Dietary restraint, ambivalence toward eating, and the valence and content of spontaneous associations with eating

Appetite 62 (2013) 150–159

Contents lists available at SciVerse ScienceDirect

Appetite journal homepage: www.elsevier.com/locate/appet

Research report

Dietary restraint, ambivalence toward eating, and the valence and content of spontaneous associations with eating Carmen Keller ⇑, Klazine van der Horst ETH Zurich, Institute for Environmental Decisions (IED), Consumer Behavior, Universitätstrasse 22, CHN J75.2, CH-8092 Zurich, Switzerland

a r t i c l e

i n f o

Article history: Received 18 July 2012 Received in revised form 2 October 2012 Accepted 17 November 2012 Available online 28 November 2012 Keywords: Restrained eating Ambivalence Health consciousness Spontaneous associations Gender Correspondence analysis

a b s t r a c t In a random sample of the German-speaking population of Switzerland (N = 1388, 730 women), the technique of spontaneous associations was used to examine the relationship between valence (positive, negative) and content of participants’ spontaneous associations with the term ‘‘eating’’ and dietary restraint, ambivalence toward eating, and health consciousness. Regression analysis revealed ambivalence to be the most important and food health consciousness the second most important predictor of restrained eating. Correspondence analysis of the content of the spontaneous associations revealed high-restrained eaters to have fewer associations with eating than unrestrained eaters. High-restrained eaters most often had negative associations with diet and positive associations with health. Unrestrained eaters mentioned a variety of positive associations, such as community, comfort and well-being, preparation of food, aesthetics, and various specific foods (e.g., starchy side dishes, and sweets). Results support the notion that the conflict between weight control and eating enjoyment is rather pronounced in high-restrained eaters, resulting in ambivalence toward eating, and the inhibition of associations with palatable foods. This was less pronounced, but still present, in medium-restrained eaters. In contrast, unrestrained eaters seemed to have a balanced and conflict-free relationship with eating. However, restrained eating also seemed to be driven by health considerations. Ó 2012 Elsevier Ltd. All rights reserved.

Introduction As food is easily accessible in Western countries, restrained eating may become an adaptive behavior strategy to limit weight gain (de Lauzon-Guillain et al., 2006). Dietary restraint is the intentional restriction of energy intake for the purpose of weight loss or weight control (Burger & Stice, 2011). However, most restrained eaters are not able to successfully reduce their weight or maintain their weight loss over long time periods (Mann & Ward, 2007). Researchers (Herman & Polivy, 1984; Stroebe, Mensink, Aarts, Schut, & Kruglanski, 2008) suggest that restrained eaters seem to be susceptible to psychological distress, such as feelings of conflict, ambivalence toward eating, and high levels of cognitive effort to control eating. The conflict model of eating suggests that the eating behavior of restrained eaters may be dominated by a conflict between the goal of weight control and eating enjoyment (Stroebe et al., 2008). Restrained eaters try to control and restrict intake of palatable foods, but at the same time, palatable foods have a positive effect (Fedoroff, Polivy, & Herman, 1997). As soon as the goal of enjoying palatable food is activated in restrained eaters, the conflicting goal of weight control is inhibited (Stroebe et al.,

⇑ Corresponding author. E-mail address: [email protected] (C. Keller). 0195-6663/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.appet.2012.11.012

2008). This mechanism is assumed to underlie the self-regulatory failure of restrained eaters in environments where attractive food is easily available (Papies, Stroebe, & Aarts, 2009; Stroebe et al., 2008). According to this goal-conflict model, the conflict between enjoying palatable food and controlling weight arises due to the more pronounced positive attitude and affective reaction toward palatable food of restrained eaters compared to unrestrained eaters; however, previous research has produced inconsistent results (for an overview, see Hofmann, van Koningsbruggen, Stroebe, Ramanathan, & Aarts, 2010). Therefore, in the present study we use the spontaneous word-association technique (Szalay & Deese, 1978) to examine the question of whether restrained and unrestrained eaters differ in their affective and intuitive responses toward eating palatable food and eating in general.

Affective and ambivalent reactions to palatable food in restrained and unrestrained eaters Restrained eaters are quick to consider the pleasure they would enjoy in eating the presented palatable food, which could influence their subsequent behavior (Papies, Stroebe, & Aarts, 2007). There is limited evidence for the hypothesis that restrained eaters are more responsive to the reward of the intake of palatable foods (Ahern, Field, Yokum, Bohon, & Stice, 2010; Burger & Stice, 2011). Other

