From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research

FRIN-05768; No of Pages 12 Food Research International xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Food Research International journal homepage: www.elsevier.com/locate/foodres

Review

From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research Egon P. Köster a,⁎, Jozina Mojet b a b

Helmholtz Institute, Utrecht University, Heidelberglaan 6, 3508 TC Utrecht, The Netherlands Wageningen University and Research, Food & Biobased Research, Bornse Weilanden 9, 6708 PD Wageningen, The Netherlands

a r t i c l e

i n f o

Article history: Received 15 December 2014 Received in revised form 29 March 2015 Accepted 2 April 2015 Available online xxxx Keywords: Emotion Eating and drinking behaviour Emotion measurement Explicit vs. implicit methods

a b s t r a c t The bi-directional influences between emotion and food consumption are discussed in view of recent efforts to find emotional factors that influence food choice and eating- and drinking behaviour independently from traditional factors as liking, wanting and appropriateness. Distinctions are made between conscious and unconscious emotions and their relative importance in food-related behaviour is discussed. In response to eating disorders like obesity, much more is known about the influence of emotion and mood on food choice and intake than about the influence of food on mood and emotion, which only recently gained prominence in food-related emotion research. This led to a number of emotion measurement methods that differ strongly in their explicit or implicit measurement approach and in the extent to which they demand conscious emotion awareness and verbal understanding on the part of the participants. These methods are critically discussed and questions are raised about the specificity of their emotional contents and about their use at different moments in time, such as before, during and at different moments after consumption. Furthermore, doubts were raised about the independency of their contributions from the traditional measurements (liking, wanting and appropriateness) and suggestions are made for improving the practical applicability of an efficient emotion measurement. © 2015 Elsevier Ltd. All rights reserved.

Contents 1. 2. 3. 4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mood and emotion: some definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . The effects of emotion and mood on food choice, food intake and food appreciation or satisfaction Effects of food on emotions and mood . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Effects of expectations and memory on emotions and mood . . . . . . . . . . . . . . 5. Measurement of the emotional effects of food . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Explicit emotion measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Non-verbal explicit emotion measurement . . . . . . . . . . . . . . . . . . . . . . 5.3. Implicit emotion measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Sensory versus remembered and prospected emotions. . . . . . . . . . . . . . . . . 6. General conclusions and practical suggestions . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Food and emotion are linked in a number of different ways. On the one hand there is the influence of mood and emotion on food choice ⁎ Corresponding author at: Wildforsterweg 4a, 3881NJ Putten, The Netherlands. Tel.: +31 577461716, +31 650632672 (mob). E-mail address: [email protected] (E.P. Köster).

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

0 0 0 0 0 0 0 0 0 0 0 0

and intake and on the other hand consuming food can have an influence on people's mood and feelings. The first of these two relationships has been investigated quite extensively in other reviews (Canetti, Bachar, & Berry, 2002; Gibson, 2006a,b; Macht, 2008), whereas the second one has only recently gained more attention as a possible aspect that might play an important role in food choice, independent of traditional factors such as liking and wanting (Berridge & Winkielman, 2003; Meiselman, 2013). Both, the influence of mood on eating behaviour

http://dx.doi.org/10.1016/j.foodres.2015.04.006 0963-9969/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

2

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

and the influence of food on mood and emotions are very complex relationships involving physiological factors such as hunger, satiation and physiological reward mechanisms, psychological factors such as age, expectations based on previous experiences, memory and habit formation, emotional coping mechanisms, restrained eating tendencies or personality traits such as manic–depressive tendencies and sociological factors such as economic status, social nature of the eating situation or eating culture. To cope with this multitude of factors seems almost impossible and like in other studies dealing with eating behaviour, most investigators have neglected this diversity and used reductionist approaches focussing on only one or a few of these aspects (Köster, 2009). Segmentation of the experimental subjects on the basis of age, sex, education and economic status and on traits like food neophobia has sometimes been taken into account, but situational factors such as eating alone in the kitchen or in front of the TV or with family and friends, which are of prime importance in both the effects of mood on food and of food on mood are almost never considered. Memories of rather personal situations evoked by eating experiences may also be more important than those related to the nature of the food itself and may provide strong links to emotional early childhood experiences, or special occasions later in life. Nevertheless, for such memories to arise it is necessary that the food experience is sufficiently similar to the original experience. Even slight changes in a common food, such as the addition of another spice or a change in texture, will readily be consciously detected and lead to pleasant surprise or may disturb the intimate memories and evoke feelings of disappointment and dissatisfaction, instead of the upcoming pleasant reminiscences raised when the food is conform to our expectations (Köster et al., 2014). Strangely enough the role of memory is almost always neglected in food-related consumer research, although it is probably much more important than the first impression experiences that are commonly investigated. The emotions, evoked by remembering a product, are essential in the expectations that guide repurchase decisions. The neglect of memory in combination with the superficiality of first impression measurements is probably the main reason for the high volatility of the market and the high flop rate of new food products (Köster & Mojet, 2007, 2012a,b). 2. Mood and emotion: some definitions Mood and emotion both reflect emotional states and are often used interchangeably in common language. In the present paper they will be distinguished according to the principles indicated by Gibson (2006a) in his overview of the sensory, physiological and psychological factors determining the role emotions play in food choice. He defines emotions as “short-term affective responses to the appraisal of particular stimuli, having reinforcement potential” (Frijda, 1999; Mathews & Deary, 1998; Rolls, 1999), whereas moods are described as “more longlasting psychological arousal states with interacting dimensions related to energy, tension and pleasure (hedonic tone) that may appear and persist in the absence of obvious stimuli and may be more covert to observers” (Thayer, 1989). Thus, pleasant mood is related to high energy and low tension and negative unpleasant mood is related to low energy and high tension. As indicated by Gibson the relationships between mood, emotions and physiological arousal may be complex, since they depend on a number of the factors mentioned above in the Introduction. Another important general point to be discussed here is the question of the conscious or unconscious nature of emotion, a topic that has been subject of serious debate in emotion theory. Berridge and Winkielman (2003) discussed the traditional view that emotions are defined as the conscious subjective experiences that accompany the affective states created by bodily sensations (Clore, 1994; Ellsworth, 1994; Frijda, 1999). They show that Ellsworth in later publications (Ellsworth 1995; Ellsworth & Scherer, 2003) mitigated his position somewhat leaving the question of the existence of unconscious emotions open. Kihlstrom, Mulvany, Tobias, and Tobis (2000) argued for the existence

of an emotional unconscious in analogy to other psychological processes that take place without conscious awareness such as is the case in implicit memory. Thus he made a distinction between explicit emotion as a person's conscious awareness of an emotion and implicit emotion as an emotional state that expresses itself in experiences, thoughts or actions without conscious awareness of that state by the person. Along with Kihlstrom et al. (2000), Berridge and Winkielman come to the conclusion that “for an emotion to be unconscious, people must not be able to report their emotional reaction at the moment it is caused. Yet there must be clear evidence of the emotional reaction in their behaviour, or physiological response or subsequent subjective impression of an affect-laden event.” They illustrate this with two experiments (Winkielman, Berridge, & Wilbarger, 2005) in which people were subliminally (and unconsciously) exposed to happy, neutral or angry faces that were immediately followed by a second supraliminal masking photograph of a neutral face that they could perceive consciously. People then rated their own subjective emotion and were asked to pour themselves as much of a fruit-flavoured drink as they wanted and to drink and evaluate it. The results showed that the exposure to the subliminally presented facial expressions influenced the people's pouring and drinking behaviour, but only when they were thirsty. In that case they poured and drank about 50% more after subliminal presentation of a happy face than after a neutral face and less than after a neutral face when they had been unconsciously exposed to a neutral face. Nevertheless the thirsty participants reported no conscious awareness of any change in their subjective emotion. Non-thirsty people showed no effect at all. These results were reconfirmed in a second, more extensive experiment along the same lines and with the same results, except for the fact that this time the non-thirsty subjects who did not change their pouring and drinking behaviour reported some impact on their subjective emotion ratings, whereas the thirsty subjects who poured and drank depending on the exposure did not report conscious emotional changes. Winkielman, Berridge, and Scher (2011) conclude that these experiments demonstrate the existence of unconscious affective reactions of which the person is simply not aware, even when explicitly asked to report on their conscious mental state. The fact that Winkielman et al. chose drinking behaviour as the means of verifying the effect of the unconscious affective manipulation is interesting in the context of the present paper, because it illustrates both the close relationship between emotion and food or drink intake and the complexity of the reactions due to the physiological state (thirsty or non-thirsty) of the participants. The possible specificity of the relationship between food and emotion is furthermore illustrated by the fact that Monahan, Murphy, and Zajonc (2000) using mere exposure to subliminal visual stimuli (Chinese ideographs) reliably found conscious mood effects, whereas the thirsty participants in Winkielman et al. did not. This might not only be due to procedural differences or to differences in the nature of the subliminal stimuli (emotional faces vs Chinese ideographs), but might above all be related to the fact that in the thirsty subjects the subliminally raised affect could be coupled directly to an action that implicated ingestion of an unknown drink. In the non-thirsty subjects the need to drink was not present and as a result the effect of the subliminally presented faces was translated in the awareness of a slight mood change. The fact that olfaction, as the only sense with a direct connection to the amygdala, was involved in the ingestion, may also have played a role. Whalen et al. (1998), using also subliminal presentation of angry and happy faces, showed that this resulted in strong stimulation effects in the amygdala without any conscious awareness of the faces or their emotional effects. It is well known that the amygdala plays an important role in fear conditioning (Aggleton, 1992; Li, 2014) and in the automatic and unconscious processing of emotional facial expressions (Adolphs, Tranel, Damasio, & Damasio, 1995; Cahill et al., 1996; Reimann, Lane, Ahorn, & Schwarz, 1997). Thus, it might be that the subliminal angry or happy faces in the experiment of Winkielman et al. provoked an automatic caution or thirst reaction in the thirsty subjects, who planned to drink a

