Use of an indigenous board game, ‘bao’, for assessing farmers’ preferences
among alternative agricultural technologies
Steven Franzel, International Centre for Research in Agroforestry. [email protected]
In: G.H.Peters and Prabhu Pingali (eds), Tomorrow's Agriculture: Incentives, Institutions, Infrastructure and
Innovations. Proceedings of the 24th International Conference of Agricultural Economists. 13-18 Aug.
2000,Berlin .Ashgate, Aldershot, 2001. pp. 416-424
Abstract: Investigations concerning farmers’ evaluations of technology are a neglected subject in the
literature on technology adoption in the developing world. Tools for obtaining such evaluations in
developed countries usually involve questionnaires asking respondents to score alternatives; these
methods have been found to be inappropriate to farmers in third world countries. Practitioners of
participatory research use visual scoring tools such as matrix ranking, but no examples were found in
the literature of such tools being used to collect quantitative data useful for hypothesis testing or
statistical analysis. This paper presents three case studies of the use of ‘bao’, an indigenous board
game found throughout Africa, Indonesia, and the Caribbean, for obtaining farmers’ evaluations of
different tree species in Burundi and Kenya. Like questionnaires in consumer evaluations, the bao
game is useful for generating quantitative data for testing hypotheses and statistical analysis. At the
same time, the game is a participatory tool that the farmer finds engaging. Moreover, because farmers
control the scoring process in the bao game, they take the exercise more seriously than in responding
to questionnaires. Finally, because the bao game is a visual tool, respondents can check their data and
members of a group can discuss differences in scores among themselves. The bao game can thus be
used for collecting quantitative data on farmers’ evaluations in an accurate, entertaining, yet
statistically rigorous manner.
5. Keywords: technology adoption, farmer assessment, qualitative data, agroforestry, participatory
research
1. Introduction
Researchers in the tropics are increasingly recognizing the farmer as a partner, not just a client or a
customer, in developing new agricultural technologies. The farming systems approach in the 1980s
helped researchers to focus on smallholder farmers’ needs and circumstances (Byerlee and Collinson,
1980; Caldwell, 1987). In the 1990s, the participatory research paradigm has helped scientists to
understand how farmers experiment on their own and to seek partnerships with them in developing
technology (Chambers, et al., 1989).
But a major weakness of participatory research is that few research tools exist that both scientists
and farmers can together use for diagnosing problems and developing new technologies (Okali, et al.,
1994). In fact, economists often view participatory research approaches, such as informal surveys, as
inimical to quantitative data analysis. On the other hand, participatory researchers often view accepted
scientific research tools, such as trial replication or questionnaire surveys, as incompatible with
farmers’ investigative processes. Conflicts are apparent in the way the two sides examine farmers’
evaluations of technology. Participatory researchers view questionnaires that ask farmers to rate
alternative technologies across selected criteria as ‘top-down’, western cultural constructs that are not
translatable to rural, third-world situations. In contrast, scientists view participatory approaches as
subjective, lacking in detail, and being qualitative, cannot be subjected to tests of statistical
significance. In fact, economists tend to shy away from investigating farmers’ evaluations of
technology altogether, preferring to calculate financial and economic returns or assess factors that
influence adoption. The neglect of farmer evaluations is indeed misplaced, as knowledge of values,
non-monetary as well as monetary, is a key element of agricultural economics research (Johnson,
1986; Agricultural Economics, 2000)
1
The objective of this paper is to present a method for obtaining data on farmers’ preferences that is
of use and is user-friendly to both farmers and researchers. The method involves the use of ‘bao’, a
traditional board game found throughout Africa, Indonesia, and the Caribbean. The method permits
the collection of quantitative data in a manner that allows the farmer to actively participate and
benefit. First the method is described. Second, three case studies, two from Kenya and one from
Burundi, are presented in which farmers use bao to evaluate technologies. Finally, bao’s advantages
and disadvantages for evaluating technology are compared with conventional scoring exercises using
questionnaires and with a common participatory research tool, matrix ranking.