C. Keller, K. van der Horst / Appetite 62 (2013) 150–159

studies using indirect measures to better grasp the spontaneous automatic nature of intuitive and affective reactions (Hofmann, Friese, & Strack, 2009) produced inconsistent results, ranging from more positive hedonic reactions to palatable foods in restrained eaters (Hoefling & Strack, 2008) to no difference in the enjoyment of high-fat palatable foods between restrained and unrestrained eaters (Roefs, Herman, MacLeod, Smulders, & Jansen, 2005). Papies et al. (2009) found that restrained eaters did not evaluate palatable foods as more positive than unpalatable or neutral foods, while unrestrained eaters did. Instead, restrained eaters were found to be more ambivalent toward palatable foods (i.e., held more pronounced evaluations of the negative aspects of such foods), which in turn affected participants’ scores on the implicit attitude measure (Papies et al., 2009; Stroebe et al., 2008). Attitudinal ambivalence refers to the simultaneous existence of strong positive and negative evaluations about the same attitude object (de Liver, van der Pligt, & Wigboldus, 2007; Sparks, Conner, James, Shepherd, & Povey, 2001). Instead of having more positive attitudes toward palatable foods, restrained eaters seem to experience more ambivalence, because the goal of dieting is incompatible with enjoying palatable food (Stroebe et al., 2008). A recent process-oriented study (Hofmann et al., 2010) revealed that under normal conditions, restrained eaters exert automatic implicit forms of self-control through the inhibition or devaluation of tempting stimuli. Through repeated confrontation with tempting cues, self-control seems to be impaired, and restrained eaters are drawn toward tempting food, while normal eaters show reduced hedonic responding over time (Hofmann et al., 2010). Thus, previous research indicate that it is still unclear whether restrained eaters (compared to unrestrained eaters) have stronger positive or negative, ambivalent, or inhibited affective responses to eating tempting foods, which may explain why restrained eaters have a goal conflict and unrestrained eaters do not. The spontaneous word-association technique to examine the affective and intuitive reaction toward palatable food and eating in general One reason for the inconsistent results or for the failure to find differences in intuitive and affective reactions between restrained and unrestrained eaters, respectively, may be that in previous research, direct (but also indirect) measures mainly focused on the valence of the affective reaction (positive versus negative) but not the content of the reaction. To examine the valence and the content of the affective reaction in the present study, we used the spontaneous word-association technique (Szalay & Deese, 1978). This technique assesses which spontaneous word associations come to people’s minds when they think about a stimulus and individuals’ affective ratings of those associations. The spontaneous word association refers to the content, while the affective rating refers to the valence of individuals’ affective reactions to a stimulus. The technique assesses individuals’ intuitive and affective imagery of a stimulus, which is theorized to guide judgments, decisions, and behaviors (Slovic, Finucane, Peters, & MacGregor, 2007). It was previously used to examine spontaneous reactions to various health-related risks evoked by, for example, adolescent healththreatening behaviors (Benthin et al., 1995), blood transfusion (Finucane, Slovic, & Mertz, 2000), global warming and climate change (Leiserowitz, 2005), and various technologies (Connor & Siegrist, 2011; Dohle, Keller, & Siegrist, 2012; Keller, Visschers, & Siegrist, 2012a, 2012b). The latter studies examined not only the valence (affective rating) but also the content of the spontaneous associations by utilizing correspondence analysis (Clausen, 1988). In the field of food and dietary behavior, correspondence analysis has been used to construct a graphical display of the relationship between the frequency of eating particular food items for males

151

and females (Guinot et al., 2001). In the present study, we used correspondence analysis to provide a visual display of the contents of the associations with eating for different levels of restrained eating and gender. Rationale of the present study In the present study, we assessed the valence and the content of participants’ spontaneous associations with the term ‘‘eating.’’ The use of the spontaneous word-association technique provides a way to assess the subjective meaning of a stimulus (for a short overview, see Leiserowitz, 2005). The technique offers an indirect and open way to capture the intuitive and affective imagery toward eating, which is assumed to underlie and direct restrained eating behavior. To assess the subjective meaning of eating as a behavior, we asked participants to associate with the wider term ‘‘eating’’ (instead of the more concrete term ‘‘food,’’ which may have resulted in a list of different foods). Using regression analysis, the first goal of the present study was to examine the influence of the valence of the spontaneous associations with eating (affective evaluation of the spontaneous word associations) and ambivalence toward eating on restrained eating controlling for demographics. Further relevant psychological determinants (e.g., emotional and external eating, health consciousness, liking and consumption of healthy and unhealthy food) were also examined or controlled for, respectively. Based on previous research that revealed that restrained eaters experience more ambivalence toward eating palatable foods than unrestrained eaters (Stroebe et al., 2008), we hypothesized that the feeling of ambivalence is the most important predictor of restrained eating. Using correspondence analysis, the second goal was to examine the relationship between the content of the spontaneous word associations and restrained eating. On the one hand, we hypothesized that the contents of the associations of restrained and unrestrained eaters would differ. In line with the goal-conflict model and previous research, we assumed restrained eaters to have associations with eating that portray the conflict between dieting and enjoying or eating palatable foods (Stroebe et al., 2008) or that illustrate the inhibition or devaluation of tempting foods, respectively (e.g., Hofmann et al., 2010; Papies et al., 2009). On the other hand, we expected unrestrained eaters to have associations with eating that indicate a conflict-free and balanced relationship with eating. In addition, we also hypothesized that the content of the associations of males and females would differ. Previous research has revealed gender differences in a various range of eating motives and attitudes (Renner, Sproesser, Strohbach, & Schupp, 2012). We therefore assumed that males and females also differ in their underlying spontaneous associations with eating. To our knowledge, the technique of spontaneous word association has not been used in the field of restrained eating behavior. The use of this technique may provide important insights into the intuitive and affective imagery underlying and directing restrained and unrestrained eating. Methods Participants Questionnaires and accompanying letters were sent to a random sample of addresses from the telephone book of the German-speaking part of Switzerland. The person in the household who was over 18 years of age and who was next in line for his/her birthday was requested to complete the questionnaire. A reminder letter was sent out 4 weeks later. A second questionnaire was sent to persons who did not respond to the letter or reminder. In total,

152

C. Keller, K. van der Horst / Appetite 62 (2013) 150–159

1388 persons sent back the questionnaire, with a response rate of 43% (730 women, 626 men, 32 missing). The mean age was 52 (SD = 16.0, 34 missing). The mean BMI was 24.52 (SD = 4.05, 40 missing). Males had higher BMIs (M = 25.48, SD = 3.61) than females (M = 23.70, SD = 4.23; see Table 1). The BMIs of males and females in the present study were similar to those assessed in the Swiss Food Panel (males: M = 25.80, SD = 3.7; females: M = 23.50, SD = 4.30), a study examining food choices in a large random sample (N = 6189) of the Swiss population (Hartmann, Siegrist, & van der Horst, 2012). The self-reported education level ranged from primary school or lower-secondary school (7.3%, n = 101) to upper-secondary vocational school (34.1%, n = 474), upper-secondary university preparation school (4.6%, n = 64), tertiary vocational school (22.3%, n = 310), and college or university (24.5%, n = 338). Of all participants, 88 indicated that they had attained none of these education levels, and 13 participants did not disclose their education level. In the sample of the present study, women, people between 40 and 64, and people with a degree of tertiary vocational school are slightly overrepresented compared to the German-speaking Swiss population (Swiss Federal Statistical Office, 2010). However, the aim of the present study was not to draw a representative sample of the German-speaking Swiss population. Instead, drawing a random sample from the general population, we aimed to cover a large variance of the most relevant demographics and personal characteristics such as gender, age, education, and BMI to obtain a large variance of spontaneous affective associations underlying eating behavior.