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

substantial amount, whereas in the non-thirsty subjects who had only the intention to taste the drink, conscious attention was directed at the hedonic effect of the drink rather than at enjoyment or caution. In any case, it is clear that much about unconscious emotions is still unknown, but that there is no doubt that they can play a role in eating and drinking behaviour and one that is rather independent from hedonic pleasure as measured by liking. In the present review, we will therefore follow the criteria defined by Kihlstrom et al. (2000) and speak of unconscious or implicit emotions when people are not able to report their emotional reaction at the moment it is caused, but show clear evidence of the emotional reaction either in their behaviour or physiological response or subsequent subjective impressions of an affect-laden event. Implicit emotions are therefore almost always measured indirectly. Explicit emotions on the other hand are feelings and moods that are consciously perceived as such and that may be directly measured and described, even if it is sometimes very hard to do so. The distinction of the two forms of emotion has direct consequences for the approach taken in assessing them. As we shall see below, most investigators measure only explicit emotion using questionnaires and rating-scales. Although such methods can certainly provide much information about the experienced effects of food, they may miss essential changes in feelings and mood that, although they escape conscious awareness, may influence later behaviour with respect to the particular food or general social behaviours of different kinds. Other methods that rely less on explicit verbalisation of feelings have also been developed and will be discussed below under measurement of emotional effects of food. Another aspect of the methods discussed there, is the degree of specificity that is used in food-related emotion measurement, which varies from just positive or negative effects to a wide and detailed gamma of emotions. 3. The effects of emotion and mood on food choice, food intake and food appreciation or satisfaction The strong interest in food-related disorders like obesity and anorexia has also raised the interest in emotions as the possible causes of the food choice and food intake behaviours that are connected with these disorders. Thus, Canetti et al. (2002) discussed the influence of emotion on food intake and the psychological and emotional effects of dieting and losing weight raised by restraining food intake. They come to the conclusion that in both normal and overweight people negative and also positive emotions (although to a lesser degree and not in all cases) increase food intake and that these effects are stronger in obese people and dieters than in non-dieters. According to Canetti et al., all of this is in line with existent theories like the psychosomatic theories of obesity like Kaplan and Kaplan's (1957) claim that obese people overeat because overeating reduces their anxiety or Bruch's (1973) theory which proposes that people overeat because they have not properly learned to recognise when they are hungry or satiated and need signals from outside to know when to eat at and how much. Emotional tension and unpleasant feelings will drive such a person to overeating. According to Canetti et al., both these theories seem to predict intake behaviour better than the theory of Schachter (1968, 1971); Schachter, Goldman, and Gordon (1968), who supposed that stressed normal weight people reacted with an increase or a decrease of food intake to physiological signs of hunger while obese people showed no such effect of internal states, but reacted only on the sight of food. Their original findings (Schachter et al., 1968) could not be confirmed by others however. Canetti et al., also discuss the influence of emotion in the light of the ‘restraint hypothesis’ (Herman & Mack, 1975; Herman & Polivy, 1980) which is based on the idea that restraint is the cognitive effort to resist the desire for food. This cognitive effort may be reduced by emotions and as a result restrained eaters may increase their food intake. Gibson (2006a) in the part of his paper on the influence of stress on eating behaviour, discusses the physiological mechanisms involved in this relationship in more detail and explains why especially sweet fatty

3

foods might help to alleviate stress. He devotes a section of his paper to the respective influences of carbohydrate and proteins on the tryptophan ratio in the brain and the effects of such changes on mood and arousal. Thus, carbohydrates will make people more calm and sleepy than protein-rich meals (Benton, 2002), but the extent of these effects depends also on individual differences in susceptibility to nutritional effects on mood, emotion and aspects of brain function (Gibson, 2006b; Gibson & Green, 2002). Variability across individuals is also an important theme in the review of Macht (2008). An overview of a large number of studies showed that “restrained eaters increase food intake, emotional eaters consume more sweet high-fat foods and binge eaters tend to binge in response to negative emotions”. Furthermore, he reports that normal eaters react to negative emotion in about equal proportions with an increase (43%) and with a decrease (39%) in food intake, while 26% do not change their intake at all according to Bellisle et al. (1990). Another source of variability in the effects on eating behaviour mentioned by Macht, 2008 is variability across emotions that differ in valence, arousal or intensity. He concludes that according to the earlier literature, higharousal and intense emotions suppress eating and that negative emotions can either increase or decrease food intake, whereas there is no clear picture of the effects of positive emotions on eating. Also little is known about the differences between the effects of negative emotions like anger, sadness and fear. Macht (2008) proposes a model describing five classes of emotioninduced changes of eating ‘that can be predicted by antecedent conditions and point to basic functional principles relating emotions and eating’. Thus: 1. Food-induced emotions like those provoked by energy-dense food can elicit craving which is accompanied by autonomic responses (Nederkoorn, Smulders, & Jansen, 2000, Nederkoorn, Smulders, Havermans, & Jansen, 2004) and can lead to binge eating (Jansen, 1998) 2. Intense emotions or chronic stress suppress food intake (Greeno & Wing, 1994) 3. Both negative and positive emotions impair cognitive eating control and increase food intake. This may be due either to a breaking down of the dieter's self-imposed restrictions in the presence of a more urgent concern (the stress) (Herman & Polivy, 1984) or to the attention demand created by the processing of the emotional stimuli which limits people's cognitive capacity to maintain their food intake restriction (Boon, Stroebe, Schut, & Jansen, 1998, Boon, Stroebe, Schut, & IJntema, 2002; Lattimore & Caswell, 2004; Lattimore & Maxwell, 2004; Vreugdenberg, Bryan, & Kemps, 2003; Wallis & Hetherington, 2004; Ward & Mann, 2000). Most of these studies suggest that the more people exert cognitive control over their intake the more vulnerable they are to overeating. 4) Negative emotions elicit eating to regulate emotions, using eating as a coping mechanism (Bruch, 1973; Kaplan & Kaplan, 1957; Slochower, 1983). Macht (1999) discusses the serotonin hypothesis stating that carbohydrate-rich meals lead to mood improvement via a meal-induced increase of brain serotonin (Benton & Donohoe, 1999; Gibson & Green, 2002; Markus et al., 1998) and the endocrinal hypothesis that supposes that ingestion of high fat and carbohydrate foods reduces Hypothalamo–pituitary adrenal activity and thus dampens stress response. Macht (1999) points out that both these effects can only occur with a delay and that in coping with negative emotions immediate reduction of their intensity would be desirable. Immediate positive affective reactions to palatable foods are the most important factor in the reduction of stress impact (Macht & Müller, 2007). 5) Emotions modulate eating in congruence with emotion features. During negative mood, negative verbal information is retrieved more readily than positive information, and conversely, positive information is retrieved more readily during positive mood.

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

4

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

Macht and Müller also demonstrate that mood improvement of experimentally induced negative mood was related to emotional eating but not to restrained eating. Since restrained eaters depend on cognitive control both positive and negative emotions will lead to increased intake, whereas emotional eaters using eating as a way of coping only eat more in response to negative emotions. Emotional eaters also experience mood improvement, and restrained eaters do and might even feel more guilty and depressed by their loss of control. 4. Effects of food on emotions and mood As mentioned above, the attention given to the effect of food on mood has not drawn as much attention as the effect of mood on food choice and food intake. Apart from the serotonin hypothesis mentioned above there have not been many other efforts to find reasonable physiological mechanisms and the serotonin hypothesis faces also many other problems than the objection of Macht (1999) that it implies a delay in the reaction. In an extensive review of it, Haddock and Dill (2000) support the conclusion of Christensen (1993) “that one of the biochemical steps required for carbohydrate ingestion to alter serotonin synthesis (Wurtman, Hefti, & Melamed, 1981; Wurtman et al., 1981) is missing, suggesting that the mood altering effects resulting from carbohydrate ingestion are probably mediated by a mechanism other than, or in addition to, the enhancement of serotonin synthesis or neurotransmission”. Furthermore, the serotonin enhancing mechanism invoked by carbohydrate ingestion may easily be disturbed by the presence of protein in the diet and meals are seldom almost protein-free. Finally, Markus et al. (1998); Markus, Panhuysen, Tuiten, and Koppeschaar (2000) showed that the effectiveness of carbohydrate-rich and protein-poor food depended on the stress proneness of the subjects and occurred only in people with a high stress proneness, thus throwing doubt on the generality of the mechanisms' functioning. Notwithstanding the lack of precise information on the physiology involved, the immediate effect of food and especially carbohydrate and sweet food on the reduction of negative feelings and mood remains well established (Macht & Dettmer, 2006; Macht & Müller, 2007). In the latter paper they compared the short-term and durable effects of palatable and unpalatable chocolate and showed that palatability is the key factor in the immediate positive effects of negative mood reduction and thus in the habit of eating to cope with stress. Nevertheless more research is needed to understand the precise mechanisms involved. Especially the role of learning and memory in the establishment of the relationship between the eating of certain foods and their effects on mood might be of interest. Strangely enough, the role of memory in food perception, and food choice is neglected in general. 4.1. Effects of expectations and memory on emotions and mood Although it is clear that expectations play an important part in the perception and appreciation of food and in food choice, there is only recent research on their emotional aspects. The influences of colourevoked expectations in fabric softeners on emotion were studied by Porcherot, Delplanque, Planchais, Gaudeau, and Cayeux (2013). They came to the conclusion that colour was not as important in raising emotional responses as odour. Comparing the differences between the emotional conceptualisations and effects of respectively blind and packaged or branded presentations of blackcurrant squashes, Ng, Chaya, and Hort (2013b) found indications “that intrinsic sensory characteristics have a stronger association with emotions, whereas extrinsic characteristics have a stronger association with abstract/functional associations”. Spinelli, Masi, Zoboli, Prescott, and Monteleone (2015) presented six hazelnut and cocoa spreads under three conditions (blind, expected from package and informed) and measured liking and emotions using the Emosemio questionnaire consisting of 23 sentences describing emotions. For two of their products, expected liking and informed liking, were higher than liking under blind presentation and this was true

only for expected liking for another product, while for still another one product the reverse was true, liking being higher in the blind than in the expected condition. These latter two products also differed in expected and informed liking. Emotion scores between the six products differed for all emotions except surprised and guilty in the blind condition and for all without exception in the informed condition. Thus, it is clear that expectations based on extrinsic factors like packaging and brand information may influence the emotional perception of products. Nevertheless, these influences are probably much weaker than the ones that are based on experiences in earlier eating situations and carry the emotional content of these earlier occasions usually with them. In fact, there is ample evidence that we do not remember the food we ate earlier with precision, but are immediately reminded of the earlier situation in which we ate it (its ambiance or the company we ate it with) or we note deviations from it as a surprise and warning: “attention, not encountered before in the same or a similar situation” (Köster et al., 2014; Morin-Audebrand et al., 2012). In a preliminary study on the “memorability” of foods – the spontaneous occurrence of food memories in non-eating situations – the present authors also obtained indications that the emotions related to the eating situations and not the properties of the food itself raise such spontaneous memories, even though the latter may have contributed to the emotional colour of the former. More research is needed to consolidate this finding however. Nevertheless, it is well in keeping with the use of situational aspects (family-, parent-, child-, lover-situations) to promote food products in advertising and marketing. Obviously marketers seem to be aware of the importance of these situational factors and use them rather than discussing the sensory qualities of, and emotions elicited by, the products themselves. 5. Measurement of the emotional effects of food The last few sentences of the preceding paragraph lead to the question what we should measure if we want to evaluate the emotional effects of food and also more specifically how and when to measure it. Should we measure the feelings that the person experiences while eating a particular product or should we measure the influence eating that product has on the persons' perception of his/her surroundings (the situation, other people present, etc.; see Macht, Meininger, & Roth, 2005; Pudel & Westenhöfer, 1991; Barrett, Mesquita, Ochsner, & Gross, 2007) and should that take place immediately or at a later time, thus involving also the function of memory? After all, what is remembered is what influences our later food choice decisions and we know that food liking and odour perception will change over time and with repeated exposure (Delplanque et al., 2008, 2009; Köster & Mojet, 2007). Moreover, if the persons' immediate feelings were measured, how should this be done? Do people actually know what they feel in a way that makes their feelings accessible for communication to others? In other words can they identify and verbalize their feelings or do we have to help them by providing lists of feelings and if so, how sure are we that they are not unduly influenced by our suggestions? Or should we rather rely on more indirect methods (physiological methods like skin conductance, heart rate) with their interpretation difficulties, on subtle psychophysiological methods (face reading, pupillometry) that can only be measured during direct contact with the food or on implicit psychological methods (reaction time measurements to emotional words or projection tests)? Other questions, like do we want to measure very specific feelings or more general mood effects and in what way does the information these measurements deliver contribute to our insight in the way we deal with the food concerned, should also be asked. In other words: do emotion measurements provide more insight in product-related food choice than traditional measurements like liking, wanting and situational appropriateness or do they also – and perhaps even predominantly – provide insight into the importance of other benefits (improving social contacts, etc.)? Does this mean that one should not measure the