2. Tools for obtaining farmers’ evaluations
In developed countries, tools for obtaining quantitative data on consumer evaluations are common
and often involve asking respondents to score from poor to excellent, or 1 to 10, across selected
criteria. A few examples exist in the literature of researchers in the tropics using such methods to ask
farmers to rate or rank technologies. (Franzel, 1983; Negassa, 1991; Polson and Spencer, 1991). Even
though they generate quantitative data these methods are generally unsatisfactory for three reasons.
First, asking farmers to rate alternatives on a 1 to x basis (1 being a poor grade and x being an
excellent grade) was generally problematic, because such verbal scoring exercises were not easily
understood by farmers. Second, where terms such as excellent, good, fair, and poor were substituted
for numerical scores, farmers were better able to understand the scoring process. However, there were
often considerable problems translating such terms into other languages and the data obtained could
only be considered as ordinal, rather than interval data (Clark and Schkade, 1974). Third, because the
farmers’ involvement was passive, that is, they merely answered questions, they quickly become
bored with the process.
Participatory researchers, on the other hand, use visual, rather than verbal or written, tools for
obtaining farmers evaluation. These methods are more suited to farmers’ conditions. For example in
matrix ranking, farmers draw a matrix on the ground and place alternative technologies (e.g., crop
varieties) along the left hand side of each row, and symbols to denote criteria (e.g., taste or drought
resistance) across the top of each column. Farmers then rank the technologies on each criterion, using
1 stone to designate first place, 2 for second place, etc. (Ashby, 1990). Farmers control the ranking
process and, along with researchers, observe the results of their ratings. In a review of 15 articles on
preference ranking, Maxwell and Bart (1995) found 3 examples in which researchers used the tool to
score rather than rank alternatives. However, none of the articles tabulated data from a sample of
farmers and used statistical methods to describe the results. In fact, participatory research emphasizes
working with groups of farmers and achieving consensus rather than collecting data from individual
farmers.
3. The Bao game
Bao players move seeds among carved-out pockets of the board, which are laid out in a matrix. The
numbers of rows and columns vary, depending on the area. To use the game in evaluating different
technological alternatives, such as tree species or crop varieties, researchers and farmers first need to
find out the criteria farmers use in assessing alternatives. This is best done by touring the farm,
viewing the different alternatives in question (e.g., tree species), and discussing their performance,
uses, advantages and disadvantages. During the discussions, researchers note the different criteria that
farmers use in evaluating and comparing the species. For example farmers may compare trees on
different end uses (straightness for timber or heat production for fuelwood) or different growth
characteristics (speed of growth, compatibility with crops, or resistance to pests). Next, researchers
and farmers find a comfortable place, and, in the case of trees, put a twig of each important tree
species next to each row of the game. Then, for each criterion the farmer mentioned during the tour,
he/she rates each species from 1 seed (performing poorly) to 5 seeds (performing well). Farmers are
asked to add further criteria if they wish and researchers may also suggest criteria. The exercise often
ends by asking farmers to give overall scores for the species, taking into account all of the criteria.
Scores are preferable to ranks because scores give interval data, whereas ranks give only ordinal data.
Since 1992, ICRAF staff and partners have used the bao game to assess farmers’ preferences among
tree species in 13 exercises in 6 countries. Franzel (2000) provides detailed guidelines on how the
game is used to obtain evaluations.
2
4. Case studies of the bao game
In the three case studies presented here, farmers used the bao game to score alternative tree species
and benefits obtained from using a new practice. In the first example, a sample of 25 farmers in
central Burundi who expressed interest in trees at farmer meetings used the bao game to evaluate the
different tree species on their farms (Franzel et al. 1995). The objective of the exercise was to involve
farmers in tree species selection; by finding out the criteria they used in evaluating trees and how trees
they knew performed on the criteria, researchers could identify trees for testing that could meet
farmers’ needs and circumstances. In the second case study, 37 members of farmers’ groups in
western Kenya who were testing newly introduced tree species in an on-farm experiment used the bao
game to score the species across criteria important to them (Franzel et al. 1999). The objective was to
find out which of the trees in the trial they wished to plant and their reasons. In the third case study,
67 farmers planting improved tree fallows (the planting of fast-growing trees and shrubs on fallow
land to improve soil fertility) in western Kenya were asked to score the different types of benefits they
obtained from the practice (Pisanelli et al. 2000). The percentage of female farmers was 32% in the
Burundi sample, 51% in the sample of farmers in western Kenya testing new trees, and 74% for the
improved fallow users.
