A Note on the Role of Emotions in Evaluation and Choice and their Calibration Ken Roberts 1, John Roberts 2, Rohan Raghavan 3, and Peter Danaher 4 Introduction The importance of modeling choice and its determinants has been well-established in marketing for over thirty years with a large body of research devoted to its study (see, for example, Guadagni and Little 1983). The dominant approach has seen the consumer viewed as “an information processing system, purposely acquiring brand information and systematically integrating it to form brand preferences; a cognitive creature” (Mittal 1988). That representation has served us well. Modeling choice using multi-attribute models has a long pedigree in marketing science (e.g., Agarwal and Malhotra 2005). In this note we propose that marketing scientists can profit by extending this approach to incorporate the role of emotions in decision making. Historically, the psychology literature has considered three aspects to consumer behavior: what the consumer thinks (cognitive), what the consumer feels (affective), and what the consumer does (connotative). See, for example, Cohen, Pham and Andrade (2008); Hilgard (1980); Bagozzi, Golpinath and Nyer (1999) and Berkowitz (1993). To quote Epstein (1994) “there is no dearth of evidence in everyday life that people apprehend reality in two fundamentally different ways, one variously labeled intuitive, automatic, natural, non-verbal narrative, and experimental, and the other analytical, deliberative, verbal, and rational.” This is also reflected in Kahneman (2003)’s System 1 (Intuition) and System 2 (Reasoning) classification of thinking. While cognitions may explain much in terms of evaluation and choice, they do not explain everything. For example, Fitzsimmons, Chartrand and Fitzsimmons (2008) show that brand primes, not retrievable from memory by the subject can cause non-conscious behavioral change (and non-conscious construct activation). We argue for the incorporation of such phenomena in our marketing science models and demonstrate a path by which that may be achieved. To understand the importance or otherwise of emotions in consumer decision making, it is useful to refer to a number of different bodies of literature; psychology, consumer behavior, neurobiology, economics and marketing science. Of these, psychology and consumer behavior have the longest history of considering the role of emotions, and the richest literature. The importance of emotions in choice in those disciplines is seen as clearly established. Many distinguished articles have, as their first sentence, a statement to this effect (including Andrade 2005; Angie et al. 2011; Avnet, Pham and Stephen 2012; Burke and Edell 1989; Cohen, Pham and Andrade 2008; Dubé, Chattopadhyay and Letarte 1996; Herr et al. 2012; Johnson and Stewart 2010; King and Janiszewski 2011; Ladhari 2007; Laros and Steenkamp 2005; Mellers, Schwartz and Ritov 1999; Pham 1998; Richins 1997; Russell 2003; Ruth 2001; and Williams 2014). Others decry the fact that the subject has not received even more attention (e.g., Shiv and Fedorikhin 1999 and Bagozzi, Golpinath and Nyer 1999). The role of emotions forms one of the Journal of Consumer Research’s eight research curations (areas of specialized focus). While emotions have long been regarded as important in 1 Chairman and CEO, Forethought Research Professor of Marketing, University of New South Wales and London Business School 3 Research Analyst, Forethought Research 4 Professor of Marketing and Econometrics, Monash University 2 1 psychology and consumer behavior, learning more about them has recently become a “hot” topic (Cohen, Pham and Andrade 2008). There are a number of reasons for this interest. Morris et al. (2002) point out that often cognitive models fail, and they fail due to a lack of inclusion of affective elements. Emotions have been shown to add explanatory power to the understanding of preference in consumer behavior studies across a wide range of categories, media and formats, usually by the use of structural equation models (e.g., Kim and Morris 2007, Morris Moreover, compared with store visits levels before the campaign, total annual visits increased by 20% over the next two and a half years, while the number of items sold increased by 42%. Last, EBIT increased by 30%. . 2002, and Allen, Machleit and Kleine 1992). Burke and Edell (1989) speak of the critical role of emotions as a complement to cognitions. Edell and Burke (1987) go so far as to suggest that in their empirical results feelings always matter in assessing the effect of advertising. In extremis, decisions may occur only using input from emotions (what Mittal 1988 calls “Affect based choice” and a number of researchers refer to as decision-making by the “How do I feel about it?” heuristic (Pham 1998)). While many authors point to the strong role that emotions have in determining choice (e.g., Mellers, Schwartz and Ritov 1999 and Pham et al. 2001), others have documented its influence on other behaviors, such as the spread of word of mouth (Howard and Gengler 2001, Ladhari 2007, and Eckler and Bolls 2011). A substantial literature also exists on the role of emotions on customer satisfaction (e.g., Oliver 1993 and Philipps and Baumgartner 2002). Angie et al. (2011) provide a useful review of the literature regarding the role of emotions in judgment and decision making. Neurobiology is a discipline which has benefited tremendously from our ability to take finer measurements of brain function and, as it has expanded, it has stressed the role of emotions or feelings. Initial studies focused on the brain activity and behavior of patients who had suffered accidents, surgery or lesions (e.g., Bechara and Damasio 2005). For example, Damasio (2005) shows that patients with ventromedial pre-frontal cortex damage lose the ability to process emotions. These patients are terrible decision makers because they don’t have an emotional marker of how important decisions are and the emotional consequences of their choice. They get stuck in analysis paralysis, even with trivial decisions such as whether to have dessert or not. More recent advances in imaging have allowed neurobiologists to trace neural activity much more closely to specific areas of the brain and, by inference, decision processes. Ledoux (1989, 1995) provides a seminal description of the early work in this area, providing a foundation for the neural basis of studying emotions. He concludes that “emotion and cognition are mediated by separate but interacting systems in the brain….. This theoretical perspective places them on equal conceptual footing as companion (somewhat parallel) processing systems of the brain.” More recently Naqvi, Shiv and Bechara (2006) have suggested that [neurobiological research] “has converged with the field of behavioural economics in showing that decision making involves not only the cold-hearted calculation of expected utility based on explicit knowledge of outcomes but also more subtle and sometimes more covert processes that depend critically on emotion.” Economics has been slow to embrace the role of emotions in decision making, despite the improved dialog that behavioral economics has brought about between economics and psychology. Economics started with the consumer as a pleasure maximizing decision maker. Since Jeremy Bentham, however, decision analysis has led expected utility maximization to dominate the subject’s 2 interest in choice (Loewenstein 2000). In 2000, Thaler suggested that the re-entry of emotions into economists’ study of consumer was imminent and Loewenstein (2000) also points to renewed interest by economists. Hannoch (2002) and Elster (1998) at the same time have joined the refrain that economists need a better understanding of emotions and Kahneman, Wakker and Sarin (1997) have lodged a plea to return to Bentham’s conceptualization of experienced utility. However, little has been happening. Bechara and Damasio (2005) decry the fact that “modern economic theory ignores the influence of emotions on decision making.” They suggest that “although the view of maximizing utility of decision making is pervasive and has a useful benchmark function, human decision makers seldom conform to it.” In a recent review on the “Role of Emotion in Economic Behavior,” Rick and Loewenstein (2008) find only three references since 2000 in economics journals, of which two refer more to heuristics and biases than emotional states. The empirics do not match the rhetoric. If economics has been slow to embrace the potential insights that an understanding of emotions could provide to decision making research and practice, then marketing science has been absent. A search on Google Scholar of “Marketing Science Emotions” or “Choice Models Emotions” provides no references in marketing journals since 2000. We argue that the findings and excitement being experienced in psychology and neurobiology suggest that this is an area where we would do well to at least investigate the opportunity to gain greater insight into choice and how marketing activity affects it. To quote Loewenstein (1996, p 289) “With all its cleverness, however, decision theory is somewhat crippled emotionally, and thus detached from the emotional and visceral richness of life.” While psychologists and consumer behavior researchers have argued that emotions are strong influences in how consumers evaluate and choose between objects and, in particular products, there is also considerable evidence that emotions also regulate how newly experienced marketing stimuli (such as brands, advertisements and store