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

5

emotions in a laboratory setting where people are either alone or eat under abnormal conditions? In other words: how important is the context and are people aware of being a subject in an experiment and therefore pay more attention than normally to the presented stimuli? In a recent experiment Porcherot, Petit, Giboreau, Gaudreau, and Cayeux (2015) investigated the self-reported affective feelings when differently flavoured aperitifs were consumed in an ecologically valid restaurant setting and although they did rather explicit experiments with questionnaires the subjects were not made aware of the different flavourings of the alcohol-wise same aperitifs they drank. One of the flavourings raised the mood of the participants more than the others, but only limited mood changes during the ensuing dinner were found. The study discusses the limits of emotion measurements in ecological settings and shows ways to improve such an approach which seems indeed preferable over measurements in laboratory settings. Over the last two decades a number of different approaches in measuring emotions have been tried and the advantages and disadvantages of them will be discussed.

emotions (boredom, disappointment and dissatisfaction) fulfilled such prominent roles (being rated higher and more quality-dependent than the others). Unfortunately, the correlations between each of the possible pairs of individual emotions in the set are not mentioned making it difficult to estimate whether they reflected relatively independent emotional aspects. Furthermore, the outcomes for certain emotions may have been limited by the choice of the foods as the authors themselves indicated. In a final discussion Desmet and Schifferstein discuss the possible sources of food emotions and illustrate five of these with examples:

5.1. Explicit emotion measurement

These possible sources of food related emotions suggest strongly that the idea, that the product-related momentary perception alone is determining its emotional effect, is rather limited. Memories and situational effects may well be at least as important. Edwards, Meiselman, Edwards, and Lesher (2003) already showed the effects of eating location on the acceptability of food, but did not specify emotional effects. Such effects were also illustrated in an elegant series of three papers by Piqueras-Fiszman and Jaeger (2014a,b,c) using evoked consumption contexts and appropriateness in these contexts to show their influence on the emotion responses. They demonstrated that imagined context (evoked either in written format or as images) influenced the majority of the emotion responses in the case of a sweet snack, but only a quarter of the emotion responses to an apple, and that the frequency of use of positive emotion terms depended on the appropriateness in different eating situations. In a second experiment they confirmed these results and showed that frequency of use of the food stimuli by the consumers did not lead to differences in the findings which were consistent over time and were more variable between contexts for emotion-laden foods than for more neutral ones. In a third experiment they compared the influences of: 1) the way in which the food stimulus was evaluated (tasting vs. seeing a food image), 2) the serving presentation of the food (food shown in isolation vs. on a plate with cutlery) and 3) the evocation method of the context (written and pictorially vs. written only). The latter two influenced the perceived appropriateness of the product and resulted in a higher frequency of negative emotion terms in less appropriate consumption contexts. All of the evoked context effects in this research have the advantage that they are person-centred, i.e., that they are not just based on objective context, but carry strong elements of personal interpretation and meaning based on previous emotional experiences by the investigated person without a demand to direct attention to the present sensory properties and to judge the emotional effects of them. Such an approach is ecologically more relevant. Rousset, Deiss, Juillard, Schlich, and Droit-Volet (2005) used a French scale developed by Juillard, 2003 to compare the emotions related to meat and other foods in high and low meat-eating women, showing differences in negative emotions like disappointment, indifference and less satisfaction and in some reactions to other foods like starchy products, pears and French beans. This once more illustrated the importance of consumer segmentation in emotion measurement. Unfortunately, most investigators have chosen for a product-centred rather than for such a more person-centred approach. Chrea et al. (2009) developed a scale for the measurement of the emotions elicited by odours. Their findings suggested “that the subjective affective experiences or feelings induced by odours are structured around a small group of dimensions that reflect the role of olfaction in well-being, social interaction, danger prevention, arousal or relaxation

Direct explicit emotion measurement is the most frequently applied method. People are requested to taste the food and to report their feelings usually on the basis of a standardized emotion lexicon provided by the experimenters. Over the last 50 years several emotion measurement scales have been developed (Mayer & Gaschke, 1968; McNair, Lorr, & Droppelman, 1971; Richins, 1997; Watson, Clark, & Tellegen, 1988; Zuckerman & Lubin, 1965, 1985), but these were of a more general nature and not especially related to food. Recently, Jiang, King, and Priniawatkul (2014) gave an excellent overview of all the efforts made to measure the relationships between food, eating behaviour and emotion. Here we will limit ourselves to some of these approaches and discuss the validity of them from a psychological point of view. Laros and Steenkamp (2005) extensively reviewed the older literature on the use of different sets of emotions in the marketing literature and came to the conclusion “that there is a wide divergence in the content of emotions studied in consumer research …. In spite of this, consumer researchers use a small number of dimensions. Among these dimensions the classification in positive and negative affect appears to be the most popular”. They then developed a hierarchical model of consumer emotions based on four types of negative affect (anger, fear, sadness and shame) and four types of positive affects (contentment, happiness, love, and pride), which each was represented by between one and nine terms to reach a total of 33 one-word emotion descriptions. They used these terms to let their subjects rate the emotions evoked by mentioning (not consuming) one of four types of food (genetically modified, functional, organic or regular food) on five point Likert scales ranging from: “I feel this emotion not at all” to “I feel this emotion very strongly”. Their results indicated that apart from positive and negative affect as the most important dimensions, important nuances may be lost if all emotions are only considered on the basis of their valence. In the following experiments in which food was really tasted such nuances might be even more important. Desmet and Schifferstein (2008) used 22 emotion descriptions that were equally divided over pleasant (P) and unpleasant (U) ones, had students rate their relevance in tasting or eating food in a first experiment and used them in a second one to evaluate three products differing in quality in each of three food categories (sweet or savoury snacks and a pasta meal). They showed not only that pleasant emotions were reported to occur more often than unpleasant ones, this difference being somewhat less outspoken in study 2 where the foods to be tasted had been selected by the experimenters instead of being chosen by the participants themselves. Some positive emotions (satisfaction, enjoyment, desire, amusement, and pleasant surprise) scored higher than others (relief, admiration and pride) and showed also a stronger dependency on the quality of the tasted products, whereas only three negative

1. Sensory attributes (amusing or surprising taste or texture; boring or disgust) 2. Experienced consequences (relief or stimulation by drinking; disappointment or dissatisfaction when still hungry) 3. Anticipated consequences (health effects, fear of obesity) 4. Personal or cultural meanings (strawberries remind me of someone) 5. Actions of associated agents (ashamed of bad table manners; contempt towards meat eaters).

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

6

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

sensations, and conscious recollection of emotional memories”. Their scale contained 36 terms unevenly divided over 6 dimensions: happiness-wellbeing (6 terms), awe-sensuality (7), disgust-irritation (8), soothing, peacefulness (5), energizing, cooling (7), and sensory pleasure (3). King and Meiselman (2010), who applied much effort in developing the explicit verbal method, should be recommended for having opened up the field and raised interest in its application in industrial settings. Here, their work is discussed as “the” illustrative example of the explicit verbal measuring methods showing the extensive work to select emotion terms and to build the final test set. This means also that any criticism against their work in the following passages applies also to the other approaches that use explicit verbal measurement. In their EsSense Profile, King and Meiselman used 39 descriptions (25 pleasant (P), 3 unpleasant (U) and 11 non-classified (NC)) that were chosen on the basis of a questionnaire of 81 terms that were evaluated (individually or clustered in groups of 2–3 items on the basis of the similarity in their definition) by an unspecified number (thousands) of consumers via the internet, central location tests and a Focus group. The terms were then used by an internet panel (N = 105) to describe their most favourite beverage, snack and dessert and their least favourite meal, dessert and snack. Unfortunately, the figure representing the results was not given, but positive terms seemed to be selected more frequently than negative terms. Finally, the 39 terms were selected by an internet panel (N = 200) on the basis of frequency of use (N20%) categorization (P, U, NC) and appropriateness to food testing. The final set was then used to measure the frequency of use in describing the emotions raised by flavoured crackers of 4 types, one of which deviated from the others in all but three emotional respects (enthusiastic, eager and free). This exceptional one was less related to positive emotions as good, happy, satisfied, pleasant, pleased, good-natured, etc. and more to emotions like active, daring, aggressive and wild. Unfortunately, the nature of the flavours of these crackers as well as their acceptance remained again unspecified, making it difficult to judge the ecological validity and practical importance of the results of this first experimental try out. Another measurement of the intensity of different emotions aroused by eating foods as different as Pizza, mashed potatoes, vanilla ice cream fried chicken and chocolate, shows surprising resemblances between the emotions related to sometimes very different product terms. In these first explorative studies there are unfortunately not yet indications of the correlations between the estimates of the different terms over the subjects and this is also not the case in the differences found between three different highly acceptable salty flavoured crackers on the rather opposite emotion pairs calm and eager or mild and aggressive. Finally, King and Meiselman show that the ratings of the emotion intensities increase with the consumption frequency of their experimental subjects. Although the authors do not mention it, these effects are rather big and in populations that are not segmented on the basis of consumption frequency, the resulting variance could obscure eventual emotion differences between products. In a follow up study King, Meiselman, and Carr (2010) showed that the extent to which the emotion measurements were correlated with overall liking of the products depended on the nature of the products (relatively frequent for herbs, spices and beverages and infrequent for protein products) and the gender of the participants (more frequent for men than for women except in the case of snacks where the men's emotions were much more often independent of overall liking than those of the women). In later research in which the EsSense method was used (Cardello et al., 2012), a number of the unverified aspects mentioned earlier, such as the reliability of the measurements and the relationship between rating emotions evoked by food names only or evoked by actually tasting foods were investigated. Rather strong correlations (+.66 to +.83) were found between the ratings in response to tasting four very well-known foods (milk and dark chocolate and regular and barbecue chips) and to their corresponding names, whereas the correlations of the ratings in the two sessions for testing the reliability were lower