Materials Free associations Utilizing the technique of spontaneous word-association technique (Szalay & Deese, 1978), at the beginning of the questionnaire, participants were asked to freely associate about the term ‘‘eating.’’ They were asked, ‘‘When you think of ‘eating’, what is the first word, image, or thought that spontaneously comes to mind?’’ Subsequently, participants were asked whether a second and third word, image, or thought came to mind and, if so, to write it down.

Affect toward eating For each of the three possible associations, participants were asked to indicate whether it evoked a negative or positive feeling. For each association, participants could answer on an 11-point scale ranging from extremely negative ( 5), to neutral (0), to

extremely positive (+5). Affect toward eating was thus measured separately for the first, second, and third associations.

Eating behavior Restrained, emotional, and external eating were measured with a German translation of the Dutch Eating Behavior Questionnaire (DEBQ) (Van Strien, Frijters, Bergers, & Defares, 1986). The DEBQRestraint Scale consists of 10 items measuring deliberate, planned weight-control behaviors. Unlike the Restraint Scale developed by Polivy, Herman, and Warsh (1978), the DEBQ-Restraint Scale is not confounded by items related to unsuccessful dieting. Typical items for restrained eating include, ‘‘When you have put on weight, do you eat less than you usually do?’’ and ‘‘Do you take into account your weight with what you eat?’’ The DEBQ-Emotional Scale (13 items) refers to eating in response to arousal states such as anger, fear, or anxiety (‘‘Do you have a desire to eat when you are irritated?’’ and ‘‘Do you have a desire to eat when you are depressed or discouraged?’’). The DEBQ-External Scale (10 items) measures eating in response to external food-related stimuli, regardless of the internal state of hunger or satiety (‘‘If food tastes good, do you eat more than usual?’’ and ‘‘If you see others eating, do you also have the desire to eat?’’). Participants answered on a five-point scale ranging from 1 = never to 5 = very often. Cronbach’s a was .89 for restrained eating, .80 for external eating, and .94 for emotional eating. The mean values of the DEBQ-Restraint Scale of the present study (M = 2.60, SD = 0.81 for people with BMI < 25, M = 2.76, SD = 0.71 for people with BMI P 25; also see Table 1 for gender differences) are similar to the Restraint mean values mentioned in other studies utilizing a sample from the general population (M = 2.51, SD = 0.85 for BMI < 25; M = 2.84, SD = 0.71 for BMI P 25; van Strien, Herman, & Verheijden, 2009) or a sample of females (M = 2.70, SD = 0.76; Anschutz, van Strien, van De Ven, & Engels, 2009). For the correspondence analysis (see Correspondence analysis), participants were divided into three groups. Low-restrained eaters (43.5%, n = 567, including answering scale points 1 = never and 2 = rarely) corresponded with Restraint Scale values between 1 and 2.49. Medium-restrained eaters (43.6%, n = 568, including answering scale point 3 = sometimes) corresponded with Restraint Scale values between 2.50 and 3.49. Finally, high-restrained eaters (13%, n = 169, including answering scale points 4 = often and 5 = very often) corresponded with Restraint Scale values between 3.50 and 5. We preferred to use the original answering scale points to create the three groups instead of using a median split to divide participants only into high- and low-restrained eaters. First, we wanted to reflect the original answers of participants. Second, the use of a median split would have resulted in lost information.

Table 1 Gender differences (M, SD).

Restrained eating External eating Emotional eating Ambivalence Liking healthy foods Liking unhealthy foods Vegetable and fruit consumption (servings per week) Sweet and snack consumption (frequency per week) BMI Health consciousness Affect toward eating Overeating 

p < .05. p < .01. p < .001.

 ***

Males

Females

2.48 (0.77, n = 592) 2.67 (0.55, n = 600) 1.67 (0.61, n = 582) 2.28 (1.12, n = 620) 5.01 (0.87, n = 614) 4.07 (1.00, n = 609) 25.43 (14.00, n = 619) 4.97 (5.05, n = 622) 25.48 (3.61, n = 622) 4.02 (0.97, n = 619) 3.15 (2.13, n = 606) 3.55 (1.64, n = 621)

2.81 (0.74, n = 686) 2.69 (0.55, n = 693) 1.98 (0.69, n = 681) 2.58 (1.18, n = 719) 5.40 (0.71, n = 713) 3.84 (1.00, n = 705) 32.19 (15.33, n = 723) 4.84 (4.98, n = 724) 23.70 (4.23, n = 724) 4.50 (0.82, n = 715) 3.23 (2.34, n = 702) 3.60 (1.68, n = 715)

t-Value 8.11*** 0.86 8.49*** 4.90*** 8.96*** 4.18*** 8.38*** 0.49 8.35*** 9.77*** 0.68 0.59

C. Keller, K. van der Horst / Appetite 62 (2013) 150–159

However, we recruited enough restrained eaters to divide participants into low-, medium-, and high-restrained eaters.