In the Burundi case study, farmers rated 8 wood producing trees across 7 criteria focusing on
management, growth, and uses for timber and firewood (table 1). Overall, there were intriguing
discrepancies between farmers' ratings and the prevalence of different tree species on farms.
Eucalyptus spp. and Grevillea robusta were the most common upper story species found on farms and
they were highly rated by farmers; eucalyptus for fast growth and firewood and grevillea for fast
growth and compatibility with crops. But two other species, Maesopsis eminii and Cedrela serrata,
were also highly rated but were not commonly grown. They were relatively new in the area and lack
of information and planting material were the biggest constraints to their adoption. On-farm trials
testing these species will help confirm their usefulness to farmers and should promote their diffusion.
The game was also useful in revealing other criteria that farmers use for rating trees. For example,
males and females appeared to have similar interests in the various species, with one important
exception. Women rated Markhamia lutea much higher than men, because they use its leaves for
preparing a medicine to treat diarrhoea in children.
In the case of new tree species in western Kenya, farmers rated 5 trees across 6 criteria including
growth characteristics and use for fodder and firewood (table 2). Grevillea was the highest rated
species on growth and firewood; Leucaena leucocephala was best for fodder. Standard deviations
were highest for farmers’ ratings on compatibility with crops, reflecting farmers’ uncertainty about
how trees performed on this criterion. Farmers were also asked to use the game to indicate which trees
in the experiment they wished to plant on their farms. The influence of selected farm and household
characteristics on farmers’ ratings of their interest in planting different species was assessed using a
linear logistic model for ordered category response data (Collet 1991). The variables considered
included wealth level, farm size, off-farm income, ethnic group, age, gender, district, and livestock
ownership. Only district emerged as a significant variable, reflecting differences in biophysical
circumstances, such as soil type, which affected tree growth. The small number of farmers that could
be monitored in this experiment, 37, limited the degree to which factors affecting adoption potential
could be rigorously examined.
As in the Burundi case, researchers learned about new criteria that farmers used in evaluating
species. Casuarina junghuhniana was introduced into the trial as a wood species but grew poorly.
Nevertheless, farmers rated it second on preference for planting because they appreciated it as an
ornamental.
In the third case study, farmers’ scores on the importance of benefits that they obtained from
improved fallows in western Kenya indicated that improved soils and crop yields were the most
important. Improved soils and crop yields received mean ratings of 4.5 (SD=0.78) and 4.4 (SD=0.82)
out of 5, respectively (Table 3). Fuelwood received a mean rating of 3.9 (SD= 1.19) and reduced
weeds (mainly Striga hermonthica), 3.6 (SD=1.26). These benefits were each mentioned by over 90%
of the farmers; other benefits, mentioned by fewer than half, included seed production and pest
resistance. Females rated improved fallows significantly higher than males on improving soils
(p=0.04) and on reducing weeds (p=0.06). Womens’ higher scores reflect Ohlsson et al’s (1998)
3
finding that women spend much more time in cropping activities than men and are thus more able to
ascertain and appreciate the effects of improved fallows on soils and weeds. The higher standard
deviations of men’s estimates suggest that they have less knowledge about the fallows and are more
uncertain about their performance.
5. Discussion
The bao game combines the strengths of conventional tools for scoring, such as questionnaires, with
those of participatory research techniques, such as matrix ranking. Like questionnaires in consumer
evaluations, the bao game is useful for generating quantitative data useful for testing hypotheses and
statistical analysis. At the same time, the bao game is a participatory tool that the farmer finds
engaging. Moreover, because farmers control the scoring process in the bao game, they take the
exercise more seriously than in responding to questionnaires. Finally, because the bao game is a visual
tool, respondents can check their data and members of a group can discuss differences in scores
among themselves. Obtaining farmers’ evaluations of agricultural technology is a neglected subject
among agricultural economists in developing countries. The bao game can be used for conducting
such evaluations in an accurate, entertaining, yet statistically rigorous manner.
