahneman, D., J. L. Knetsch, and R. H. Thaler. 1991. Anomalies - the endowment effect, loss aversion, and status-quo bias. Journal of Economic Perspectives 5: 193-206. Kahneman, D., and A. Tversky. 1979. Prospect theory - analysis of decision under risk. Econometrica 47: 263-291. Kaiser, M. 2005. Assessing ethics and animal welfare in animal biotechnology for farm production. Revue Scientifique Et Technique-Office International Des Epizooties 24: 75-87. Kanis, E., A. F. Groen, and K. H. De Greef. 2003. Societal concerns about pork and pork production and their relationships to the production system. Journal of Agricultural & Environmental Ethics 16: 137-162. Kendall, H. A., L. M. Lobao, and J. S. Sharp. 2006. Public concern with animal well-being: Place, social structural location, and individual experience. Rural Sociology 71: 399-428. Kennedy, O. B., B. J. Stewart-Knox, P. C. Mitchell, and D. I. Thurnham. 2004. Consumer perceptions of poultry meat: a qualitative analysis. Nutrition and Food Science 34: 122-129. Kiley-Worthington, M. 1989. Ecological, ethological and ethically sound environments for animals: towards symbiosis. Journal of Agricultural & Environmental Ethics 2: 223-247. Kilgour, R. 1978. The application of animal behavior and the humane care of farm animals. Journal of Animal Science 46: 1478-1486. Kim, Y., and S. M. Choi. 2005. Antecedents of green purchase behaviour: an examination of collectivism, environmental concern and PCE. Advances in Consumer Research 32: 592-597. Kjaernes, U., E. Roe, and B. Bock. 2007. Societal concerns on farm animal welfare. Powerpointpresentation on the Second Stakeholder Conference ‘Assuring animal welfare: from societal concerns to implementation’, Welfare Quality®, 3-4 May, Berlin, Germany. Kling-Eveillard, F., A. C. Dockes, and C. Souquet. 2007. Attitudes of French pig farmers towards animal welfare. British Food Journal 109: 859-869. 228 References Knierim, U., and C. Wincler. 2009. On-farm welfare assessment in cattle: validity, reliability and feasibility issues and future perspectives with special regard to the Welfare Quality® approach. Animal Welfare 18: 451-458. Koch, V. W. 2009. American veterinarians' animal welfare limitations. Journal of Veterinary Behavior-Clinical Applications and Research 4: 198-202. Korte, S. M., B. Olivier, and J. M. Koolhaas. 2007. A new animal welfare concept based on allostasis. Physiology & Behavior 92: 422-428. Korthals, A. 2001a. Ethical dilemmas in sustainable agriculture. International Journal of Food Science and Technology 36: 813-820. Korthals, M. 2001b. Taking consumers seriously: Two concepts of consumer sovereignty. Journal of Agricultural & Environmental Ethics 14: 201-215. Korthals, M. 2002. Voor het eten: Filosofie en ethiek van voeding. Amsterdam: Uitgeverij Broom, 250 pages. Korthals, M. 2003. Kunnen consumenten iets meer dan eten? In: H. Dagevos and L. Sterrenberg (eds.) Burgers en consumenten: tussen tweedeling en twee-eenheid. p 43-56. Wageningen, The Netherlands: Wageningen Academic Publishers. Korthals, M. 2005. Ethics of food production and consumption. Available on http://www.app.wur.nl/NR/rdonlyres/E96C8F3F-A0A4-48C3-956C355C676B1836/24283/MichielKorthalsEthicsofFoodProductionandConsumptio.doc (last consulted: 18/09/2010). Korzen, S., and J. Lassen. 2010. Meat in context. On the relation between perceptions and contexts. Appetite 54: 274-281. Krueger, R. A. 1988. Focus groups: a practical guide for applied research. Newbury Park, CA: Sage Publications, 197 pages. Krystallis, A., M. D. de Barcellos, J. O. Kuegler, W. Verbeke, and K. G. Grunert. 2009. Attitudes of European citizens towards pig production systems. Livestock Science 126: 46-56. Kubberød, E., Ø. Ueland, Å. Tronstad, and E. Risvik. 2002. Attitudes towards meat and meat-eating among adolescents in Norway: a qualitative study. Appetite 38: 53-62. Lagerkvist, C., F. Carlsson, and D. Viske. 2006. Swedish consumer preferences for animal welfare and biotech: a choice experiment. AgBioforum 9: 51-58. Laskova, A. 2007. Perceived consumer effectiveness and environmental concerns. Proceeding of the 13th Asia Pacific Management Conference (APMC), 18-20 November, Melbourne, Australia. p 206-209. Lassen, J., P. Sandøe, and B. Forkman. 2006. Happy pigs are dirty! Conflicting perspectives on animal welfare. Livestock Science 103: 221-230. Lawrence, A. B., and A. W. Stott. 2009. Profiting from animal welfare: an animal-based perspective. Available on http://www.fao.org/fileadmin/user_upload/animalwelfare/lawrence%2009.pdf (last consulted: 18/09/2010). 229 References Lea, E., and A. Worsley. 2002. The cognitive contexts of beliefs about the healthiness of meat. Public Health Nutrition 5: 37-45. Lewis, K. A., J. Tzilivakis, A. Green, D. Warner, and A. Coles. 2008. Farm assurance schemes: can they improve farming standards? British Food Journal 110: 1088-1105. Liljenstolpe, C. 2008. Evaluating animal welfare with choice experiments: an application to Swedish pig production. Agribusiness 24: 67-84. Lindeman, M., and M. Väänänen. 2000. Measurement of ethical food choice motives. Appetite 34: 55-59. LNV Consumentenplatform. 2005. Dierenwelzijn, willen we dat weten? Beleidsdossier, Den Haag, 10 juni 2005, 15 pages. Available on http://www.minlnv.nl (last consulted: 15/09/2010). Lund, V., G. Coleman, S. Gunnarsson, M. C. Appleby, and K. Karkinen. 2006. Animal welfare science working at the interface between the natural and social sciences. Applied Animal Behaviour Science 97: 37-49. Lundstrom, K., K. R. Matthews, and J. E. Haugen. 2009. Pig meat quality from entire males. Animal 3: 1497-1507. MacArthur Clark, J. 2008. Making sustainable improvements in animal welfare. Available on http://www.daff.gov.au/__data/assets/pdf_file/0003/1046712/60-judy-macarthurclark.pdf (last consulted: 18/09/2010) . Mackinnon, J. D., and M. C. Pearce. 2007. ImprovacTM (Pfizer Animal Health): an immunological product for the control of boar taint in male pigs. Boar taint and its control and the mode of action, safety and efficacy of ImprovacTM. The Pig Journal 59: 29-67. Makatouni, A. 1999. What motivates consumers to buy organic food in the UK? British Food Journal 104: 345-352. Malhotra, N. K., and M. Peterson. 