ambience) affect those evaluation processes (Batra and Ray 1986). As Lutz (1975) rather nicely puts it “feelings are perceived not as qualities of the object [the product] but as states of the subject [the consumer].” One of these major stimuli is the brand name itself. Relationships between consumers and their brands (as manifested by physical product characteristics and other elements of the marketing mix) in many ways mirror relationships between people (Fournier 1998). When speaking of brands, Heath, Brandt and Nairn (2006) go so far as to suggest “it is the emotional not the rational content in communications that drives relationships.” It is unsurprising then that feelings between consumers and their brands are important in shaping the consumer-brand bond. Keller (2003) advances feelings as a key concept in his construct of consumer “brand knowledge.” Thomson, MacInnis and Park (2005) identify three dimensions of brand emotional attachment (affection, connection, and passion) and demonstrate their convergent and discriminant validity. Similarly, advertisements may evoke feelings which, in turn, can (either directly or through thoughts and/or feelings), influence product evaluation and choice (MacInnes and Jaworski 1989). Emotions may also lead to the censoring of advertisements through zapping and zipping (Olney, Holbrook and Batra 1991). Based on the literature, MacKenzie, Lutz and Belch (1986) test four models of how feelings might affect intentions to purchase, all through attitude to the ad, AAd. Their data favored a dual mediation model; AAd affects both cognitions about the brand and the brand 3 attitude directly. Pham, Geuens and de Pelsmacker (2013) also tested 1070 television commercials on 1576 consumers. Advertisement evoked feelings had strong effects on brand attitudes both directly and mediated by AAd, across different levels of involvement and in different categories, but more strongly with hedonic goods. Holbrook and Batra (1987) and Stayman and Aaker (1988) come to similar conclusions. Finally, emotions have been empirically shown to be important in the retail environment, as hypothesized by Johnson and Stewart (2010). Ladhari (2007) summarizes a series of studies that indicate their role in the amount purchased, time spent in store, number of items purchased and liking of store. See also Sherman, Mathur and Smith (1997), Donovan and Rossiter (1982) and Donovan et al. (1994). Background on Emotions and Affect The nature of emotions Emotions are often thought of as a pre-cognitive reaction to stimuli that stems in an evolutionary sense from the need for fast “fight or flight decisions (Loewenstein 2000 and Griskevicius, Shiota, and Nowlis 2010). They also act as gatekeeper mechanisms for goals, warning when cognitive evaluations “don’t feel right” (Johnson and Stewart 2010). It is when we are in a state of disequilibrium that emotions have the ability to play an important regulatory role (Johnson and Stewart 2010). They signify a state of readiness and, as such, should be of particular interest to marketers. See Bagozzi, Golpinath and Nyer (1999) who provide an interesting and insightful discussion of emotions, including a contrast with their cognitive counterparts. Before proceeding it is useful to establish the meaning of some commonly used terms. Historically, an analysis of emotions has been made difficult by the fact that different researchers have used the same terms with different meanings (e.g., Russell 2003). However that problem seems to be reducing as a consensus on usage begins to emerge. The term “Affect” is used as an umbrella term to describe a state of wellbeing (or otherwise) that a person may experience (Cohen Pham and Andrade 2008 and Herr et al. 2012). Affect is usually described by its mood and emotion elements. Mood is a generalized state of being, normally not specific to any object and usually somewhat longer lasting than an emotion. Emotions, the subject of this background note, are specific in relation to an object and more transitory. Most authors in consumer behavior explicitly use the terms “emotions” and “feelings” interchangeably (e.g., Burke and Edell 1989 and Ruth 2001) and we shall follow that practice as well. It is useful to consider some aspects of emotions that will assist in our measurement and modeling of them. Below we list some of their more interesting qualities. • There is some debate as to whether emotions must precede cognitive processing, with Zajonc (1980, 1984) arguing in favour and Lazarus (1982) against. This would not appear to be a fruitful debate (at least not for our purposes), since in a dynamic system with continuing feedback, the literature does appear to be quite clear that cognitions can inform later feelings and conversely (Ledoux 1989). The continuous temporal interaction between the two systems makes this debate at best a fine one. 4 • • • • • • • Feelings are subjective. They are specific to the person experiencing them (although Loewenstein 2000 suggests that there is a lot more agreement amongst consumers about their feelings towards a product than for cognitive beliefs). Affective states are self-focused (Mittal 1988) and are personalized representations of the world (Ledoux 1995). Feelings are also experienced with respect to a specific object: one feels a given set of emotions about something (Johnson and Stewart 2010). Finally, feelings are generally valanced (Bagozzi, Golpinath and Nyer 1999). Feelings are heterogeneous. They vary from person to person and for a given person, they vary from situation to situation (Slovic et al. 2004). Consumers likely to rely more heavily on emotions in decision making include those with a low need for cognition (Caccioppo et al. 1996), those with a low cognitive ability (Atkins and Ozanne 2005), those with a strong affect orientation, highly visual and sensory people (Pham 1998), impulsive people (Shiv and Fedorikhin 1999), and those for whom feelings provide a strong fit with their social identity (Coleman and Williams 2013). Situations in which a given consumer is likely to rely more on emotions include buying low involvement, expressive or hedonic products (Mittal 1988 and Pham 1998), non-comparable alternatives (Pham 1998), products requiring low cognitive resources (Shiv and Fedorikhin 1999), situations where feelings are readily accessible and diagnostic (Avnet Pham and Stephen 2012 and Dubé, Chattpopadhyay and Letarte 1996), in cases when attitudes are weakly held (Smith, Haugtvedt and Petty 1994) and when the consumer is promotion focused rather than prevention focused (Cohen, Pham and Andrade 2008). Emotions determine trade-offs in decision making (Loewenstein 2000) While emotions may be transitory (Zeelenberg and Pieters (2006), their resultant effect on evaluation may be enduring (e.g., Baumgartner, Sujan and Padgett 1997 and Loewenstein 2000) Emotions can be either experienced (looking backwards) which has been the focus of much psychological research or anticipatory (looking forwards), more the domain of economists (Loewenstein 2000). They may be either instrumental (central to the evaluation of the focal object) or incidental (Cohen, Pham and Andrade 2008 and Coleman and Williams 2013). Affect may influence evaluations and decisions either directly or by its possible effect in shaping cognitions (Cohen, Pham and Andrade 2008). That is, emotions may act either as a cue or as a resource to cognitions (Bakamitsos 2006). In terms of the latter, there is an extensive literature on “affect as information” (e.g., Herr et al. 2012, Schwartz and Clore 1983, and Avnet Pham and Stephen 2012). Emotions may also help the brain overcome some of its limitations due to bounded rationality (Hanoch 2002). The focus of this note is descriptive, rather than normative. However, there has been considerable work, conceptually and empirically, as to whether emotions lead to improved or degraded decision making (e.g., Herr et al. 2012). Historically emotions were perceived as “a destructive force in human behavior” (Loewenstein 2000). However, there is growing evidence, neurobiological and psychological) of their role in increasing evaluative efficiency (Hanoch 2002, Ledoux 1989, and Zeelenberg and Pieters 2006). For example, Damasio (2005) observed that patients with damage to their amygdala (a home for much emotional reaction) were unable to make decisions about the simplest issues. Chung (2007) suggests that this may not be symmetric: positive emotions are more likely to lead to improved decision making than negative ones. 5 For a more formal treatment of these concepts in psychology, see Russell (2003) Given the importance accorded to emotions in the psychology and consumer behavior literatures, one may expect that considerable effort has gone into understanding how best to represent them. At its simplest, researchers think in terms of emotions in terms of their valence (pleasurable/discomforting). However, Watson and Tellegen (1985) observed that negatively valenced states are not just the opposite of positively valenced ones. They developed their circumplex model of emotions in which each of the eight resulting sub-quadrants represents a different emotional state (with associated differential behavioral outcomes). Based on this idea Watson, Clark and Tellegen (1988) developed and validated a PANAS scale (Positive and Negative Affect Scale) with each dimension consisting of ten items. A number of variations of this 2x2 framework have emerged. Other dimensions introduced include appetitive/aversive and high/low arousal. The use of three underlying dimensions has also been