INCREASING HEALTHY FOOD CHOICES:
The effect of a diet goal prime on healthy food choices and the moderating role of implicit attitude
Institution: Graduate School of Communication
Course: Master’s Thesis
Program: Master’s programme Communication Science
Student: Olle Sjoerd Stendert de Wit, 10773258
Supervisor: Gert-Jan de Bruijn
Date: February 3rd, 2017
Version: Final
1
ABSTRACT
In our daily lives, we are seduced by many unhealthy food temptations, which could be one of the causes for lots of people becoming overweight and obese. This study looks at these
temptations as a contributing factor in overweight and obesity, by examining whether a diet goal prime could increase the proportion of healthy food choices compared to unhealthy food choices. In addition, this is the first study that takes the implicit attitude towards the healthy hedonic goal and health goal into account, as moderating factors on the effect of a diet goal prime on the proportion of healthy food choices. A laboratory experiment was conducted, exposing participants to product folders with either a diet goal prime or a control prime. Furthermore, two implicit association tests were conducted to access the implicit attitude towards (a) the healthy hedonic goal (high vs. low) and (b) the health goal (high vs. low). A food choice-task was conducted to measure the proportion of healthy compared to unhealthy food choices. This study shows that a diet goal prime as opposed to a control prime, does not increase the proportion of healthy food choices directly. However, it was found that a diet goal prime increased the proportion of healthy food choices but only in the case of
participants with a high implicit attitude towards the health goal as distinct from a low one. These findings could provide an explanation as to why most studies only found healthy behaviour to be affected by a diet goal prime, when participants were either overweight or restrained eaters, suggesting that implicit attitude towards the health goal could be an underlying mechanism.
INTRODUCTION
Obesity is currently one of the largest health problems in the world. In 2014, more than 39% of the world’s adult population was overweight and 13% was obese (World Health
2 all kinds of media, making them difficult to avoid. These ads, with attractive unhealthy food, are environmental cues that activate strong short-term hedonic goals in the brain, such as “the rewarding feeling of pleasure”. This makes it difficult for people to control their impulses and pursue their long-term goals of, for example, staying healthy (Burger & Stice, 2011; Friese, Hofmann, & Wänke, 2008; Ferriday & Brunstorm, 2011). In fact, exposure to food ads on television increased food intake (Harris, Bargh & Brownell, 2009; Levitsky & Pacanowski, 2012), and in addition effected, the food choice, as it was found that exposure to food ads affected children’s and adult’s food preference (Borzekowski & Robinson, 2001; Harris, Bargh, & Brownell, 2009). There is therefore reason to believe that environmental cues, such as food ads, are one of the influencing factors in people becoming overweight or obese.
Protecting people from overweight and obesity caused by environmental cues, makes it relevant to examine intervention techniques that could increase healthy food choices. Most existing interventions, however, are information-based (e.g., showing the negative outcomes of being overweight) focusing on conscious processes, that have been shown to be ineffective in reducing overweight (Ng et al., 2014). One of the causes could be that many of our food choices made in the grocery store are taken impulsively (Friese, Hofmann, & Schmitt, 2008), that is making a quick nonconscious choice based on associations (e.g., taste, texture,
feelings) of a stimulus stored in the memory (Hofmann, Friese, & Wiers, 2008). Thus, when seeing a tempting chocolate bar for example, automatic associations (e.g., I like this chocolate bar) will direct your choice towards the chocolate bar via a nonconscious process. In addition, more and more literature suggests that interventions should focus on nonconscious processes by decreasing the impact of environmental cues and hedonic goals (Gortmaker et al., 2011; Hill, Wyatt, Reed, & Peters, 2003; Marteau, Hollands, & Fletcher, 2012).
Goal priming is an intervention technique which focus on nonconscious processes. Priming is the activation of associations using a certain stimulus (e.g., seeing an ad for a
3 chocolate bar) which subsequently affects information processing and potentially behaviour. If a prime motivates a person to pursue a certain goal (e.g., reduce weight), in order to obtain a certain behavioural outcome (e.g., not eating a chocolate bar), it is called a “goal prime” (Papies, 2016). Previous research showed that a diet goal prime increased healthy food choices in a restaurant (Papies & Veling, 2013), and decreased snack purchases in a grocery store (Papies, Potjes, Keesman, Schwinghammer, & van Koningsbruggen, 2014). This indicates that goal priming can be an effective intervention technique to increase healthy behaviour.
However, most significant effects of goal priming were found when participants were overweight (Papies, Potjes, Keesman, Schwinghammer, & van Koningsbruggen, 2013) or restrained eaters (Buckland, Finlayson, & Hetherington, 2013; Coelho, Polivy, Herman, & Pliner, 2009; Papies & Hamstra, 2010; van Koningsbruggen, Stroebe, & Aarts, 2011). Implicit
attitudes could be an underlying mechanism that could explain why only those people are
affected by the goal prime. Implicit attitudes are automatic affective (i.e., feelings towards a product) and cognitive (i.e., beliefs about a product) associations people have of a relevant stimulus (Sheeran, Gollwitzer, & Bargh, 2013). For example, when you see a chocolate bar you could think “I love chocolate” (affective) or “chocolate makes me fat” (cognitive). Although findings show that implicit attitudes influence behaviour directly (e.g., Conner, Perugini, O’Gorman, Ayres, & Prestwich, 2007; Friese, Hofmann, & Wänke, 2008; Hofmann, Gschwendner, Friese, Wiers, & Schmitt, 2008), there is reason to believe that implicit
attitudes moderate the effect of a goal prime. This is because a person subject to the goal prime should value the same goal as used in the prime (Papies, 2016). Hence, when a person has a positive implicit attitude towards healthy living and a goal prime motivates this person to do so, it can influence behaviour.
4 the effect of diet goal priming on the proportion of healthy food choices, combined with implicit attitude towards the healthy hedonic goal and health goal as a moderating variable. This study can therefore, give a further explanation of the underlying mechanisms of a diet goal prime by adding implicit attitude towards the healthy hedonic goal and health goal as a moderating variable. The aim of this study is to examine if a diet goal prime could increase the proportion of healthy food choices among individuals, and to examine the moderating role of implicit attitude towards the healthy hedonic goal and health goal. The main question of this study is to what extent does a diet goal prime increase the proportion of healthy food
compared to unhealthy food choices, and to what extent does a high implicit attitude compared to a low one towards the healthy hedonic goal and health goal, moderate this effect?
THEORETICAL FRAMEWORK
This chapter will give a theoretical foundation to form the hypotheses that could answer the main question of this study. First, underlying theories and mechanisms of conscious and nonconscious processes that influence behaviour will be discussed. This is followed by a review of studies examining the effect of goal priming on health behaviour, which will form the first hypothesis. Finally, the moderating role of the implicit attitude towards the healthy hedonic goal and the health goal will be discussed and the second hypothesis will be formulated.
Conscious processes that influence behaviour
Up until two decades ago, most health interventions that had been studied and developed, focussed on conscious processes that potentially influence behaviour (Papies, 2016). The
theory of planned behaviour is a commonly used theory for interventions and studies
5 McKellar, Reynolds, Lean & Mela, 1998; Matvienko, 2007). This theory assumes that
behaviour is influenced by intentions, which are formed via attitudes towards the behaviour,
subjective norms and self-efficacy (Azjan, 1991). Interventions using the theory of planned
behaviour are, usually, information-based interventions that try to influence one or more of these variables to eventually influence behaviour (Marteau, Hollands, & Kelly, 2015). In many occasions this is information about the consequences or positive outcomes of a certain behaviour (Hollands, Marteau, & Fletcher, 2016). Information-based interventions have proven, for example, to be effective in decreasing fat intake and increasing fruit intake (Irvine, Ary, Grove, & Gilfillan-Morton, 2004), and influencing the food preference, such that a person chooses fruit rather than snacks (Anderson et al., 1998; Matvienko, 2007).