2005. Basic marketing research: a decision-making approach. Upper Saddle River: Prentice Hall, 672 pages. Mann, S. 2005. Ethological farm programs and the "market" for animal welfare. Journal of Agricultural & Environmental Ethics 18: 369-382. Manning, L., R. N. Baines, and S. A. Chadd. 2006. Ethical modelling of the food supply chain. British Food Journal 108: 358-370. Maria, G. A. 2006. Public perception of farm animal welfare in Spain. Livestock Science 103: 250256. Marie, M. 2006. Ethics: the new challenge for animal agriculture. Livestock Science 103: 203-207. Marusic, A. 2009. Can journal editors police animal welfare? Three Es for three Rs in scientific journals. American Journal of Bioethics 9: 66-67. Maslow, A. H. 1943. A theory of human motivation. Psychological Review 50: 370-396. Mason, G., and M. Mendl. 1993. Why is there no simple way of measuring animal welfare? Animal Welfare 2: 301-319. 230 References Mayfield, L. E., R. M. Bennett, R. B. Tranter, and M. J. Wooldridge. 2007. Consumption of welfarefriendly food products in Great Britain, Italy and Sweden, and how it may be influenced by consumer attitudes to, and behaviour towards, animal welfare attributes. International Journal of Sociology of Agriculture and Food 15: 59-73. McCarthy, M., M. de Boer, S. O'Reilly, and L. Cotter. 2003. Factors influencing intention to purchase beef in the Irish market. Meat Science 65: 1071-1083. McDonald, B. 2000. "Once you know something, you can't not know it" - an empirical look at becoming vegan. Society & Animals 8: 1-23. McEachern, M. G., and M. J. A. Schröder. 2002. The role of livestock production ethics in consumer values towards meat. Journal of Agricultural & Environmental Ethics 15: 221-237. McGlone, J. J. 2001. Farm animal welfare in the context of other society issues: Toward sustainable systems. Livestock Production Science 72: 75-81. McInerney, J. 1991. A socioeconomic perspective on animal-welfare. Outlook on Agriculture 20: 5156. McInerney, J. 2004. Animal welfare, economics and policy. Report on a study undertaken for the Farm & Animal Health Economics Division of DEFRA. Available on http://www.defra.gov.uk/evidence/economics/foodfarm/reports/documents/animalwelfar e.pdf (last consulted: 18/09/2010). McMillan, F. D. 2000. Quality of life in animals. Journal of the American Veterinary Medical Association 216: 1904-1910. Mejdell, C. M. 2006. The role of councils on animal ethics in assessing acceptable welfare standards in agriculture. Livestock Science 103: 292-296. Mench, J. 2004. Management, handling, and transport of farm animals. Proceedings of the Global Conference on Animal Welfare: an OIE initiative, 23-25 February, Paris, France, p 149-155. Menghi, A. 2007. Italian pig producers' attitude toward animal welfare. British Food Journal 109: 870-878. Meulenberg, T. 2003. 'Consument en burger', betekenis voor de markt van landbouwproducten en voedingsmiddelen. Tijdschrift voor Sociaal Wetenschappelijk Onderzoek van de Landbouw 18: 43-54. Meuwissen, M., and I. Van der Lans. 2004. Trade-offs between consumer concerns: an application for pork production. Proceedings of the 6th International Conference on Chain and Network Management in Agribusiness and the Food Industry, 27-28 May, Ede, The Netherlands, 9 pages. Miele, M. 1998. Literature review and policy aspects. Italy, 1st report, ‘Consumer concerns about animal welfare and the impact on food choice’ (EU FAIR CT 98-3678), published by University of Pisa. Miele, M. 2001. Italian survey report, ‘Consumer concerns about animal welfare and the impact on food choice’ (EU FAIR CT 98-3678), published by University of Pisa.. 231 References Miele, M., and A. Evans. 2005. European consumers' views about farm animal welfare. Proceedings of the First Stakeholder Conference ‘Science and society improving animal welfare’, Welfare Quality®, 9-25 November, Brussels, Belgium. Milne, C. E., G. E. Dalton, and A. W. Stott. 2007. Integrated control strategies for ectoparasites in Scottish sheep flocks. Livestock Science 106: 243-253. Mirabito, L., J. Capdeville, A. C. Dockès, F. Kling Eveillard, and A. Brule. 2008. Assessment of cattle welfare on farms: an example of cooporation with consumers. Proceedings of the 4th International Workshop on Animal Welfare Assessment at Farm or Group Level (WAFL), 1013 September, Ghent, Belgium, p 61. Mittal, B., and M. S. Lee. 1989. A causal model of consumer involvement. Journal of Economic Psychology 10: 363-389. Mondelaers, K., W. Verbeke, and G. Van Huylenbroeck. 2009. Importance of health and environment as quality traits in the buying decision of organic products. British Food Journal 111: 1120-1139. Morgan-Davies, C., A. Waterhouse, C. E. Milne, and A. W. Stott. 2006. Farmers' opinions on welfare, health and production practices in extensive hill sheep flocks in Great Britain. Livestock Science 104: 268-277. Napolitano, F., C. Pacelli, A. Girolami, and A. Braghieri. 2008. Effect of information about animal welfare on consumer willingness to pay for yogurt. Journal of Dairy Science 91: 910-917. Nelson, K. A. 2004. Consumer decision making and image theory: understanding value-laden decisions. Journal of Consumer Psychology 14: 28-40. Nibert, D. 1994. Animal rights and human social issues. Society and Animals 2: 115-124. Niesten, E., J. Raymaekers, and Y. Segers. 2003. Lekker dier!? Dierlijke productie en consumptie in de 19de en 20ste eeuw. Leuven: CAG, 192 pages. Nilsson, T., K. Foster, and J. L. Lusk. 2006. Marketing opportunities for certified pork chops. Canadian Journal of Agricultural Economics-Revue Canadienne D Agroeconomie 54: 567583. Nocella, G., L. Hubbard, and R. Scarpa. 2010. Farm animal welfare, consumer willingness to pay, and trust: results of a cross-national survey. Applied Economic Perspectives and Policy 32: 275297. Nordenfelt, L. 2006. Animal and human health and welfare: a comparative philosophical analysis. Wallingford, UK: CABI Publishing, 204 pages. Nussbaum, M. C. 2006. Frontiers of justice: disability, nationality, species membership. Cambridge, UK: Harvard University Press, 520 pages. Olesen, I., F. Alfnes, M. B. Røra, and K. Kolstad. 2010. Eliciting consumers' willingness to pay for organic and welfare-labelled salmon in a non-hypothetical choice experiment. Livestock Science 127: 218-226. Oliver, R. L. 1980. A cognitive model of the antecedents and consequences of satisfaction decisions. Journal of Marketing Research 17: 460-469. 