suggested (for example, Russell (1980)’s Pleasure/displeasure, arousal/non arousal, and dominance/submissiveness (PAD) and Edell and Burke (1987)’s Upbeat, Negative and Warm dimensions). While these two (and three) dimensional models are admirable for their parsimony, they are limited in their explanatory power. Cohen, Pham and Andrade (2008) point out that many two or three dimensional models assume that negative emotions are the converse of positive emotions (as opposed to potentially different constructs) and they frequently do not allow the respondent the option of a neutral point. Angie et al. (2011) observe that emotions are inherently multi-faceted. Multidimensional representations will be more diagnostic and they will also enable us to better understand the role of mixed emotions (Williams and Aaker 2002). While two dimensions may be adequate when there is no variance in the appraisal conditions, often different negative emotions may bring about extremely different behavioral outcomes (Bagozzi, Golpinath and Nyer 1999 and Lerner and Keltner 2000). For example, the difference between sadness and anger may stem from differences in the attribution of the cause of the source. Sadness may well lead to negative word of mouth, while anger may lead to defections (Angie et al. 2011). Zeelenberg and Pieters 2006, for example, argue passionately for the adoption of multidimensional representations of emotions states. They suggest that the main reason for the continued existence of simple representations is their “parsimony, communicability, and measurability.” There is an enormous literature identifying different individual and groups of emotions that might be at play in consumer decision making (see Laros and Steenkamp 2005 and Ladhari 2007 for reviews). Perhaps the most influential has been that of Richins (1997) who used a multi-stage project to develop a 16 item battery of consumption emotions. While many emotions crop up on almost every list, the list identified in the literature is huge. Laros and Steenkamp (2005, Table 2) identify 173 negatively-charged emotions and 140 positively charged ones. Research has focused not just on emotions in the consumer evaluation process of products, but also the feelings that marketing stimuli arouse with respect to those products. Most notable among these is the work done on the role of feelings with respect to advertising. See, for example, Batra and Holbrook (1990), Edell and Burke (1987), Batra and Ray 1986, and Bagozzi, Golpinath and Nyer 1999. Scales developed in this area are similar to those used to calibrate emotions with respect to objects and that gives us confidence to use the same measurement approach to consider feelings 6 evoked when a consumer considers a brand or product and the feelings they may experience when reacting to a stimulus that relates to that product, such as a marketing mix element. In lauding the parsimony of 2x2 scales and the diagnosticity of multidimensional ones, Laros and Steenkamp (2005) wondered whether some intermediate representation might occupy a position on the efficient frontier of the diagnosticity-parsimony trade off. Laros and Steenkamp identify four positive (Contentment, Happiness, Love, Pride) and four negative (Anger, Fear, Sadness, Shame) emotions. The four negative emotions correspond to four of Richins’ eight, while two of the four positive ones are the same as two of her eight. One is very similar (Laros and Steenkamp’s Happiness is semantically close to Richins’ “Joy”). Surprisingly, for some reason, “Pride” does not appear to figure in any of the progressive winnowing down of scales by Richins, despite its importance being identified by other scholars (e.g., Griskevicius, Shiota, and Nowlis 2010 and Roseman 1991). Laros and Steenkamp demonstrate the desirable measurement properties of their scales. Barriers to Incorporating Affect in Choice Models In view of the historical and even stronger recent interest in the role of feelings in judgment and decision making, on the face of it, it is curious that this subject has not attracted more attention from either marketing scientists or managers who aim to influence consumer behavior. By identifying the reasons for this, we are likely to be better equipped in our endeavor to remedy it. There are three major reasons for the lack of study of emotions in choice modeling, given their established importance in decision making: differences in culture, measurement difficulties, and modeling challenges. Cultural differences. Culturally, Kahneman (2003) points out that economists (and by extension perhaps, marketing scientists) “often criticize psychological research for its propensity to generate lists of errors and biases, and for its failure to offer a coherent alternative to the rational-agent model.” However, he suggests that they are on shaky ground when their “rational models are psychologically unrealistic.” Loewenstein (2000) suggests that emotions are unappreciated and regarded by economists as unpredictable (a belief that he holds to be in significant error). Bechara and Damasio (2005) attribute a lot of these cultural issues to measurement and modelling issues: “disagreement on how to define them, disagreement on what they are for, and what to include in them.” With the rise of behavioral economics, we could argue that many of these communications problems should be reduced in magnitude. Measurement challenges. There are many challenges to the measurement of emotions. The most obvious is that “affective evaluations do not align with [an object’s] discrete, intrinsic attributes” (Mittal 1988). It is difficult to measure emotions because our most prevalent way of eliciting individuals’ brain activity, namely asking them about it, perforce requires them to cognitively retrieve those emotions (Morris et al. 2002). While that is not necessarily disastrous, it has the potential to introduce considerable measurement problems. A number of measurement approaches to assessing emotions have been suggested, many of which attempt to address this difficulty. Because feelings are generally believed to often occur at a pre-cognitive stage of evaluation (Zajonc 1984), it is difficult to elicit them by forcing the respondent into a thinking mode. We 7 consider the methodology that we have developed to quantitatively assess feelings to be quite radical as it allows for the measurement of responses in a non-verbal, pre-cognitive manner and the data collected through this process are amenable to modeling and other quantitative exercises. Modeling challenges. There is also very little guidance as to how to represent emotions in choice models. While emotions have been related to behavioral intent extensively using regression and structural equations approaches (e.g., Kim and Morris 2007), we know of no research that has examined choice models, and only little that has considered non-linearities and interactions. Before we attempt to measure emotions, it is useful to examine previous efforts to do so and to build on the strengths of those that have met with some success. We classify measurement approaches to emotions as being physiological, direct, or indirect. While work in the modeling of emotions is at best in its nascent stage, there is an extensive body of literature regarding their calibration. For excellent reviews, see Mauss and Robinson (2009), Poels and Dewitte (2006), and Erevelles (1998). Physiological approaches to emotion measurement Calibration of emotions related to marketing stimuli using physiological reactions has a long history in marketing. Use of skin conductance (galvanic skin response or GSR) based on the observation that perspiration increases with heightened arousal has been used for over forty years (e.g., Kong et al. 2013). More recently, advances in neurophysiology have allowed the profession to move beyond the broad measures of brain activity detected by techniques such as electroencephalography (EEG) to identify very specific areas of arousal. For example, functional magnetic resonance imaging (fMRI) by measuring oxygen uptake to different areas of the brain has been able to trace neural activity in those parts associated with emotions, such as the amygdala. These approaches have been demonstrated to discriminate between different marketing stimuli such as television commercials (e.g., Ohme at al. 2009 and Keil et al. 2003). Another approach to measuring emotions using physiological reactions includes the use of facial recognition software. Facial response to external stimuli registers in a number of ways. For example, surprise is often accompanied by blinking and the electromyogram (EMG) can detect the extent of movement in the lower eyelid to act as a “startle probe” (Mauss and Robinson 2009). More sophisticated facial recognition software can identify other emotions (such as the positive valanced feelings associated with a smile). There is some question as to whether such reactions transcend cultural and ethnic boundaries (with Ekman and Friesen 1971 arguing in favor and Russell 1994 expressing doubt). Perhaps of more concern is the fact that Ledoux (1989) and Cohen, Pham and Andrade (2008) suggest that facial recognition approaches lack the requisite specificity to be diagnostic in discriminating between the different emotions that drive various behaviors. In summary, physiological approaches have a number of advantages. They do not require the respondent to retrieve information and are thus less subject to cognitive filtering and they are very good at tracing emotional reactions over time (over the 30 seconds of a television commercial, for