153

Results Gender differences

Overeating Overeating was measured using the item, ‘‘How often do you eat or drink too much?’’ Participants answered on a six-point scale ranging from 1 = seldom or never to 6 = several times daily. Responses were as follows: 1 = seldom or never (n = 275, 19.8%), 2 = several times per year (n = 186, 13.4%), 3 = several times per month (n = 176, 12.7%), 4 = several times per week (n = 296, 21.3%), 5 = daily (n = 294, 21.2%), 6 = several times daily (n = 138, 9.9%), missing cases (n = 23, 1.7%). Ambivalence Experienced ambivalence toward eating was measured with an adapted version of the ambivalence scale (Stroebe et al., 2008) consisting of seven items (e.g., ‘‘I would enjoy tasty food more if it did not contain so many calories,’’ ‘‘A good meal tastes better if you forget that it makes you gain weight’’). Participants answered on a six-point scale ranging from 1 = does not apply at all to 6 = applies a lot. Cronbach’s a was .86. Food health consciousness This variable was measured with an adapted version of the Food Health Consciousness Scale of Schifferstein and Ophuis (1998) and consisted of five items (e.g., ‘‘It is important to me that I eat a healthy diet,’’ ‘‘I eat what I like; it doesn’t matter whether it is healthy or not’’). Participants answered on a six-point scale ranging from 1 = does not apply at all to 6 = applies a lot. Cronbach’s a was .75. Liking healthy/unhealthy food Liking healthy food referred to liking cooked/uncooked vegetables, fruits, whole grains, and so on. Liking unhealthy food referred to liking chocolates, salted snacks (nuts and chips), cookies, cakes, pastries, and so on. Participants answered on a six-point scale ranging from 1 = do not like at all to 6 = like a lot. Cronbach’s a for liking healthy and unhealthy food was .69 and .76, respectively. Fruit/vegetable consumption This was assessed on a five-point answering scale: ‘‘daily’’ (coded as 7 times/week), ‘‘4–6 times/week’’ (coded as 5 times/ week), ‘‘1–3 times/week’’ (coded as 2 times/week), ‘‘1–3 times/ month’’ (coded as 0.5 times/week) and ‘‘less or never’’ (coded as 0). Additionally, participants were asked how many portions of vegetables (one portion = a handful, or 50 g) and fruit (one piece or one handful) they usually eat. The items relating to portions were combined to indicate ‘‘servings’’ of vegetables and ‘‘pieces’’ of fruits consumed per week (Hartmann et al., 2012). We used the sum score of the consumption of cooked/uncooked vegetables, salad, and fruit. Cronbach’s a for fruit consumption was .51. Sweet and snack consumption Participants answered on a six-point scale. For statistical calculation, ‘‘several times per day’’ was assumed to be 2 times/day (coded as 14 times/week), ‘‘daily’’ was assumed to be 1 time/day (coded as 7 times/week), ‘‘several times per week’’ was assumed to be 3 times/week), and ‘‘several times per month’’ was assumed to be 3 times/month (coded as 0.75 times/week). Finally, ‘‘several times per year’’ and ‘‘less or never’’ were considered negligible (coded as 0). We used the sum score of the consumption of cookies, chocolates, and pastries to indicate consumption frequency of these foods per week (Hartmann et al., 2012). Cronbach’s a for snack consumption was .74.

A t-test revealed significantly higher levels of restrained eating, ambivalence, emotional eating, and health consciousness in females compared to males. There were no gender differences in affective evaluations of the associations with the term ‘‘eating.’’ The full results are presented in Table 1.

Regression analysis with restrained eating as dependent variable The correlation table (Table 2) shows that ambivalence and restrained eating as well as food health consciousness and restrained eating are highly correlated. No correlations were found between restrained and external eating or between restrained eating and affect toward eating. Ambivalence was significantly negatively correlated with liking unhealthy food. Correlations and regression coefficients between affect toward eating and further variables are only presented for the first mentioned association, but results were similar for the second and third associations. We submitted the data to a hierarchical regression analysis with restrained eating as the dependent variable. The independent variables were z-standardized (Aiken & West, 1991). In the first step, we included the socio-demographic variables gender, age, and education as well as BMI, affect toward eating, emotional and external eating, overeating, liking and consumption of healthy and unhealthy foods, and food health consciousness as independent variables. To examine the increment to R2 due to ambivalence over and above the other predictors, we included ambivalence in the second step. We further aimed to explore whether BMI moderated the relationship between ambivalence and restrained eating. Therefore, to examine the increment to R2 due to the interaction term over and above the other predictors (Cohen, Cohen, West, & Aiken, 2003), in the third step, we included the interaction term between the z-standardized BMI and ambivalence. In Table 3, the hierarchical regression analysis with the significant predictors is presented. The whole model explained 52.3% (F(6/1243) = 227.14; p < .001) of the variance in restrained eating. The results presented in Table 3 shows that ambivalence was the most important predictor (b = .50). It explained 20.3% (F(1/1244) = 506.75; p < .001) of the variance in restrained eating over and above the other predictors. Food health consciousness was the second most important predictor (b = .34) of restrained eating. The higher the experienced ambivalence with eating palatable food and the higher the food health consciousness, the higher the restrained eating. BMI was also significant (b = .09). The higher the BMI, the higher the restrained eating. The effects of ambivalence and BMI were in addition qualified by a significant negative interaction between them (b = 0.15), indicating that the impact of ambivalence on restrained eating is higher when BMI is low. Accordingly, the simple slope analysis (Aiken & West, 1991) revealed for participants with a low BMI (one SD below the mean BMI; i.e., a BMI of ca. 20.4) a significant effect of b = .64 (p < .001) of ambivalence on restrained eating. Conversely, for participants with a high BMI (one SD above the mean BMI; i.e., a BMI of ca. 28.4), a significant effect of b = .35 (p < .001) was found. However, explaining only 2% (F(1/ 1243) = 54.92; p < .001) of the variance in restrained eating, this interaction effect was rather small. Furthermore, overeating was also associated with restrained eating, and females had higher levels of restrained eating than males. Emotional and external eating, liking and consumption of healthy or unhealthy food, and affect toward eating were not significant predictors. There were no significant interaction effects with gender.

154

C. Keller, K. van der Horst / Appetite 62 (2013) 150–159

Table 2 Correlations between eating behavior, ambivalence, liking food and consumption, BMI, and age (N = 1122). 1 1. Restraint 2. External 3. Emotional 4. Ambivalence 5. Liking healthy foods 6. Liking unhealthy foods 7. Veg. and fruit consumption 8. Sweet and snack consumption 9. BMI 10. Food health consciousness 11. Affect toward eating 12. Overeating 13. Age * ** ***

2

3

4

5

6

7

8

9

10

11

12

– .02 .22*** .62*** .21*** .18*** .23*** .12*** .14*** .49*** .04 0.11*** .12***

.59*** .17*** .07* .32*** .01 .23*** .07* .18*** .05 .45*** .29***

.34*** .03 .18*** .11*** .22*** .13*** .01 .06* .35*** .18***

.04 .10*** .11*** .02 .19*** .27*** .12*** .18*** .07**

.00 .38*** .05 .06* .43*** .14*** .10** .18***

.08** .28*** .02 .24*** .04 .17*** .31***

.03 .10*** .36*** .10** .02 .01

.01 .16*** .03 .08** .07*

.12*** .06* .17*** .23***

.05 .14** .15***

.01 .03

.23***

Sig. (1-tailed): p < .05. Sig. (1-tailed): p < .01. Sig. (1-tailed): p < .001.