(+.50 to +.70 for food names and +.37 to +.70 for tasted products). In discussing the difference between these two sets of correlations, Cardello et al. suppose that food names reflect stable and broader cognitive images and associations, whereas tasted foods vary over time due to perceptual variability, changing expectations and preceding appetitive contexts. Thus, they make implicit statements on the roles of memory and perception in the emotional effects. The question then arises what is meant by stable and broader images and what types of associations (situational experiences during the first or later emotional encounters with the product, remembrances of special eating events and habits or relationships with other people) are evoked by the food name. As mentioned above, it is clear that we do not remember the products we have eaten before by the recollection of something like a detailed internal representation (see Morin-Audebrand et al., 2012 for an overview), but that our food memory is directed at novelty and change detection. Furthermore, food names and tasted foods both evoke associated memories of previous eating occasions and are thus linked to emotions experienced on these occasions. Of course these remembered emotions may also be based on experienced food satisfaction expressed in expectations (Cardello, 2007), but they will also strongly be related to situational mood factors (the pleasantness of the surroundings and company). Obviously, in the case of the very common products used in the experiments of Cardello et al. (2012), the specific relationship to one or a few emotional key occasions may have been lost to a great extent, but the memory of the type of occasions for eating chocolate or chips may have contributed at least as much to the difference in evoked emotions as the sensory differences between the products. This might be illustrated by the fact that emotions raised by tasting the different products in the same food category (i.e., milk- or dark chocolate; regular or barbecue chips) do hardly differ from each other in the extents of the evoked emotions while the emotions raised by the product names are much stronger in the case of chocolate and somewhat weaker in the case of chips, but are highly correlated with those of the tasted products in both cases. This might mean that remembering chocolate evokes stronger emotions than eating it and this seems to be especially true for strongly person- or situation-related memories (satisfied, good, happy, pleased, energetic, active, enthusiastic, merry, adventurous, loving, affectionate, eager and guilty) with perhaps a few exceptions (good-natured, friendly, calm, nostalgic) and not at all for emotions that might be more product-related (pleasant, whole, steady, secure, mild, bored, tender). In the case of the chips only two meaningful differences between the memory-based and the tasted-product-related emotions are found: one stronger for food names (guilty) that was clearly person-related and one weaker for food names (interested) that might be more product-related. Furthermore, the emotions satisfied, good, happy, pleased, glad, pleasant, good-natured and friendly, tended to be slightly less strong in memory evoked by the names than those evoked by both chips when actually eaten. Nevertheless, again the three measurements were highly correlated and they are of course not independent of each other since the memories are based on earlier experiences with the product and in turn influence the perception of the product by the expectations they bring to the tasting. The question then arises whether these expectations are emotional and associative or sensory in nature. In view of the fact that there is ample evidence that our recognition memory for food is based on change detection rather than on recollection (Köster et al., 2014; Morin-Audebrand et al., 2012) the latter is highly unlikely. Thus, in keeping with our preliminary findings on the ‘memorability’ of foods it seems that the emotions are more related to the eating situations than to the food itself. This may also explain why in the experiments discussed here, the emotions related to chocolate that is mostly eaten alone and often as a consolation during negative mood moments, are much stronger than the emotions to chips which are usually consumed rather inattentively as a side dish to drinks in larger gatherings of people. In summary, like odours (Wiemers, Sauvage, & Wolf, 2014; Zucco, Aiello, Turuani, & Köster, 2012), food

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

names and even tasted foods are effective retrieval cues for earlier associated situations, but not for precise sensory recollection of the unchanged food itself (Köster et al., 2014; Møller, Mojet, & Köster, 2007; Morin-Audebrand et al., 2012). Basing themselves on the idea that people eat the things they like, King and Meiselman (2010) and Cardello et al. (2012) have assumed that foods elicit mainly positive emotions, constructing the EsSense scale with a predominance of positive emotion words. The fact that so far, emotion measurement was mainly used with pleasant snack items and favourite foods has probably fortified the idea of the predominance of positive emotional effects. The idea that pleasant emotions prevail in eating was questioned by Ng et al. (2013) and more recently by Spinelli et al. (2015). As mentioned above, the latter authors asked themselves whether the negative emotions might not be of more importance as acceptance indicators and predictors of future consumer behaviour than the positive ones. After all it is well known that feelings of discontent grow over time and positive feelings wear off with repeated exposure (Köster & Mojet, 2007). Similar concerns regarding the imbalance of positive and negative emotion measures were expressed by Kuesten, Chopra, Bi, and Meiselman (2014) who used PANAS scales, which are based on an even distribution of positive (PA) and negative (NA) terms, with success measuring consumer emotions associated with aromas of phytonutrient supplements in four rather different countries (US, Japan Korea, Germany). Ng et al., who compared the EsSense profile, the content of which was predetermined by the investigators, with a consumer defined lexicon based on the CATA (Check All That Applies) approach which contained many more and probably important negative words (disappointment, annoyed, unpleasant surprise, resentment) and some positive ones (trust, desire, refreshed, reminiscence, pleasant surprise) that could provide information about the relationship between the memory-based expectations and the actual taste of the product which in the experience of the present writers is an essential factor in the prediction of market success. Furthermore, Ng et al., provide detailed information about the correlations between the sets of emotion terms and the liking for their eleven black current squashes representing the variation in sensory characteristics on the UK market. In the 39-item EsSense list only four emotion terms (calm, daring, tame and tender) correlated lower than 0.80, and 15 correlated higher than 0.90 with liking. In the 42-item list twelve terms (angry, bored, cautious, confused, curious, desire, reminiscence, sceptical, sickly, warm, and worried) correlated lower than 0.80 and 11 correlated higher than 0.90 with pleasure. On the basis of these results Ng et al., rightly ask the question whether emotional data go beyond liking. In their set of data they find nevertheless some cases in which products or sets of products were further discriminated using emotional data. Thus, although similarly liked, compared to products with added sugar, their non-added sugar product evoked less ‘pleasant surprise’ and ‘trust’ than the sugar-added ones, whereas the latter were characterised as evoking less ‘disgust’, ‘displeasure’, ‘uncomfortable’, and ‘unpleasant surprise’ responses than the nosugar-added one. In order to check this, they analysed and compared the emotion data of pairs of products with similar liking rates and found that emotions ‘tame’ and ‘daring’ that were not correlated with liking and formed the extremes on the engagement/activation axis in the PCA plot provided indeed additional insight into product acceptance. CATA also discriminated products of similar liking and did so to a greater extent. Other authors using adapted verbal methods have also found a number of cases of emotional differences between products in the same category that were not directly related to liking. Spinelli, Masi, Dinnella, Zoboli, and Monteleone (2014) used 23 full sentences instead of just words in their EmoSemio questionnaire in order to reduce ambiguity in the measurement due to differences in understanding of the terms between the participants. They included 16 positive and 7 negative emotions. The negative terms were considered important in order to be able to distinguish between product failure due to indifference or to disappointment for instance. This distinction can indeed be of great importance for the prediction of market success or failure

7

and for the determination of the rather different ways to cure the problem before a market flop arises (Köster & Mojet, 2007, 2012a,b). Spinelli et al. found very significant differences between the six chocolatehazelnut spreads they used on all of their EmoSemio sentences except the ones involving guilty and surprised. When tested with the EsSense profile the spreads were also significantly distinguished on 33 of the 39 descriptors, aggressive, eager, guilty, nostalgic, tame, and wild being the six non-distinguishing ones. When the authors compared the emotional reactions to the highly liked and the least liked products the EmoSemio questionnaire with its negative terms differentiated better between the products than the EsSense scale, providing more insight in the nature of the dislike in the least liked products. Unfortunately, no insight in the correlations between liking and each of the terms or sentences used in the methods was given, but, judging from the parallel lines in the profile maps for the differently liked products, the correlations were probably high, especially in the EsSense profile. Bhumiratana, Adhkari, and Chambers (2014) developed an emotion lexicon for the coffee drinking experience. They reduced a list of 86 emotion terms to 44 coffee drinking experience emotions (CDE) and tested them on 6 different coffees in 6 liking-based consumer clusters. This approach showed large differences between groups of consumers in liking for the different products and in the differences in emotional coffee drinking experiences that were involved in the cluster segmentation. The authors concluded that: “the coffee drinking experience can be explained by two main dimensions: the positive–negative and the high–low energy dimensions”. Furthermore, they stressed the fact that “the consumers had not only varying preferences for the coffees, but they also sought different emotion experiences from the beverage”. According to the present authors, the segmentation of the consumers on the basis of their preference is a major step forward and should be recommended in all food-related emotion research. This becomes even more clear when people start collecting massive amounts of data via internet questionnaires. As often pointed out “The Consumer” does not exist (Köster, 1996) and segmentation on the basis of gender, age and social status is less important than segmentation on ways of dealing with the products and the expectations of their benefits. The clearest data on the independence of emotional aspects from liking measurements was provided by Porcherot et al. (2010) using not food but the odours of different fragranced and flavoured products as stimuli and a new strongly reduced Geneva Emotion and Odor Scale (GEOS) questionnaire (Chrea et al., 2009) consisting of six series of three terms instead of the original 36 terms to be rated in the original form (Chrea et al., 2009). They showed that for pairs of odours that did not significantly differ in liking, differences in three of their six emotional dimensions (1. energetic–invigorated–clean; 2. nostalgic– amusement–mouth-watering; 3. romantic–desire–in love) which are probably more independent of liking could be found but not in the other three dimensions (4. happiness–well-being–pleasantly surprised; 5. disgusted–irritated unpleasantly surprised; 6. relaxed, serene, reassured) which are probably more linked to liking. Much effort was spent by King, Meiselman, and Carr (2013) to check the influences of a number of experimental details: response format (rating scales vs. a CATA (Check All That Applies)) approach applied to the EsSense profile terms; emotion list order (alphabetical vs. random); acceptability question (before or after emotion measurement); product context (name only, aroma only, flavour of tasted product); time of day (noon vs late afternoon), and number of products in a session. On the basis of tests with vanilla vs. favourite ice creams they came to the conclusion that rating scales were better at differentiating low level emotions whereas CATA showed more differentiation at high levels. Time of day of the measurement had no significant influence and neither did alphabetical or randomised word presentation with either method, but overall liking was clearly higher when asked before than when asked after the emotion measurement. This seems no wonder after the task of rating 39 emotions, many of which the participant probably