example). However, they are costly (often leading to small samples) and not highly diagnostic. While we would not go so far as Johnson and Stewart (2010) in saying that they are “cumbersome, very intrusive, and do not differentiate many of the emotions [and thus] provide little insight into the causes or consequences of emotional response,” we would feel that their role is a quite specific 8 one. A detailed assessment of their potential and limitations is provided by the Advertising Research Foundation (2011). Direct elicitation solutions The most common approach to measuring emotions in consumer behavior is by direct elicitation (Burke and Edell 1989). For example, Edell and Burke (1987) develop, test and validate a 56 item feelings scale to calibrate consumer response to advertising. Bagozzi, Golpinath and Nyer (1999) review the extensive array of different scales proposed for this purpose. Most verbal elicitation approaches measure a large number of (related) emotions and then use some combination of theory and data reduction to develop a taxonomy of higher level emotions (Cohen, Pham and Andrade 2008 and Aaker, Stayman and Vezina 1988). This approach has the advantage that it is very easy to administer, particularly in conjunction with the collection of other information such as cognitions, preferences, and behavior. For example, Voss, Spanenberg and Grohmann (2003) develop a combined hedonic/utilitarian 10 item scale. It is also highly diagnostic because inventories of emotions can be customized to the application in hand and administered in as much detail as is costeffective. However, it does have a number of drawbacks. By asking the respondent to verbalize his or her emotions, it is likely that cognitive processing mechanisms will be engaged (Poels and Dewitte 2006). While it may be possible for respondents to recall emotions (and thus it may be still emotions that are being measured), they will effectively be “filtered” through cognitions or lost in translation. Additionally, a number of emotions may have no cognitive equivalent. For example, the German word for the feeling “schadenfreude” (meaning taking pleasure from another’s misfortune) has no English equivalent (Johnson and Stewart 2010). Coulter and Zaltman (1994) note that “brand images have a strong non-verbal component. It is clear that there are some evaluative aspects of the object or brand that do not align with its discrete, intrinsic attributes.” While there are those who defend the use of verbal elicitation (e.g. Batra and Ray 1986), it is clear that such measured emotions may be diagnostic, but are also likely to be subject to systematic error. Indirect measurement solutions Difficulties in measuring emotions physiologically or directly have led to a number of indirect approaches. This line of research also has a long pedigree in marketing, starting off with projective techniques such as storytelling, sentence completion, thought listing, scenario analyses (Aaker, Kumar and Day 2008). Indirect approaches aim to understand emotions evoked by analogy (to a related experience, to a different person, to a similar concept, etc.). One highly systematic way of understanding the structure of emotions and how they relate to specific stimuli is the Implicit Association Test (e.g., Greenwald McGhee and Schwartz 1998). To illustrate this approach imagine a subject was asked to say “Hello” every time they were given a man’s name and “Goodbye” every time they saw a woman’s name. In a second task, respondents are asked to say “Hello” every time they were given a picture of a man and “Goodbye” every time they saw a picture of a woman. Response times in the latter task will be quite short because of the related association between men’s/women’s names and faces. For a second set of respondents, the second task is reversed: respondents are asked to say “Hello” every time they were given a picture of a woman and “Goodbye” every time they saw a picture of a man. This is now a more difficult task. The difference in response times (latencies) between the second sub-sample’s responses to the second task and those of the first sub-sample may be taken as a measure of the degree of association between 9 men’s/women’s faces and names. See Greenwald McGhee and Schwartz (1998) for an example of the application of this technique to judge the association between different feelings and different types of peoples. One form of eliciting the relationship between people and objects, including the associations involved, is by the use of metaphors. Zaltman and Coulter (1995), for example, argue in favor of the use of metaphors, suggesting that non-verbal cues tend to be more credible than verbal ones. They give a detailed overview of a methodology to understand concepts associated with different products or other objects. By asking respondents to collect pictures, articles and ideas related to a stimulus and then working with that material to understand it, build on it, and interpret it, their proprietary product, ZMET, “surfaces the models that drive consumer thinking and behavior.” ZMET has the advantages of being highly idiosyncratic (that is, it can be customized to each individual’s market perceptions), revealing considerably more depth than other methods, and being very diagnostic. However, it is not readily quantifiable, scalable, objective, or reproducible. For a theoretical framework that demonstrates how metaphors (or “conceptual blending”) can be used by the brain to create feelings without engaging cognitive processing facilities see Slingerland (2005). These metaphors can be quite simple, such as a pictorial representation of a person experiencing a given feeling and, as such, indicate one possible path by which we might be able to gauge feelings. While indirect methods of identifying the association between objects (such as products) and reactions (such as feelings) appears to have promise, the question still arises as to how to design stimuli to represent feelings in a valid yet cost effective way, and how to gauge respondents’ reactions to those stimuli. Since Morris et al. (2002) point out the dangers of using words for calibrating emotional reactions, we turn to visual representations. Bradley and Lang (1994) proposed the use of three sets of pictures to measure respondent emotional reactions to stimuli on Russell (1980)’s Pleasure/Arousal/Dominance classification. They call this the Self Assessment Manikin (SAM) and demonstrated its convergent and discriminant validity against a semantic differential directly elicited measure. Keil et al. (2003) showed the usefulness of this approach in evaluating advertisements. Using a similar approach, Desmet developed a series of cartoon characters (Product Emotion Measure or PrEmo) that have been applied to advertisement evaluation (Desmet, Hekkert and Jacobs 2000). Eighteen different animated characters are used to represent nine positive and nine negative emotions, from which “an emotions map” plotting both specific emotions and the position of products relative to them can be derived. While SAM captures a limited range of emotions and PrEmo only detects emotions dichotomously (with no measure of degree), visual representations of emotions do appear to provide a mechanism to gauge emotions by establishing an association between the representation and the feeling experienced. There is some debate about the degree to which SAM eliminates cognitive processing, with Morris et al. (2002) suggesting that it is completely eliminated and Poels and Dewitte (2006) being more circumspect, stating “we agree that SAM reduces introspection and cognitive process when compared to a verbal self-report. However, it does not completely eliminate it.” The degree to which pictorial representations of emotions can capture pre-cognitive emotions depends to some extent not only on the pictures, but also on the way in which respondents’ reactions to them are elicited. 10 “Pen and pencil” approaches (including choice amongst verbally-expressed options using on line surveys – even yes/no) are likely to lead to more cognitive engagement than the click of a mouse or other physical reaction in which no words are involved. Aaker, Stayman and Hagerty (1986)’s “warmth monitor” which requires horizontal movements of a pencil as it moves down the page provides a one dimensional example of such an instrument. See also Pham et al. (2001)’s comparison of turning a dial versus pencil and paper answers to elicit emotional responses. Emotions measurement summary Since each of the above approaches to emotions measurement has its strengths and weaknesses, one possible solution would be to combine them. One could gain accurate, dynamic measures of arousal using physiological approaches, for example, and then combine them with direct elicitation methods to reveal deeper insights as to which specific emotions are being aroused. Alternatively, since in many applications this will not be cost effective, Micu and Plummer (2010) suggest that choice of methodology should be a matter of “horses for courses.” Given the different strengths of different approaches, the choice of measurement instrument should depend on the management objectives it is addressing. In our application, the management application requires an approach that is cost-effective, allows comparability of reactions to the product and marketing material about the product, is diagnostic in identifying the role of a range of different emotions, and which has scaling properties that enable it to be used in models of evaluation and choice. A Proposed Solution Bagozzi, Golpinath