Table 3 Regression analysis with restrained eating as dependent variable (N = 1250). B

SE

Beta

Step 1 Constant Gender BMI Overeating Food health consciousness

2.54 0.23 0.15 0.12 0.36

0.03 0.04 0.02 0.02 0.02

0.15*** 0.20*** 0.15*** 0.47***

Step 2 Constant Gender BMI Overeating Food health consciousness Ambivalence

2.61 0.12 0.07 0.05 0.26 0.39

0.02 0.03 0.02 0.02 0.02 0.02

0.08*** 0.08*** 0.06** 0.34*** 0.49***

Step 3 Constant Gender BMI Overeating Food health consciousness Ambivalence Ambivalence  BMI

2.63 0.13 0.06 0.05 0.25 0.39 0.12

0.02 0.03 0.02 0.02 0.02 0.02 0.02

0.09*** 0.09*** 0.06** 0.34*** 0.50*** 0.15***

Note: R2 = 29.9% for Step 1, DR2 = 20.3% for Step 2 (p < .001), DR2 = 2.1% for Step 3 (p < .001). Gender (0 = male, 1 = female). *** p < .001. ** p < .01.

Content and affective evaluation of spontaneous associations The majority of respondents indicated a first association (n = 1289, 99 missing), a second association (n = 1206, 182 missing), and a third association (n = 1045, 343 missing). The 3540 associations were assigned to one of the 30 categories depicted in Table 4, which resulted from a qualitative analysis procedure. After the initial inspection of the associations, a codebook, including preliminary categories, was defined. The associations were then independently coded by two researchers. Associations that were not assigned to the same category by the two researchers were discussed. If it was not possible to clearly assign an association to a category, a new category was defined, and all associations were recoded using the revised categories. This procedure was repeated until all associations could be definitively assigned to a category by common consent. To examine inter-rater reliability, every fifth association (20%, n = 692) was coded and assigned to the categories by a third rater who was not previously involved. Inter-rater reliability (Cohen’s kappa) was very good (j = .89). A

complete list of the categories and their frequencies is provided in Table 4. Table 4 shows that the associations with enjoyment and hunger were most dominant (together 26%, n = 928). The association with enjoyment went along with a very positive affective evaluation (M = 4.2, SD = 1.26), whereas images with hunger were evaluated less positively but still slightly above neutral. Eating with family and friends (community, n = 257), preparation of food (n = 231), having time and gemütlichkeit (comfort and well-being: n = 154), and health-related associations (n = 141) were also quite prevalent. Only 20% (n = 471) of the associations with eating included particular foods or food categories, such as meat (n = 137), bread (n = 91), starchy side dishes (n = 75), sweets (72), various dishes (n = 46), dairy products (n = 31), sausages (n = 14), or muesli (n = 5), with vegetables and fruits (n = 218) being the most dominant. The majority of associations with eating were rated very positively (see Table 4). On the scale ranging from very negative ( 5) to neutral (0) to very positive (5), only 5.3% (n = 189) of the associations (N = 3540) were rated slightly negatively ( 1.4 for dieting and 0.3 for distributional justice). Of all associations, 13.6% (n = 482) were rated neutrally or slightly positively (0.75 for purchase, 1.05 for diversity of foods, 1.62 for hunger), 26.7% were rated very positively with a value higher than 4 (4.37 for community, 4.29 for aesthetics, 4.2 for enjoyment, 4.13 for dairy products), and 55.4% had an affect value between 2 and 3.9. First, the affective ratings of the three associations as judged by participants were analyzed (total number of associations N = 3540). Participants’ affective ratings of the first (M = 3.26, SD = 2.20, n = 1283), second (M = 3.15, SD = 2.35, n = 1201), and third associations (M = 2.74, SD = 2.81, n = 1039) were clearly positive. In total, 17 associations were not affectively rated and were therefore missing cases. According to a linear contrast revealed through a repeated measures ANOVA for participants who provided three associations, the affective rating of the associations decreased significantly from the first to the third association (F(1,1038) = 40.82, p < .001). Log-linear model To test our assumption that the content of people’s spontaneous associations with eating varied according to the level of restrained eating and gender, we conducted a three-way log-linear analysis with gender, restrained eating, and associations as primary factors. The three associations were taken together in an overall list of associations. Only categories with n > 10 were included in the log-linear model (n = 3535 associations).

155

C. Keller, K. van der Horst / Appetite 62 (2013) 150–159 Table 4 Affective imagery of respondents. Total n 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Total

Enjoyment (pleasure, desire, deliciousness) Hunger (being hungry, hunger) Community (eating with family and friends, sociability) Preparation of food (cooking, preparing meals) Vegetables and fruits Comfort and well-being (having time, comfort, gemütlichkeit) Health (healthiness, healthy food, organic, balanced nutrition) Meat (grilled, cooked, raw, menus with meat) Diet (dieting, getting big, eating (not) too much, calories, fat, overweight, reservation) Basic need (vital, essential) Eating facilities (cutlery, table, plate, table arrangement) Aesthetics (beautiful table and plate arrangement, presentation) Bread (sorts of bread) Positive descriptions (culinary, fresh, good, fine) Taste (flavor, aroma) Liking eating (looking forward to eating, what shall be eaten, eating is important) Starchy side dishes (pasta, potatoes) Eating out (restaurant, Chinese, Italian, Asian) Sweets (chocolates, cakes, deserts) Purchases (buying and storing food) Beverages (wine, beer, drinks) Meal (dinner, lunch, breakfast, menu) Distributional justice (malnutrition, famine in developing countries, abundance in developed countries) Eating place (kitchen, eating in the garden, picnic) Dish (particular recipes such as pizza, fondue, meat cut into strips and mashed potatoes, schnitzel, and French fries) Diversity (variety and range of food, variation) Food (food products, foodstuffs) Dairy products (milk, butter, cheese) Sausage (various sorts of sausages, lunchmeat, salami) Muesli

Males %

n

Females %

n

Affect

%

M

SD

556 372 257 231 218 154 141 137 132 114 100 100 91 88 79 75 75 74 72 69 69 61 57

15.7 10.5 7.3 6.5 6.2 4.4 4.0 3.9 3.7 3.2 2.8 2.8 2.6 2.5 2.2 2.1 2.1 2.1 2.0 1.9 1.9 1.7 1.6

244 182 87 72 68 62 58 99 60 59 41 36 42 42 37 43 38 44 24 15 47 33 26

15.5 11.6 5.5 4.6 4.3 3.9 3.7 6.3 3.8 3.8 2.6 2.3 2.7 2.7 2.4 2.7 2.4 2.8 1.5 1.0 3.0 2.1 1.7

301 185 164 151 146 88 78 38 68 52 57 60 48 46 42 30 37 29 45 53 18 25 28

15.9 9.8 8.7 8.0 7.7 4.7 4.1 2.0 3.6 2.7 3.0 3.2 2.5 2.4 2.2 1.6 2.0 1.5 2.4 2.8 1.0 1.3 1.5

4.2 1.62 4.37 2.28 3.86 3.66 3.57 3.93 1.4 2.11 3.11 4.29 3.88 3.73 3.76 2.35 3.86 2.88 3.61 0.75 3.35 2.86 0.3

1.26 2.47 1.27 2.66 1.6 2.13 1.84 1.48 2.95 2.4 1.89 1.25 1.34 1.79 1.75 2.4 1.65 2.29 1.96 2.39 1.79 1.77 4.22

50 46

1.4 1.3

22 30

1.4 1.9

27 15

1.4 0.8

3.42 3.63

2.26 2.1

18 16 14 11 2 1572

1.1 1.0 0.9 0.7 0.1 100.0

23 15 17 2 3 1891

1.2 0.8 0.9 0.1 0.2 100.0

1.05 3.37 4.13 3.64 3.8 3.07

3.78 1.75 1.31 2.13 1.79 2.45

41 31 31 14 5 3540

1.2 0.9 0.9 0.4 0.1 100

Note: Associations with n < 10 (Muesli) were included neither in the correspondence analysis nor in the log-linear analysis.

The analysis revealed that the main effects were significant (Pearson v2(31) = 4352.39; p < .001), as were the two-way interactions (Pearson v2(86) = 423.54; p < .001). However, the three-way interaction (Pearson v2 (56) = 45.36; p = .84) was not significant. Each of the three two-way interactions was significant (gender  association v2(28) = 155.43; p < .001; gender  restrained eating v2(2) = 132.77; p < .01; and association  restrained eating v2(56) = 114.77; p < .001). Thus, the significant two-way interactions indicated that, in line with our hypothesis, the associations significantly differed between participants who were low, medium, and high in restrained eating. In addition, and in line with our hypothesis, the associations differed between males and females. As we already know from testing gender differences, gender and restrained eating are also significantly associated. The non-significant three-way interaction indicates that the associations of males and females with low, medium, or high restrained eating, respectively, do not differ. For example, males and females with low-restrained eating did not have significantly different associations. Correspondence analysis Correspondence analysis (Clausen, 1988) is closely related to principal component analysis (PCA) or multidimensional scaling (MDS). While PCA is used to analyze interval data and MDS to analyze ordinal data, correspondence analysis is used to analyze nominal data. Because in our study, the database was a contingency table (including nominal data), we used correspondence analysis, which examines the relationship between two or more categorical variables (contingency tables) by representing the categories of the variables as points in space with the lowest possible number of dimensions. Categories with similar distributions are represented

as points that are close in space, while categories that have very dissimilar distributions are positioned far apart (Clausen, 1988). In the present study, only categories with n > 10 associations were included in this analysis (n = 3535 associations). A two-way contingency table with the association categories as columns (29 columns) and gender combined with restrained eating as rows (6 rows: gender (female, male)  restraint (1 = low, 2 = medium, 3 = high)) was used as input for the correspondence analysis (see Materials for information on the creation of these groups). The variables emotional eating and ambivalence were used as supplementary elements. Supplementary elements do not determine the solution space, but rather support the interpretation of the geometric orientation and serve as a validity check (Clausen, 1988). The answering scale points of the emotional eating and ambivalence items were used to divide participants into three groups as follows: emotional eating: low (83.7%, n = 1078, scale points 1 and 2), medium (14.4%, n = 186, scale point 3), and high (1.9%, n = 24, scale points 4 and 5) (100 missing); ambivalence: low (56.6%, n = 774, scale points 1 and 2), medium (37.2%, n = 509, scale points 3 and 4), and high (6.2%, n = 85, scale points 5 and 6) (20 missing). The graphical display depicted in Fig. 1 illustrates the relationship between the content of the associations and the different levels of restrained eating combined with gender. Ambivalence and emotional eating are also included in the graph. The first dimension explained 52% of the variance, while the second explained 28%. The third dimension and further dimensions explained only 10% and less; therefore, a two-dimensional solution was used for interpretation. The first dimension illustrates the differences in the content of the associations in relation to gender. Figure 1 shows that males

156

C. Keller, K. van der Horst / Appetite 62 (2013) 150–159

Fig. 1. Content of associations in relation to restrained eating and gender.

more often associate eating with liking to eat and high-calorie foods and beverages, such as ‘‘sausages,’’ ‘‘meat,’’ and ‘‘dishes.’’ In contrast, females associated eating with specific food items less often but had positive associations such as ‘‘well-being,’’ ‘‘community,’’ and ‘‘preparation of food.’’ Only ‘‘vegetables and fruits’’ represented a clearly female association with eating. The association with dieting did not differ between males and females. The second dimension illustrates the differences between low, medium, and high restrained eating. Figure 1 indicates that highrestrained eaters have fewer associations with eating than unrestrained eaters, who have more diverse associations. Unrestrained eaters mentioned different associations such as ‘‘community,’’ ‘‘comfort and well-being,’’ ‘‘preparation of food,’’ ‘‘aesthetics,’’ ‘‘meat,’’ ‘‘starchy side dishes,’’ ‘‘sweets,’’ and various specific foods (e.g., ‘‘sausages,’’ and ‘‘dairy products’’), whereas high-restrained eaters most often mentioned diet and health. In Fig. 1, only the second dimension is labeled as ‘‘restraint,’’ because restrained eating as well as the affective connotation of the associations can be interpreted in a dimensional way (high to low and positive to negative, respectively). The first dimension is not labeled. Gender is a categorical variable to classify the associations. The content of the associations is not interpretable along the first dimension of Fig. 1. It is common not to label the axes of the graphical display when using correspondence analysis (Clausen, 1988; Le Roux & Rouanet, 2010).

Considering emotional eating and ambivalence as supplementary elements, the correspondence analysis revealed that participants who were low in emotional eating and who were low in ambivalence toward eating also expressed ‘‘comfort’’ and ‘‘well-being’’ more often. In contrast, participants who were high in emotional eating and ambivalence mainly expressed associations related with diet. As found earlier, restrained eating was correlated with emotional eating (van Strien, Herman, Engels, Larsen, & van Leeuwe, 2007; van Strien et al., 2009) and ambivalence (Stroebe et al., 2008). Accordingly, this result provides evidence for the validity of the results of the correspondence analysis. We also included BMI (18– 23 = low, 23–25 = normal, 25–30 = high, >30 very high) and overeating (median split into low and high) as supplementary elements in the correspondence analysis. All BMI and overeating groups were located in the center of the graph, indicating that there was no relationship between the associations, BMI, and overeating. For clarity, we did not present overeating and BMI groups in Fig. 1.

Discussion Referring to the goal-conflict model (Stroebe et al., 2008) and using the spontaneous word-association technique (Szalay & Deese, 1978), the aim of the present study was to examine whether restrained eating was related with feelings of ambivalence. Another goal was to determine whether high-restrained and

C. Keller, K. van der Horst / Appetite 62 (2013) 150–159

unrestrained eaters differ with regard to their affective and intuitive responses toward eating in general and in particular toward eating palatable food.

157

evaluations of the spontaneous associations (affect toward eating) were not related to restrained eating. Spontaneous associations underlying restrained eating

Predictors of restrained eating We first examined the relationship between restrained eating, the valence (in terms of positive versus negative affect toward the spontaneous associations with eating), and ambivalence toward eating. In line with our hypotheses, the results of the regression analysis with restrained eating as the dependent variable showed that ambivalence was the most important predictor and health consciousness was the second most important predictor of restrained eating, even when further relevant variables were controlled for. The present results are in line with previous studies by Stroebe et al. (2008) and Papies et al. (2009), indicating that restrained eaters experienced more ambivalence with eating palatable foods than unrestrained eaters. Using a larger sample than previous studies, the present results support the conclusion that as the chronic goal of dieting is incompatible with enjoying palatable food, restrained eaters seem to experience more ambivalence toward eating palatable food than normal eaters (Stroebe et al., 2008). Furthermore, results of the present study suggest that this relationship is slightly moderated by individuals’ BMI, indicating that the impact of ambivalence on restrained eating is higher when BMI is low. These results might be interpreted in the context of recent research that examined the different psychological processes underlying successful and unsuccessful restrained eating (Koningsbruggen, Stroebe, & Aarts, 2011; Ouwehand & Papies, 2010). Successful restrained eaters (characterized by normal weight) were found to activate their dieting goal in response to tempting food, suggesting that in successful restrained eaters, tempting food activated processes of successful self-regulation (Ouwehand & Papies, 2010). Conversely, overweight restrained eaters activated their hedonic goal and seemed to forget about their dieting goal in tempting food situations (Ouwehand & Papies, 2010). Similarly, although both successful and unsuccessful restrained eaters were found to be tempted by palatable food, tempting food primes increased accessibility of the dieting goal in successful restrained eaters but decreased its accessibility in unsuccessful restrained eaters (Koningsbruggen et al., 2011). This suggests that successful restrained eaters resist the temptation of eating palatable food by keeping their dieting goal in mind (Koningsbruggen et al., 2011). Our results suggest that in people with low BMI, restrained eating seems to be more strongly influenced by the conflict between dieting goal and eating enjoyment than in people with high BMI. It might be that the activated diet goal as a response to palatable and tasty food in restrained eaters with low BMI amplifies the influence of the conflict between the goal of dieting and eating enjoyment on restrained eating. As with the diet goal, corresponding processes of self-regulation and selfrestriction are also activated. In contrast, in restrained eaters with high BMI, the hedonic goal is activated as a response to palatable food. This might attenuate the influence of the perceived conflict on restrained eating, as processes of self-regulation and selfrestriction might be impaired. Thus, restrained eaters with high BMI may sometimes restrain eating and sometimes they may not. However, caution should be exercised with this interpretation, as the interaction effect was rather small. Further research should more deeply examine the different processes underlying restrained eating, self-regulatory dieting success or failure, and ambivalence toward eating palatable food. Apart from these results, the present results further suggest that health considerations also seem to be drivers for restrained eating: Restrained eaters seem to follow a healthy dietary pattern to control their weight. However, in the regression analysis affective

Using correspondence analysis in the second step, we further analyzed the content of the participants’ spontaneous word associations with the term ‘‘eating.’’ In line with our hypothesis, the present results indicate that high, medium, and unrestrained eaters had different spontaneous associations with eating. Interestingly, highrestrained eaters had fewer associations with eating than unrestrained eaters, who had various and diverse associations with eating. Unrestrained eaters mentioned different positively evaluated associations, such as ‘‘community,’’ ‘‘comfort and well-being,’’ ‘‘preparation of food,’’ ‘‘aesthetics,’’ and various specific foods, which they also positively evaluated (e.g., ‘‘meat,’’ ‘‘sausages,’’ ‘‘dairy products,’’ ‘‘starchy side dishes,’’ and ‘‘sweets’’). In contrast, high-restrained eaters most often mentioned associations with diet and health, whereby they evaluated diet negatively. Thus, on the one hand, these results further support the findings of the regression analysis, which indicate that high-restrained eaters also seem to restrain eating due to health considerations. On the other hand, these results show that high-restrained eaters often associate eating with dieting, which is accompanied by negative feelings. In addition, it shows that high-restrained eaters do not spontaneously associate eating with high-calorie foods. These results thus indicate that in high-restrained eaters, the dieting goal seems to be active, whereas associations with tempting high-calorie foods seem to be inhibited. High-restrained eaters could have mentioned associations with palatable food tagged with a negative value (negative affect toward the association with palatable food). This would have pointed to the underlying mechanism of devaluation of palatable food. However, high-restrained eaters did not spontaneously mention palatable or other specific foods, which rather points to the underlying mechanism of inhibiting palatable food. In contrast, the low-restrained eaters often mentioned associations with specific palatable foods and positively evaluated them. This supports results of previous research (Papies et al., 2009) that found that high-restrained eaters inhibit associations with palatable food and unrestrained eaters value palatable foods. It is also in line with another study (Hofmann et al., 2010) that revealed that in the default condition (which is also assessed in the present study using a postal questionnaire), restrained eaters exert automatic implicit forms of self-control through the inhibition of tempting stimuli. This supports the idea that restrained eaters exert their dieting goal by inhibiting rather than devaluating goal-incompatible food. Normal eaters, however, seem not to be biased by a dieting goal and thus appear to be hedonically drawn toward tempting food under normal conditions (Stroebe et al., 2008). The present results thus nicely illustrate the benefit of the spontaneous word-association technique, which is an indirect measure that differentiates between the valence and content of affective and intuitive reactions toward foods. They also illustrate the benefit of the correspondence analysis for analyzing the content of the spontaneous association with eating underlying the different levels of restrained eating. In addition, the inclusion of the supplementary variables ambivalence and emotional eating into the correspondence analysis further indicates that high-restrained eaters who often mention the negatively evaluated association with diet are also those who have high levels of ambivalence and emotional eating. This suggests that high-restrained eaters may be driven by emotional eating, resulting in feelings of ambivalence toward eating palatable food due to their goal of dieting to control weight. Emotional eating and restrained eating were also correlated in the present study. However, only in the second step, when ambivalence was also controlled for, was emotional eating not a significant predictor of

158

C. Keller, K. van der Horst / Appetite 62 (2013) 150–159

restrained eating anymore, indicating that ambivalence possibly mediates the influence of emotional eating on restrained eating. Previous research has demonstrated a relationship between restrained eating and emotional eating (Anschutz et al., 2009) and binge eating (van Strien et al., 2007). Thus, further research should seek to clarify whether high-restrained eaters—especially the unsuccessful—may be driven by emotional eating resulting in the binge eating of palatable food. This in turn may result in higher feelings of ambivalence toward eating palatable food, as, due to emotional eating, the goal of dieting to control weight may be impaired. In summary, the results of the correspondence analysis of the spontaneous associations with eating illustrate that the goal conflict between dieting and eating (Stroebe et al., 2008) seems to be rather pronounced in high-restrained eaters. They seem to be susceptible to psychological distress (e.g., negative feelings of conflict) and ambivalence toward eating, and they appear to use cognitive effort to control eating and inhibit their associations with palatable foods. This conflict seems to be less pronounced, but still present, in medium-restrained eaters. In contrast, unrestrained eaters seem to have a balanced and conflict-free relationship with eating. Most importantly, restrained eating also seems to be driven not only by weight concerns but also by health considerations. In addition, the present results revealed females to have higher levels of restrained and emotional eating, ambivalence, health consciousness, and consumption of healthy food than males and to have associations with different contents. Besides previously known differences with regard to food preferences (e.g., females preferring vegetables and fruits and males preferring meat) (Jensen & Holm, 1999; Wardle et al., 2004), the present study found females to mention associations with community or food preparation more often, whereas males mentioned eating out or liking to eat (partially pointing to gender-specific lifestyles) more often. Certain limitations and critical arguments must be discussed. The present study was cross-sectional in nature. Longitudinal research may clarify the question of causality; that is, the question of whether spontaneous associations direct eating behavior at a later point in time. In addition, one might argue that ambivalence is the main predictor of restrained eating, because both constructs measure the same thing. The scale to measure restrained eating, however, assesses restrained eating as a means to control weight gain, whereas the ambivalence scale assesses the temptation of eating tasty and palatable food and the co-occurring fear of eating too many calories and putting on weight. In conclusion, the present study provides important insights into the differences in the psychological precondition of eating and food perception between high- and low-restrained eaters as well as the differences between males and females. It also illustrates that the spontaneous word-association technique and the analysis of the content of the associations (not only of the valence of the associations) using correspondence analysis is an effective means of providing more insights into the affective and intuitive imagery underlying and directing restrained eating. References Ahern, A. L., Field, M., Yokum, S., Bohon, C., & Stice, E. (2010). Relation of dietary restraint scores to cognitive biases and reward sensitivity. Appetite, 55(1), 61–68. http://dx.doi.org/10.1016/j.appet.2010.04.001. Aiken, L. S., & West, S. G. (1991). Multiple regression. Testing and interpreting interactions. Thousand Oaks, CA: Sage. Anschutz, D. J., van Strien, T., van De Ven, M. O. M., & Engels, R. C. M. E. (2009). Eating styles and energy intake in young women. Appetite, 53(1), 119–122. http://dx.doi.org/10.1016/j.appet.2009.03.016. Benthin, A., Slovic, P., Moran, P., Severson, H., Mertz, C. K., & Gerrard, M. (1995). Adolescent health-threatening and health-enhancing behaviors – A study of word-association and imagery. Journal of Adolescent Health, 17(3), 143–152. 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. Neuroimage, 55(1), 233–239. http://dx.doi.org/10.1016/ j.neuroimage.2010.12.009. Clausen, S. E. (1988). Applied correspondence analysis. An introduction. Thousand Oaks: Sage. Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/ correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Lawrence Erlbaum Associates. Connor, M., & Siegrist, M. (2011). The power of association. Its impact on willingness to buy GM food. Human and Ecological Risk Assessment, 17(5), 1142–1155. http://dx.doi.org/10.1080/10807039.2011.605725. de Lauzon-Guillain, B., Basdevant, A., Romon, M., Karlsson, J., Borys, J. M., & Charles, M. A. (2006). Is restrained eating a risk factor for weight gain in a general population? American Journal of Clinical Nutrition, 83(1), 132–138. de Liver, Y., van der Pligt, J., & Wigboldus, D. (2007). Positive and negative associations underlying ambivalent attitudes. Journal of Experimental Social Psychology, 43(2), 319–326. http://dx.doi.org/10.1016/j.jesp. 2006.02.012. Dohle, S., Keller, C., & Siegrist, M. (2012). Mobile communication in the public mind. Insights from free associations related to mobile phone base stations. Human and Ecological Risk Assessment, 15(4), 435–446. