Research reportApproach avoidance training in the eating domain: Testing the effectiveness across three single session studies☆
Introduction
The steep rise of overweight in large parts of the world demonstrates that a healthy lifestyle has become exceedingly difficult with palatable, energy-dense food being easily available and often too tempting to resist (Ng et al, 2014, World Health Organization, 2012). The over-consumption of high-calorie food can cause severe health problems (e.g., heart disease, diabetes), which raises the acute question of how to help people change their eating patterns effectively. Research from outside the eating domain suggests that people can be trained to make healthier choices if they learn to automatically respond with avoidance movements to temptations (see Marteau, Hollands, & Fletcher, 2012). In the present studies we tested whether the training version of the approach avoidance task (AAT-training; Wiers, Eberl, Rinck, Becker, & Lindenmeyer, 2011) is effective in modifying automatic response tendencies toward tempting but unhealthy food.
Over the past years, dual-process models of health behavior have become increasingly influential in explaining why people often fail to control their behavior in accordance with their long-term health goals (Hofmann et al, 2008a, Strack, Deutsch, 2004). According to those models, unhealthy choices stem from impulsive (i.e., automatically activated approach) behavior toward rewarding stimuli (e.g., a chocolate cookie) which unfolds too fast to be controlled on time by the relatively slower and more effortful reflective processes (Gladwin, Figner, Crone, & Wiers, 2011). Indeed, the more rewarding people find a particular food the stronger their approach behavior (Brignell et al, 2009, Veenstra, de Jong, 2010), and the more likely they are to actually consume it (Hofmann et al, 2008b, Nederkoorn et al, 2010).
The dual-process conceptualization inspired a new class of interventions aimed at modifying automatic precursors of behavior (e.g., Friese et al, 2011, Veling et al, 2011, Wiers et al, 2013). One of those interventions is the AAT-training which has proven successful at changing unhealthy approach biases in the alcohol dependency domain. In the AAT-training, participants repeatedly make avoidance movements (e.g., pushing a joystick) in response to pictures depicting critical stimuli (e.g., alcohol beverages) and approach movements (e.g., pulling a joystick) to stimuli from a contrast category (e.g., soft drinks). Wiers and colleagues (Wiers, Rinck, Kordts, Houben, & Strack, 2010) showed that after a single session of AAT-training, student participants had developed an avoidance bias to alcohol pictures (i.e., relatively faster avoidance than approach responses), and successfully trained participants also consumed less alcohol in a subsequent taste test. The effect was replicated in a sample of alcohol-dependent patients and extended with the finding that training led to lower relapse rates at one year follow-up (Eberl et al, 2013, Wiers et al, 2011).
In the food domain, the effectiveness of AAT-training has yet to be established. Some indirect evidence comes from Kemps and colleagues (Kemps, Tiggeman, Martin, & Elliott, 2013) who trained participants to associate words related to avoidance with chocolate pictures and found decreases in implicit preferences and craving for chocolate. No behavioral measures were taken though. A more direct application of AAT-training using joysticks (Fishbach & Shah, 2006, Study 5) found that participants subsequently made healthier snack choices. This latter finding is qualified, however, by a small sample, lack of baseline measures and other methodological weaknesses (e.g., approach-avoidance responses were not disambiguated by a perceptual zooming feature; see below). Taken together, existing evidence is suggestive rather than conclusive and calls for additional and more rigorous testing.
In the present studies, we apply AAT-training to the domain of healthy eating, and investigate whether training to avoid high-calorie food stimuli and approach low-calorie food stimuli changes (implicit and explicit) food preferences and eating behavior. Given the successful application of AAT-training against alcohol addiction (Wiers et al., 2011) one might hypothesize that training can also change food preferences. However, preferences for fatty and sweet food have a very strong innate basis (Drewnowski, 1997) and might thus be less easily adjusted than preferences for alcohol. Moreover, the category ‘unhealthy food’ is much fuzzier than ‘alcoholic beverage’ which implies that in the latter case stimulus response associations are more easily processed (Rosch, Simpson, & Miller, 1976). Accordingly, successful AAT-training of food preferences should not be taken for granted but needs to be empirically tested. Irrespective of the outcome, such tests will have high informative value to all researchers and clinicians who want to target automatic preferences for unhealthy food.
Section snippets
General method
The present studies investigated the extent to which AAT-training can modify participants' implicit and explicit food preferences and eating behavior. To keep the sample homogenous with respect to general eating behavior and food-related attitudes, we only tested women (Rolls, Fedoroff, & Guthrie, 1991). In three single session studies participants were randomly assigned to experimental AAT-training in which they learned to avoid pictures of unhealthy food or to sham AAT-training without
Participants
Fifty-two students participated for course credit or €10. Sample size was determined based on Fishbach and Shah (2006). Due to a technical error the data of one participant were unusable, reducing the sample to fifty-one (Mage = 20.47, SD = 2.34, nexperimental = 26).