and Nyer (1999) call for more research on emotions measurement as a precursor to a better understanding of their role in the decision making process. Similarly, Pham (1998) sees a better understanding of the information as a useful bridge between “the overly cold literature on consumer decision making and the growing literature on hot consumer behavior.” In an academic sense we aim to provide an attempt at such a bridge. In a managerial sense, our objective is to develop a set of emotions, a stimulus with which to represent them, and a measurement instrument with which to calibrate their intensity for individual consumers when presented with specific marketing stimuli. Those measured emotions need to be sufficiently diagnostic to focus management communications and service design, but not so onerous on respondents or complex for managers as to be intractable. In terms of the specific emotions that we use, we are drawn to Laros and Steenkamp (2005)’s argument for an intermediate level of measurement as being on the efficient frontier of diagnosticity and parsimony. Given the testing and validation of their four positive and four negative emotions (and their demonstrated nested relationship to both finer and coarser classifications) we adopt their identified emotions of Happiness, Love, Pride, Contentment, Anger, Sadness, Anxiety, and Shame. In addition, we add a neutral valanced emotion, “Surprise.” Laros and Steenkamp exclude it not because they argue that it is not important, but rather because they wish to relate their emotions to the higher level classification previously discussed and positive/negative (or pleasurable/unpleasant). A number of authors have argued for the consideration of surprise (e.g., Roseman 1991) and thus we include it to bring our emotion inventory up to nine items. 11 Stimulus development for the affective measure Having selected the verbal representations of the emotions that we believe will be influential in choice, we proceed to develop stimuli that can represent them to respondents while provoking minimal cognitive processes in their interpretation or response to them. The literature on measurement, summarized above, suggested to us that the most appropriate representation of each emotion (what Slingerland 2005 calls a blend between the object being evaluated, in this case Kmart, and the respondent’s feelings) was a metaphor for the emotion in pictorial form. We call these visual representations “avatars” because they use a picture of a person experiencing the emotion to an extreme degree. We initially based our corresponding candidate metaphors and associated images on a review of established research. In particular, Kovecses (2000) describes the possible primary metaphors that have been found to relate to primary emotions. For example, the emotional experience of love is often correlated with a physical experience of warmth, which comprises a metaphorical concept which for most people becomes embedded within the neural pathways of the brain at a very early stage in life. However, we also conducted additional original research studies to identify suitable candidate metaphors and images. We first ran five focus groups in which respondents were asked to describe the image that came to their minds in relation to a list of emotions, including the nine identified emotions listed above. From a total of approximately 50 participants, a list of approximately 250 images was generated. We categorized these according to the metaphorical domains associated with the images. We found that similar underlying metaphorical themes were expressed across the images described by a range of different individuals for each targeted emotion. For example, the emotion of happiness was very commonly associated with “being up”, as well as with “warm and sunny” images. These observations led to the selection of a final list of three candidate images per emotion. We then tested three candidate images for each emotion to select a single final image for each emotion. The initial pilot test of these images involved 31 participants, with the goals of finding the most effective design format and identifying difficulties encountered in completing the required tasks, including ease of understanding of the concepts behind the candidate images. The test methodology was based on the Implicit Association Test (IAT) previously described (Greenwald, McGhee, and Schwartz 1998). We programmed the candidate images using the Inquisit™ software package from Millisecond Software™ (www.millisecond.com). The three candidates for each emotion were also tested in a further study, in which the objective was to use a reaction time task (RTT) or response latency to identify which of the three images would be most strongly associated with the targeted emotion (Bluemke and Friese 2008). Each respondent was required to complete both the RTT and a series of self-report questions. The study was conducted on line using a panel of 750 respondents, recruited by a digital data collection agency, GMI (www.gmi-mr.com). Each panel member was sent a random selection of five Inquisit™ test links, with each RTT having a quota of 150 respondents out of the total panel of 750. The data gathered during the study were validated and cleaned according to procedures set out for the IAT. In particular, trials with a reaction time of greater than 10 seconds were deleted, as were respondents for whom more than 10 percent of RTT trials exhibit a latency of less than 300 milliseconds. 12 The results were processed via an analysis of variance (ANOVA) model. For emotions where no avatar provided a strong and unambiguous representation, a combination of visual concepts drawn from groups with the lowest latency and highest self-report scores were used. Animated visual scales were then developed by which respondents could change the animations using a sliding scale operated with a computer mouse to reflect their feelings towards a given stimulus. We used Adobe™ Flash™/Shockwave™ to facilitate web-based delivery of the interactive visual scales. This resulted in 11 pictures for each avatar ranging from the neutral position all the way to extreme emotion. An example of three states of transition for the emotion of “anger” is provided in Figure 1. Neutral Mid range Extreme Figure 1: Example of three frames of a dynamic avatar representing ‘anger’ Measurement of emotions The dynamic avatars above in which the respondent sees different levels of the emotion as he or she moves the cursor across the bottom of the picture, mean that the measurement of respondents’ reaction to each one requires a minimum level of engagement to complete the task. When the cursor is in a position with which he or she feels comfortable, the respondent simply has to click the mouse which is moving the cursor. We believe, in keeping with the literature discussed above, that this form of elicitation reduces the chances of cognitive conditioning of responses. Validation and testing As part of the validation process the scales were included in six studies, across five industries with a database of approximately 4500 respondents. Results indicated that the animated, nonverbal scales were effective in accurately capturing respondents’ feelings. No significant differences were found between those who saw the feelings with labels versus those that had no labels revealed to them. This indicated the adequacy of the nine scales in capturing feelings towards brands. Nomological validity. In line with the previous discussion about the properties of the Semantic Animated Manikin, the question obviously arises as to the degree to which we have captured emotions directly and the degree to which respondents have thought about and then consciously recalled their emotions. (Note that the latter is not ruinous, and is the most popular form of eliciting emotions. However, the less measured emotions are confounded with cognitive appraisals, the 13 greater will be the ability to identify any incremental explanatory power that they have the potential to provide.) The most recognized way to test this is by the use of response latencies (e.g., Olofsson et al. 2008). The timing of pre-conscious processing is subject to debate. Greenwald, McGhee and Schwartz (1998) suggest that it does not begin until 300 milliseconds after the stimulus and to be largely complete in three seconds. As a part of this research, we also elicit perceptions of Kmart on attributes that respondents identify as being important to their choice of store, a cognitive measure. Pham et al. (2001) suggest that emotions should be able to be elicited faster than cognitions. We are thus able to compare the response latencies of the two sets of measures, which we report in Figure 2 below. An examination of the response latencies represented on the horizontal axis in milliseconds shows that by the end of three seconds (3000 milliseconds) 85% of feelings have been reported, while less than half of the cognitions have been. The median time for reporting feelings is well under 1 second, while for cognitions it is over three seconds. This gives us confidence that cognitive processing is at a minimum using this dynamic avatar approach. Figure 2: Comparison of response latencies (response times) for thought and feelings Convergent and discriminant validity. While the above figure supports the proposition that we have measured emotions, we do need to establish which emotions we have measured. We have already pointed out that there is not a perfect correspondence between emotions as felt and the words that we use for them. However, even with that imperfect correspondence, researchers still find it useful to attach certain words to specific affective reactions (and we have done that in selecting our nine emotions). Therefore, we need to determine whether the meaning that we have put on our nine dynamic avatars corresponds to that which respondents would when asked to identify the emotion that they thought that they were expressing when exposed to each avatar. To establish convergent and discriminant validity we asked 405 respondents to identify the nine feelings that they associated with each avatar (as well as the options of other or none). We eliminated respondents who did not move the cursor, though the substantial findings are not changed if they are included. This led to an average of 353 observations for each avatar. We undertook a principal components factor analysis 14 with varimax rotation on the resultant 9x9 matrix which enables us to see which emotions load on which avatar. The results are included in the Table 1 below. Convergent validity is clearly established. For no avatar is the primary factor loading less than 0.95 (demonstrated in bold in the above table) demonstrating a strong relationship between the avatar and the underlying emotion. Discriminant validity is also strong. In no case is the factor loading of an avatar more than 0.17 of an emotion with which it was not intended to be associated. This series of tests give us confidence that the basis of our measurement approach is valid and reliable (although the task still remains to demonstrate that it measures emotions that drive choice and that it provides managerial insights that will be helpful in focusing marketing activity). For a more detailed description of the development of this measurement instrument see Roberts, Wong and Stein (2011). FACTOR FEELING 1 2 3 4 5 6 7 8 9 1 PRIDE 0.02 -0.06 0.03 0.10 0.97 0.16 0.13 -0.09 0.04 2 SURPRISE 0.04 -0.05 0.02 0.12 0.16 0.96 0.11 -0.09 0.10 3 ANGER 0.06 -0.05 -0.05 0.13 0.13 0.11 0.96 -0.07 0.13 4 LOVE 0.07 0.02 0.04 0.97 0.10 0.11 0.12 0.03 0.13 5 HAPPY 0.11 0.07 0.17 0.13 0.04 0.10 0.13 0.05 0.95 6 SHAME 0.98 0.06 0.06 0.07 0.02 0.03 0.05 0.07 0.10 7 SADNESS 0.06 0.07 0.97 0.04 0.03 0.02 -0.05 0.15 0.16 8 CONTENTMENT 0.06 0.98 0.07 0.02 -0.06 -0.05 -0.05 0.13 0.06 9 ANXIETY 0.07 0.13 0.15 0.03 -0.09 -0.09 -0.07 0.96 0.05 Table 1: Convergent and discriminant validity of the dynamic avatars with respect to the emotions that they measure Representation (Modeling) Even if we are able to measure emotions, the question arises as to how to represent them in a model. Given the paucity of studies incorporating emotions into models of choice, it is unsurprising that the literature is largely silent on the functional form by which this should be achieved. Most of the modeling of the effect of emotions on intent uses structural equation modeling (with an associated assumption of linearity). However, there is reason for some optimism. Loewenstein (2000) suggests that emotions are “not only systematic, but amenable to formal modeling.” We can gain some idea from research, conceptual and empirical, on the psychological processes by which emotions are hypothesized to influence attitude and choice. Three major approaches emerge. First, much of the literature suggests that these are two separate processes and, as such, should be modeled separably. Zeelenberg and Pieters (2006) review the emotions literature and conclude that the most defensible conceptualization places cognitions (thoughts) and emotions (feelings) on the same level. He enters a plea for researchers to focus on the effects of emotions on behavior rather than their effect on 15 cognitions. Burke and Edell (1989) suggest that emotions will not be felt through cognitions. At the other extreme, considerable work had gone into testing whether cognitions fully mediate the effect of emotions on evaluation and choice. For example, Agarwal and Malhotra (2005) test the presence of an interaction term. However, full mediation does not attract a lot of support in the psychology literature on the subject. Finally, some middle ground is possible where cognitions may partially mediate the effect of emotions on evaluation, but there may be a direct effect too. For example MacKenzie, Lutz and Belch (1986) and Bodur, Brinberg and Coupey (2000) find both direct and indirect paths to occur. That suggests that the two most appropriate candidates to represent emotions (E) and cognitions (C) in a utility or brand attitude model (U) will be of the form: U = f(E) + g(C) + ε (1) or U = f’(E) + g’(C) + h(E*C) + υ (2) where f, g, f’, g’, and h are functions and ε and υ are error terms. Of course, pragmatic considerations such as multicollinearity may lead us to use the more parsimonious form in equation (1), since we would expect the three sets of independent variables in equation (2) to be highly correlated. Summary We have argued that incorporating emotions in models of evaluation and choice may both increase predictive ability and increase their diagnostic capability (Bodur, Brinberg and Coupey 2000). Indeed, it was a managerial need to create an emotional point of difference, with its accompanying need to calibrate the emotional impact of marketing activity that led us on the path to calibrate and model emotions. The application for which this note is written provides an example of the impact that can be realized by such an approach. However, while talking about emotions in general and their role in choice specifically, it is useful to consider how this approach might be used more broadly, possible limitations to its application, and how it might be developed further. Applications and Usefulness of Methodology We have identified two uses for this methodology. One is to conduct brand audits to understand the emotional (and cognitive) drivers of evaluation and choice and to monitor the position of the organization’s products on those key variables. That may obviously be extended to competitive monitoring (particularly in the case of new entrants). The second major use of the approach is regarding the effect of marketing activity. The most obvious (and most common in practice) is assessment of marketing mix elements, whether pre-launch or in the field for tracking and control. Perhaps a more interesting application is in marketing mix construction. Ideas for new product features, design and packaging, advertising and point of sale material can be systematically generated by using identified emotional reactions and needs as a source of concept seeding. The approach has the potential to be applied to all forms of marketing activity. Typical examples include the emotional fit of brand extensions, the viability of co-branding strategies, and congruence of push and pull strategies at different levels of the channel. Where any consumer evaluation of a company’s marketing strategies takes place, the literature suggests that more often than not these will have a significant emotional component. 16 Limitations There are of course potential limitations of our approach. Just as we identified situations in which emotions are likely to be strong drivers of consumer choice, so there will be situations where their consideration will not add significant insight (for example for utilitarian, high involvement products for which there is easily accessible objective and comparable information). We tested our emotions for relevance across a broad range of product categories. However, there may be specific product categories for which other emotions may be more relevant. For example, with a product whose perceived value includes a high reliance on the opinion of others, “social acceptability” may be a more insightful feeling than “pride.” Future extensions Our approach has already had considerable commercial impact across a range of industries and management applications. However, it may be refined with further experience in a number of ways. Some of these include measurement, modeling, and conceptual extensions. Measurement While the measurement methodology appears reasonably robust across applications, we have already used its affective measures combined with cognitive elicitation. This hybrid approach could be extended. For example, it would be possible to combine our affective measures with metaphor elicitation or fMRI to get Coulter and Zaltman (1994)’s “deep understanding” in the first case and dynamic frame-by-frame ad diagnostics in the second. Models Our modeling tests both main effects models and partial mediation models for the effect of emotions on brand attitude. There is a large variety of other models that may be relevant to different choice models. These include two phase choice models (e.g., Roberts and Lattin 1991), models of incoherence between emotions and cognitions (e.g., Kayande et al. 2007), and models with non-linearities and non-compensatory models (Swait 2001). Conceptual extensions The approach that we have adopted also lends itself to incorporate conceptual extensions. For example, Colemann and Williams (2013) suggest that the importance of emotions may be a function of the social identity of potential consumer. It may be possible to undertake emotional profiling of consumers and potential consumers to ensure a fit between the emotional messages being conveyed to store card holders (via, for example, customized catalogs) and their social identity. This would represent an affective equivalent to the highly successful behavioral targeting described for example by Humby and Hunt (2003) in their work for Tesco in the U.K. We have cited references (particularly in psychology and consumer behavior) that speak to the importance of emotions in choice. We have also noted a lack of such considerations in marketing science choice models, in part because of the methodological difficulties in calibrating emotions. Given a management problem that required such an approach, we have developed a way to overcome these barriers and demonstrated its effectiveness in focusing marketing activities with a substantial impact on the bottom line of the organization involved. 17 References Aaker, David A., Vineet Kumar, and George S. Day. Marketing research. John Wiley & Sons, 2008. Aaker, David A., Douglas M. Stayman, and Michael R. Hagerty. "Warmth in advertising: Measurement, impact, and sequence effects." Journal of Consumer Research (1986): 365-381. Aaker, David A., Douglas M. Stayman, and Richard Vezina. "Identifying feelings elicited by advertising." Psychology & Marketing 5, no. 1 (1988): 1-16. Adkins, Natalie Ross, and Julie L. Ozanne. "The low literate consumer." Journal of Consumer Research 32, no. 1 (2005): 93-105. Advertising Research Foundation NeuroStandards Project Whiter Paper (2011) New York: Advertising Research Foundation Agarwal, James, and Naresh K. Malhotra. "An integrated model of attitude and affect: theoretical foundation and an empirical investigation." Journal of Business research 58, no. 4 (2005): 483-493. Allen, Chris T., Karen A. Machleit, and Susan Schultz Kleine. "A comparison of attitudes and emotions as predictors of behavior at diverse levels of behavioral experience." Journal of consumer research (1992): 493-504. Andrade, Eduardo B. "Behavioral consequences of affect: Combining evaluative and regulatory mechanisms." Journal of Consumer Research 32, no. 3 (2005): 355-362. Angie, Amanda D., Shane Connelly, Ethan P. Waples, and Vykinta Kligyte. "The influence of discrete emotions on judgement and decision-making: A meta-analytic review." Cognition & Emotion 25, no. 8 (2011): 1393-1422. Avnet, Tamar, Michel Tuan Pham, and Andrew T. Stephen. "Consumers’ Trust in Feelings as Information." Journal of Consumer Research 39, no. 4 (2012): 720-735. Bagozzi, Richard P., Mahesh Gopinath, and Prashanth U. Nyer. "The role of emotions in marketing." Journal of the Academy of Marketing Science 27, no. 2 (1999): 184-206. Bakamitsos, Georgios A. "A cue alone or a probe to think? The dual role of affect in product evaluations." Journal of Consumer Research 33, no. 3 (2006): 403-412. Batra, Rajeev, and Morris B. Holbrook. "Developing a typology of affective responses to advertising." Psychology & Marketing 7, no. 1 (1990): 11-25. Batra, Rajeev, and Michael L. Ray. "Affective responses mediating acceptance of advertising." Journal of consumer research (1986): 234-249. Baumgartner, Hans, Mita Sujan, and Dan Padgett. "Patterns of affective reactions to advertisements: The integration of moment-to-moment responses into overall judgments." Journal of Marketing Research (1997): 219-232. Bechara, Antoine, and Antonio R. Damasio. "The somatic marker hypothesis: A neural theory of economic decision." Games and economic behavior 52, no. 2 (2005): 336-372. Berkowitz, Leonard. "Towards a general theory of anger and emotional aggression: Implications of the cognitive-neoassociationistic perspective for the analysis of anger and other emotions" in Advances in Social Cognition 6 (1993). 18 Bluemke, Matthias, and Malte Friese. "Reliability and validity of the Single‐Target IAT (ST‐IAT): assessing automatic affect towards multiple attitude objects." European Journal of Social Psychology 38, no. 6 (2008): 977-997. Bodur, Onur H., David Brinberg, and Eloise Coupey. "Belief, affect, and attitude: Alternative models of the determinants of attitude." Journal of Consumer Psychology 9, no. 1 (2000): 17-28. Bradley, Margaret M., and Peter J. Lang. "Measuring emotion: the self-assessment manikin and the semantic differential." Journal of behavior therapy and experimental psychiatry 25, no. 1 (1994): 4959. Burke, Marian Chapman, and Julie A. Edell. "The impact of feelings on ad-based affect and cognition." Journal of Marketing Research (1989): 69-83. Cacioppo, John T., Richard E. Petty, Jeffrey A. Feinstein, and W. Blair G. Jarvis. "Dispositional differences in cognitive motivation: The life and times of individuals varying in need for cognition." Psychological Bulletin 119, no. 2 (1996): 197. Chuang, Shih-Chieh. "Sadder but wiser or happier and smarter? A demonstration of judgment and decision making." The Journal of psychology 141, no. 1 (2007): 63-76. Cohen, Joel B., Michel T. Pham, and Eduardo B. Andrade. "The nature and role of affect in consumer behavior." Handbook of consumer psychology (2008): 297-348. Coleman, Nicole V. and Patti Williams Feeling like My Self: Emotion Profiles and Social Identity” Journal of Consumer Research (2013) 40 (August) 203-222 Coulter, Robin Higie and Gerald Zaltman "Using the Zaltman metaphor elicitation technique to understand brand images." Advances in Consumer Research 21 (1994): 501-507. Damasio, Antonio. Descartes' error: Emotion, reason, and the human brain. Penguin. com, 2005. Desmet, Peter MA, Paul Hekkert, and Jan J. Jacobs. "When a car makes you smile: Development and application of an instrument to measure product emotions." Advances in consumer research 27 (2000): 111-117. Donovan, Robert J., and John R. Rossiter. "Store atmosphere: an environmental psychology approach." Journal of retailing 58, no. 1 (1982): 34-57. Donovan, Robert J., John R. Rossiter, Gilian Marcoolyn, and Andrew Nesdale. "Store atmosphere and purchasing behavior." Journal of retailing 70, no. 3 (1994): 283-294. Dubé, Laurette, Amitava Chattopadhyay, and Anick Letarte. "Should advertising appeals match the basis of consumers' attitudes?" Journal of Advertising Research 36, no. 6 (1996): 82-89. Eckler, Petya, and Paul Bolls. "Spreading the virus: Emotional tone of viral advertising and its effect on forwarding intentions and attitudes." Journal of Interactive Advertising 11, no. 2 (2011): 1-11. Edell, Julie A., and Marian Chapman Burke. "The power of feelings in understanding advertising effects." Journal of Consumer research (1987): 421-433. Ekman, Paul, and Wallace V. Friesen. "Constants across cultures in the face and emotion." Journal of personality and social psychology 17, no. 2 (1971): 124. Elster, Jon. "Emotions and economic theory." Journal of economic literature 36, no. 1 (1998): 47-74. 19 Erevelles, Sunil. "The role of affect in marketing." Journal of Business Research 42, no. 3 (1998): 199215. Fitzsimons, Gráinne M., Tanya L. Chartrand, and Gavan J. Fitzsimons. "Automatic effects of brand exposure on motivated behavior: how apple makes you “think different”." Journal of Consumer Research 35, no. 1 (2008): 21-35. Fournier, Susan. "Consumers and their brands: developing relationship theory in consumer research." Journal of consumer research 24, no. 4 (1998): 343-353. Greenwald, Anthony G., Debbie E. McGhee, and Jordan LK Schwartz. "Measuring individual differences in implicit cognition: the implicit association test." Journal of personality and social psychology 74, no. 6 (1998): 1464. Griskevicius, Vladas, Michelle N. Shiota, and Stephen M. Nowlis. "The Many Shades of Rose‐Colored Glasses: An Evolutionary Approach to the Influence of Different Positive Emotions." Journal of consumer research 37, no. 2 (2010): 238-250. Guadagni, Peter M., and John DC Little. "A logit model of brand choice calibrated on scanner data." Marketing science 2, no. 3 (1983): 203-238. Hanoch, Yaniv. "“Neither an angel nor an ant”: Emotion as an aid to bounded rationality." Journal of Economic Psychology 23.1 (2002): 1-25. Heath, Robert, David Brandt, and Agnes Nairn. "Brand relationships: Strengthened by emotion, weakened by attention." Journal of Advertising Research-New York- 46, no. 4 (2006): 410. Herr, Paul M., Christine M. Page, Bruce E. Pfeiffer, and Derick F. Davis. "Affective influences on evaluative processing." Journal of Consumer Research 38, no. 5 (2012): 833-845. Hilgard, Ernest R. "The trilogy of mind: Cognition, affection, and conation." Journal of the History of the Behavioral Sciences 16, no. 2 (1980): 107-117. Holbrook, Morris B., and Rajeev Batra. "Assessing the role of emotions as mediators of consumer responses to advertising." Journal of consumer research (1987): 404-420. Howard, Daniel J., and Charles Gengler. "Emotional contagion effects on product attitudes." Journal of Consumer Research 28, no. 2 (2001): 189-201. Humby, Clive, and Terry Hunt. Scoring Points: How Tesco is winning customer loyalty. Buy now from Kogan Page, 2003 Johnson, Allison R. and Stewart, David W. "A reappraisal of the role of emotion in consumer behavior: traditional and contemporary approaches." Review of marketing research 1 (2010): 3-34. Kahneman, Daniel. "Maps of bounded rationality: Psychology for behavioral economics." The American economic review 93, no. 5 (2003): 1449-1475. Kahneman, Daniel, Peter P. Wakker, and Rakesh Sarin. "Back to Bentham? Explorations of experienced utility." The Quarterly Journal of Economics 112, no. 2 (1997): 375-406. Kayande, Ujwal, John H. Roberts, Gary L. Lilien, and Duncan KH Fong. "Mapping the bounds of incoherence: How far can you go and how does it affect your brand?." Marketing Science 26, no. 4 (2007): 504-513. Keil, Andreas, Thomas Gruber, Matthias M. Müller, Stephan Moratti, Margarita Stolarova, Margaret M. Bradley, and Peter J. Lang. "Early modulation of visual perception by emotional arousal: evidence 20 from steady-state visual evoked brain potentials." Cognitive, Affective, & Behavioral Neuroscience 3, no. 3 (2003): 195-206. Keller, Kevin Lane. "Brand synthesis: The multidimensionality of brand knowledge." Journal of consumer research 29, no. 4 (2003): 595-600. Kim, Jooyoung, and Jon D. Morris. "The power of affective response and cognitive structure in product-trial attitude formation." Journal of Advertising 36, no. 1 (2007): 95-106. King, Dan, and Chris Janiszewski. "Affect-Gating." Journal of Consumer Research 38, no. 4 (2011): 697-711. Kong, Wanzeng, Xinxin Zhao, Sanqing Hu, Giovanni Vecchiato, and Fabio Babiloni. "Electronic evaluation for video commercials by impression index." Cognitive Neurodynamics (2013): 1-5. Kovecses, Z. (2000) Metaphor and Emotion – Language, Culture, and Body in Human Feeling, Cambridge University Press, UK Ladhari, Riadh. "The effect of consumption emotions on satisfaction and word‐of‐mouth communications." Psychology & Marketing 24, no. 12 (2007): 1085-1108. Laros, Fleur JM, and Jan-Benedict EM Steenkamp. "Emotions in consumer behavior: a hierarchical approach." Journal of business Research 58, no. 10 (2005): 1437-1445. Lazarus, Richard S. "Thoughts on the relations between emotion and cognition." American Psychologist 37, no. 9 (1982): 1019-24. Ledoux, Joseph E. "Cognitive-emotional interactions in the brain." Cognition & Emotion 3, no. 4 (1989): 267-289. Ledoux, Joseph E. "Emotion: Clues from the brain." Annual review of psychology 46, no. 1 (1995): 209-235. Lerner, Jennifer S., and Dacher Keltner. "Beyond valence: Toward a model of emotion-specific influences on judgement and choice." Cognition & Emotion 14, no. 4 (2000): 473-493. Loewenstein, George. "Out of control: Visceral influences on behavior." Organizational behavior and human decision processes 65, no. 3 (1996): 272-292. Loewenstein, George. "Emotions in economic theory and economic behavior." The American Economic Review 90, no. 2 (2000): 426-432. Lutz, Richard J. "Affective and cognitive antecedents of attitude toward the ad: A conceptual framework." Psychological processes and advertising effects (1985): 45-63. Maclnnis, Deborah J., and Bernard J. Jaworski. "Information processing from advertisements: Toward an integrative framework." The Journal of marketing (1989): 1-23. MacKenzie, Scott B., Richard J. Lutz, and George E. Belch. "The role of attitude toward the ad as a mediator of advertising effectiveness: A test of competing explanations." Journal of marketing research (1986): 130-143. Mauss, Iris B., and Michael D. Robinson. "Measures of emotion: A review." Cognition and emotion 23, no. 2 (2009): 209-237. Mellers, Barbara, Alan Schwartz, and Ilana Ritov. "Emotion-based choice." Journal of Experimental Psychology: General 128, no. 3 (1999): 332. 21 Mittal, Banwari. "The role of affective choice mode in the consumer purchase of expressive products." Journal of Economic Psychology 9, no. 4 (1988): 499-524. Morris, Jon D., Chongmoo Woo, James A. Geason, and Jooyoung Kim. "The power of affect: Predicting intention." Journal of Advertising Research 42, no. 3 (2002): 7-18. Naqvi, Nasir, Baba Shiv, and Antoine Bechara. "The role of emotion in decision making a cognitive neuroscience perspective." Current Directions in Psychological Science 15, no. 5 (2006): 260-264. Ohme, Rafal, Dorota Reykowska, Dawid Wiener, and Anna Choromanska. "Analysis of neurophysiological reactions to advertising stimuli by means of EEG and galvanic skin response measures." Journal of Neuroscience, Psychology, and Economics 2, no. 1 (2009): 21. Oliver, Richard L. "Cognitive, affective, and attribute bases of the satisfaction response." Journal of consumer research (1993): 418-430. Olney, Thomas J., Morris B. Holbrook, and Rajeev Batra. "Consumer responses to advertising: The effects of ad content, emotions, and attitude toward the ad on viewing time." Journal of Consumer Research (1991): 440-453. Olofsson, Jonas K., Steven Nordin, Henrique Sequeira, and John Polich. "Affective picture processing: an integrative review of ERP findings." Biological psychology 77, no. 3 (2008): 247-265. Pham, Michel Tuan. "Representativeness, relevance, and the use of feelings in decision making." Journal of consumer research 25, no. 2 (1998): 144-159. Pham, Michel Tuan, Joel B. Cohen, John W. Pracejus, and G. David Hughes. "Affect monitoring and the primacy of feelings in judgment." Journal of consumer research 28, no. 2 (2001): 167-188. Pham, Michel Tuan, Maggie Geuens, and Patrick De Pelsmacker. "The Influence of Ad-Evoked Feelings on Brand Evaluations: Empirical Generalizations from Consumer Responses to More Than 1,000 TV Commercials." International Journal of Research in Marketing (2013), 383-94 Phillips, Diane M., and Hans Baumgartner. "The role of consumption emotions in the satisfaction response." Journal of Consumer Psychology 12, no. 3 (2002): 243-252. Poels, Karolien, and Siegfried Dewitte. "How to capture the heart? Reviewing 20 years of emotion measurement in advertising." (2006). Richins, Marsha L. "Measuring emotions in the consumption experience." Journal of consumer research 24, no. 2 (1997): 127-146. Rick, Scott and George Loewenstein The role of emotion in economic behavior” in Lewis, HavilandJones and Barrett, eds., Handbook of Emotions Third Edition Guildford, U.K.: Guildford Press 138-56 Roberts, John H., and James M. Lattin. "Development and testing of a model of consideration set composition." Journal of Marketing Research (1991): 429-440 Roberts, Ken, Elaine Wong and Darren Stein “Feelings, Nothing More than Feelings: Rules for Emotion-based Communication” Proceedings of the National Conference (2011), Sydney: Australian Market and Social Research Society Roseman, Ira J. "Appraisal determinants of discrete emotions." Cognition & Emotion 5, no. 3 (1991): 161-200. 22 Russell, James A. “A Circumplex Model of Affect” Journal of Personality and Social Psychology 39 (1980) 1161-78 Russell, James A. "Is there universal recognition of emotion from facial expressions? A review of the cross-cultural studies." Psychological bulletin 115, no. 1 (1994): 102. Russell, James A. "Core affect and the psychological construction of emotion." Psychological review 110, no. 1 (2003): 145. Ruth, Julie A. "Promoting a brand's emotion benefits: the influence of emotion categorization processes on consumer evaluations." Journal of Consumer Psychology 11, no. 2 (2001): 99-113. Schwarz, Norbert, and Gerald L. Clore. "Mood, misattribution, and judgments of well-being: Informative and directive functions of affective states." Journal of personality and social psychology 45, no. 3 (1983): 513-23 Sherman, Elaine, Anil Mathur, and Ruth Belk Smith. "Store environment and consumer purchase behavior: mediating role of consumer emotions." Psychology & Marketing 14, no. 4 (1997): 361-378. Shiv, Baba, and Alexander Fedorikhin. "Heart and mind in conflict: The interplay of affect and cognition in consumer decision making." Journal of consumer Research 26, no. 3 (1999): 278-292. Slingerland, Edward G. "Conceptual blending, somatic marking, and normativity: a case example from ancient Chinese." Cognitive Linguistics 16, no. 3 (2005): 557-584. Slovic, Paul, Melissa L. Finucane, Ellen Peters, and Donald G. MacGregor. "Risk as analysis and risk as feelings: Some thoughts about affect, reason, risk, and rationality." Risk analysis 24, no. 2 (2004): 311-322. Smith, Stephen M., Curtis P. Haugtvedt, and Richard E. Petty. "Attitudes and recycling: does the measurement of affect enhance behavioral prediction?." Psychology & Marketing 11, no. 4 (1994): 359-374. Stayman, Douglas M., and David A. Aaker. "Are all the Effects of Ad-Induced Feelings Mediated by A ad?." Journal of Consumer Research (1988): 368-373. Swait, Joffre. "A non-compensatory choice model incorporating attribute cutoffs." Transportation Research Part B: Methodological 35, no. 10 (2001): 903-928. Thaler, Richard H. "From homo economicus to homo sapiens." The Journal of Economic Perspectives 14, no. 1 (2000): 133-141. Thomson, Matthew, Deborah J. MacInnis, and C. Whan Park. "The ties that bind: Measuring the strength of consumers’ emotional attachments to brands." Journal of Consumer Psychology 15, no. 1 (2005): 77-91. Voss, Kevin E., Eric R. Spangenberg, and Bianca Grohmann. "Measuring the hedonic and utilitarian dimensions of consumer attitude." Journal of marketing research (2003): 310-320. Watson, David, Lee A. Clark, and Auke Tellegen. "Development and validation of brief measures of positive and negative affect: the PANAS scales." Journal of personality and social psychology 54, no. 6 (1988): 1063. Watson, David, Lee A. Clark, and Auke Tellegen. "Development and validation of brief measures of positive and negative affect: the PANAS scales." Journal of personality and social psychology 54, no. 6 (1988): 1063. 23 Williams, Patti “Emotions and Consumer Behavior” Journal of Consumer Research 40 February (2014) pp. viii-xi Williams, Patti, and Jennifer L. Aaker. "Can mixed emotions peacefully coexist?" Journal of Consumer Research 28, no. 4 (2002): 636-649. Zajonc, Robert B. "Feeling and thinking: Preferences need no inferences." American psychologist 35, no. 2 (1980): 151-75. Zajonc, Robert B. "On the primacy of affect." American Psychologist February (1984): 117-23. Zaltman, Gerald, and Robin Higie Coulter. "Seeing the voice of the customer: Metaphor-based advertising research." Journal of advertising research 35, no. 4 (1995): 35-51. Zeelenberg, Marcel and Rik Pieters “Feeling is for Doing: A Pragmatic Approach to the Study of Emotions in Economic Behavior” in De Cremer, David, Marcel Zeelenberg, and J. Keith Murnighan, eds. Social psychology and economics (2006) Mahwah, NJ: Lawrence Erlbaum Associates, 117-37 24
© Copyright 2025 Paperzz