References
Agricultural Economics, 2000. General Information: Aims and scope. Agricultural
Economics 7 (1992): inset page
Ashby, J.A., 1990. Evaluating technology with farmers: A handbook. Publication no. 187.
International Centre for Tropical Agriculture (CIAT, Cali, Columbia.
Byerlee, D. and Collinson, M., 1980. Planning technologies appropriate to farmers. Mexico,
DF: CIMMYT.
Caldwell, J. S., 1987. An overview of farming systems research and development. In:
Gittenger, J.P., Leslie, J., and Hoisington, C. (Eds) Food policy: Integrating supply,
distribution, and consumption. pp. 165-178, Johns Hopkins University Press, Baltimore, MD.
Chambers, R, Pacey, A, and Thrupp, L.A., 1989. Farmer first: farmer innovation and
agricultural research. Intermediate Technology Publications, London.
Clark, C.T. and Schkade, L.L., 1974. Statistical analysis for administrative decisions.
Southwestern Publ. Co., Cincinnatti, OH
Collett, D. 1991. Modelling binary data. Chapman & Hall, London.
Franzel, S., 1983. Planning an Adaptive Production Research Program for Small Farmers: A
Case Study of Farming Systems Research in Kirinyaga District, Kenya. Unpubl. PhD thesis,
Department of Agricultural Economics, Michigan State University, East Lansing, MI.
Franzel, S., 2000. Use of an indigenous board game, ‘bao’, for assessing farmers’ preferences
among alternative agroforestry technologies. Working paper. ICRAF, Nairobi.
Franzel, S., Hitimana, L.,and Akyeampong, E., 1995. Farmer participation in on-station tree
species selection for agroforestry: a case study from Burundi. Experimental Agriculture,
31:27-38.
4
Franzel, S., Ndufa, J.K., and Obonyo, C., 1999. Farmer-designed agroforestry tree trials:
Farmers' experiences with newly introduced tree species in Western Kenya. AFRENA Report
no. 131, ICRAF, Nairobi.
Johnson, G.L., 1986. Research methodology for economists: Philosophy and practice.
Macmillan, New York.
Maxwell, S. and Bart, C. 1995. Beyond ranking: Exploring relative preferences in P/RRA.
PLA Notes No. 22. London: International Institute for Environment and Development.
Negassa, A., Tolessa, B., Franzel, S., Gedeno, G., and Dadi, L., 1991. The introduction of an
early maturing maize variety to a mid-altitude farming system in Ethiopia, Experimental
Agriculture, 27:375-383.
Ohlsson, E., Shepherd, K.O., and David, S. 1998. A study of farmers’ soil fertility
management practices on small-scale mixed farms in western Kenya. Interna Publikationer
25, Sveriges Lantbruks Universitet. Uppsala, Sweden
Okali, C., Sumberg, J., and Farrington, J., 1994. Farmer Participatory Research: Rhetoric and
reality. Intermediate Technology, London.
Pisanelli, A., Franzel, S., DeWolf, J., Rommelse, R., and Poole, J. 2000. The adoption of
improved tree fallows in western Kenya: farmer practices, knowledge, and
perception.Working paper. ICRAF, Nairobi.
Polson , R.A. and Spencer, D.S.C., 1991. The technology adoption process in subsistence
agriculture: the case of cassava in Southwestern Nigeria. Agricultural Systems, 36:65-78
van Veldhuizen, L., Waters-Bayer, A., and de Zeeuw, H., 1997. Developing technologies
with farmers: A trainer’s guide for participatory learning. Zed Books, London.