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

8

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

would never have thought of in connection to the presented product(s). Furthermore, emotional response to spices based on name only produced stronger emotion responses to most of the terms related to the active dimension and to some person-related emotions like affectionate, guilty, nostalgic and worried. Smelling the stimuli did not lead to any differentiation and the tasting of the flavours of low spice concentrations elicited somewhat stronger emotion responses. The spice naming and tasting differed in some common ways from the aroma testing. Thus, the spices evoked emotions as glad, good, happy, interested, joyful, loving, merry, mild, pleasant, pleased, tender and whole more strongly than the aroma did. It should be remarked that according to Ng et al. (2013) all except four of these terms [happy (r = 0.890), tender (r = 0.784), whole (r = 0.864), and mild (r = unknown)] all of these terms correlated well over r = 0.900 with overall liking, thus raising doubts about their specificity and independence. It could be asked whether the indicated strengths of these emotions are not more based on the association of the emotional concepts described with liking, than on specific feelings. This question is also brought forward by Thomson, Crocker, and Marketo (2010) who write: “Measuring the emotional consequences engendered by unbranded products is often futile because they may be subtle, may occur some time later and may not be immediately apparent to the person concerned. As a consequence, most emotional measurement tools don't access emotional consequences but emotional conceptualisations (or emotional associations). This means that when someone tells us that a product makes them feel ‘happy’, ‘passionate, etc. it's more likely that they are reflecting what the product is communicating to them (emotional conceptualisations) rather than doing to them (emotional consequences)”. Thomson & Crocker, 2013 then describe three problems in measuring conceptual associations: 1) Certain conceptualisations are readily accessible, but others are not and are replaced by logical rational associations that sometimes have little bearing on reality. 2) Liking is more immediate and it may be the ‘easiest way out’ for the participants to associate positive and negative conceptualisations with respectively likes and dislikes. 3) Some important conceptualisations may be counterintuitive (e.g., trustworthiness in dark chocolate). On the basis of the fact that a scale makes people think about the meaning of the words and leads to rational thought processes, whereas in choice and purchase behaviour often irrational and non-cognitive influences play a major role and because rational processes makes the individual focus on literal instead of metaphorical meaning, Thomson & Crocker (2013) decided to use best–worst scaling or maximum difference scaling (Finn & Louvière, 1992) thus avoiding external measurement scales. Instead of presenting one emotion word at a time, they presented four or five at a time and asked their participants to indicate which of these they felt fitted best and which fitted worst to their experience. From a psychological point of view this method has the advantage that it does not force the participants to judge the strength of each and every emotion separately with the risk that they fall prey to the old psychological adage: “whenever you ask a subject a question, you will always get an answer” and start reporting things they would never have felt when just eating the product or thinking of its name. Eating consciously and attentively or explicitly thinking about a food as source of emotion are already very unnatural behaviours and even with best–worst scaling the attention to suggested words and the necessity to make a choice may severely distort the actual experience. Furthermore, best–worst scaling is considerably more timeconsuming and it may thus take too long to catch the rather fleeting and transient immediate emotions evoked during consumption. In fact, the necessity of explicit attention is the major handicap with all methods that use verbal material. Most of our emotions, and often not the least important ones, remain unconscious. Finally, the verbal methods discussed here have the additional handicap that they are hard to translate in different languages without loss of precision in their meaning.

5.2. Non-verbal explicit emotion measurement This is the reason why some researchers have looked for more implicit and more non-verbal methods of measurement. Thus, Desmet (2003) developed the Product Emotion Measuring instrument (PrEmo), using cartoon figures expressing different emotions. The idea was based on the fact that people from different cultures all recognise emotions in facial and vocal (pitch) expressions and in gestures and body postures via mechanisms involving empathic reactions (Ekman, 1972, 1994; Ekman & Friesen, 1971). Seven pleasant (desire, pleasant surprise, inspiration, amusement, admiration, satisfaction, fascination) and seven unpleasant (indignation, contempt, disgust, unpleasant surprise, dissatisfaction, disappointment, boredom) are depicted. Each of these is accompanied by a three point scale (I do feel the emotion; to some extent I feel the emotion; I do not feel the emotion expressed by this animation) which only appears after clicking on the particular image. In some versions the degree of strength of each emotion was also pictured in a series of increasing expressions of the emotion. The major advantage of the method seems to be that it avoids the translation of emotion words into feelings, which is a necessity in all of the verbal methods. Since we use emotion words normally without much thinking we tend to forget that we have to realise their meaning before we can rate them and that this process is a rather unusual one. In most cases the rather minor emotions evoked by food are just felt and not named. Looking for the presence of each of 39 named emotions in ones' own experience as in the EsSense list or even realising which of five named emotions fits best or worst (or neither of these two), is an utterly strange and perhaps even estranging task. No wonder that it is better to ask overall liking before and not after this. The effects of this might also be reflected in strong order effects when two or more products are compared, unless the subjects have opted indeed for the easiest way out and just verified how well the words fitted their liking for the product as is suggested by the high correlations with overall acceptation. Unfortunately, there is little verification of product order effects in the literature. The same holds for a verification of the unanimity of the meaning of the emotion words among the participants. No indications are given about the variability in meaning of emotion words over subjects, who may vary in cultural background and education level. Since facial and bodily emotion expressions are directly understood without translation and independent of the educational level of the person, the use of the pictured emotional expressions in the PrEmo cartoons or in the photographs of emotion reflecting faces at least avoids the unnatural word translation problem. Whether the 14 emotions are well chosen for use with food products (they were mainly tested on design of technical products) and are so well depicted by the artist that they do better than photographs of emotional faces might perhaps still be questionable, but their use has at least the extra advantage to be appreciated by the participants as a pleasurable and not very bothersome task. On the other hand, this carries the risk that the pleasantness of the task unduly influences the emotions felt and reported. 5.3. Implicit emotion measurement A still more natural and less intervening and disturbing method of emotion measurement is direct reading of the behavioural emotion expression of the person having the experience (reviews: Mauss & Robinson, 2009; Ahn, Bailenson, Fox, & Jabon, 2010; Drozdova, 2014). Mauss and Robinson also discuss autonomic measures of emotion (electro-dermal and cardiovascular responses) as measured by short level skin conductance responses or heart rate, blood pressure, total peripheral resistance cardiac output pre-ejection period, and heart rate variability. However, they conclude on the basis of a meta-analysis made by Cacioppo, Berntson, Larsen, Poehlmann, and Ito (2000) that there was little specificity in the autonomic responses and that they probably are mainly useful in indicating such broad dimensions as arousal, although this meta-analysis did show response to emotional

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

valence by blood pressure, cardiac output, heart rate and skin conductance response duration. Furthermore, looking at combinations of autonomic responses may in some cases show a greater degree of autonomic specificity, but it should also be remembered that autonomic responses also depend on task demands, coping mechanisms and motor behaviour and not on emotion alone. Brain states can also serve as a measure of emotion that does not involve explicit behaviour on the side of the experimental subject. Mauss and Robinson discuss both electro-encephalography (EEG) with its excellent temporal resolution but poor localisation, and functional magnetic resonance imaging (fMRI) with its specific localisation but lack of temporal resolution. With regard to EEG, they mainly follow Davidson's (Davidson et al., 1990; Davidson, 1999) view that “frontal asymmetry”, contrasting alpha power (8–13 Hz band) in the left frontal region with that in the right frontal region is particularly important with respect to emotion. Greater left-sided activation seemed to be related to pleasant emotional experiences and is definitely positively related to approach, whereas right-sided dominance is related to avoidance. These latter behavioural tendencies seem to be more important effects than emotional valence (Davidson, 1999). This might therefore be too global a measurement to deduct sufficient emotional content from, and especially content that is independent from overall liking. Similar objections can be made to the use of fMRI or of positron emission tomography (PET), another technique for precise localisation of brain activity. Although complex reactions such as emotional states are likely to involve circuits rather than isolated brain regions (Kagan, 2007; LeDoux, 2000). Mauss and Robinson cite two meta-studies (Murphy, Nimmo-Smith, & Lawrence, 2003; Phan, Wager, & Liberzon, 2002) that show that localisation studies may be meaningful in identifying particular regions that may be more or less involved in specific emotions like fear (related to the amygdala), disgust (related insula activation), sadness (either related to the medial prefrontal cortex (Phan et al., 2002) or to the anterior cingulate cortex (Murphy et al., 2003) or to a circuit involving both anger and happiness (both located even less clearly)). Mauss and Robinson conclude that “although there has been some progress in understanding the neural correlates of fear, disgust and potentially sadness, the discrete-emotions perspective has yet to produce strong replicable findings”. In their discussion of behaviour as a measure of emotion, Mauss and Robinson distinguish vocal, facial and whole body behaviour and come to the conclusion that these three forms of behaviour highlight different aspects of emotions. Thus, vocal characteristics appear to convey mainly levels of arousal of the emotion, higher levels of pitch and volume being related to higher arousal, but do not seem to be linked to emotional valence or discrete emotions (Bachorowski, 1999; Johnstone & Scherer, 2000), although Banse and Scherer (1996) found some rather complex indications of links between rather special emotions (e.g., elation or anger) and aspects of the vocal pattern such as low frequency energy (medium in elation and low in anger) and change in pitch over time (increasing in elation and decreasing in anger). However, as Mauss and Robinson point out, these complex findings might need confirmation in other cultures and languages. Facial behaviour measurement is divided by Mauss and Robinson in two sections: Observer ratings of videotaped facial expressions and electromyography (EMG), i.e., measuring the electrical potential from facial muscles via electrodes on the face, while the more advanced techniques of automatic facial emotion registration via interpretation by computer programmes of the development of camera observed facial expressions remain unmentioned. When trained observers are used, they most often use the Facial Action Coding System (FACS) developed by Ekman (Ekman & Friesen, 1971) which takes 44 different muscle movements into account. With such measurements occasionally very high correlations (r N .80) between valence and liking have been found (Mauss, Levenson, McCarter, Wilhelm, & Gross, 2005), but emotion specificity of facial behaviour is not always evident. In measuring food-related emotions, the situation is furthermore complicated in the case of tasted foods (not with food names) when