Materials and procedure
Participants signed informed-consent, reported their demographic information, their height and weight, and completed the following tasks in the respective order. The entire study was computer-based and took place in individual
Study 2
Study 2 was set up as a replication of Study 1 with an increased sample size and a number of methodological improvements. First, we measured implicit preferences between the (sham) training phase and the post-assessment phase. Second, as a measure of implicit preferences we replaced the IAT with an affective priming task, because responses in the IAT are more strongly influenced by the evaluation of category labels than by the evaluation of specific stimuli (De Houwer, 2001). Third, to improve
Study 3
Study 2 did not reveal any signs of effectiveness of AAT-training in changing food preferences. In Study 3, we changed the set-up another time to create what we considered optimal circumstances for the effect to occur. First, we used pictures of chocolate as unhealthy stimuli and pictures of stationery objects (e.g., envelopes, pencils) as control stimuli to create less fuzzy training categories. Second, we only recruited participants with a strong desire for chocolate and intentions to reduce
General discussion
The aim of the present research was to test the effectiveness of AAT-training in the eating domain. AAT-training has been successfully employed in the alcohol domain (Eberl et al, 2013, Wiers et al, 2010, Wiers et al, 2011) and there is some preliminary evidence suggesting that it might also have beneficial effects on eating behavior (Fishbach, Shah, 2006, Kemps et al, 2013). Overall, across three single session studies using rigorous methods and sufficient statistical power we could not find
References (45)
- et al.
The development of the attitudes to chocolate questionnaire
Personality and Individual Differences
(1998) Food reward. Brain substrates of wanting and liking
Neuroscience & Biobehavioral Reviews
(1996)- et al.
Attentional and approach biases for pictorial food cues. Influence of external eating
Appetite
(2009) A structural and process analysis of the Implicit Association Test
Journal of Experimental Social Psychology
(2001)- et al.
Approach bias modification in alcohol dependence. Do clinical effects replicate and for whom does it work best?
Developmental Cognitive Neuroscience
(2013) Resistance can be futile. Investigating behavioural rebound
Appetite
(2008)- et al.
Addiction, adolescence, and the integration of control and motivation
Developmental Cognitive Neuroscience
(2011) - et al.
The power of food scale. A new measure of the psychological influence of food in the environment
Appetite
(2009) - et al.
Approach and avoidance in fear of spiders
Journal of Behaviour Therapy and Experimental Psychiatry
(2007) - et al.
Movement direction or change in distance? Self- and object-related approach-avoidance movements
Journal of Experimental Social Psychology
(2008)
Experimentally induced chocolate craving leads to an attentional bias in increased distraction but not in speeded detection
Appetite
Restrained eaters show enhanced automatic approach tendencies towards food
Appetite
Using stop signals to inhibit chronic dieters’ responses to palatable foods
Behaviour Research and Therapy
Well-being and the accessibility of pleasant and unpleasant concepts
European Journal of Personality
Taste preferences and food intake
Annual Review of Nutrition
Procedures for the analysis of classroom tests
Educational and Psychological Measurement
On the automatic activation of attitudes
Journal of Personality and Social Psychology
Self-control in action. Implicit dispositions toward goals and away from temptations
Journal of Personality and Social Psychology
On taming horses and strengthening riders. Recent developments in research on interventions to improve self-control in health behaviors
Self and Identity
Measuring individual differences in implicit cognition the implicit association test
Journal of Personality and Social Psychology
Restrained eating
Cited by (97)
Disentangling go/no-go from motivational orientation to foods: Approaching is more than just responding
2023, Food Quality and PreferenceResponse inhibition training and measures of explicit and implicit food valuation
2022, Current Opinion in Behavioral Sciences‘Don't stop believing’: The role of training beliefs in cognitive bias modification paradigms
2022, AppetiteCitation Excerpt :This may suggest that CBM as an isolated intervention is not robust enough to elicit changes to explicit measures of unhealthy food value and preference, with the observed effects here appearing to be at least partially dependent on the presentation of the manipulation message, irrespective of the actual training content itself. While previous research has suggested that CBM can positively influence food choice and value ((e.g., Chen et al., 2018; Hollands et al., 2011; Oomen et al., 2018), in the current study, there was only limited evidence to suggest that the CBM training independently influenced unhealthy food value and choice, supporting the findings of previous work that did not find evidence to support the use of CBM training within food contexts (e.g., Becker et al., 2015; Carbine et al., 2021; Masterton et al., 2021). The inconsistent outcomes reported throughout the literature in relation to training effectiveness could indicate that factors external to training (and not consistently measured (such as beliefs or expectations in relation to training impact)) may play an important role in successful intervention outcomes.
Too hot to handle: Mood states moderate implicit approach vs. avoidance tendencies toward food cues in patients with obesity and active binge eating disorder
2021, Journal of Psychiatric ResearchCitation Excerpt :Indeed, the AAT has also been used as a training tool, for example, to reduce food craving (Brockmeyer et al., 2015), chocolate consumption (Schumacher et al., 2016), and for creating an avoidance bias for alcohol-related cues (Eberl et al., 2013), which positively affected relapse rates at a 1-year follow-up in patients with alcohol dependency. Despite these so-called implicit processes interventions gaining in popularity (Aulbach et al., 2019), AAT-based trainings are not always successful (Becker et al., 2015). However, existing procedures have not yet taken triggering conditions of disordered eating into consideration.
Targeting executive function for weight loss in adults with overweight or obesity
2021, Physiology and Behavior
- ☆
Acknowledgements: The stimulus material of this research has kindly been provided by Lisette Charbonnier, Image Science Center, UMC Utrecht (Full4Health project, funded by the European Union Seventh Framework Programme (FP7/2007–2013) under grant agreement no. 266408). The third author is supported by VICI grant 453-08-001 of the Dutch National Science Foundation (NWO). We also thank Linda Olde Dubbelink, Jonas Dalege and Imca Hensels for assisting with data collection.