Elsevier

Appetite

Volume 104, 1 September 2016, Pages 33-43
Appetite

Distinguishing the affective and cognitive bases of implicit attitudes to improve prediction of food choices

https://doi.org/10.1016/j.appet.2015.10.005Get rights and content

Highlights

  • Implicit attitudes are driven by automatic affective and cognitive reactions.

  • Affective and cognitive bases of implicit attitudes directly influence food choice.

  • The two bases of implicit attitudes uniquely drive choice in different conditions.

  • Automatic perceived tastiness uniquely drives food choice under cognitive load.

  • Automatic perceived healthiness only influences choice of low impulsive people.

Abstract

Eating behaviors largely result from automatic processes. Yet, in existing research, automatic or implicit attitudes toward food often fail to predict eating behaviors. Applying findings in cognitive neuroscience research, we propose and find that a central reason why implicit attitudes toward food are not good predictors of eating behaviors is that implicit attitudes are driven by two distinct constructs that often have diverging evaluative consequences: the automatic affective reactions to food (e.g., tastiness; the affective basis of implicit attitudes) and the automatic cognitive reactions to food (e.g., healthiness; the cognitive basis of implicit attitudes). More importantly, we find that the affective and cognitive bases of implicit attitudes directly and uniquely influence actual food choices under different conditions. While the affective basis of implicit attitude is the main driver of food choices, it is the only driver when cognitive resources during choice are limited. The cognitive basis of implicit attitudes uniquely influences food choices when cognitive resources during choice are plentiful but only for participants low in impulsivity. Researchers interested in automatic processes in eating behaviors could thus benefit by distinguishing between the affective and cognitive bases of implicit attitudes.

Introduction

Human behaviors in general, and eating behaviors, in particular, are largely influenced by automatic processes (Bargh, 1997, Rangel, 2013, Strack and Deutsch, 2004, Wiers et al., 2010). Imagine having to choose between an apple and a chocolate bar for dessert. The associations that spontaneously come to your mind for each food item are likely to influence your decision (Raghunathan et al., 2006, Strack and Deutsch, 2004, Stroebe et al., 2008). However, although implicit attitudes toward food (i.e., the automatic evaluative reaction toward the item) are shown to sometimes influence food choices and behaviors (e.g., Friese, Hofmann, & Wänke, 2008), implicit attitudes often fail to predict eating behaviors (for a review see Roefs et al., 2011). In particular, implicit attitudes fail to predict the behavior of obese people (Craeynest et al., 2007, Roefs et al., 2005a, Roefs and Jansen, 2002) and restrained eaters (e.g., Papies et al., 2009, Roefs et al., 2005b).

In this article, we argue that one central reason why implicit attitudes are not always good predictors of eating behaviors is because they are not only driven by the automatic hedonic or affective reactions to food (e.g., tastiness) but also by the automatic utilitarian or cognitive reactions to food (e.g., healthiness). And these automatic affective and cognitive reactions to food often do not have the same evaluative consequences. A piece of chocolate cake will usually be perceived as tasty (which has a positive evaluative outcome) and, at the same time, as unhealthy (which has a negative evaluative outcome). Automatic perceived tastiness and healthiness can even be negatively correlated (Keller and van der Horst, 2013, Raghunathan et al., 2006).

Based on cognitive neuroscience research on systems of implicit learning and memory (Amodio and Devine, 2006, Amodio and Ratner, 2011, Rangel, 2013), we propose that the affective and cognitive bases of implicit attitudes towards a food item are distinct constructs that independently build the conventional overall implicit attitude toward the item. More importantly, we propose that each basis directly and uniquely influences eating behaviors under different conditions. In line with recent theoretical developments stressing the importance of palatability in explaining food intake (e.g., Lowe and Butryn, 2007, Rangel, 2013), we expect the affective basis of implicit attitudes (i.e., automatic hedonic reactions to food) to be the main driver of eating behaviors, and the only driver when cognitive resources during choice are limited (Mann & Ward, 2007). We expect the cognitive basis of implicit attitudes (i.e., automatic cognitive reactions to food) to predict food choice only when cognitive resources during choice are plentiful and only for participants good at self-control (Hare et al., 2009, Rangel, 2013).

We explore these predictions through two studies. Our first study examines the influence of affective and cognitive automatic reactions to food on overall implicit attitudes towards the food items and shows that both types of reactions independently influence overall implicit attitudes. Our second study investigates how affective and cognitive automatic reactions to food explain food choices under different conditions. This study also shows that implicit attitudes are not good predictors of food choices because they mix both affective and cognitive automatic reactions. We begin with a brief review of prior work providing evidence for affective and cognitive bases of implicit attitudes and for the role of both bases in eating behaviors. We then present our studies and discuss theoretical and practical implications of this research.

Eating behaviors are not only the result of controlled processes but also and predominantly the result of automatic processes (Rangel, 2013, Wiers et al., 2010). Automatic processes are fast, unintentional and effortless (Bargh, 1994). To account for the importance of automatic processes and to better understand eating behaviors, researchers focused on assessing automatic or implicit attitudes (Chen and Bargh, 1999, Roefs et al., 2011, Wiers et al., 2010). Yet, implicit attitudes often fail to predict eating behaviors (e.g., Karpinski and Hilton, 2001, Olson and Fazio, 2004, Spruyt et al., 2007; for a review, see Roefs et al., 2011). We suggest that a key reason why implicit attitudes are not good predictors of eating behaviors is that multiple memory systems actually contribute to automatic evaluative processes (Amodio and Ratner, 2011, Amodio and Devine, 2006, Rangel, 2013, Stanley et al., 2008). In particular, there is evidence 1) that affective and cognitive forms of information processing can be performed automatically, 2) that both processes are distinct and 3) that both processes can influence implicit attitudes (Amodio and Devine, 2006, Amodio and Mendoza, 2010). We briefly review such evidence in the next part.

While traditional models of information processing assume that a uniform processing mode characterizes automatic processes (e.g., Sloman, 1996, Smith and Decoster, 2000, Strack and Deutsch, 2004), findings in cognitive neuroscience research clearly support the view of distinct automatic systems (Amodio and Ratner, 2011, Poldrack and Foerde, 2008). In particular, affective learning and memory involve the amygdala and its related subcortical circuits (LeDoux, 2000) while cognitive (or semantic) learning and memory are associated with activity in evolutionary newer network of neocortical structures (e.g. left prefrontal cortex; Martin, 2007, Rissman et al., 2003). Importantly, both affective and cognitive processes can operate automatically. It is well known that Pavlovian learning (i.e., classical conditioning), one of the fundamental systems for the learning of affective associations in the food domain, can operate automatically (Rangel, 2013). Yet, the mechanisms producing cognitive associations can also operate automatically. This can, for instance, be observed in implicit semantic priming tasks (Rissman et al., 2003). Seeing a word such as “chocolate” spontaneously activates part of the semantic network of the word. It has even been shown that affective associations are not necessarily more accessible than semantic associations (Giner-Sorolla, 2004). The two systems, affective and cognitive, can thus operate automatically. Interestingly, they have also been shown to be distinct because (1) amygdala-based learning is not dependent on semantic associations and (2) semantic associations can be learned without involving the amygdala (Bechara, Damasio, & Damasio, 2003).

It is well established that affective associations influence implicit attitudes. Indeed, the amydgala is strongly associated with automatic evaluation of stimuli (Cunningham and Zelazo, 2007, Stanley et al., 2008). Some researchers even adopt the view that implicit attitudes “represent the affective component attributed to attitudes” (Gawronski & Bodenhausen, 2006, p. 694). However, besides the amygdala, neural components associated to cognitive learning and memory are also involved in automatic evaluations. In particular, the dorsolateral prefrontal cortex is involved in the cognitive regulations of implicit attitudes (Stanley et al., 2008). Moreover, such cognitive regulation can be performed automatically in a few hundred milliseconds (Cunningham & Zelazo, 2007). Thus, “an implicit evaluation (i.e., attitude) may reflect a combination of affective and semantic (i.e., cognitive) associations” (Amodio & Mendoza, 2010, p. 367; see Eagly & Chaiken (2007) for a similar view).

In research on traditional explicit (i.e., deliberative) attitudes, the conceptual distinction between the affective and cognitive bases of attitudes was useful to improve the prediction of behaviors, and this also in the food domain (Dubé et al., 2003, Millar and Tesser, 1986). We suggest that a similar distinction between the affective and cognitive bases of attitudes at an implicit level should improve the understanding of the relationship between implicit attitudes toward food and eating behaviors. We next define the affective and cognitive bases of implicit attitudes toward food.

Following research on explicit attitudes toward food (Dubé et al., 2003), the affective component of an implicit attitude toward a food item corresponds to the automatic hedonic reactions to the item (its spontaneous palatability) and the cognitive component contains the automatic beliefs about the item (e.g., spontaneous perceived healthiness, calories, fattiness, dieting effect). Using implicit measures, some researchers have independently assessed either the affective basis (e.g., Papies et al., 2007, Roefs et al., 2005b) or part of the cognitive basis (e.g., Stroebe et al., 2008, Werrij et al., 2009) of implicit attitudes toward food. Yet, to the best of our knowledge and quite surprisingly, no study concerns the prediction of eating behaviors. Because the affective and cognitive bases of implicit attitudes reflect independent memory systems, each basis should uniquely influence eating behaviors under different conditions (Amodio & Devine, 2006). We next detail such conditions.

There is ample theoretical and empirical evidence that food decisions are largely driven by hedonic reactions (e.g. tastiness) (Lowe and Butryn, 2007, Pinel et al., 2000). Indeed, palatability is one of the main drivers of food intake and is one major reason the homeostatic regulation of hunger does not well explain eating behaviors (Herman and Polivy, 2014, Rangel, 2013, Wansink and Chandon, 2014). In particular, the Pavlovian system, based on affect, often controls food decisions (Rangel, 2013). This system dominates when food decisions are made under cognitive load (Mann & Ward, 2007). We thus expect the affective basis of implicit attitudes to be the only driver of eating behaviors when cognitive resources are limited.

The Pavlovian system can be inhibited by a goal-directed system that computes the overall value of a food choice based on the attributes of the food, such as sweetness, water content, or health consequences (Rangel, 2013). Yet, such inhibition requires correct evaluations of food value; specifically, perceived healthiness should be used when computing the value of food. Often, this is not the case (Rangel, 2013). In most situations, value is based on basic attributes with immediate outcomes, above all, perceived tastiness. Only some individuals good at self-control spontaneously use more abstract attributes with delayed outcomes, such as health considerations, to compute the value of food (Hare et al., 2009). We thus expect the cognitive basis of implicit attitudes to be consistently used in food decisions only by some individuals, those having cognitive resources available and not having the tendency to act on immediate urges, such as those low in impulsivity (Deyoung, 2010, Rangel, 2013). Moreover, because the neural systems underlying the two bases of implicit attitudes influence behavioral responses independently, we expect each basis of implicit attitudes to uniquely influence behaviors under the above-specified conditions (Amodio & Devine, 2006).

A direct consequence of the two bases of implicit attitudes uniquely influencing behaviors under different conditions is that we do not expect overall implicit attitudes to be good predictors of eating behaviors. Indeed, implicit attitudes mix both affective and cognitive automatic reactions and these reactions often do not have similar evaluative consequences. For instance, high-fat palatable food will spontaneously be perceived as tasty (which has a positive evaluative outcome) and, at the same time, as unhealthy (which has a negative evaluative outcome) (Papies et al., 2007, Stroebe et al., 2008). Moreover, several studies suggest that the relation between the affective and cognitive bases of implicit attitudes is often negative (Keller and van der Horst, 2013, Raghunathan et al., 2006, Stroebe et al., 2008). The healthier a food item, the worst its spontaneous tastiness (Raghunathan et al., 2006). This could explain the modest predictive validity of implicit attitudes in past studies (Roefs et al., 2011).

In the present research we propose to show 1) that the affective and cognitive bases of implicit attitudes toward food are distinct constructs that independently build the overall implicit attitude toward food and 2) that each basis of implicit attitudes directly and uniquely influences food choices under different circumstances. Because both affective and cognitive reactions should drive overall implicit attitudes, we do not expect overall implicit attitudes to be good predictors of choice. Fig. 1 summarizes the theoretical model of this research.

To test these predictions, we used the implicit associations test (IAT) because it is extensively used to assess implicit attitudes toward food (Roefs et al., 2011) and because it is one of the most reliable implicit measures (Hofmann, Gawronski, Gschwendner, Le, & Schmitt, 2005). The IAT is a method for indirectly measuring the strength of associations between concepts (Greenwald, McGhee, & Schwartz, 1998). We designed separate IATs to assess the affective and cognitive bases of implicit attitude, and also overall implicit attitude. Because the IAT is a relative measure, we actually assessed implicit preferences between apples and chocolate. To assess the individual impact of each basis of implicit attitudes on eating behavior, we considered food choice between apples and chocolate bars. This should ensure direct comparison with previous findings because numerous studies on the influence of implicit attitudes on eating behaviors used food choices with similar products (e.g., choice between fruit and chocolate in Conner et al., 2007, Friese et al., 2008). In study 1, we test our hypothesis that the cognitive and affective bases of implicit attitudes are distinct constructs that independently drive overall implicit attitudes. We also assess the construct validity of the two bases of implicit attitudes by comparing the scores of high and low-restrained eaters. In study 2, we specifically examine how cognitive and affective bases of implicit attitudes and the overall implicit attitude predict food choice.

Section snippets

Participants and procedure

283 French undergraduate students from a major business school (mean age = 20.6 years, 57.6% female, mean body mass index (BMI) = 21.43, SD = 2.50) participated in this correlational study in exchange for course credit. Participants first performed three IAT measures assessing the affective basis of implicit preference, the cognitive basis of implicit preference (order of these 2 IATs were counterbalanced between participants to minimize common method variance) and the overall implicit

Study 2

In this study, we examined the degree to which the two bases of implicit attitudes influence actual food choice. We used a real choice task between apples and chocolate bars. We expected the affective basis of implicit attitudes to be the main driver of choice. We manipulated cognitive load during choice because we hypothesized that the affective basis of implicit attitudes would be the only predictor of choice under cognitive load. We also hypothesized that the cognitive basis of implicit

General discussion

Two main sets of findings are produced by this research. First, in both studies we find that implicit attitudes toward food have distinct affective and cognitive bases and that both bases drive implicit attitudes independently. Consistent with findings in cognitive neuroscience research on systems of implicit learning and memory (e.g., Amodio & Ratner, 2011), overall implicit attitudes toward food are not only driven by automatic perceived tastiness (i.e., a measure of the affective basis of

References (76)

  • E.K. Papies et al.

    Healthy dining: subtle diet reminders at the point of purchase increase low-calorie food choices among both chronic and current dieters

    Appetite

    (2013)
  • R.A. Poldrack et al.

    Category learning and the memory systems debate

    Neuroscience and Biobehavioral Reviews

    (2008)
  • A. Roefs et al.

    At first sight: how do restrained eaters evaluate high-fat palatable foods?

    Appetite

    (2005)
  • A. Roefs et al.

    The environment influences whether high-fat foods are associated with palatable or with unhealthy

    Behaviour Research and Therapy

    (2006)
  • A. Roefs et al.

    Early associations with food in anorexia nervosa patients and obese people assessed in the affective priming paradigm

    Eating Behaviors

    (2005)
  • A. Spruyt et al.

    On the predictive validity of indirect attitude measures: prediction of consumer choice behavior on the basis of affective priming in the picture–picture naming task

    Journal of Experimental Social Psychology

    (2007)
  • S. Stieger et al.

    Personalizing the IAT and the SC-IAT: Impact of idiographic stimulus selection in the measurement of implicit anxiety

    Personality and Individual Differences

    (2010)
  • W. Stroebe et al.

    Why dieters fail: testing the goal conflict model of eating

    Journal of Experimental Social Psychology

    (2008)
  • B. Wansink et al.

    Slim by design: redirecting the accidental drivers of mindless overeating

    Journal of Consumer Psychology

    (2014)
  • C.O.C. Werle et al.

    Unhealthy food is not tastier for everybody: the “healthy=tasty” French intuition

    Food Quality and Preference

    (2013)
  • M.Q. Werrij et al.

    Early associations with palatable foods in overweight and obesity are not disinhibition related but restraint related

    Journal of Behavior Therapy and Experimental Psychiatry

    (2009)
  • L.S. Aiken et al.

    Multiple regression: Testing and interpreting interactions

    (1991)
  • D.M. Amodio et al.

    Stereotyping and evaluation in implicit race bias: evidence for independent constructs and unique effects on behavior

    Journal of Personality and Social Psychology

    (2006)
  • D.M. Amodio et al.

    Implicit intergroup bias: cognitive, affective, and motivational underpinnings

  • D.M. Amodio et al.

    A memory systems model of implicit social cognition

    Current Directions in Psychological Science

    (2011)
  • J.A. Bargh

    The four horsemen of automaticity: awareness, intention, efficiency, and control in social cognition

  • J.A. Bargh

    The automaticity of everyday life

  • A. Bechara et al.

    Role of the amygdala in decision-making

    Annals of the New York Academy of Sciences

    (2003)
  • M. Chen et al.

    Consequences of automatic evaluation: immediate behavioral predispositions to approach or avoid the stimulus

    Personality and Social Psychology Bulletin

    (1999)
  • M.T. Conner et al.

    Relations between implicit and explicit measures of attitudes and measures of behavior: evidence of moderation by individual difference variables

    Personality & Social Psychology Bulletin

    (2007)
  • F.R. Conrey et al.

    Separating multiple processes in implicit social cognition: the quad model of implicit task performance

    Journal of Personality and Social Psychology

    (2005)
  • C.G. Deyoung

    Impulsivity as a personality trait

  • A.H. Eagly et al.

    The advantages of an inclusive definition of attitude

    Social Cognition

    (2007)
  • F. Foroni et al.

    Picture -IAT versus word -IAT: level of stimulus representation influences on the IAT

    European Journal of Social Psychology

    (2010)
  • M. Friese et al.

    When impulses take over: moderated predictive validity of explicit and implicit attitude measures in predicting food choice and consumption behaviour

    The British Journal of Social Psychology/The British Psychological Society

    (2008)
  • B. Gawronski et al.

    Associative and propositional processes in evaluation: an integrative review of implicit and explicit attitude change

    Psychological Bulletin

    (2006)
  • D.T. Gilbert et al.

    The trouble of thinking: activation and application of stereotypic beliefs

    Journal of Personality and Social Psychology

    (1991)
  • R. Giner-Sorolla

    Is affective material in attitudes more accessible than cognitive material? The moderating role of attitude basis

    European Journal of Social Psychology

    (2004)
  • Cited by (36)

    • To look tasty, let's show the ingredients! Effects of ingredient images on implicit tasty–healthy associations for packaged products

      2021, Journal of Retailing and Consumer Services
      Citation Excerpt :

      In contrast, the number of FOP ingredient images for unhealthy products does not alter product preferences. Our results thus confirm that implicit evaluations inform explicit choices (Trendel and Werle, 2016). For hedonic, unhealthy products, our results contrast with the explicit positive impact of the number of FOP ingredient images on perceived taste (Thomas and Capelli, 2018).

    • Consumers’ implicit attitudes toward corporate social responsibility and corporate abilities: Examining the influence of bank governance using the implicit association test

      2021, Journal of Retailing and Consumer Services
      Citation Excerpt :

      For ease of interpretation, we present the results in milliseconds (ms). Following Trendel and Werle (2016), we eliminated trial responses for which the latencies were less than 300 ms or greater than 3,000 ms as better data validity can be obtained. The significance levels remained unchanged if we retained all participants.

    View all citing articles on Scopus
    View full text