A meta-analysis, however, showed that information-based interventions only had a small to medium effect on behaviour, and that these interventions are mainly effective when (a) the behaviour and the circumstances are not susceptible to habit formation, (b) they are performed under low cognitive and emotional load, (c) when it is not within a social context, and (d) when an individual has high levels of self-control or working memory capacity (Webb & Sheeran, 2006). These conditions are not ideal in many domains of changing health
behaviour, such as eating behaviour, alcohol consumption, or physical activity (Papies, 2016). Furthermore, interventions focussing on conscious processes are effective for influencing behaviour directly after the intervention, but the effects of these interventions disappear in the long term. One study, for example, found that an information-based intervention with the aim of improving a healthy diet among participants, increased the fruit and vegetable consumption in the short term, but had no effect in the long term (Stadler, Oettingen, & Gollwitzer, 2010). Furthermore, studies also show that interventions focussing on nonconscious processes have a longer effect on behaviour than interventions focussing on conscious processes (Sheeran, Gollwitzer, & Bargh, 2013). It was found, for example, that an intervention focussing on
6 nonconscious processes had a stronger effect on the alcohol consumption one year after the intervention, than an intervention focussing on nonconscious processes. (Wiers, Eberl, Rinck, Becker, & Lindenmeyer, 2011). One of the reasons could be, that many of our diet choices are taken impulsively, via a nonconscious process rather than a conscious process (Dijksterhuis, 2007; Perugini, 2005a). Studies also indicate that our behaviour is not actually a conscious driven process, influenced by information about certain behavioural outcomes, but a nonconscious driven process activated by stimuli in the environment (Bargh & Morsella, 2008; Dijksterhuis, Smith, Baaren, & Wigboldus, 2005; Neal, Wood, & Quinn, 2006; Marteau, Hollands, & Fletcher, 2012). This is confirmed by, for example, findings that that environmental cues, such as the scent of cleaning fluids, increased the cleaning behaviour of participants (Holland, Hendriks, & Aarts, 2005), or the proportion between healthy and unhealthy snacks at the checkout counter in a grocery store directed the participants snack choice (van Kleef, Otten, & van Trijp, 2012). Neither finding relies on conscious processing of information that influence behaviour, but instead nonconscious processing of
environmental cues. As a result, a growing amount of literature suggests that interventions should focus more on nonconscious processes that influence behaviour (Hollands, Marteau, & Fletcher, 2016; Marteau, Hollands, & Fletcher, 2012; Sheeran, Gollwitzer, & Bargh, 2013).
This study will therefore focus on nonconscious processes that influence health behaviour. The following paragraph will further explain how nonconscious processes influence behaviour in relation to conscious processes.
Nonconscious processes that influence behaviour
The reflective-impulsive model (Strack & Deutsch, 2004) can give a further explanation how nonconscious and conscious processes determine our behaviour, and why it is better to develop interventions focussing on nonconscious processes to influence health behaviour.
7 According to this model, behaviour is guided by two interacting systems, the reflective and the impulsive system. The reflective system affects behaviour based on knowledge about facts and values (conscious process), while the impulsive system affects behaviour via associative links stored in memory (nonconscious process). When people’s reflective system is activated, they will use thoughtful decision making to guide their behaviour. For example, when a person sees a television commercial for “M&M’s”, the person will weigh the pros and cons of consuming M&M’s, which may lead to the person not wanting to consume M&M’s, because he or she believes it will make him or her fat. Thus, the reflective system generated thoughtful decision making for behaviour execution. The impulsive system, on the other hand, depends on automatic processing (i.e., nonconscious processing of information without the awareness of being activated; Papies & Aarts, 2016) of environmental cues (e.g., text, music, pictures, smell) via associations stored in memory through experiential learning (de Houwer, Thomas, & Baeyens, 2001). When confronted with a temptation these associations are activated and further guide attention, information processing and finally behaviour. The impulsive system is related to the approach-avoidance theory (Friese, Hofmann, & Wiers, 2011). A person who has positive associations with a certain stimulus will have an approach reaction towards the stimulus, while a person who has negative associations with a certain stimulus will have an avoidance reaction away from the stimulus. In other words, the approach-avoidance
encompasses both the motivation (the initial trigger of behaviour) and direction (towards or away) of the behaviour (Elliot, 2006). This was also found in a study were participants had to push (avoid) and pull (approach) a joystick away or towards themselves when they saw words related to healthy and unhealthy food (Fishbach & Shah, 2006). In one group, participants had to avoid words describing unhealthy food by pushing the joystick forward and approach words related to healthy food by pulling the joystick backward, and this was reversed for the other group. The results show that the group who had to push away words describing
8 unhealthy food were less likely to choose unhealthy food items as a reward for their
participations than the reversed group. Similar results were found in a study examining approach-avoidance tendencies among hazardous drinkers (Wiers, Rinck, Kordts, Houben, & Strack, 2010). Participants in the avoid alcohol condition drink less alcohol than participants in the approach alcohol condition. Although these results indicate that automatic associations direct behaviour, it does not explain why interventions should focus on nonconscious
processes rather than conscious processes.
A possible explanation could be, that at many occasions in our daily life, such as grocery shopping, we make food choices based on impulses (i.e., an automatic and immediate approach reactions towards positive stimuli; Friese, Hofmann, & Schmitt, 2008), because we do not have enough cognitive capacity to make thoughtful decisions for every choice we encounter (Neal, Wood, & Quinn, 2006). The impulsive system generates automatic and quick suggestions for behaviour execution. The reflective system, on the other hand, is slower and often more effortful (i.e., needs more cognitive capacity) than the impulsive system (Strack & Deutsch, 2004). When a person is unable or unwilling to make the effort, the impulsive system will be activated (Strack & Deutsch, 2004; Vohs, 2006). For example, seeing a bowl of nuts on the bar could immediately trigger you to grab and eat some. According to Barsalou (2009) impulses are formed by previous health-related behaviour experiences. Representations of these experiences have been stored in memory and are defined as situated conceptualisations. These conceptualisations of a relevant stimulus could contain information about goals one was pursuing, the cognitive or affective responses one was having, contextual information (e.g., with who, when, and where), sensory information (e.g., taste and texture of the food one was eating) or one’s feeling of reward. The more one performs a behaviour repeatedly, in the same context and pursuing the same goal, the more anchored these conceptualisations are stored in memory. Triggering one of the elements of a
9 well anchored conceptualisation by an environmental cue may re-active all other elements of this conceptualisation, and could eventually lead to nonconscious effects on behaviour (Papies & Barsalou, 2015). Studies found, for example, that exposure to food ads on television
increased food intake (Harris, Bargh & Brownell, 2009; Levitsky & Pacanowski, 2012), and food preference (Borzekowski & Robinson, 2001; Harris, Bargh, & Brownell, 2009). These findings indicate that our behaviour is largely influenced by impulses (a nonconscious process) rather than thoughtful decision making (a conscious process), and subsequently, it weakens the effects of information-based interventions.
Interventions should thus focus on nonconscious rather than conscious processes. An intervention technique that focus on nonconscious processes is goal priming, and will be further discussed in the following paragraph.
Diet goal priming
Priming refers to “the activation of situated conceptualisations by using a stimulus, which
affects information processing and potentially behaviour” (Papies, 2016, pp. 6). If the prime is goal-directed (i.e., it motivates a person to pursue a behaviour that has a reward value), it can be defined as goal priming. (Papies, 2016). Goals are situated conceptualisations of a certain behaviour or behavioural outcomes that have a rewarding value (Custers & Aarts, 2005; Fishbach & Ferguson, 2007). When a goal is activated by a prime, it can influence the behaviour of a person without the person being consciously aware of it (Custers & Aarts, 2010). For example, one particular study found that participants who had to solve word puzzles with words related to “achievement”, were more motivated to perform an intellectual task than the control participants, because the “achievement goal” of these participants was activated without them being consciously aware of it (Bargh, Gollwitzer, Lee-Chai,
10 is translated into behaviour through a nonconscious process.
Multiple studies show that goal priming can positively influence health behaviour. Fishbach, Friedman and Kruglanski (2003) found, for example, that a diet goal prime compared to a control prime increased the amount of healthy food choices among
diet-concerned participants. Also, Papies and Veling (2013) found that a restaurant menu with diet related words (goal prime condition) compared to a menu with non-diet related words (control condition) increased the amount of choices for low-calorie food in comparison with high-calorie food, among chronic and current dieters. Another experiment showed, that a diet goal prime compared to a control prime reduced snack purchases in a supermarket among
overweight and obese participants, but not for participants with normal weight (Papies, Potjes, Keesman, Schwinghammer and Van Koningsbruggen, 2014). Furthermore, Papies and
Hamstra (2010) found that participants exposed to the scent of grilled chicken at a local butcher, and thus activating heuristic goals, consumed less when being goal primed than participants who were not. This indicates that the impact of the environment cue (the smell of grilled chicken) and the heuristic goal (e.g., the pleasure of eating grilled chicken) was
reduced by the diet goal prime. In addition, Papies, Stroebe and Aarts (2008) found that when chronic dieters are exposed to related words, it can decrease their attention to
diet-incongruent foods. These findings are also confirmed in an eye-tracking study, as it was found that health primes increased the time spent looking at healthy food which predicted the food choices of the participants (van der Laan, Papies, Hooge, & Smeets, 2017). Hence, this indicates that a diet goal prime could increase healthy food choices. Therefore, the following hypothesis is formulated:
H1: A diet goal prime compared to a control prime will increase the proportion of healthy
11
The moderating effect of implicit attitude
However, when a person does not value the primed goal such as staying healthy, it could decrease the effect of the diet goal prime on the proportion of healthy food choices (Papies 2016). These values can be determined by measuring the implicit attitude towards the healthy hedonic goal and health goal. Implicit attitudes are automatic cognitive and affective
associations of a relevant stimulus that stay outside the conscious awareness of the individual, and can be measured by an implicit association test (IAT; Sheeran, Gollwitzer, & Bargh, 2013). In other words, implicit attitudes are nonconscious associations, such as beliefs about a product (e.g., an apple is healthy) and feelings towards a product (e.g., I like the taste of an apple), which are activated when a person encounters a relevant stimulus (e.g., an apple). The
affective (i.e., feelings towards a certain stimulus) and cognitive (i.e., beliefs about a certain
stimulus) bases independently affect the implicit attitude towards a relevant stimulus (Trendel & Werle, 2015). For example, when one encounters a chocolate bar in a grocery store, one could have a positive implicit attitude towards the chocolate bar based on one’s affective reaction towards the chocolate bar (e.g., consuming a chocolate bar gives me pleasure), and at the same time having a negative implicit attitude based on one’s cognitive reaction towards the chocolate bar (e.g., consuming a chocolate bar makes me fat). This was also found in a study of Trendel and Werle (2015), who measured the implicit attitude towards the hedonic
goal (affective basis) and the health goal (cognitive basis) of participants by conducting two
IAT’s, and found that the affective and cognitive bases directly and uniquely influenced the amount of healthy compared to unhealthy food choices of the participants. The implicit attitude towards the hedonic goal was found to be the main driver, and only driver when participants had limited cognitive resources during the food choice (healthy vs. unhealthy). The implicit attitude towards the health goal was found to be the main driver when
12 choice.
Although it is not always clear which basis (affective or cognitive) studies use to measure implicit attitudes, it has been found that implicit attitudes affect health behaviour. Correlations, for example, have been found between the implicit attitude towards healthy and unhealthy food and the snack consumption (Conner, Perugini, O’Gorman, Ayres, &
Prestwich, 2007; Friese, Hofmann, & Wänke, 2008; Hofmann, Gschwendner, Friese, Wiers, & Schmitt, 2008), low-calorie food intake habits (Maison, Greenwald, & Bruin, 2001), and food choice (Conner et al., 2007; Friese et al., 2008; Perugini, 2005). One particular study found, that the implicit attitude towards healthy food correctly predicted the food choice of participants presented with the options of buying chocolate, fruit or not buying at all (Prestwich, Hurling, & Baker, 2011). This could explain why most studies mainly found significant effects of a diet goal prime on health behaviour for participants either overweight or obese (Papies, Potjes, Keesman, Schwinghammer, & van Koningsbruggen, 2013), or who were restrained eaters (Buckland, Finlayson, & Hetherington, 2013; Coelho, Polivy, Herman, & Pliner, 2009; Papies & Hamstra, 2010; van Koningsbruggen, Stroebe, & Aarts, 2011), because restrained eaters and individuals with overweight or obese probably have a higher positive implicit attitude towards the health goal and healthy food than patricians who are not. Fishbach and Ferguson (2007), for instance, found that participants with a high implicit attitude towards the goal of being thin, displayed more goal directed behaviour, such as resisting high-fat food and consuming less fattening food in a taste test, than participants with a low implicit attitude. These findings thus indicate that implicit attitudes can influence healthy behaviour, and therefore, the following hypothesis is formulated:
H2: A high compared to a low (a) implicit attitude towards the healthy hedonic goal, and (b)
a high compared to a low implicit attitude towards the health goal will increase the effect of a diet goal prime on the proportion of healthy food choices.
13
Figure 1. Conceptual model hypothesis 1 and hypothesis 2a
Figure 2. Conceptual model hypothesis 1 and hypothesis 2b
METHOD AND MATERIALS Design
The design of this quasi-experiment is a 2 (diet goal prime vs. control) x 2 (positive implicit attitude towards the healthy hedonic goal and health goal: (high vs. low) between-subject factorial design. Participants were randomly divided into one of the two groups (diet goal prime or control prime). Based on the IAT scores, participants were classified as having a high or low positive implicit attitude towards the healthy hedonic goal and a high or low positive implicit attitude towards the health goal via a median split.
Participants
In total, 128 persons, mostly undergraduate students at the “University of Amsterdam” (UvA) in the Netherlands, participated and completed the experiment. All participants could read and speak Dutch, considering the whole experiment was carried out in Dutch. Participants were
Implicit attitude towards the healthy hedonic goal
(High/Low)
Diet goal prime vs. Control prime
(High/Low)
Proportion healthy food choices +
+
Implicit attitude towards the health goal (High/Low)
Diet goal prime vs. Control prime
(High/Low)
Proportion healthy food choices +
14 recruited by Facebook, flyers, posters and via the science portal of the UvA with a cover
story, asking them to participate in a study examining their awareness towards products. In
order to participate in this experiment people had to voluntary subscribe on the science portal of the UvA. In addition, one participation point or a five-Euro reward was given in exchange for participating in this experiment. The sample of this study consist of 34 male and 90 female participants, with a mean age of Mage = 24,84 (SD = 8.92), ranging between 18 and 67, and a
mean BMI of MBMI = 21.92 (SD = 2.59) ranging between 17.72 and 30.85.
Four of the additional participants were excluded from the experiment, because one participant only pressed the “E” key during the choice-task, another participant pressed the wrong key during the IAT, and two participants did not see the prime due to technical issues. Furthermore, after the questionnaire, the researcher asked the participants if they could guess what the aim of the experiment was. If the participants could guess that the aim of the
experiment or indicate that the experiment tried to influence their food choice, they were removed from the experiment. None of the participants correctly answered this question, but some participants mentioned that they were primed, but thought the prime was hidden in the IAT, and moreover, did not think the prime was used to influence their food choices. This, indicates that the participants were not aware of the aim of this study and being primed in the folders, and thus, none of the participants was removed from this study.
Stimulus materials
In total, 20 product folders (10 for each condition) were created as stimulus material for this study (see Appendix A). All folders contained four non-visible blocks, with one or two visible food or non-food items in each block. Furthermore, these folders had different blocks with different products in different orders.
15 product folders. In total, 22 undergraduate (20 female, 2 male) students at the UvA completed an online-questionnaire on Qualtrics, who were recruited on Facebook with a link to the online questionnaire, and were subsequently excluded from the final experiment. The top five most named healthy and unhealthy food items in the pre-test were used in the stimulus
materials. In the unhealthy food item category, the most named items were: crisps, Snickers, Mars, M&M’s, and a stroopwafel (a traditional Dutch biscuit). The most named healthy food items were: apples, bananas, grapes, kiwi’s and tangerines. Furthermore, two control non-food items were imbedded in the product folders to obscure the purpose of the experiment and to make the product folders look more realistic. In addition, the design of the product folders was based on a bonus folder of a well-known grocery store in the Netherlands to make the folders more credible, such that it was being perceived as an actual product folder. All ten folders in the diet goal prime condition and the control condition were the same, with the only exception being the diet goal prime used in the folders.
The 10 folders in the diet goal prime condition contained two small blue text blocks at the top of the folder with the sentences, “gezond en weinig calorieën” [healthy and low in calories] and “goed voor de lijn” [good for a slim-figure]. The sentences of the diet goal prime used in this study were based on multiple studies (Papies & Hamstra, 2010; Papies, Potjes, Keesman, Schwinghammer, & van Koningsbruggen, 2014; Papies & Veling, 2013), that already proved the effectiveness of this diet goal prime. Furthermore, the colour blue was used in the diet goal prime because this has been shown to be associated with healthiness (Gelici-Zeko, Lutters, ten Klooster, & Weijzen, 2013; van der Laan, De Ridder, Viergever, & Smeets, 2012), and was also used in the, previously mentioned, study of Van Der Laan, Papies, Hooge and Smeets (2017). The 10 folders in the control condition contained two similar text blocks at the top of the folder, only the sentences and colour varied from the prime condition. The sentences used in the control condition were, “veel nieuwe producten”
16 [lots of new products] and “probeer ze deze week” [try them this week], and the colour orange was used for the text blocks. Both, the colour and sentences were unrelated to health, and were also used in the study of Van Der Laan and colleagues (2017).
Implicit attitude
The implicit attitude towards the healthy hedonic goal and health goal was measured via an IAT with the computer program Inquisit 5, which has been shown to be an effective tool in accessing the implicit attitude (Gawronski & Bodenhausen, 2006; Hollands, Prestwitch, & Marteau, 2011). Participants were randomly allocated to first conduct a goal or hedonic IAT, measuring the implicit attitude towards the healthy hedonic goal and health goal. This because, implicit attitudes consist of two independent factors, and therefore should be measured separately (Trendel & Werle, 2015). The IAT accesses how quickly participants sort attributes relating to healthy and unhealthy food into four categories (two target categories: fruit and snack; two attribute categories: positive and negative) that are paired (snack and negative; fruit and positive), using the keys “E” and “I” on a keyboard. A red cross would appear on the computer screen if participants made a wrong categorization. Participants then had to corrected their mistake by pressing “E” or “I” in order to proceed with the test. In total, two IAT’s consisting of seven blocks were conducted. The hedonic and goal related word attributes of both IAT’s were selected such that they clearly belong to only one of the categories (see Appendix B). Furthermore, the healthy and unhealthy food images used in the stimulus material were also embedded in both IAT’s.
The first three blocks and block five and six were practice blocks of 20 trails, beginning with categorizing the images into the target categories (snack or fruit). Then, participants had to categorize the words into the attribute categories (positive or negative), and in the final practice block they had to categorize the words and images into incompatible
17 pairings of the target and attribute categories (snack or positive; fruit or negative). These practice blocks were followed by the actual incompatible pairing test, which was the same as the last practice block only now there were 40 trails instead of 20 trails. After the test block, participants had to practice again with categorizing the images into switched target categories (fruit or snack), and subsequently with categorizing the images and words into compatible pairings (fruit or positive; snack or negative), followed by the actual compatible paring test of 40 trails. The scores, ranging between +2 and -2, of both IAT’s were based on Greenwald, Nosek and Banaji’s (2003) algorithm. Negative scores reflect a more favourable implicit attitude towards the healthy hedonic goal and health goal. Positive scores reflect a more favourable implicit attitude towards the unhealthy hedonic goal and unhealthy goal.
Participants were then divided into having a high or low implicit attitude towards the healthy hedonic goal and a high or low implicit attitude towards the health goal, based on a median split of the final scores for each IAT.
Food choice
The food choice was measured via a choice-task with the computer program Presentation 19. To simulate impulsive behaviour, participates were instructed (see Appendix C) to choose as quickly as possible between two products items within three seconds, by pressing “E” and “I” on a keyboard. The same product items used in the product folders were used in the choice-task. In total, participants had to make 75 product choices, containing 25 choices in each choice category (snack or fruit; non-food or snack; non-food or fruit), which appeared randomly on the computer screen. Non-food items were used in order to hide the actual purpose of the test. The overall proportion of fruit compared to snack choices made within the category “snack or fruit” is the dependent measure. This choice-task is based on the
18 to be an effective tool to access impulsive food choices (Custers & Aarts, 2005; Finlayson, King, & Blundell, 2007; Ouwehand & Papies, 2010).
Procedure
First, participants were picked up at the waiting room by the researcher and were put seated at a computer. Then, they were asked by the researcher if they would participate in exchange for points or a five-Euro participation reward and had to sign an informed consent. After this question and signing of the informed consent, an instruction sheet for the first IAT was given by the researcher. This could be the hedonic or goal IAT as it was randomly allocated. When participants completed the first IAT they would call the researcher who then gave the second instruction sheet, and opened the next IAT on the computer. After completion of the second IAT, participants conducted a filler task in order to reduce potential bias of the IAT tests. The filler task required the participant to name three characteristics of a soap and a package of kitchen papers displayed in the filler task. Next, participants were asked to sit in front of an eye-tracker were the product folders of one of the two conditions (diet goal prime/control prime) were shown. The participants were instructed to skip through the folders at their own pace, like they would do in real-life. When the participant had seen all the folders, they were asked to go back to the other computer, and then received instructions for the food
choice-task. After completion of the food choice-task, they had to fill in a questionnaire on the same
computer. Finally, the researcher asked the participants what the aim of the experiment was, and noted down their answer. Participants were then thanked for their participation and received, where applicable, their five-Euro reward.
RESULTS Randomisation check
19 randomization of age was successful in the prime condition (diet goal prime vs. control
prime) and in the implicit attitude towards the health goal condition (high vs. low), with age as dependent variable and the two conditions as independent variables. Result show no significant difference between age in the prime condition (F(1, 121) = .23, p = n.s.), and implicit attitude towards the health goal condition (F(1, 121) = .45, p = n.s.). The second two-way ANOVA analysis was conducted to test if randomization of age was successful in the
prime condition and in the implicit attitude towards the healthy hedonic goal condition (high
vs. low), with age as dependent variable and the two conditions as independent variables. Again, results show no significant difference between age in the prime condition (F(1, 121) = .08, p = n.s.), and in the implicit attitude towards the healthy hedonic goal condition (F(1, 121) = 3.68, p = n.s.). Thus, randomization of age was successful in all conditions.
The third two-way ANOVA analysis was conducted to test if randomization of BMI was successful in the prime condition and in the implicit attitude towards the health goal
condition with BMI as dependent variable and the two conditions as independent variables.
Results show no significant difference between BMI in the prime condition (F(1, 121) = .90,
p = n.s.), and implicit attitude towards the health goal condition (F(1, 121) = 1.01, p = n.s.).
The final two-way ANOVA analysis was conducted to test if randomization of BMI was successful in the prime condition and in the implicit attitude towards the healthy hedonic goal
condition, with BMI as dependent variable and the two conditions as independent variables.
Again, result show no significant difference between BMI in the prime condition (F(1, 121) = .99, p = n.s.), and in the implicit attitude towards the healthy hedonic goal condition (F(1, 121) = .52, p = n.s.). Thus, randomization of BMI was successful in all conditions.
Finally, a chi-square test was conducted to measure if randomization of sex was successful in all conditions. Result show a significant difference between sex in the prime condition (χ2 = 7.33, p = .007), and in the implicit attitude towards the health goal condition
20 (χ2 = 4.51, p = .034). No significant difference was found between sex in the implicit attitude towards the healthy hedonic goal condition (χ2 = 3.22, p = .073). This indicates the
randomization of sex was partial unsuccessful, and thus, has to be controlled for in further analyses.
Confounding variables analysis
A correlation analysis was performed to check if there were any confounding variables. The correlation analysis measured the correlation between the dependent variable proportion of
healthy food choices and the independent variables age, BMI, and sex. No significant
correlations were found between the proportion of healthy food choices and age (r = .16, n = 124, p = .077), and BMI (r = -.05, n = 124, p = .542). Hence, age and BMI will not be used as covariate in further analyses. However, a moderate positive significant correlation was found between the proportion of healthy food choices and sex (r = .37, n = 124, p < 0.001). This means sex should be controlled for in further analyses.
Main analyses
Two two-way ANCOVA analyses were conducted (see Appendix E) to test if a diet goal prime compared to a control prime increases the proportion of healthy food choices (H1), and if (a) a high compared to a low implicit attitude towards the healthy hedonic goal and (b) a high compared to a low implicit attitude towards the health goal will increase the effect of a diet goal prime on the proportion of healthy food choices (H2). Furthermore, because implicit attitude towards the healthy hedonic goal and health goal is measured with two IAT’s each will be used separately in the two-way ANCOVA analyses.
First, a two-way ANCOVA analysis was conducted with proportion of healthy food
21
hedonic goal condition as independent variables, and sex as covariate. The two-way
ANCOVA analysis shows no significant main effect of a diet goal prime compared to the control prime on the proportion of healthy food choices after controlling for sex, F(1, 119) = .03, p = .876. There was no difference in the proportion of healthy food choices between participants in the diet goal prime condition (M = .76, SD = .23) and participants in the control condition (M = .74, SD = .23). However, a small to medium significant main effect of implicit attitude towards the healthy hedonic goal on the proportion of healthy food choices was found after controlling for sex, F(1, 119) = 6.96, p = .009, η2 = .05. The pairwise comparisons shows that participants with a high implicit attitude towards the healthy hedonic goal (M = .82, SD = .17) had a higher proportion of healthy food choices than participants with a low implicit attitude towards the healthy hedonic goal (M = .69, SD = .26), Mdiffrence = .10, p = .009, 95%CI
[.03, .18]. No significant interaction effect between the prime condition and implicit attitude towards the healthy hedonic goal condition on the proportion of healthy food choices was found after controlling for sex, F(1, 119) = .01, p = .941. Thus, a high implicit attitude towards the healthy hedonic goal did not increase the effect of a diet goal prime on the proportion of healthy food choices. This means H2a, is rejected.
In the second two-way ANCOVA analysis, implicit attitude towards the health goal was used instead of implicit attitude towards the healthy hedonic goal, all other variables were the same as in the first two-way ANCOVA analysis. Again, the two-way ANCOVA analysis showed no significant main effect of a diet goal prime compared to the control prime on the proportion of healthy food choices after controlling for sex, F(1, 119) = .45, p = .504. A diet goal prime (M = .76, SD = .23) did not increase the proportion of healthy food choices
comparted to the control prime (M = .74, SD = .23). Hence, because both two-way ANCOVA analyses do not show a significant effect of a diet goal prime compared to the control prime on the proportion of healthy food choices, H1 is rejected.
22 Furthermore, no significant effect was found of implicit attitude towards the health goal on the proportion of healthy food choices after controlling for sex, F(1, 119) = .36, p = .769. Participants with a high implicit attitude towards the health goal (M = .76, SD = .21) did not have a higher proportion of healthy food choices than participants with a low implicit attitude towards the health goal (M = .74, SD = .25). However, a small to medium significant interaction effect between the prime condition and implicit attitude towards the health goal condition on the proportion of healthy food choices was found after controlling for sex, F(1, 119) = 5.21, p = .024, η2 = .04. The pairwise comparisons shows that participants in the diet goal prime condition with a high implicit attitude towards the health goal (M = .81, SD = .19) had a significantly higher proportion of healthy food choices than participants with a low implicit attitude towards the health goal (M = .71, SD = .26), Mdiffrence = .08, p = .168, 95%CI [-.03, .19]. This interaction effected is also illustrated by the line chart (see Appendix E), showing almost no difference in the proportion of healthy food choices between participants with a low and high implicit attitude towards the health goal in the control condition, while there is a difference in the proportion of healthy choices between participants with low and high implicit attitude towards the health goal in the diet goal prime condition. This means,
H2b is accepted.
Sex was used as covariate in both ANCOVA analyses. The first (F(1, 119) = 14.65, p < .001, η2 = .10) and second (F(1, 119) = 20.46, p < .001, η2 = .14) ANCOVA analysis both
show a medium to strong significant effect of sex on the proportion of healthy food choices.
CONCLUSION AND DISCUSSION
The research question of this study is to what extent does a diet goal prime increase
the proportion of healthy food compared to unhealthy food choices, and to what extent does a high implicit attitude compared to a low one towards the healthy hedonic goal and health
23
goal, moderate this effect? This study shows that a diet goal prime in compression with a
control prime does not increase the proportion of healthy food choices directly (H1). However, a significant interaction effect was found between a diet goal prime and implicit attitude towards the health goal on the proportion of healthy food choices (H2b). Participants in the diet goal prime condition had a higher proportion of healthy food choices when they had a high compared to a low implicit attitude towards the health goal. Hence, to answer the research question, a diet goal prime increases the proportion of healthy food choices
compared to unhealthy food choices, but only in the case of participants with a high implicit attitude towards the health goal as distinct from a low one. This is in line with the argument of Papies (2016) that a person should value the health goal used in the diet goal prime, as this studies shows that participants with a low implicit attitude towards the health goal were not affected by the diet goal prime, while participants with a high implicit attitude towards the health goal were affected. Furthermore, these findings indicate that implicit attitude towards the health goal could be an underlying mechanism that explains why most studies (Buckland, Finlayson, & Hetherington, 2013; Coelho, Polivy, Herman, & Pliner, 2009; Papies, Potjes, Keesman, Schwinghammer, & van Koningsbruggen, 2013; Papies & Hamstra, 2010) only found effects of a diet goal prime for participants who were overweight, obese, or restrained eaters. However, this was not directly tested in this study and could be taken into account in future research, by examining if there is a correlation between the implicit attitude towards the health goal and BMI, and between the implicit attitude towards the health goal and the level of restrained eating.