5
Table 1. Farmers' mean ratings, using the bao game, of selected tree species that grow on their farms,
central Burundi
Use for timber
Management &
growth
Species/
Ratings *
Compatibility with
crops
Speed
of
Growth
Resistance to
Insects
Wood
Appearance
Straightness
Use for firewood
Quick in
drying
Durability
of fire
(mean rating and standard deviation in parentheses)
Maesopsis
eminii
3.8 (0.5)
4.7
(0.4)
4.2 (1.0)
4.8 (0.4)
4.2 (0.6)
3.1 (0.8)
3.5 (0.9)
Cedrela
serrata
4.6 (0.5)
4.3
(0.7)
4.5 (0.8)
5.0 (0)
5.0 (0)
-
-
Grevillea
robusta
4.9 (0.3)
4.6
(0.5)
2.5 (0.7)
2.6 (0.9)
4.1 (0.8)
3.2 (1.0)
2.8 (1.2)
Casuarina
cunninghamiana
1.0 (0.0)
2.2
(0.8)
4.2 (1.0)
4.1 (1.1)
3.9 (0.9)
3.0 (0.6)
3.8 (0.7)
Markhamia
lutea
3.7 (1.0)
1.9
(0.9)
4.5 (0.8)
4.3 (0.6)
1.8 (0.8)
2.3 (1.1)
4.2 (0.7)
Eucalyptus
spp.
1.1 (0.3)
4.3
(0.8)
4.0 (1.2)
2.5 (0.5)
3.6 (1.0)
4.7 (0.5)
5.0 (0.0)
Cupressus
lusitanica
1.0 (0.0)
3.2
(0.8)
4.5 (0.8)
4.6 (0.5)
3.9 (0.8)
4.6 (0.5)
3.5 (1.0)
Albizia
chinensis
4.0 (1.2)
3.5
(1.3)
1.3 (0.7)
-
1.3 (0.4)
2.3 (0.8)
3.3 (0.7)
Source: Franzel, et al. 1995.
Notes: Twenty-five persons were interviewed; the number rating a specific species on a particular criterion varies
from 5 to 20. The rating of 1 to 5 refers to the score in number of seeds the farmers gave to a species on a particular
criteria. A rating of 5 was considered excellent, a rating of 1, poor. For some species certain criteria are irrelevant
e.g., C. serrata is never used for firewood, and A. chinesis is never used for timber.
6
Table 2. Farmers' mean ratings of species in an on-farm trial using the ‘bao’ game, 30 months after
planting, western Kenya
Species
Growth
Biomass
production
Compatability
with crops
Fodder
Firewood
% farmers
preferring for
future planting
(mean rating and standard deviation in parentheses)
Grevillea
robusta
4.4 (0.9)
-
4.0 (1.3)
-
Casuarina
junghuhniana
3.2 (1.1)
-
4.5 (0.7)
-
4.1 (1.0)
-
73
46
Leucaena
leucocephala
-
3.4 (0.8)
3.8 (1.8)
4.0 (1.4)
3.8 (1.0)
29
Leucaena
diversifolia
-
3.7 (0.9)
3.6 (1.6)
3.4 (1.5)
3.8 (0.7)
24
Calliandra
calothyrsus
-
4.9 (0.2)
3.3 (1.8)
4.1 (1.3)
4.1 (1.1)
41
3.6 (1.2)
27
Eucalyptus spp
4.3 (1.0)
-
1.4 (0.9)
-
Notes: Data from 37 farmers. The rating of 1 to 5 refers to the score in number of seeds the farmers gave to a
tree on a particular criterion. A rating of 5 was excellent, a rating of 1, poor.
7
Table 3. Farmers’ mean scores on the importance of different benefits of improved fallows
using the bao game, western Kenya (standard deviations in parentheses)
Benefit
Overall score
No. of
farmers
Improved crop
yields
Improved
soils
Reduced
weeds
Fuelwood
production
Seed production
Pest
reduction
Males
Mean No. of
score farmers
Females
Mean
score
No. of
farmers
Mean
score
Difference Standard
in means error of the
(Females- differences
Males)
Signifi95%
cance of confidence
difference interval for
(p-value)
the
difference
63
4.4 (0.82)
16
4.2 (0.98)
45
4.5 (0.73)
0.3
0.23
0.169
(-0.14; 0.78)
63
4.5 (0.78)
17
4.1 (0.99)
45
4.6 (0.66)
0.5
0.22
0.037
(0.02; 0.90)
62
3.6 (1.26)
16
3.1 (1.29)
44
3.7 (1.20)
0.7
0.36
0.059
(-0.03; 1.14)
67
3.9 (1.19)
17
3.6 (1.23)
48
3.9 (1.18)
0.3
0.34
0.333
(-0.35; 1.01)
31
3.7 (1.05)
13
4.0 (0.86)
17
3.4 (1.12)
- 0.5
0.37
0.183
(-1.27; 0.25)
10
3.3 (1.25)
4
3.8 (1.50)
6
3.0 (1.10)
- 0.7
0.82
0.384
(-2.64; 1.14)
Notes. Sixty-seven farmers were involved in the evaluation. ‘No. of farmers’ refers to numbers mentioning a
particular benefit. Numbers of males and females do not sum to totals because in some cases, males and females
preferred to score benefits together.
8
Annex 1. Tips for obtaining farmer evaluations using the bao game
The bao game requires some practice and can generate inaccurate results if not implemented
carefully. The following discussion of guidelines and issues is intended to help researchers to use the
tool effectively so as to obtain accurate farmer evaluations.
1. Before conducting the game, walk around the farm with the farmer. Try to get all of the major
decision makers (e.g., male head of household and wife or wives) to accompany you. Observe the
species that were planted in the trial and discuss the performance of each with them. The farm visit
serves two purposes. First, it provides the farmers with orientation for playing the game -- the
performance and characteristics of each species will be fresh in their minds. Second, in the farm tour,
you can learn the criteria that farmers use in evaluating the species. For example, in visits to three
farms in the same area, the following criteria were brought up by farmers:
i Growth/management criteria
a. Speed of growth
b. Survival
c. Resistance to drought
d. Resistance to termites
e. Speed of growth under shade of hedge or other trees
f. Compatibility when grown with crops
ii. End uses: Use of species for
a. Poles
b. Timber
c. Medicine
d. Firewood
e. Fruit
f. Manure/mulch to improve crop production
g. Fodder
It may be useful to break down some of these criteria. For example, the value of a species for poles
may depend on both the straightness of the tree and the durability of the wood. Therefore, it would be
useful to rate species first on straightness, then on durability, and finally a composite rating on
usefulness for poles.
During the farm tour, it is important to note down each criterion as soon as the farmer mentions it.
Prompt the farmer with such questions as "how do you compare the performance of these two
species?" or "what do you like and dislike about this species?
2. It may be best to put the game on a table, so that participants do not have to keep bending over to
reach the beans in the pockets. Alternatively, they can sit or kneel on the ground around the game. To
set up the game, put a branch of each species next to each row of "pockets". The order of species in
the game should be the same as the order of species on your form/questionnaire. Explain the
objectives and procedures of the game and show farmers how to score species by putting the beans in
the pocket next to the species. Discuss one criterion at a time, and ask the farmer to score each species
on the selected criterion, 5 beans is an "excellent" rating and 1 bean is a "poor" rating. Leaving the
pocket empty means the farmer does not know how the species performs on the criteria.
3. When the farmer has completed rating all the species on a criterion, ask him/her questions in order
to check that he/she agrees with the scores he/she has given. For example, if two species receive the
same score on resistance to termites, ask, "do these two species have about the same degree of
resistance to termites?". Encourage the farmers to make corrections, adding or subtracting the number
of beans in pockets, in order to make sure that the beans placed reflect his true preferences. These
check questions are especially important when more than three species are being compared in a single
game.
9
4. Do not fill in your chart on the farmers' scores until you are sure he/she has completed scoring and
modifying his scores for a particular criterion. Farmers will usually make some modifications during
the discussion that follows initial placement of the beans. The best time to record the responses is
when it is evident that it is time to discuss the next criterion.
5. Word all criteria in a positive way -- instead of discussing termite damage discuss resistance to
termite damage. This way, the scoring on each criterion will be highest for those species that perform
best from the point of view of the farmer.
6. The issue of the role of outsiders, e.g., neighbors visiting the farm, depends on the objectives of the
exercise. If your objective is to learn what criteria are important to farmers in evaluating species, it
may be advantageous to involve a group of farmers in playing the game rather than using individuals.
In group interviews, you obtain information from more farmers (increasing your sampling efficiency)
and more important, you can learn a great deal from the interaction among farmers -- e.g., their
discussions, disagreements, and reasons for disagreeing. However, if your objective is to associate
farm or farmer characteristics with species performance or preferences, you need to obtain the
preferences of individual farmers without interference from others. In this case you have to find a
diplomatic way of excluding outsiders from participating. Also, beware of extension agents or local
"politicians" espousing ‘official’ views, e.g., denouncing Eucalyptus. Even if these people keep quiet,
their presence when the game is played may keep farmers from presenting their true opinions.
7. Farmers should be encouraged to leave a pocket blank if they are not sure of how the species
performs on a particular criterion. For example, during a field visit, a farmer suspected that his
Mimosa scabrella had not survived because of poor resistance to drought; during the game he/she
decided not to rank the species on this criterion because he/she was not sure. The interviewer in this
case should leave the rating blank but note below that the farmer suspects Mimosa was not resistant to
drought.
8. Interviewers have to try to keep the farmer's focus on the particular criterion at hand; otherwise
farmers may give a species a rating with another criterion in mind. For example in rating species on
rapidity of growth, a farmer gave Maesopsis a high rating mentioning that it grows very straight; in
this case the farmer has to be reminded that the criterion we are discussing was rapidity of growth,
straightness could be handled next.
9. In rating a species, farmers should rate it relative to other species, not relative to what they expect
from that species. For example one farmer gave Markhamia lutea a 5 on rapidity of growth, even
thought it was growing much slower than the other species. When asked why, he told us that he gave
it a 5 because it grew fast compared to other Markhamia lutea seedlings he had planted in the past,
not with the other species in the trial.
10. There are different ways to involve different household members in the game; For example, where
different members have different preferences it may be useful to gather the same information from
each, e.g., each can have a different column of the game to score the species on the same criterion.
Alternatively, you can ask them to try to come to a consensus -- however, if their opinions are actually
conflicting this will not yield useful data. Also, recognize that different household members will have
knowledge about different aspects. For example, in Africa, women tend to be more knowledgeable
about the quality of firewood whereas men may know more about timber.
11. Once you have finished obtaining ratings for all the species for all the criteria, ask the farmer if
there are any other important criteria. Finally, ask farmers to give a general rating of each species
across all criteria. There are different ways you can do this. One is to ask which trees are you most
interested in planting in the future? The farmer can rate 5 beans for the ones he/she is most interested,
1 for the ones of least interested in, etc. The problem with asking this question is that after planting
250 trees for our farmer-designed trial they may no longer be interested in planting trees! So perhaps
10
it is better to ask which trees would you most highly recommend to a neighbour who wanted to plant
more trees.
12. Try to find out from farmers which criteria are the most important ones. The best way to do this is
probably to ask farmers which criteria are very important and which are of medium importance. It is
probably not possible to actually rank criteria or even to break them down into more than two groups
but we can experiment with this. We could also use a tool called conjoint analysis to regress the
criteria values on the dependent variable, overall rating, -- this would give us a quantitative value for
weighting of each criterion.
13. Given the complexity of the game, it is probably necessary that at least two interviewers be
present. One is responsible for conducting the interview and assisting the farmer in playing the game,
the other records the scores and notes farmers’ reasons for giving high or low scores.
14. We have found it possible to have farmers rate up to 12 species at a time. Rating 12 species is not
much more complicated than rating 4 species although it does take more time.
15. We have used the game for rating farmers' existing trees on their farms as well as newly
introduced species that farmers have recently planted. Of course, farmers have much more knowledge
about species they have used for a long time. Nevertheless, in on-farm trials, the game is useful for
getting early feedback on newly introduced species.
11