9

the facial expressions are disturbed by the mouth movements. This also seriously hampers the registration of emotion during tasting by the automatic registration systems (Mojet, Dürrschmid, Danner, Joch, & Köster, submitted for publication). It means that the existing methods are mainly useful to measure the reactions to food names and not to tasted food. Electromyography is also not a very practical method because of the necessity to involve quite large numbers of experimental subjects and in view of the emotional effects on ordinary consumers of having their faces plastered with electrodes. Eye-tracking and pupillometry, are certainly also interesting methods for measuring memory- and expectation based emotions, but are limited to the visual inspection of products or packages. Nevertheless, they can provide very useful information about the change of emotional reactions from before to after consumption (Mojet et al., submitted for publication). The role of body posture effects of emotion has not been investigated extensively according to Mauss and Robinson. It has been suggested that body postures are mostly related to social-status-related emotions (App, McIntosh, & Reid, 2007) such as pride and embarrassment which play no (or at most a very minor) role in food-related emotions. Thus, in summary it can be said that behavioural measurements have the advantage that they do not demand verbal emotion interpretation or explicit attention to the emotion on the part of the participants, but that they have the handicap that the most applicable facial measurements are severely disturbed by mouth movements when food is actually eaten. Thus, they are mainly useful in measuring reactions to remembered or prospected emotions. Nevertheless, when they are applied in such cases they are highly effective, because they can analyse the quickly changing facial expressions with much more precision and high time resolution (Ahn et al., 2010; Drozdova, 2014). The implicit methods described thus far use physiological reactions and could be used without explicit awareness of the relation between the measurement and the food eaten although in most cases the subjects are aware of the relationship. They are mainly implicit because the subject has no explicit power over his/her reaction. Furthermore, even if he/she is not explicitly told, the experimental situation may lead the person to wonder what is the relationship between the measurement and the stimulation with food or food packages or food names and to make the connection. Fully implicit measurements are therefore very rare. Nevertheless, in psychological experiments it has been shown that the unnoticed presence of an odour may influence behaviour just as strong or more strongly than when it is consciously perceived and explicitly recognised (Holland et al. 2005; Degel, Piper, and Köster, 2001; Degel and Köster, 1999). In analogy to these experiments in which the subjects were completely unaware of the experimental goal, it has recently been tried to measure the emotional and mood effects of food using a seemingly completely independent test. It uses projection of the participants' positive and negative feelings on the photographs of others and thus has the advantage to measure mood effects of external stimulations without any explicit awareness of the relationship to its cause by the participating subjects. Thus, in recent experiments (Mojet et al., submitted for publication to this journal issue) after eating a version of a product, participants were asked to take part in a completely different psychological experiment characterising photographs of other people by rating the fit of a random sequence of six positive and six negative character traits, that were selected and tested for their contribution to different mood aspects (e. g., friendly and enterprising on the positive and arrogant and depressed on the negative side). This procedure was repeated after one week with another version of the same product and a different but carefully selected equivalent set of photographs. Using a reverse engineering approach (Moskowitz, 2000) three groups of subjects each judged the photographs after eating one pair of yoghurts similar in liking but differing strongly in market success in their country. The results showed significant differences in the positive and negative evaluations of the photographs after eating the yoghurts in two of the three pairs and these differences were not

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

10

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

significantly related to overall liking. This “emotive projection test” (developed by the present authors for the assessment of the unconscious influence of ambient odours on mutual human relations and tested for the equivalence of the two sets of photographs) was also used to show the influence of the presence of flowers in eating situations (Mojet et al., submitted for publication) and in an experiment on the influence of ambient odours (De Wijk & Zijlstra, 2012). Even if this test seems a rather indirect and unspecific method, because it melts all specific feelings and convictions (e.g., organic, industrial fair trade, etc.) about the product together in one seemingly unrelated mood effect, this might be another advantage, because it does not rest on fixed ideas of the experimenter about the ways in which such feelings and convictions interact, but rests on the emotional mood result of the interaction of all these factors in each individual participant (Köster, 2009) and the lack of correlation with liking shows that the mood changes, how diverse they may be, are not just a reflection of the way in which the products are liked by the consumers like in many of the cases of specific emotions measured by explicit verbal methods. More research with larger groups of subjects is still needed to validate this method, however.

all three for food memory). This latter comparative information may then be used in the prediction of market success of new products. Both very large positive (pleasant surprise) and negative (due to disappointment or growing indifference) discrepancies in this relative memory are to be avoided, the first indicating that the memory of the product has deteriorated since the last encounter with it which may negatively influence repurchase, whereas the latter means that the actual product experience in the repeated measurement does not live up to the expectations that had been formed under the influence of other (situational or post-ingestional) factors than its sensory experience. Mild pleasant surprise and positive recognition of the product are important indicators of success and these should be related to positive mood effects which can perhaps be tested best in a way that is seemingly completely independent of the tasting experience. As also indicated by Spinelli et al. (2014), disappointment or indifference and boredom are indeed very important negative emotions which might have much stronger and long-lasting effects on product acceptance than the emotions elicited by immediate sensory impressions and which demand rather different product adjustments (Köster & Mojet, 2007),

5.4. Sensory versus remembered and prospected emotions

6. General conclusions and practical suggestions

This raises once more the general question, how important the actual sensorially evoked emotions are in comparison to the remembered ones and the prospected ones, based on expectations. Much depends on the purpose of the exercise in answering this question. If product purchase or even menu choice are involved it is clear that memory and expectations are the only factors that count. And although the memories of a food are certainly related to previous sensory experiences with the product they are also heavily influenced by the emotions experienced after the meal (feelings of unpleasant heaviness or pleasant lightness, bodily satisfaction and energetic feelings or sleepiness) and by the emotional situational context of the meal (the pleasantness of surroundings and the company or absence of other persons) which may be completely unrelated to the characteristics of the products eaten. Furthermore it should be remembered that the sensory qualities of the product actually do not come into play directly in later food choice, but may lead to emotions of disappointment or (un)pleasant surprise at the moment of tasting when they are confronted with the memorybased expectations of the consumer (see also Cardello, 2007; Schifferstein, Mojet, & Kole, 1999). It seems indeed that these latter emotions are very important in tasted products, and that therefore specificity of food evoked emotions should be tested primarily in memory. At the same time it is necessary to find ways to separate the sensory and the situational elements in the expected emotions. The above cited papers by Piqueras-Fiszman and Jaeger (2014a,b,c) already go a long way in that direction. Situational analysis, as earlier proposed (Köster, 1996) and based on the different situations in which people eat their food the frequency of occurrence of these situations and the emotions linked to them, may take it even further. This is in line with the strategy of the present authors in their applied and confidential work where they test discrepancies between remembered and actual experiences with the target stimuli and slight variations of these on aspects that are suspected to be emotionally relevant. They let people eat a meal containing hidden target items in a casual way as a service to subjects who come for other experiments and without any focus on the items (incidental learning phase) and when the subjects come back later (days or weeks) they ask them a number of questions about their situational memories of that meal and present them in a monadic way with the original targets and a number of slight variations of these targets and ask whether the subjects recognise the foods they ate before and whether they now judge them to be more, less or equally pleasant as the ones they ate during the previous meal (Mojet & Köster, 2002, 2005 for food flavour and texture memory, and Köster et al., 2004; Laureati et al., 2008; Morin-Audebrand et al., 2012,

In view of its important role as a factor in the development of obesity, the influence of emotional states on food choice and eating behaviour has been studied quite extensively and has led to insights in the role that eating may play in reducing strong and unpleasant emotions. To what degree and by what precise physiological and/or psychological mechanisms food consumption influences emotions and mood is less clear and further research on this topic might be both of scientific and practical value. For the time being, this research is perhaps still most hampered by the lack of clarity about what to know and how to measure it. It has been realised already for a long time that merely measuring liking of products is insufficient for the prediction of their acceptance at any given moment and notions like appropriateness and wanting have been introduced. These have indeed helped to explain the momentary behaviour, but contributed only modestly to the insight in long term preference and longing for the product. These seemed to be mainly based on the adequateness of the arousing properties and the emotional valence of the products. Although both are perhaps equally important, more attention is nowadays given to the latter and much effort is devoted to emotion measurement. Unfortunately, most methods were developed without much discussion about the “what and how” of the measurement. Do people really experience a wide gamma of emotions or would it be sufficient to measure mood change and perhaps some very specific food- situation-, or person-related feelings (e.g., disgust, disappointment, guilt) that are not closely related to liking? When should one measure the emotions, before during or after (and in that case how long after) eating? From a psychological point of view and as a prediction of product success, the remembered emotions might be more important, because they contain elements of the experienced emotions during the eating of the product but in combination with the aftereffects of the meal and the situational experience all weighted and integrated by the person in his/her own particular way. It is this emotional content and not the momentary eating experience that forms the basis for the persons' expectations. This might mean also that one should not just average the ratings of the emotions over subjects, but rather segment the subjects on the basis of the way they weigh the different successive emotion elements and integrate them in their expectations. As it is done now in most of the methods (verbal or other) people are not segmented even on the way they consume the product or on the extent of their product liking and important segments of the population may thus remain hidden. After all one should remember that in every family one person that has bad feelings about a product might severely reduce and even annihilate the products' chances of being repurchased. Knowing to what extent people differ in their emotional reactions to the product may therefore be essential.

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

Hitherto all measurements have neglected this and have fallen to the fallacy of uniformity (Köster, 2003), just averaging over subjects as if all subjects were the same in essence and only differed in the degree to which they had these emotions. Apart from these proposed changes there is a need for good verification of the real contribution that emotion measurement may make. This seems to be done best with a reverse engineering approach. In such an approach similar products (same brand, variations of a same type of product and similar liking) but with strongly divergent market success, should be tested for the emotions they evoke before, during, a few hours after, and a week after consumption to check whether emotional reaction patterns could be located that are directly related to market acceptance and are not linked to momentary sensory liking of the product In the eyes of the present authors there is a great need for this type of information, before food emotion measurement can be taken seriously as an extra element independent of liking, wanting and situational appropriateness. References Adolphs, R., Tranel, D., Damasio, H., & Damasio, A. (1995). Fear and the human amygdala. Journal of Neuroscience, 75(g), 5879–5891. Aggleton, J. P. (1992). In J. P. Aggleton (Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 485–503). New York: Wiley-Liss. Ahn, S. J., Bailenson, J., Fox, J., & Jabon, M. (2010). Using automated facial expression analysis for emotion and behaviour prediction. The Routledge handbook of emotions and mass media. New York: Taylor and Francis, 349–369. App, B., McIntosh, D., & Reid, C. A. (2007). A social–functional approach to emotion communication. Poster presented at the Annual Meeting of the Society for Personality and Social Psychology, 2007. Bachorowski, J. A. (1999). Vocal expression and perception of emotion. Current Directions in Psychological Science, 8(2), 53–57. Banse, R., & Scherer, K. R. (1996). Acoustic profiles in vocal emotion expression. Journal of Personality and Social Psychology, 70(3), 614–636. Barrett, L. F., Mesquita, B., Ochsner, K. N., & Gross, J. J. (2007). The experience of emotion. Annual Review of Psychology, 58, 373–403. Bellisle, F., Louis–Sylvestre, J., Linet, N., Rocaboy, B., Dalle, B., Cheneau, F., et al. (1990). Anxiety and food intake in man. Psychosomatic Medicine, 52(4), 452–457. Benton, D. (2002). Carbohydrate ingestion, blood glucose and mood. Neuroscience Biobehavioral Reviews, 26, 293–308. Benton, E., & Donohoe, R. T. (1999). The effects of nutrients on mood. Public Health Nutrition, 2, 403–409. Berridge, K. C., & Winkielman, P. (2003). What is an unconscious emotion? (the case for unconscious “liking”). Cognition and Emotion, 17(2), 181–211. Bhumiratana, N., Adhkari, K., & Chambers, E., IV (2014). The development of an emotion lexicon for the coffee drinking experience. Food Research International, 61, 83–92. Boon, B., Stroebe, W., Schut, H., & IJntema, R. (2002). Ironic processes in the eating behaviour of restrained eaters. British Journal of Health Psychology, 7, 1–10. Boon, B., Stroebe, W., Schut, H., & Jansen, A. (1998). Food for thought: Cognitive regulation of food intake. British Journal of Health Psychology, 3, 27–40. Bruch, H. (1973). Eating disorder: Obesity and anorexia nervosa. Riverside: Behavior Science Book Service. Cacioppo, J. T., Berntson, G. G., Larsen, J. T., Poehlmann, K. M., & Ito, T. A. (2000). The psychophysiology of emotion. In M. Lewis, & J. M. Haviland-Jones (Eds.), The handbook of emotion (pp. 173–191). New York: Guildford Press. Cahill, L., Haier, R. J., Fallon, J., Alkire, M. T., Tang, C., Keator, D., et al. (1996). Amygdala activity at encoding correlated with long-term, free recall of emotional information. Proceedings of the National Academy of Sciences of the United States of America, 93(15), 8016–8021. Canetti, L., Bachar, E., & Berry, E. M. (2002). Food and emotion. Behavioural Processes, 60, 157–164. Cardello, A. V. (2007). Measuring consumer expectations to improve food product development. In H. J. H. MacFie (Ed.), Consumer-led food product development (pp. 223–261). Abbington Cambridge UK: Woodhead Publishing. Cardello, A. V., Meiselman, H. L., Schutz, H. G., Graig, C., Given, Z., Lesher, L. L., et al. (2012). Measuring emotional responses to foods and food names using questionnaires. Food Quality and Preference, 24, 243–250. Chrea, C., Grandjean, D., Delplanque, S., Cayeux, I., Le Calvé, B., Aymard, L., et al. (2009). Mapping the semantic space for the subjective experience of emotional responses to odours. Chemical Senses, 34, 49–62. Christensen, L. (1993). Effects of eating behaviour on mood: A review of the literature. International Journal of Eating Disorders, 14(2), 171–183. Clore, G. L. (1994). Why emotions are never unconscious. In P. Ekman, & R. J. Davidson (Eds.), The nature of emotion: Fundamental questions (pp. 285–290). New York: Oxford University Press. Davidson, R. J. (1999). Neuropsychological perspectives on affective styles and their cognitive consequences. In T. Dalgleish, & M. J. Power (Eds.), Handbook of cognition and emotion (pp. 103–123). New York: Wiley. Davidson, R. J., Ekman, P., Saron, C. D., Senulis, J. A., & Friesen, W. V. (1990). Approachwithdrawal and cerebral asymmetry: Emotional expression and brain physiology. Journal of Personality and Social Psychology, 58(2), 330–341.