232 References Oma, K. A. 2010. Between trust and domination: social contracts between humans and animals. World Archaeology 42: 175-187. Osborne, N. J., D. Payne, and M. L. Newman. 2009. Journal editorial policies, animal welfare, and the 3 Rs. American Journal of Bioethics 9: 55-59. Ouedraogo, A. et al. 2009. Survey on the attitudes of stakeholders in relation to piglet castration and alternatives in different regions of Europe. Report of PIGCAS, WP1: Attitude, PIGCASproject, sine loco. Passantino, A., F. Conte, and M. Russo. 2008. Animal welfare labelling and the approach of the European Union: an overview on the current situation. Journal Für Verbraucherschutz Und Lebensmittelsicherheit – Journal of Consumer Protection and Food Safety 3: 396-399. Patterson, R. L. S. 1968. 5alpha-androst-16-ene-3-1: compound responsible for taint in boar fat. Journal of the Science of Food and Agriculture 19: 31-37. Paul, E. S., and A. L. Podberscek. 2000. Veterinary education and students' attitudes towards animal welfare. Veterinary Record 146: 269-272. Peek, C. W., C. C. Dunham, and B. E. Dietz. 1997. Gender, relational role orientation, and affinity for animal rights. Sex roles 37: 905-920. Petty, R. E., and J. T. Cacioppo. 1986. The elaboration likelihood model of persuasion. Advances in Experimental Social Psychology 19: 123-205. Phan-Huy, S. A., and R. B. Fawaz. 2003. Swiss market for meat from animal-friendly production responses of public and private actors in Switzerland. Journal of Agricultural & Environmental Ethics 16: 119-136. Plous, S. 1993. Psychological mechanisms in the human use of animals. Journal of Social Issues 49: 11-52. Postollec, G., E. Boilletot, R. Maurice, and V. Michel. 2008. The effect of pen size and an enrichment structure (elevated platform) on the performances and the behaviour of fattening rabbits. Animal Welfare 17: 53-59. Pouta, E., J. Heikkila, S. Forsman-Hugg, M. Isoniemi, and J. Makela. 2010. Consumer choice of broiler meat: the effects of country of origin and production methods. Food Quality and Preference 21: 539-546. Prabha, C., L. S. Connaway, L. Szewski, and L. R. Jenkins. 2007. What is enough? Satisficing information needs. Journal of Documentation 63: 74-89. Prunier, A., and M. Bonneau. 2006. Alternatives to piglet castration. Productions Animales 19: 347356. Prunier, A. et al. 2006. A review of the welfare consequences of surgical castration in piglets and the evaluation of non-surgical methods. Animal Welfare 15: 277-289. Quintili, R., and G. Grifoni. 2004. Consumer concerns for animal welfare: from psychosis to awareness. Proceedings of the Global Conference on Animal Welfare: an OIE initiative, 2325 February, Paris, France, p 93-96. Ransom, E. 2007. The rise of agricultural animal welfare standards as understood through a neoinstitutional lens. International Journal of Sociology of Agriculture and Food 15: 26-44. 233 References Regan, T. 1983. The case for animal rights. California, US: University of California Press, 425 pages. Ritzer, G. 2001. Explorations in the sociology of consumption: fast food, credit cards and casinos. London, Sage Publications, 274 pages. Roberts, J. A. 1995. Profiling levels of socially responsible consumer behavior: a cluster analytic approach and its implications for marketing. Journal of Marketing Theory and Practice 3: 97118. Robinson, R., and C. Smith. 2003. Associations between self-reported health conscious consumerism, body-mass index, and attitudes about sustainably produced foods. Agriculture and Human Values 20: 177-187. Rollin, B. E. 1981. Animal rights and human mortality. New York, NY: Prometheus Books, 182 pages. Rushen, J. 2003. Changing concepts of farm animal welfare: bridging the gap between applied and basic research. Applied Animal Behaviour Science 81: 199-214. Rushen, J., and A. M. B. Depassille. 1992. The scientific assessment of the impact of housing on animal-welfare - a critical-review. Canadian Journal of Animal Science 72: 721-743. Safran Foer, J. 2009. Eating animals. New York, NY: Little, Brown and Company, Hachette Book Group, 341 pages. Salaün, Y., and K. Flores. 2001. Information quality: meeting the needs of the consumer. International Journal of Information Management 21: 21-37. Sandøe, P., and H. B. Simonsen. 1992. Assessing animal welfare: where does science end and philosophy begin? Animal Welfare 1: 257-267. Schlozman, K. L., N. Burns, S. Verba, and J. Donahue. 1995. Gender and citizen participation: Is there a different voice? American Journal of Political Science 39: 267-293. Scipioni, R., G. Martelli, and L. A. Volpelli. 2009. Assessment of welfare in pigs. Italian Journal of Animal Science 8: 117-137. Seamer, J. H. 1998. Human stewardship and animal welfare. Applied Animal Behaviour Science 59: 201-205. Serpell, J. 1996. In the company of animals. A study of the human-animal relationships. Cambridge, UK: Cambridge University Press, 320 pages. Serpell, J. 1999. Sheep in wolves' clothing? Attitudes to animals among farmers and scientists. In: F. L. Dolins (ed.) Attitudes to animals: Views in animal welfare. p 26-33. Cambridge, UK: Cambridge University Press. Serpell, J. A. 2004. Factors influencing human attitudes to animals and their welfare. Animal 13. Sevi, A. 2009. Animal-based measures for welfare assessment. Italian Journal of Animal Science 8: 904-911. Shepherd, R., M. Magnusson, and P. O. Sjödén. 2005. Determinants of consumer behavior related to organic foods. Ambio 34: 352-359. Shrum, L. J. 1995. Buyer characteristics of the green consumer and their implications for advertising strategy. Journal of Advertising 24: 71-82. 234 References Singer, P. 1975. Animal liberation: A new ethics for our treatment of animals. New York, NY: New York Review/Random House, 301 pages. Skarstad, G. Å., L. Terragni, and H. Torjusen. 2007. Animal welfare according to Norwegian consumers and producers: definitions and implications. International Journal of Sociology of Agriculture and Food 15: 74-90. Sørensen, J. T., and D. Fraser. 2010. On-farm welfare assessment for regulatory purposes: issues and possible solutions. Livestock Science 131: 1-7. Spaargaren, G. 2003. Duurzaam consumeren of ecologisch burgerschap? In: H. Dagevos and L. Sterrenberg (eds.) Burgers en consumenten: tussen tweedeling en twee-eenheid. p 70-84. Wageningen, The Netherlands: Wageningen Academic Publishers. Spinu, M., S. Benveneste, and A. A. Degen. 2003. Effect of density and season on stress and behaviour in broiler breeder hens. British Poultry Science 44: 170-174. Spoolder, H. A. M., S. A. Edwards, and S. Corning. 1999. Effects of group size and feeder space allowance on welfare in finishing pigs. Animal Science 69: 481-489. Spring, P., T. Kupper, and C. Pauly. 2009. Proschwein: Alternatives to the conventional castration of piglets. Agrarforschung 16: 16-21. Stafleu, F. R., F. J. Grommers, and J. Vorstenbosch. 1996. Animal welfare: evolution and erosion of a moral concept. Animal Welfare 5: 225-234. Storstad, O., and H. Bjørkhaug. 2003. Foundations of production and consumption of organic food in Norway: common attitudes among farmers and consumers? Agriculture and Human Values 20: 151-163. Sundrum, A. 2007. Conflicting areas in the ethical debate on animal health and welfare. In: W. Zollitsch, C. Winckler, S. Waiblinger and A. Haslberger (eds.) Sustainable food production and ethics: Preprints of the 7th Congress of the European Society for Agricultural and Food Ethics (EURSAFE). p 257-262. Wageningen, The Netherlands: Wageningen Academic Publishers. Swinnen, J., J. McCluskey, and N. Francken. 2005. Food safety, the media and the information market. Agricultural Economics 32: 175-188. Tannenbaum, J. 1991. Ethics and animal-welfare - the inextricable connection. Journal of the American Veterinary Medical Association 198: 1360-1376. Taylor, N., and T. D. Signal. 2009. Willingness to pay: Australian consumers and "on the farm" welfare. Journal of Applied Animal Welfare Science 12: 345-359. Te Velde, H. T., N. Aarts, and C. Van Woerkum. 2002. Dealing with ambivalence: farmers' and consumers' perceptions of animal welfare in livestock breeding. Journal of Agricultural & Environmental Ethics 15: 203-219. Thompson, P., C. Harris, D. Holt, and E. A. Pajor. 2007. Livestock welfare product claims: the emerging social context. Journal of Animal Science 85: 2354-2360. Tilly, C. 1999. Durable inequality. Berkley, CA: University of California Press, 316 pages. Turkstra, J. A. et al. 2002. Performance of male pigs immunised against GnRH is related to the time of onset of biological response. Journal of Animal Science 80: 2953-2959. 235 References Tuyttens, F. et al. 2008. Broiler chicken health, welfare and fluctuating asymmetry in organic versus conventional production systems. Livestock Science 113: 123-132. Tuyttens, F., F. Vanhonacker, E. Van Poucke, and W. Verbeke. 2010a. Quantitative verification of the correspondence between the Welfare Quality® operational definition of farm animal welfare and the opinion of the Flemish farmers, citizens and vegeterians. Livestock Science 131: 108-114. Tuyttens, F. A. M. et al. 2010b. Level of information provisioning affects consumers' attitude towards surgical castration of male piglets and alternative strategies for dealing with boar taint. Research in Veterinary Science submitted. Ubilava, D., K. A. Foster, J. L. Lusk, and T. Nilsson. 2008. Effects of income and social awareness on heterogeneous consumers' WTP for informational attributes. Journal of Agricultural and Resource Economics 33: 504-504. Uyeki, E. S., and L. J. Holland. 2000. Diffusion of pro-environment attitudes. American Behavioral Scientist 43: 646-662. van Bruchem, C. 2003. Hoe krijgen we burger en consument op één lijn? In: H. Dagevos and L. Sterrenberg (eds.) Burgers en consumenten: tussen tweedeling en twee-eenheid. p 111122. Wageningen, The Netherlands: Wageningen Academic Publishers. Van Calker, K. J., P. B. M. Berentsen, G. W. J. Giesen, and R. B. M. Huirne. 2005. Identifying and ranking attributes that determine sustainability in Dutch dairy farming. Agriculture and Human Values 22: 53-63. Van de Velde, L., W. Verbeke, M. Popp, and G. Van Huylenbroeck. 2010. The importance of the message frame in the case of information provision related to energy and environmental problems. Energy Policy 38: 5541-5549. Van Huik, M., and B. Bock. 2006. Attitudes of organic pig farmers towards animal welfare. In: H. M. Kaiser and M. Lein (eds.) Ethics and the politics of food: Preprints of the 6th Congress of the European Society for Agricultural and Food Ethics (EURSAFE). p 550-554. Wageningen, The Netherlands: Wageningen Academic Publishers. Van Poucke, E. et al. 2006. Defining the concept of animal welfare: integrating the opinion of citizens and other stakeholders. In: H. M. Kaiser and M. Lien (eds.) Ethics and the politics of food: Preprints of the 6th Congress of the European Society for Agricultural and Food Ethics (EURSAFE). p 555-559. Wageningen, The Netherlands, Wageningen Academic Publishers. Van Tichelen, S. 2009. An NGO view. Proceedings of the Final Stakeholder Conference ‘Delivering animal welfare and quality: transparancy in the food production chain’, Welfare Quality®, 89 October, Uppsala, Sweden, p 83-84. Vanhonacker, F., and W. Verbeke. 2009. Buying higher welfare poultry products? Profiling Flemish consumers who do and do not. Poultry Science 88: 2702-2711. Vanhonacker, F., W. Verbeke, and F. A. M. Tuyttens. 2009. Belgian consumers' attitude towards surgical castration and immunocastration of piglets. Animal Welfare 18: 371-380. Vanhonacker, F., W. Verbeke, E. Van Poucke, and F. A. M. Tuyttens. 2007. Segmentation based on consumers' perceived importance and attitude toward farm animal welfare. International Journal of Sociology of Agriculture and Food 15: 91-107. 236 References Vanhonacker, F., W. Verbeke, E. Van Poucke, and F. A. M. Tuyttens. 2008. Do citizens and farmers interpret the concept of farm animal welfare differently? Livestock Science 116: 126-136. Veissier, I., A. Butterworth, B. Bock, and E. Roe. 2008. European approaches to ensure good animal welfare. Applied Animal Behaviour Science 113: 279-297. Veissier, I., and A. Evans. 2007. Rationale behind the Welfare Quality® assessment of animal welfare. Proceedings of the Second Stakeholder Conference ‘Assuring animal welfare: from societal concerns to implementation’, Welfare Quality®, 3-4 May, Berlin, Germany, p 19-22. Velarde, A. et al. 2008. The effect of immunocastration on the behaviour of pigs. EAAP-meeting ‘Production and Utilisation of Meat from Entire Male Pigs’, 26-27 March, Gerona, Spain. Verbeke, W. 2002. A shift in public opinion. Pig Progress 18: 25-27. Verbeke, W. 2005. Agriculture and the food industry in the information age. European Review of Agricultural Economics 32: 347-368. Verbeke, W. 2009. Stakeholder, citizen and consumer interests in farm animal welfare. Animal Welfare 18: 325-333. Verbeke, W., L. J. Frewer, J. Scholderer, and H. F. De Brabander. 2007a. Why consumers behave as they do with respect to food safety and risk information. Analytica Chimica Acta 586: 2-7. Verbeke, W., F. J. A. Perez-Cueto, M. D. de Barcellos, A. Krystallis, and K. G. Grunert. 2010. European citizen and consumer attitudes and preferences regarding beef and pork. Meat Science 84: 284-292. Verbeke, W., and I. Vackier. 2004. Profile and effects of consumer involvement in fresh meat. Meat Science 67: 159-168. Verbeke, W., and I. Vackier. 2005. Individual determinants of fish consumption: Application of the theory of planned behaviour. Appetite 44: 67-82. Verbeke, W., F. Vanhonacker, I. Sioen, J. Van Camp, and S. De Henauw. 2007b. Perceived importance of sustainability and ethics related to fish: a consumer behavior perspective. Ambio 36: 580-585. Verbeke, W., I. Vermeir, and K. Brunso. 2007c. Consumer evaluation of fish quality as basis for fish market segmentation. Food Quality and Preference 18: 651-661. Verbeke, W., and J. Viaene. 1999a. Beliefs, attitude and behaviour towards fresh meat consumption in belgium: empirical evidence from a consumer survey. Food Quality and Preference 10: 437-445. Verbeke, W., and J. Viaene. 1999b. Consumer attitude to beef quality labeling and associations with beef quality labels. Journal of International Food and Agribusiness Marketing 10: 45-65. Verbeke, W., and J. Viaene. 2000. Ethical challenges for livestock production: meeting consumer concerns about meat safety and animal welfare. Journal of Agricultural & Environmental Ethics 12: 141-151. Verbeke, W., J. Viaene, and O. Guiot. 1999. Health communication and consumer behavior on meat in Belgium: from BSE until dioxin. Journal of Health Communication 4: 345-357. 237 References Verbeke, W., and R. W. Ward. 2001. A fresh meat almost ideal demand system incorporating negative TV press and advertising impact. Agricultural Economics 25: 359-374. Verhille, B. 2008. Bevraging van Vlaamse varkenshouders over hun houding t.o.v. chirurgische castratie zonder verdoving en zijn alternatieven. Thesis to obtain the degree of Master in bio-sciences, option agriculture, Hogeschool Ghent. Verhue, D., and D. Verzeijden. 2003. Burgeroordelen over de veehouding, uitkomsten publieksonderzoek. Onderzoeksrapport, Amsterdam. Vermeir, I., and W. Verbeke. 2006. Sustainable food consumption: exploring the consumer "attitude - behavioral intention" gap. Journal of Agricultural & Environmental Ethics 19: 169-194. VILT (Vlaams infocentrum land- en tuinbouw). 2009. Alternatieve eieren halen marktaandeel van 86 procent. VILT nieuwsbrief (8/04/2009). VILT (Vlaams infocentrum land- en tuinbouw). Belg eet in 2009 iets minder vlees, maar varieert meer. VILT nieuwsbrief (18/03/2010). VILT (Vlaams infocentrum land- en tuinbouw). Veggie-organisatie EVA analyseert de diverse sectoren. VILT nieuwsbrief (22/09/2010). Vold, E. 1970. Fleischproduktionseigenschaften bei Ebern und Kastraten. IV. Organoleptische und gaschromatografische Untersuchungen Wasserdampfflüchtiger Stoffe des Rückenspeckes von Ebern. Meldinger Nordlandbrukshoegskole 49: 1-25. Von Alvensleben, R. 1997. Consumer behaviour. In: D. Padberg, C. Ritson and L. Albisu (eds.) Agrofood marketing. p 209-224. Wallingford, UK: CABI Publishing. von Borell, E. et al. 2009. Animal welfare implications of surgical castration and its alternatives in pigs. Animal 3: 1488-1496. von Borell, E., M. Oliver, B. Fredriksen, S. Edwards, and M. Bonneau. 2008. Standpoints , practices and state of information for pig castration in Europe (PIGCAS) - project targets and first results. Journal Für Verbraucherschutz und Lebensmittelsicherheit-Journal of Consumer Protection and Food Safety 3: 216-220. Walstra, P., and G. Maarse. 1970. Onderzoek geslachtsgeur van mannelijke mestvarkens, Zeist, The Netherlands. Watson, D. 1992. Correcting for acquiescent response bias in the absence of a balanced scale - an application to class-consciousness. Sociological Methods & Research 21: 52-88. Webster, J. 2005a. Animal welfare: limping towards eden. Oxford, UK: Blackwell Publishing, 304 pages. Webster, J. 2005b. The assessment and implementation of animal welfare: theory into practice. Revue Scientifique Et Technique-Office International Des Epizooties 24: 723-734. Wells, D. L., and P. G. Hepper. 1997. Pet ownership and adults' views on the use of animals. Society & Animals 5: 45-63. Wilson, T. 1981. On user studies and information needs. Journal of Documentation 37: 3-15. 238 References Windschnurer, I., C. Schmied, X. Boivin, and S. Waiblinger. 2008. Reliability and inter-test relationship of tests for on-farm assessment of dairy cows' relationship to humans. Applied Animal Behaviour Science 114: 37-53. Winter, M., C. Fry, and S. P. Carruthers. 1998. European agricultural policy and farm animal welfare. Food Policy 23: 305-323. Worsley, A., and E. Lea. 2003. Consumers' personal values and sources of nutrition information. Ecology of Food and Nutrition 42: 129-151. Zaichkowsky, J. L. 1985. Measuring the involvement construct. Journal of Consumer Research 12: 341-352. Zajonc, R. B. 1980. Feeling and thinking: preferences need no inferences. American Psychology 35: 151-175. Zamaratskaia, G. et al. 2008. Long-term effect of vaccination against gonadotropin-releasing hormone, using ImprovacTM, on hormonal profile and behaviour of male pigs. Animal Reproduction Science 108: 37-48. Zamaratskaia, G., and E. J. Squires. 2009. Biochemical, nutritional and genetic effects on boar taint in entire male pigs. Animal 3: 1508-1521. Zander, K., and U. Hamm. 2010. Consumer preferences for additional ethical attributes of organic food. Food Quality and Preference 21: 495-503. Zimmerman, P. H. et al. 2006. The effect of stocking density, flock size and modified management on laying hen behaviour and welfare in a non-cage system. Applied Animal Behaviour Science 101: 111-124. 239 Curriculum vitae CURRICULUM VITAE Personal information First names: Filiep Pieter Remi Last name: Vanhonacker Adress: Zwevegemsestraat 114 8530 Harelbeke (S) Date of birth: April 19, 1982 Place of birth: Kortrijk Nationality: Belgian Sex: Male Marital status: Married to Joke Pannecoucque Telephone: 003256295796 GSM: 0032485425398 Email: [email protected] [email protected] Education 2000: Secondary education, Sint-Jozefscollege, Kortrijk 2003: Bachelor degree in Bioscience engineering, Ghent University (distinction) 2005: Master degree in Bioscience engineering, Ghent University (great distinction) Dissertation: Role and impact of communication on fish consumption behaviour, 106 pages, promotor: Prof. Wim Verbeke, Department of Agricultural Economics Additional diplomas 2006: Advanced Methods of Market Research I, Ghent University 2007: Advanced Methods of Market Research II, Ghent University 240 Curriculum Vitae Career September 2005 – October 2008 Doctoral researcher, Dept. Agricultural Economics, Ghent University October 2008 – present Full time assistant, Dept. Agricultural Economics, Ghent University Involved in exercises and practicals of the courses: Food and Nutrition Policies Food Marketing and Consumer Behaviour Applied Rural Economic Research Methods Kwaliteitsbeheer en Risico-analyse Consumentengedrag en Marketing van Bio-Industriële Producten Publications Articles in peer-reviewed international journals included in the Science Citation Index Verbeke, W., F. Vanhonacker, I. Sioen, J. Van Camp, and S. De Henauw S. 2007. Perceived importance of sustainability and ethics related to fish: a consumer behaviour perspective. Ambio, 36(7): 580-585. Verbeke, W., F. Vanhonacker, L. Frewer, I. Sioen, S. De Henauw, and J. Van Camp. 2008. Communicating risks and benefits from fish consumption: impact on Belgian consumers’ perception and intention to eat fish. Risk Analysis 28(4): 951-967. Vanhonacker, F., W. Verbeke, E. Van Poucke, and F. A. M. Tuyttens. 2008. Do citizens and farmers interpret the concept of farm animal welfare differently? Livestock Science 116: 126-136. Sioen, I., J. Van Camp, F. Verdonck, W. Verbeke, F. Vanhonacker, J. Willems, and S. De Henauw. 2008. Probabilistic intake assessment of multiple compounds as a tool to quantify the nutritional-toxicological conflict related to seafood consumption. Chemosphere 71: 10561066. Vanhonacker, F., W. Verbeke, E. Van Poucke, S. Buijs, and F. A.M. Tuyttens. 2009. Societal concern related to stocking density, pen size and group size in farm animal production. Livestock Science 123: 16-22. Guerrero, L., M.D. Guardia, J. Xicola, W. Verbeke, F. Vanhonacker, S. Zakowska-Biemans, M. Sajdakowska, C. Sulmont-Rossé, S. Issanchou, M. Contel, M. L. Scalvedi, B. S. Granli, and M. Hersleth. 2009. Consumer-driven definition of traditional food products and innovation in traditional foods. A qualitative cross-cultural study. Appetite 52: 345-354. 241 Curriculum vitae Guerrero, L., A. Claret, W. Verbeke, G. Enderli, S. Zakowska-Biemans, F. Vanhonacker, S. Issanchou, M. Sajdakowska, B. S. Granli, L. Scalvedi, M. Contel, and M. Hersleth. 2009. Perception of traditional food products in six European regions using free word association. Food Quality and Preference 21(2): 225-233. Pieniak, Z., W. Verbeke, F. Vanhonacker, L. Guerrero, and M. Hersleth. 2009. Association between traditional food consumption and motives for food choice in six European countries. Appetite 53: 101-108. Vanhonacker, F., W. Verbeke, and F. A. M. Tuyttens. 2009. Belgian consumers’ attitude towards surgical castration and immunocastration of piglets. Animal Welfare 18: 371-380. Vanhonacker, F., and W. Verbeke. 2009. Buying higher welfare poultry products? Profiling Flemish consumers who do and do not. Poultry Science 88: 2702-2711. Tuyttens, F.A.M., F. Vanhonacker, E. Van Poucke, and W. Verbeke. 2010. Quantitative verification of the correspondence between the Welfare Quality® operational definition of farm animal welfare and the opinion of Flemish farmers, citizens and vegetarians. Livestock Science 131(1): 108-114. Kühne, B., F. Vanhonacker, X. Gellynck, and W. Verbeke. 2010. Innovation in traditional food products in Europe: Do sector innovation activities match consumers’ acceptance? Food Quality and Preference 21(6): 629-638. Altintzoglou, T., W. Verbeke, F. Vanhonacker, and J. Luten. 2010. The image of fish from aquaculture among Europeans: impact of exposure to balanced information. Journal of Aquatic Food Product Technology 19(2): 103-119. Vanhonacker, F., T. Altintzoglou, J. Luten, and W. Verbeke. 2010. Does fish origin matter to European consumers? Insights from a consumer survey in Belgium, Norway and Spain. British Food Journal 112(8): 871-886. Vanhonacker, F., E. Van Poucke, F. A. M. Tuyttens, and W. Verbeke. 2010. Citizens’ views on farm animal welfare and related information provision: exploratory insights from Flanders, Belgium. Journal of Agricultural & Environmental Ethics 23(6): 551-569. Vanhonacker, F., W. Verbeke, L. Guerrero, A. Claret, M. Contel, L. Scalvedi, S. Zakowska-Biemans, K. Gutkowska, C. Sulmont-Rossé, J. Raude, B. S. Granli, and M. Hersleth. 2010. How European consumers define the concept of traditional food: evidence from a survey in six countries. Agribusiness 26(4): 453-476. Vanhonacker, F., V. Lengard Almli, M. Hersleth, and W. Verbeke. 2010. Profiling European traditional food consumers. British Food Journal 112(8-9): 871-886. Lengard Almli, V., W. Verbeke, F. Vanhonacker, T. Naes, and M. Hersleth. 2011. General image and attribute perception of traditional foods in six European countries. Food Quality and Preference 22: 129-138. Molnár, A., X. Gellynck, F. Vanhonacker, T. Gagalyuk, and W. Verbeke. 2010. Are chain goals matching with consumer perceptions? The case of the traditional food sector in the EU. Agribusiness doi: 10.1002/agr.20260. 242 Curriculum Vitae Vanhonacker, F., W. Verbeke, E. Van Poucke, Z. Pieniak, and F. A. M. Tuyttens. 2010. The concept of farm animal welfare: citizen perceptions and stakeholder opinion in Flanders, Belgium. Journal of Agricultural & Environmental Ethics doi 10.1007/s10806-010-9299-6. Vanhonacker, F., and W. Verbeke. 2011. Consumer reactions towards the possible use of a vaccine method to control boar taint versus physical piglet castration with anaesthesia: A quantitative pan-European study. Animal accepted. Altintzoglou, T., F. Vanhonacker, W. Verbeke, and J. Luten. 2011. Association of health involvement and attitudes towards eating fish on farmed and wild fish consumption in Belgium, Norway and Spain. Aquaculture International accepted. Tuyttens, F. A. M., F. Vanhonacker, B. Verhille, D. De Brabander, and W. Verbeke. 2011. Pig producer attitude towards surgical castration of piglets without anaesthesia versus alternative strategies. Research in Veterinary Science resubmitted. Tuyttens, F. A. M., F. Vanhonacker, K. Langendries, M. Aluwé, S. Millet, K. Bekaert, and W. Verbeke. 2011. Level of information provisioning affects attitude towards surgical castration of male piglets and alternative strategies for dealing with boar taint. Research in Veterinary Science resubmitted. Articles in peer-reviewed international journals not included in the Science Citation Index Sioen, I., J. Van Camp, F. Verdonck, N. Van Thuyne, P. A. Vanrolleghem, F. Vanhonacker, W. Verbeke, and S. De Henauw. 2006. Risk-benefit analysis regarding seafood consumption: a tool for combined intake assessment. Organohalogen Compounds 68: 379-382. Vanhonacker, F., W. Verbeke, E. Van Poucke, and F. A. M. Tuyttens. 2007. Segmentation based on consumers’ perceived importance and attitude toward farm animal welfare. International Journal of Sociology of Food and Agriculture 15(3): 91-107. Molnár, A., X. Gellynck, F. Vanhonacker, and W. Verbeke. 2010. Towards the development of innovative strategies for traditional food chains in the EU. International Journal of Food System Dynamics 1: 1-12. Vanhonacker, F., Z. Pieniak, and W. Verbeke. 2010. Fish market segmentation base don consumers’ motives, barriers and risk perception in Belgium. Journal of Food Products Marketing 16: 166183. Articles in national journals (A3) Sioen, I., J. Van Camp, F. Verdonck, F. Vanhonacker, W. Verbeke and S. De Henauw. 2006. Nutritional-toxicological conflict of seafood consumption: a tool for combined intake assessment. Communications in Agricultural and Applied Biological Sciences 71(1): 263-266. Vanhonacker, F., W. Verbeke, L. Guerrero, A. Claret, M. Contel, M. L. Scalvedi, S. Zakowska-Biemans, K. Gutkowska, C. Sulmont-Rossé, J. Raude, B. S. Granli, and M. Hersleth. 2009. I prodotti alimentary tradizonali dal punto di vista del consumatore: un’indagine in sei paesi europei. Agriregionieuropa 5: 16. 243 Curriculum vitae Conference proceedings covered in the Science Citation Index (P1) Vanhonacker, F., W. Verbeke, and I. Sioen. 2006. Consumer perception about ethical and sustainability issues of fish. In: H. M. Kaiser and M. Lien (eds.) Ethics and the politics of food: Preprints of the 6th Congress of the European Society for Agricultural and Food Ethics (EURSAFE). p 464-469. Wageningen, The Netherlands, Wageningen Academic Publishers. Van Poucke, E., F. Vanhonacker, G. Nijs, J. Braeckman, W. Verbeke, and F. A. M. Tuyttens. 2006. Defining the concept of animal welfare: integrating the opinion of citizens and other stakeholders. In: H. M. Kaiser and M. Lien (eds.) Ethics and the politics of food: Preprints of the 6th Congress of the European Society for Agricultural and Food Ethics (EURSAFE). p 555559. Wageningen, The Netherlands, Wageningen Academic Publishers. Vanhonacker, F., E. Van Poucke, G. Nijs, J. Braeckman, F. A. M. Tuyttens, and W. Verbeke. 2006. Defining animal welfare from a citizen and consumer perspective: exploratory findings from Belgium. In: H. M. Kaiser and M. Lien (eds.) Ethics and the politics of food: Preprints of the 6th Congress of the European Society for Agricultural and Food Ethics (EURSAFE). p 580-582. Wageningen, The Netherlands, Wageningen Academic Publishers. Book chapters (B2) Vanhonacker, F., W. Verbeke, V. Lengard Almli, L. Guerrero, and M. Hersleth. 2008. Consumerbased definition and general image of traditional foods in Europe. In: Perspectives of Traditional Food Supply Chains on the European Market. Banterle, A. and Gellynck, X. (Eds.), pp 13-29. Roma, Italy, Aracne editrice. Attended international conferences and seminars with oral or poster presentation Vanhonacker, F., and W. Verbeke. 2006. Relevance of ethics and sustainability as consumer issues in fish consumption decisions. Oral presentation on AQUA 2006, Florence, Italy, May 9-13, 2006. Vanhonacker, F., W. Verbeke, and I. Sioen. 2006. Content and source impact of risk/benefit information on consumer attitude and behaviour towards fish. Poster presentation on AQUA 2006, Florence, Italy, May 9-13, 2006. Vanhonacker, F., W. Verbeke, and I. Sioen. 2006. Consumer perception about ethical and sustainability issues of fish. Oral presentation on the 6th Congress of the European Society for Agricultural and Food Ethics (EURSAFE), Oslo, Norway, June 22-24, 2006. Vanhonacker, F., E. Van Poucke, G. Nijs, J. Braeckman, F. A. M. Tuyttens, and W. Verbeke. 2006. Defining animal welfare from a citizen and consumer perspective: exploratory findings from Belgium. Poster presentation on the 6th Congress of the European Society for Agricultural and Food Ethics (EURSAFE), Oslo, Norway, June 22-24, 2006. Van Poucke, E., F. Vanhonacker, G. Nijs, J. Braeckman, W. Verbeke, and F. A. M. Tuyttens. 2006. Defining the concept of farm animal welfare. Oral presentation on 6th Congress of the European Society for Agricultural and Food Ethics (EURSAFE), Oslo, Norway, June 22-24, 2006. (presented by Els Van Poucke) Vanhonacker, F., W. Verbeke, E. Van Poucke, and F. A. M. Tuyttens. 2007. Consumers versus producers: a different view on farm animal welfare? Oral presentation on the 7th Congress of 244 Curriculum Vitae the European Society for Agricultural and Food Ethics (EURSAFE), Vienna, Austria, September 13-15 September 2007. Vanhonacker, F., W. Verbeke, E. Van Poucke, and F. A. M. Tuyttens. 2007. Market segmentation based on perceived importance and evaluation of farm animal welfare. Oral presentation on the 7th Congress of the European Society for Agricultural and Food Ethics (EURSAFE), Vienna, Austria, September 13-15 September 2007. Vanhonacker, F., and W. Verbeke. 2007. Content and source impact of risk-benefit information on consumer perception and behaviour towards fish consumption. Oral presentation on the World Seafood Congress, Dublin, Ireland, September 25-27, 2007. Vanhonacker, F., Z. Pieniak, and W. Verbeke. 2007. Fish market segmentation based on motives, barriers and risk perception in Flanders. Oral presentation on 37th WEFTA Annual Meeting, Lisbon, Portugal, October 24-27, 2007. Vanhonacker, F., and W. Verbeke. 2007. Are aquaculture products adapted to consumer demands? Oral presentation on PROFET workshop, Athens, Greece, November 22-23, 2007. Tuyttens, F.A.M., B. Verhille, M. Van Oeckel, S. Isebaert, D. De Brabander, and F. Vanhonacker. 2008. Attitudes of Flemish pig producers concerning the surgical castration of piglets and its alternatives. Poster presentation on EAAP Working Group Meeting, Girona, Spain, March 2628, 2008. (presented by Frank Tuyttens) Altintzoglou, T., F. Vanhonacker, J. Luten, and W. Verbeke. 2008. Testing balanced messages among European consumers about health, safety and sustainability from aquaculture. Presentation on CONSENSUS Final Stakeholder Meeting, Ostend, Belgium, April 23-25, 2008. (co-presented with Themis Altintzoglou) Vanhonacker, F., Verbeke, W., Lengard, V., Guerrero, L., Hersleth, M., 2008. Consumer-based definition and general image of traditional foods in Europe. Oral presentation on the 12th EAAE Congress, Ghent, Belgium, August 26-29, 2008. Vanhonacker, F., V. Lengard Almli, L. Guerrero, M. L. Scalvedi, J. Raude, S. Zakowska-Biemans, M. Hersleth, and W. Verbeke. 2008. Europeans and traditional foods: definition and image from the consumers’ perspective. Poster presentation on the 12th EAAE Congress, Ghent, Belgium, August 26-29, 2008. Vanhonacker, F., Z. Pieniak, and W. Verbeke. 2008. Fish market segmentation based on consumers’ motives, barriers and risk perception in Belgium. Poster presentation on the 5th SEAFOODplus Conference, Copenhagen, Denmark, June 8-11, 2008. Vanhonacker, F., W. Verbeke, and F. A. M. Tuyttens. 2008. Consumer attitude towards the surgical castration of piglets and immunocastration. Oral presentation on the 4th International Workshop on the Assessment of Animal Welfare at Farm and Group Level (WAFL), Ghent, Belgium, September 10-13, 2008. Vanhonacker, F., E. Van Poucke, F. A. M. Tuyttens, and W. Verbeke. 2008. What is the relevancy of space allowance in communicating farm animal welfare to the consumer? Poster presentation on the 4th International Workshop on the Assessment of Animal Welfare at Farm and Group Level (WAFL), Ghent, Belgium, September 10-13, 2008. 245 Curriculum vitae Vanhonacker, F., W. Verbeke, E. Van Poucke, and F. A. M. Tuyttens. 2008. Farm animal welfare through the eyes of citizens and producers: similarities and dissimilarities. Poster presentation on the 4th International Workshop on the Assessment of Animal Welfare at Farm and Group Level (WAFL), Ghent, Belgium, September 10-13, 2008. Vanhonacker, F., F. A. M. Tuyttens, J. Scholderer, K. Grunert, and W. Verbeke. 2010. European consumers’ attitude towards surgical castration and immunocastration of piglets. Oral presentation on the EAAP Working Group Meeting, Bristol, UK, March 18-19, 2010. Vanhonacker, F., W. Verbeke, T. Altintzoglou, and J. Luten. 2010. Do European consumers care about fish origin? Oral presentation on the 2nd International Feed for health conference, Tromsø, Norway, June 14-15, 2010. Altintzoglou, T., W. Verbeke, F. Vanhonacker, and J. Luten. 2010. How Europeans think of fish from aquaculture after exposure to balanced information. Oral presentation on the 2nd International Feed for health conference, Tromsø, Norway, June 14-15, 2010. (presented by Themis Altintzoglou) Vanhonacker, F., W. Verbeke, and F. A. M. Tuyttens. 2010. Consumer attitude and acceptance of immunocastration as an alternative for surgical castration. Oral presentation on the 14th International Conference on Production Diseases in Farm Animals, Ghent, Belgium, June 2024, 2010. Attended national seminars, conferences or events Vanhonacker, F., W. Verbeke, and F. A. M. Tuyttens. 2009. Attitude of Flemish citizens towards surgical castration and immunocastration of piglets. Poster presentation on BAMST studiedag, Melle, Belgium, October 8, 2009. Vanhonacker, F., and W. Verbeke. 2009. Society’s view on farm animal welfare: Defining the concept of animal welfare and case studies. Poster presentation on the 15th Symposium on Applied Biological Sciences, Leuven, Belgium, November 6, 2009. Verbeke, W., and F. Vanhonacker. 2009. Consumenten versus het vermijden van berengeur: Kennis en acceptatie? Presentation for Pfizer in Sint-Laureins, Belgium, December 15, 2009. (presented by Wim Verbeke) Vanhonacker, F. 2010. Castreren of niet: de mening van consumenten. Presentation at the press event ‘Varkensbedrijf Wolkenhoeve gaat voor niet-gecastreerde varkens’, Geel, Belgium, May 20, 2010. Attended workshops IOPS course on Structural Equation Modeling, Leiden, The Netherlands, February 12-13, 2007. Seminar ‘Applying PLS Path Modeling: Introduction and Extensions’, Hamburg, Germany, June 2021, 2008. 246 Curriculum Vitae Supervision of master students Olivier Buyse (2005-2006). Impact van informatie op het consumentengedrag t.a.v. groenten en fruit. Promotor: Prof. W. Verbeke. Thesis to obtain the degree of Bio-Engineer, master Agriculture. Ken De Backer (2006-2007). Traditionele voeding en innovatie gedefinieerd door consumenten. Promotor: Prof. W. Verbeke. Thesis to obtain the degree of Bio-Engineer, master Agriculture. Davy Furniere (2007-2008). Consumentenperceptie ten aanzien van immunocastratie als alternatief voor chirurgische castratie bij biggen. Promotor: Prof. W. Verbeke. Thesis to obtain the degree of Bio-Engineer, master Agriculture. Ilse Delcour (2008-2009). Het imago van aquacultuur en consumenteninteresse in een duurzaamheidslabel. Promotor: Prof. W. Verbeke. Thesis to obtain the degree of BioEngineer, master Agriculture. Laurens Delameilleure (2009-2010). De effectiviteit van communicatie van kwaliteitslabels. Promotor: Prof. W. Verbeke. Thesis to obtain the degree of Bio-Engineer, master Agriculture. Allen Katale (2009-2010). Cross country comparison of consumers’ perception towards farmed fish, sea bass and sea bream. Promotor: Prof. W. Verbeke. Thesis to obtain the degree of Master of Human Nutrition and rural development. Shanshan Chen (2009-2010). Consumers’ perception and attitudes towards farm animal welfare. Promotor: Prof. W. Verbeke. Thesis to obtain the degree of Master of Human Nutrition. 247
© Copyright 2024 Paperzz