Furthermore, while an interaction effect was found between the diet goal prime and the implicit attitude towards the health goal, no interaction effect was found between a diet goal prime and implicit attitude towards the healthy hedonic goal (H2a). This could be
24 diet goal prime. Papies (2016) argues that a person should value the same goal as used in the diet goal prime, the implicit attitude towards the healthy hedonic goal does not measure the value of the goal but of the food, which can be different from the implicit attitude towards the health goal (Trendel & Werle, 2015). It is therefore assumable, that only the implicit attitude towards the health goal moderates the effects of a diet goal prime, as was found in this study. Future research should, therefore, always measure the implicit attitude towards the health goal instead of the implicit attitude towards the healthy hedonic goal as moderating factor of a diet goal prime; this because the implicit attitude towards the healthy hedonic goal does not relate with a diet goal prime.
Another interesting finding is that the implicit attitude towards the healthy hedonic goal directly affected the proportion of healthy food choices by the participants, while this direct effect was not found for the implicit attitude towards the health goal. Participants with a high implicit attitude towards the healthy hedonic goal had a higher proportion healthy food choices than participants with a low implicit attitude towards the healthy hedonic goal. This goes against the findings of Fishbach and Ferguson (2007), who found that people with a high compared to a low implicit attitude towards the goal of being thin consumed less fatting food. However, it is in line with the findings of Trendel and Werle (2015) who found that implicit affect is the main driver of food choices, while implicit cognition only effects food choices when people are low in impulsivity and have enough cognitive resources. This could also explain why this study did not find a direct effect of implicit attitude towards the health goal on the proportion of healthy food choices. The food choice-task stimulated impulsive
behaviour, and therefore, it is assumable that most participants were high in impulsivity, and thus, used their implicit attitude towards the healthy hedonic goal as main driver to make their food choices. Future research should therefore control for impulsivity in order to check if this assumption is true.
25 This study has some limitation. First, the interaction effect between the diet goal prime and implicit attitude towards the health goal only had a small to medium significant effect. This could be due to used sample size and composition. This study had a small sample size, containing mostly female students with a healthy BMI. Results show that almost all
participants had a positive implicit attitude towards the healthy hedonic goal and health goal. Furthermore, almost all participants made more healthy food choices than unhealthy food choices. This indicates there were only small differences between the participants, and thus, using a more diverse (e.g., more difference in age, BMI, education) and a larger sample size could improve the differences between the participant and eventually lead to a higher effect size. Moreover, the findings of this study are not generalizable to the entire Dutch population. This because, the used sample is not diverse in age, BMI and sex, which is a threat to the external validity. Future research should, therefore, use a more diverse and larger sample.
Second, randomization of sex was unsuccessful and had a significant correlation with the proportion of healthy food choices, therefore, it was used as covariate in the main
analyses. Previous research on diet goal priming also found a correlation between sex and the number of snacks eaten (Papies & Hamstra, 2010), which could explain the correlation between sex and the proportion of healthy compared to unhealthy food choices in this study. However, the found correlation between sex and the proportion of healthy food choices could also be caused by the large difference between the proportion of male and female participants in the sample. Therefore, to test if there really is significant difference between male and female participants, future research should contain the same proportion female and male participants in the sample.
This study has some implications for science and society. First, this is the first study examining the moderating role of the implicit attitude towards the healthy hedonic goal and health goal on the effect of a diet goal prime on the proportion of healthy food choices. It
26 gives a further explanation of the underlying mechanisms of a diet goal prime by adding implicit attitude towards the healthy hedonic goal and health goal as moderating variable. Second, it gives multiple suggestions for future research. Third, advertisers can use this study to develop intervention campaigns to increase healthy food choices. It shows that individuals should have a high implicit attitude towards the health goal in order to be influenced by the diet goal prime, and thus, should be the most effective target group for an intervention campaign.
To conclude, a diet goal prime could be an effective intervention technique to improve healthy behaviour via a nonconscious process. This does not mean we should only develop health interventions focusing on nonconscious processes, it should rather be used as an extension of interventions focusing on conscious processes. Associations can be learned via a conscious process and activated by a nonconscious process. This synergy, will hopefully lead to greater and longer effects on healthy behaviour than both processes do independently.
REFERENCES
Anderson, A.S., Cox, D.N., Mckellar, S., Reynolds, J., Lean, M.E.J., & Mela, D.J. (1998). Take Five, a nutrition education intervention to increase fruit and vegetable intakes: impact on attitudes towards dietary change. Britisch Journal of Nutrition, 80(2), 133-140. doi:10.1017/S0007114598001032
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human
Decision Processes, 50, 179–211. doi:10.1016/0749-5978(91)90020T
Bargh, J. A., Gollwitzer, P. M., Lee-Chai, A., Barndollar, K., & Trötschel, R. (2001). The automated will: nonconscious activation and pursuit of behavioral goals. Journal of
Personality and Social Psychology, 81, 1014-1027. doi:10.1037//0022-3514.81.6.1014
Bargh, J. A., & Morsella, E. (2008). The unconscious mind. Perspectives on psychological
27 Barsalou, L. W. (2009). Simulation, situated conceptualization, and prediction. Philosophical
Transactions of the Royal Society B: Biological Sciences, 364(1521), 1281–1289.
doi:10.098/rstb.2008.0319
Borzekowski, D. L., & Robinson, T. N. (2001). The 30-second effect: an experiment revealing the impact of television commercials on food preferences of
preschoolers. Journal of the American Dietetic Association, 101, 42-46. doi:10.1016/S0002-8223(01)00012-8
Buckland, N. J., Finlayson, G., & Hetherington, M. M. (2013). Pre-exposure to diet-congruent food reduces energy intake in restrained dieting women. Eating behaviors, 14(3), 249-254. doi:10.1016/j.eatbeh.2013.03.003
Burger, K. S., & Stice, E. (2011). Relation of dietary restraint scores to activation of reward related brain regions in response to food intake, anticipated intake, and food pictures.
Neuro Image, 55, 233–239. doi:10.1016/j.neuroimage.2010.12.009
Coelho, J. S., Polivy, J., Herman, C. P., & Pliner, P. (2009). Wake up and smell the cookies. Effects of olfactory food-cue exposure in restrained and unrestrained
eaters. Appetite, 52(2), 517-520. doi:10.1016/j.appet.2008.10.008
Conner, M. T., Perugini, M., O’Gorman, R., Ayres, K., & Prestwich, A. (2007). Relations between implicit and explicit measures of attitudes and measures of behavior: Evidence of moderation by individual difference variables. Personality and Social
Psychology Bulletin, 33, 1727–1740. doi:10.1177/0146167207309194
Custers, R., & Aarts, H. (2010). The unconscious will: How the pursuit of goals operates outside of conscious awareness. Science, 329(5987), 47–50.
28 Dallas, R., Field, M., Jones, A., Christiansen, P., Rose, A., & Robinson, E. (2014). Influenced
but unaware: Social influence on alcohol drinking among social
acquaintances. Alcoholism: Clinical and Experimental Research, 38, 1448-1453. doi:10.1111/acer.12375
de Houwer, J., Thomas, S., & Baeyens, F. (2001). Associative learning of likes and dislikes: A review of 25 years of research on human evaluative conditioning. Psychological
Bulletin, 127, 853–869. doi:10.1037/0033-2909.127.6.853
Dijksterhuis, A. (2007). Het slimme onbewuste: Denken met gevoel. Amsterdam: Bert Bakker Dijksterhuis, A., Smith, P. K., Van Baaren, R. B., & Wigboldus, D. H. (2005). The
unconscious consumer: Effects of environment on consumer behavior. Journal of
Consumer Psychology, 15(3), 193-202. doi:10.1207/s15327663jcp1503_3
Elliot, A. J. (2006). The hierarchical model of approach-avoidance motivation. Motivation
and emotion, 30(2), 111-116. doi:10.1007/s11031-006-9028-7
Ferriday, D., & Brunstrom, J. M. (2011). ‘I just can’t help myself’: effects of food-cue exposure in overweight and lean individuals. International Journal of Obesity, 35, 142–149.
doi:10.1038/ijo.2010.117
Finlayson, G., King, N., & Blundell, J. E. (2007). Is it possible to dissociate “liking” and “wanting” for foods in humans? A novel experimental procedure. Physiology &
Behavior, 90, 36–42. doi:10.1016/j.physbeh.2006.08.020
Fishbach, A., & Ferguson, M. J. (2007). The goal construct in social psychology. In W. Arie & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (pp. 490-515), New York, NY: Guilford Press. doi:10.1007/BF02333407
Fishbach, A., Friedman, R. S., & Kruglanski, A. W. (2003). Leading us not into temptation: Momentary allurements elicit overriding goal activation. Journal of Personality and
29
Social Psychology, 84(2), 296–309. doi:10.1037/0022-3514.84.2.296
Fishbach, A., & Shah, J. Y. (2006). Self-control in action: Implicit dispositions toward goals and away from temptations. Journal of Personality and Social Psychology, 90, 820-832. doi:10.1037/0022-3514.90.5.820
Friese, M., Hofmann, W., & Wänke, M. (2008). When impulses take over: Moderated predictive validity of implicit and explicit attitude measures in predicting food choice and consumption behavior. British Journal of Social Psychology, 47, 397-419. doi:10.1348/014466607X241540
Friese, M., Hofmann, W., & Wiers, R. W. (2011). On taming horses and strengthening riders: Recent developments in research on interventions to improve self-control in health behaviors. Self and Identity, 10(3), 336-351. doi:10.1080/15298868.2010.536417 Friese, M., Hofmann, W., & Schmitt, M. (2008). When and why do implicit measures predict
behavior? Empirical evidence for the moderating role of opportunity, motivation, and process reliance. European Review of Social Psychology, 19, 285–338.
doi:10.1080/10463280802556958
Gawronski, B., & Bodenhausen, G. V. (2006). Associative and propositional processes in evaluation: an integrative review of implicit and explicit attitude change.
Psychological Bulletin, 132(5), 692-731. doi:10.1037/0033-2909.132.5.692
Gelici-Zeko, M. M., Lutters, D., ten Klooster, R., & Weijzen, P. L. G. (2013). Studying the influence of packaging design on consumer perceptions using categorizing and perceptual mapping. Packaging Technology and Science, 26, 215–228.
doi:10.1002/pts.1977
Gortmaker, S. L., Swinburn, B. A., Levy, D., Carter, R., Mabry, P. L., Finegood, D. T., ... & Moodie, M. L. (2011). Changing the future of obesity: Science, policy, and action. The
30 Greenwald, A. G, Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the
implicit association test: I. an improved scoring algorithm. Journal of Personality and
Social Psychology, 85, 197-216. doi:10.1037/0022-3514.85.2.197
Harris, J. L., Bargh, J. A., & Brownell, K. D. (2009). Priming effects of television food advertising on eating behavior. Health psychology, 28(4), 404-413.
doi:10.1037/a0014399
Hill, J. O., Wyatt, H. R., Reed, G. W., & Peters, J. C. (2003). Obesity and the environment: where do we go from here?. Science, 299(5608), 853-855.
doi:10.1126/science.1079857
Hofmann, W., Friese, M., & Wiers, R. W. (2008). Impulsive versus reflective influences on health behavior: A theoretical framework and empirical review. Health Psychology
Review, 2(2), 111–137. doi:10.1080/17437190802617668
Hofmann, W., Gschwendner, T., Friese, M., Wiers, R. W., & Schmitt, M. (2008). Working memory capacity and self-regulatory behavior: Toward an individual differences perspective on behavior determination by automatic versus controlled processes.
Journal of personality and social psychology, 95, 962-977. doi:10.1037/a0012705.
Hollands, G. J., Marteau, T. M., & Fletcher, P. C. (2016). Non-conscious processes in changing health-related behaviour: A conceptual analysis and framework. Health
Psychology Review, 10(4), 381-394. doi:10.1080/17437199.2015.1138093
Irvine, A. B., Ary, D. V., Grove, D. A., & Gilfillan-Morton, L. (2004). The effectiveness of an interactive multimedia program to influence eating habits. Health Education Research,
19(3), 290-305. doi: 10.1093/her/cyg027
Levitsky, D. A., & Pacanowski, C. R. (2012). Free will and the obesity epidemic. Public
31 Maison, D., Greenwald, A. G., & Bruin, R. (2001). The implicit association test as a measure
of implicit consumer attitudes. Polish Psychological Bulletin, 2, 61–79. doi://10.1066/S10012010002
Marteau, T. M., Hollands, G. J., & Fletcher, P. C. (2012). Changing human behavior to prevent disease: the importance of targeting automatic processes. Science, 337(6101), 1492-1495. doi:10.1126/science.1226918
Marteau, T. M., Hollands, G. J., & Kelly, M. P. (2015). Changing population behavior and reducing health disparities: Exploring the potential of “choice architecture”
interventions. In R. M. Kaplan, M. Spittel, & D. H. David (Eds.), Emerging
Behavioral and Social Science Perspectives on Population Health (pp. 105–126).
Bethesda, MD: National Institutes of Health/Agency for Healthcare Research and Quality.
Matvienko, O. (2007). Impact of a Nutrition Education Curriculum on Snack Choices of Children Ages Six and Seven Years. Journal of Nutrition Education and Behavior,
39(5), 281-285. doi:10.1016/j.jneb.2007.01.004
Neal, D. T., Wood, W., & Quinn, J. M. (2006). Habits: A repeat performance. Current
Directions in Psychological Science, 15(4), 198-202.
doi:10.1111/j.14678721.2006.00435.x
Ng, M., Fleming, T., Robinson, M., Thomson, B., Graetz, N., Margono, C., ... & Gakidou, E. (2014). Global, regional, and national prevalence of overweight and obesity in
children and adults during 1980–2013: A systematic analysis for the global burden of disease study 2013. The Lancet, 384(9945), 766–781.
32 Ouwehand, C., & Papies, E. K. (2010). Eat it or beat it: The differential effect of food
temptations on overweight and normal-weight restrained eaters. Appetite, 55, 56–60. doi:10.1016/j.appet.2010.04.009
Papies, E. K. (2016). Health goal priming as a situated intervention tool: How to benefit from nonconscious motivational routes to health behavior. Health psychology review, 10(4), 408-424. doi:10.1080/17437199.2016.1183506
Papies, E. K., & Aarts, H. (2016). Automatic self-regulation: From habit to goal pursuit. In K. D. Vohs & R. F. Baumeister (Eds.), Handbook of self-regulation: Research, theory,
and applications (3rd ed.). New York, NY: Guilford Press.
Papies, E. K., & Barsalou, L. W. (2015). Grounding desire and motivated behavior: A theoretical framework and review of empirical evidence. In W. Hofmann & L. F. Nordgren (Eds.), The psychology of desire (pp. 36–60). New York, NY: Guilford Press.
Papies, E. K., & Hamstra, P. (2010). Goal priming and eating behavior: enhancing self-regulation by environmental cues. Health Psychology, 29(4), 384-389.
doi:10.1037/a0019877
Papies, E. K., Potjes, I., Keesman, M., Schwinghammer, S., & Van Koningsbruggen, G. M. (2014). Using health primes to reduce unhealthy snack purchases among overweight consumers in a grocery store. International Journal of Obesity, 38(4), 597-602. doi:10.1038/ijo.2013.136
Papies, E. K., Pronk, T. M., Keesman, M., & Barsalou, L. W. (2015). The benefits of simply observing: Mindful attention modulates the link between motivation and
behavior. Journal of Personality and Social Psychology, 108, 148-170. doi:10.1037/a0038032
33 Papies, E. K., Stroebe, W., & Aarts, H. (2008). The allure of forbidden food: On the role of
attention in self-regulation. Journal of Experimental Social Psychology, 44, 1283– 1292. doi:10.1016/j.jesp.2008 .04.00
Papies, E. K., & Veling, H. (2013). Healthy dining: Subtle diet reminders at the point of purchase increase low-calorie food choices among both chronic and current dieters.
Appetite, 61, 1-7. doi:10.1016/j.appet.2012.10.025
Perugini, M. (2005). Predictive models of implicit and explicit attitudes. British Journal of
Social Psychology, 44, 29–45. doi:10.1348/014466604X23491
Prestwich, A., Hurling, R., & Baker, S. (2011). Implicit shopping: Attitudinal determinants of the purchasing of healthy and unhealthy foods. Psychology & Health, 26(7), 875-885. doi:10.1080/08870446.2010.509797
Sheeran, P., Gollwitzer, P. M., & Bargh, J. A. (2013). Nonconscious processes and` health. Health Psychology, 32(5), 460-473. doi:10.1037/a0029203
Stadler, G., Oettingen, G., & Gollwitzer, P. M. (2010). Invertention effects of information and self-regulation on eating fruits and vegetables over two years. Health Psychology,
29(3),274-283. doi:IO.1037/aOOI8644
Strack, F., & Deutsch, R. (2004). Reflective and impulsive determinants of social behavior.
Personality and Social Psychology Review, 8, 220–247.
doi:10.1207/s15327957pspr0803_1
Trendel, O., & Werle, C. O. (2015). Distinguishing the affective and cognitive bases of implicit attitudes to improve prediction of food choices. Appetite, 104, 33-43. doi:10.1016/j.appet.2015.10.005
van der Laan, L. N., De Ridder, D. T., Viergever, M. A., & Smeets, P. A. (2012). Appearance matters: Neural correlates of food choice and packaging aesthetics. PLoS ONE, 7, e41738. doi:10.1371/ journal.pone.0041738
34 van der Laan, L. N., Papies, E. K., Hooge, I. T., & Smeets, P. A. (2017). Goal-Directed Visual
Attention Drives Health Goal Priming: An Eye-Tracking Experiment. Health
Psychology, 36, 82-90. doi:10.1037/hea0000410
van Kleef, E., Otten, K., & van Trijp, H. (2012). Healthy snacks at the checkout counter: A lab and field study on the impact of shelf arrangement and assortment structure on consumer choices. BMC Public Health, 12, 1072. doi:10.1186/1471-2458-12-1072 van Koningsbruggen, G. M., Stroebe, W., & Aarts, H. (2011). Through the eyes of dieters:
Biased size perception of food following tempting food primes. Journal of
Experimental Social Psychology, 47(2), 293-299. doi:10.1016/j.jesp.2010.10.012
Vohs, K. D. (2006). Self-regulatory resources power the reflective system: Evidence from five domains. Journal of Consumer Psychology, 16(3), 217-223.
doi:10.1207/s15327663jcp1603_3
Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions engender behaviour change? A meta-analysis of the experimental evidence. Psychological Bulletin, 132(2), 249- 268. doi:10.1037/0033-2909.132.2.249.
Wiers, R. W., Eberl, C., Rinck, M., Becker, E. S., & Lindenmeyer, J. (2011). Retraining automatic action tendencies changes alcoholic patients’ approach bias for alcohol and improves treatment outcome. Psychological science, 22(4), 490-497.
doi:10.1177/0956797611400615
Wiers, R. W., Rinck, M., Kordts, R., Houben, K., & Strack, F. (2010). Retraining automatic action-tendencies to approach alcohol in hazardous drinkers. Addiction, 105, 279–287. doi:10.1111/j.1360-0443.2009.02775.x
World Health Organisations (2016, June). Fact sheets: Obesity and overweight. Retrieved December 1, 2016, from http://www.who.int/mediacentre/factsheets/fs311/en/
35
Appendix A: Stimulus material Diet goal prime condition
36
37
Appendix B: IAT word list
Table 1
Affective and cognitive IAT word attributes
Hedonic IAT Goal IAT
Positive Negative Positive Negative
Smaakvol [Tasteful] Smakeloos [Tasteless] Gezond [Healthy] Ongezond [Unhealthy] Heerlijk [Delicious] Smerig [Nasty] Weinig calorieën [Low calories] Veel calorieën [High calories] Verrukkelijk [Delightful] Vies [Dirty]
Goed voor de lijn [Good for a slim-figure]
Dikmakend [Fattening] Hemels [Heavenly] Ranzig [Rancid] Gezond gewicht [Healthy weight] Overgewicht [Overweight] Lekker [Tasty] Saai [Boring] Slank [Slim] Zwaarlijvig [Obese]
38
Appendix C: Instruction Choice Task
The final computer task is a small adjustment of the previous third computer task. In this computer task, you will see two different products on the screen. The aim is that you choose, as quickly as possible, for one of these two products.
In this task, you choose a product on the left side of the screen with the E-key; If you want to choose the product on the right side of the screen, you should press the I-key.
Note - the products disappear after a random time, but the products never stay on the screen longer than three seconds. In other words – you will quickly have to make a choice. Do not think about the choice you want to make - choose spontaneously one of the two products!
This final task will take about three to four minutes.
If you have any questions about computer task 4, just ask the researcher BEFORE you start the task.
39
Appendix D: Randomization check summery tables
Table 2
Two-way ANOVA Summary Table: DV = Age
Source SS df MS F
PC 18.71 1 18.71 .23
IA HG 35.94 1 35.94 .45
Within groups 9726.41 121 80.38
Total 9782.77 123
Note. PC = prime condition; IA HG = implicit attitude health goal.
Table 3
Two-way ANOVA Summary Table: DV = Age
Source SS df MS F
PC 6.58 1 6.58 .08
IA HHG 287.96 1 287.96 3.67
Within groups 9474.38 121 78.30
Total 9782.77 123
Note. PC = prime condition; IA HHG = implicit attitude healthy hedonic goal.
Table 4
Two-way ANOVA Summary Table: DV = BMI
Source SS df MS F
PC 6.10 1 6.10 .94
IA HG 6.81 1 6.81 1.01
Within groups 816.89 121 6.75
Total 829.40 123
Note. PC = prime condition; IA HG = implicit attitude health goal.
Table 5
Two-way ANOVA Summary Table: DV = BMI
Source SS df MS F
PC 6.68 1 6.68 .99
IA HHG 3.50 1 3.50 .52
Within groups 820.20 121 6.78
Total 829.40 123