11

Degel, J., & Köster, E. P. (1999). Odors: implicit memory and performance effects. Chemical Senses, 24, 317–325. Degel, J., Piper, D., & Köster, E. P. (2001). Implicit learning and implicit memory for odors: The influence of odor identification and retention time. Chemical Senses, 26, 267–280. De Wijk, R. A., & Zijlstra, S. (2012). Differential effects of exposure to ambient vanilla and citrus aromas on mood, arousal and food choice. Flavour, 1, 24. Delplanque, S., Grandjean, Chrea, C., Aymard, L., Cayeux, I., Le Calvé, B., et al. (2008). Emotional processing of odors: Evidence for a nonlinear relation between pleasantness and familiarity evaluations. Chemical Senses, 33, 469–479. Delplanque, S., Grandjean, D., Chrea, C., Coppin, G., Aymard, L., Cayeux, I., et al. (2009). Sequential unfolding of novelty and pleasantness appraisals of odors: Evidence from facial electromyography and autonomic reactions. Emotion, 9(3), 316–328. Desmet, P. M. A. (2003). Measuring emotions. In M. A. Blythe, K. Overbeeke, A. F. Monk, & P. C. Wright (Eds.), Funology: From usability to enjoyment (pp. 111–125). Dordrecht, The Netherlands: Kluwer. Desmet, P. M. A., & Schifferstein, H. N. J. (2008). Sources of positive and negative emotions in food experience. Appetite, 50, 290–301. Drozdova, N. (2014). Measuring emotions in marketing and consumer behaviour: Is face reader an applicable tool? Master thesis Bergen, Norway: Norwegian School of Economics. Edwards, J. S. A., Meiselman, H. L., Edwards, A., & Lesher, L. (2003). The influence of eating locations on the acceptability of identically prepared foods. Food Quality and Preference, 14, 647–652. Ekman, P. (1972). Universals and cultural differences in facial expression of emotion. In J. R. Cole (Ed.), Nebraska symposium on motivation, Vol. 19. (pp. 207–283). Lincoln: University of Nebraska Press. Ekman, P. (1994). Strong evidence for universals in facial expressions: A reply to Russel's mistaken critique. Psychological Bulletin, 115(2), 268–287. Ekman, P., & Friesen, W. P. (1971). Constants across cultures in the face and emotion. Journal of Personality and Social Psychology, 17(2), 124–129. Ellsworth, P. C. (1994). Levels of thought and levels of emotion. In P. Ekman, & R. J. Davidson (Eds.), The nature of emotion: Fundamental questions (pp. 192–196). New York: Oxford University Press. Elsworth, P. C. (1995). The right way to study emotion. Psychological Inquiry, 6, 213–216. Ellsworth, P. C., & Scherer, K. R. (2003). Appraisal processes in emotion. In R. J. Davidson, H. H. Goldsmith, & K. R. Scherer (Eds.), Handbook of the affective sciences (pp. 572–595). New York/ Oxford, UK: Oxford University Press. Finn, A., & Louvière, J. J. (1992). Determining the appropriate response to evidence of public concern: The case for food safety. Journal of Public Policy and Marketing, 11(1), 12–25. Frijda, N. H. (1999). Emotions and hedonic experience. In D. Kahneman, E. Diener, & N. Schwarz (Eds.), Well-being: The foundations of hedonic psychology (pp. 190–210). New York: Russell Sage Foundation. Gibson, E. L. (2006a). Emotional influences on food choice: Sensory, physiological and psychological pathways. Physiology and Behavior, 89, 53–61. Gibson, E. L. (2006b). Mood, emotions and food choice. In R. Shepherd, & M. Raats (Eds.), Psychology of food choice (pp. 113–140). Wallingford: CAB International. Gibson, E. L., & Green, M. W. (2002). Nutritional influences on cognitive function: Mechanisms of susceptibility. Nutrition Research Review, 15, 169–206. Greeno, G. G., & Wing, R. R. (1994). Stress-induced eating. Psychological Bulletin, 115, 444–464. Haddock, C. K., & Dill, P. L. (2000). The effects of food on mood and behavior: Implications for the addictions model of obesity and eating disorders. Physiology and Behavior, 70, 333–342. Herman, C. P., & Mack, D. (1975). Restraint and unrestrained eating. Journal of Personality, 43, 647–660. Herman, C. P., & Polivy, J. (1980). Restrained eating. In A. B. Stunkard (Ed.), Obesity (pp. 208–225). Philadelphia: Saunders. Herman, C. P., & Polivy, J. (1984). A boundary model for the regulation of eating. Eating Disorders, 141–156. Holland, R. W., Hendriks, M., & Aarts, H. (2005). Smells like clean spirit. Nonconscious effects of scent on cognition and behaviour. Psychol Science, 16, 689–693. Jansen, A. (1998). A learning model of binge eating: Cue reactivity and cue exposure. Behaviour Research and Therapy, 36, 257–272. Jiang, Y., King, J. M., & Priniawatkul, W. (2014). A review of measurement and relationships between food, eating behavior and emotion. Trends in Food Science Technology, 7(36), 15–28. Johnstone, T., & Scherer, K. R. (2000). Vocal communication of emotion. In M. Lewis, & J. M. Haviland-Jones (Eds.), Handbook of emotions (pp. 220–235). New York: Guilford Press. Juillard, E. (2003). Lexique émotionnel et consommation alimentaire: le cas plus spécifique des produits carne´s auprès de la population féminine (Emotional lexicon and food consumption: the specific case of meat products in women). Doctoral dissertation Clermont-Ferrand: University of Psychology (2003). Kagan, J. (2007). A trio of concerns. Perspectives on Psychological Science, 2(4), 361–376. Kaplan, H. I., & Kaplan, H. S. (1957). The psychosomatic concept of obesity. Journal of Nervous and Mental Diseases, 125, 181–201. Kihlstrom, J. F., Mulvany, S., Tobias, B. A., & Tobis, J. P. (2000). The emotional unconscious. In E. Eich (Ed.), Cognition and emotion. New York: Oxford University Press. King, S. C., & Meiselman, H. L. (2010). Development of a method to measure consumer emotions associated with foods. Food Quality and Preference, 21, 168–177. King, S. C., Meiselman, H. L., & Carr, B. T. (2010). Measuring emotions associated with foods in consumer testing. Food Quality and Preference, 21, 1114–1116. King, S. C., Meiselman, H. L., & Carr, B. T. (2013). Measuring emotions associated with foods: Important elements of questionnaire and test design. Food Quality and Preference, 28, 8–16. Köster, E. P. (1996). The consumer? The quality? Production industrielle & qualité sensorielle; Agoral 96. Huitièmes rencontres scientifiques et technologiques des industries alimentaires 2-3 Avril 1996 — Dijon, 11–19.

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006

12

E.P. Köster, J. Mojet / Food Research International xxx (2015) xxx–xxx

Köster, E. P. (2003). The psychology of food choice: Some often encountered fallacies. Food Quality and Preference, 14, 359–373. Köster, E. P. (2009). Diversity in the determinants of food choice: A psychological perspective. Food Quality and Preference, 20(2), 70–82. Köster, E. P., & Mojet, J. (2007). Boredom and the reasons why some new products fail. In H. J. H. MacFie (Ed.), Consumer-led food product development (pp. 262–280). Abington, Cambridge, UK: Woodhead Publishing Limited. Köster, E. P., & Mojet, J. (2012a). Analysis of flops: A useful tool for future innovations. Part 1: The diagnostic phase. Agro Food Industry hi-tech, 23(1), 6–9. Köster, E. P., & Mojet, J. (2012b). Analysis of flops: A useful tool for future innovations. Part 2: The reduction of future flop risks. Agro Food Industry hi-tech, 23(2), 6–10. Köster, E. P., Møller, P., & Mojet, J. (2014). A “misfit theory of spontaneous conscious odor perception (MITSCOP)”: Reflections on the role and function of odor memory in everyday life. Frontiers in Psychology, 5, 64. http://dx.doi.org/10.3389/fpsyg.2014.00064. Köster, M. A., Prescott, J., & Köster, E. P. (2004). Incidental learning and memory for three basic tastes in food. Chemical Senses, 29, 441–453. Kuesten, C., Chopra, P., Bi, J., & Meiselman, H. L. (2014). A global study using PANAS (PA and NA) scales to measure consumer emotions associated with aromas of phytonutrient supplements. Food Quality and Preference, 33, 86–97. Laros, F. J. M., & Steenkamp, J. -B. E. M. (2005). Emotions in consumer behaviour: A hierarchical approach. Journal of Business Research, 58, 1437–1445. Lattimore, P., & Caswell, N. (2004). Differential effects of active and passive stress on food intake in restrained and unrestrained eaters. Appetite, 42, 167–173. Lattimore, P., & Maxwell, L. (2004). Cognitive load, stress, and disinhibited eating. Eating Behaviors, 5, 315–324. Laureati, M., Morin-Audebrand, L., Pagliarini, E., Sulmont-Rossé, C., Köster, E. P., & Mojet, J. (2008). Influence of age and liking on food memory: An incidental learning experiment with children, young and elderly people. Appetite, 51, 273–282. LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience, 23, 155–184. Li, W. (2014). Learning to smell danger: Acquired associative representation of threat in the olfactory cortex. Frontiers in Behavioral Neuroscience, 8, 98. Macht, M. (1999). Characteristics of eating in anger, fear, sadness, and joy. Appetite, 33, 129–139. Macht, M. (2008). How emotions affect eating: A five-way model. Appetite, 50, 1–11. Macht, M., & Dettmer, D. (2006). Everyday mood and emotions after eating a chocolate bar or an apple. Appetite, 46, 332–336. Macht, M., Meininger, J., & Roth, J. (2005). The pleasure of eating. A qualitative analysis. Journal of Happiness Studies, 6, 137–160. Macht, M., & Müller, J. (2007). Immediate effects of chocolate on experimentally induced mood states. Appetite, 49(3), 667–674. http://dx.doi.org/10.1016/j. appet.2007.05.004. Markus, C. R., Panhuysen, G., Tuiten, A., & Koppeschaar, H. (2000). Effects of food on cortisol and mood in vulnerable subjects under controllable and uncontrollable stress. Physiology and Behavior, 70, 333–342. Markus, C. R., Panhuysen, G., Tuiten, A., Koppeschaar, H., Fekkes, D., & Peters, M. L. (1998). Does carbohydrate-rich, protein-poor food prevent a deterioration of mood and cognitive performance of stress-prone subjects when subjected to a stressful task? Appetite, 31, 49–65. Mathews, G., & Deary, I. J. (1998). Personality traits. Cambridge: Cambridge University Press. Mauss, I. B., Levenson, R. W., McCarter, L., Wilhelm, F. H., & Gross, J. J. (2005). The tie that binds? Coherence among emotion experience, behavior, and physiology. Emotion, 5(2), 175–190. Mauss, I. B., & Robinson, M. D. (2009). Measures of emotion: A review. Cognition and Emotion, 23(2), 209–237. Mayer, J. D., & Gaschke, Y. N. (1968). The experience and meta-experience of mood. Journal of Personality and Social Psychology, 55, 102–111. McNair, D. M., Lorr, M., & Droppelman, L. F. (1971). Profile of mood states. San Diego: Educational and Industrial Testing Service. Meiselman, H. L. (2013). The future in sensory/consumer research: …… evolving to a better science. Food Quality and Preference, 27, 208–214. Mojet, J., Dürrschmid, K., Danner, L., Joch, M., & Köster, E. P. (2014n). Are implicit emotion measurements independent from liking? (submitted for publication: this issue). Mojet, J., & Köster, E. P. (2002). Texture and flavour memory in foods: An incidental learning experiment. Appetite, 38, 110–117. Mojet, J., & Köster, E. P. (2005). Sensory memory and food texture. Food Quality and Preference, 16, 251–266. Møller, P., Mojet, J., & Köster, E. P. (2007). Incidental and intentional flavour memory in young and older subjects. Chemical Senses, 32(6), 557–567. Monahan, J. L., Murphy, S. T., & Zajonc, R. B. (2000). Subliminal mere exposure: Specific, general and diffuse effects. Psychological Science, 11(6), 462–466. Morin-Audebrand, L., Mojet, J., Chabanet, C., Issanchou, S., Møller, P., Köster, E. P., et al. (2012). The role of novelty detection in food memory. Acta Psychologica, 139, 233–238. Moskowitz, H. (2000). Inter-relating data sets for product development: The reverse engineering approach. Food Quality and Preference, 11(s 1–2), 105–119. Murphy, F. C., Nimmo-Smith, I., & Lawrence, A. D. (2003). Functional neuroanatomy of emotions: A meta-analysis. Cognitive, Affective, & Behavioral Neuroscience, 3(3), 207–233. Nederkoorn, C., Smulders, F., Havermans, R., & Jansen, A. (2004). Exposure to binge food in bulimia nervosa: Finger pulse amplitude as a potential measure of urge to eat and predictor of food intake. Appetite, 42, 125–130. Nederkoorn, C., Smulders, F. T. Y., & Jansen, A. (2000). Cephalic phase responses, craving and food intake in normal subjects. Appetite, 35, 45–55. Ng, M., Chaya, C., & Hort, J. (2013). Beyond liking: Comparing the measurement of emotional response using EsSense profile and consumer defined check-all-that-apply methodologies. Food Quality and Preference, 28, 193–205.

Phan, K. L., Wager, T. D., & Liberzon, I. (2002). Functional neuroanatomy of emotion: A meta-analysis of emotion activation studies in PET and fMRI. NeuroImage, 16, 331–348. Piqueras-Fiszman, B., & Jaeger, S. R. (2014a). The impact of evoked consumption contexts and appropriateness on emotion responses. Food Quality and Preference, 32, 277–288. Piqueras-Fiszman, B., & Jaeger, S. R. (2014b). Emotion responses under evoked consumption contexts: A focus on the consumers' frequency of product consumption and the stability of responses. Food Quality and Preference, 35, 24–41. Piqueras-Fiszman, B., & Jaeger, S. R. (2014c). The impact of the means of context evocation on consumer's emotion associations towards eating occasions. Food Quality and Preference, 37, 61–70. Porcherot, C., Delplanque, S., Planchais, A., Gaudeau, N., & Cayeux, I. (2013). Seeing, smelling, feeling! Is there an influence of color on subjective affective responses to perfumed fabric softeners? Food Quality and Preference, 27, 161–169. Porcherot, C., Delplanque, S., Ravot-Derrien, S., Le Calvé, B., Chréa, C., Gaudreau, N., et al. (2010). How do you feel when you sell this ? Optimization of a verbal measurement of odor-elicited emotions. Food Quality and Preference, 21, 938–947. Porcherot, C., Petit, E., Giboreau, A., Gaudreau, N., & Cayeux, I. (2015). Measurement of self-reported feelings when an aperitif is consumed in an ecological setting. Food Quality and Preference, 39, 277–284. Pudel, V., & Westenhöfer, J. (1991). Ernährungspsychologie. Göttingen: Hogrefe Verlag. Reimann, E. M., Lane, R. D., Ahorn, G. L., & Schwarz, G. E. (1997). Neuroanatomical correlates of externally and internally generated human emotion. American Journal of Psychiatry, 154, 918–925. Richins, M. L. (1997). Measuring emotions in the consumption experience. Journal of Consumer Research, 24, 127–146. Rolls, E. T. (1999). The brain and emotion. Oxford: Oxford University Press. Rousset, S., Deiss, V., Juillard, E., Schlich, P., & Droit-Volet, S. (2005). Emotions generated by meat and other food products in women. British Journal of Nutrition, 94, 609–616. Schachter, S. (1968). Obesity and eating. Science, 161, 751–756. Schachter, S. (1971). Some extraordinary facts about obese humans and rats. American Psychologist, 26, 129–144. Schachter, S., Goldman, R., & Gordon, A. (1968). Effects of fear, food deprivation, and obesity on eating. Journal of Personality and Social Psychology, 10, 91–97. Schifferstein, H. J. N., Mojet, J., & Kole, A. P. W. (1999). Asymmetry in the disconfirmation of expectations for natural yoghurt. Appetite, 32, 307–329. Slochower, J. (1983). Excessive eating: The role of emotions and environment. New York: Human Sciences Press. Spinelli, S., Masi, C., Dinnella, C., Zoboli, G. P., & Monteleone, E. (2014). How does it make you feel? A new approach to measuring emotions in food experience. Food Quality and Preference, 37, 109–122. Spinelli, S., Masi, C., Zoboli, G. P., Prescott, J., & Monteleone, E. (2015). Emotional responses to branded and unbranded foods. Food Quality and Preference, 42, 1–11. Thayer, R. E. (1989). The biopsychology of mood and arousal. New York: Oxford University Press. Thomson, D. M. H., & Crocker, C. (2013). A data-driven classification of feelings. Food Quality and Preference, 27, 137–152. Thomson, D. M. H., Crocker, C., & Marketo, C. G. (2010). Linking sensory characteristics to emotions: An example using dark chocolate. Food Quality and Preference, 21, 1117–1125. Vreugdenberg, L., Bryan, J., & Kemps, E. (2003). The effect of self-initiated weight-loss dieting on working memory: The role of preoccupying cognitions. Appetite, 41, 291–300. Wallis, D. J., & Hetherington, M. M. (2004). Stress and eating: The effects of ego-threat and cognitive demand on food intake in restrained and emotional eaters. Appetite, 43, 39–46. Ward, A., & Mann, T. (2000). Don't mind if I do: Disinhibited eating under cognitive load. Journal of Personality and Social Psychology, 78, 753–763. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070. Whalen, P. J., Rauch, S. L., Etcoff, N. L., McInerney, S. C., Lee, M. B., & Jenike, M. A. (1998). Masked presentations of emotional facial expressions modulate amygdala activity without explicit knowledge. Journal of Neuroscience, 18(1), 411–418. Wiemers, U. S., Sauvage, M. M., & Wolf, O. T. (2014). Odors as effective retrieval cues for stressful episodes. Neurobiology of Learning and Memory, 112, 230–235. Winkielman, P., Berridge, K., & Scher, S. (2011). Emotion, consciousness and social behavior. In J. Decety, & J. T. Cacioppo (Eds.), Handbook of social neuroscience (pp. 195–211). Oxford: Oxford University Press. Winkielman, P., Berridge, K. C., & Wilbarger, J. (2005). Unconscious affective reactions to masked happy versus angry faces influence consumption behavior and judgments of value. Personality and Social Psychology Bulletin, 31(1), 121–135. Wurtman, R. J., Hefti, F., & Melamed, E. (1981a). Precursor control of neurotransmitter synthesis. Pharmacological Reviews, 32, 315–335. Wurtman, J. J., Wurtman, R. J., Growdon, J. H., Henry, P., Lipscomb, A., & Zeisel, S. H. (1981b). Carbohydrate craving in obese people: Suppression by treatments affecting serotoninergic transmission. International Journal of Eating Disorders, 1, 2–15. Zucco, G. M., Aiello, L., Turuani, L., & Köster, E. P. (2012). Odor-evoked autobiographical memories: Age and gender differences along the life span. Chemical Senses, 37, 179–189. Zuckerman, M., & Lubin, B. (1965). Manual for the multiple affect adjective check list. San Diego: Educational and Industrial Testing Service. Zuckerman, M., & Lubin, B. (1985). The multiple affect adjective check list revised. San Diego: Educational and Industrial Testing Service.

Please cite this article as: Köster, E.P., & Mojet, J., From mood to food and from food to mood: A psychological perspective on the measurement of food-related emotions in consumer research, Food Research International (2015), http://dx.doi.org/10.1016/j.foodres.2015.04.006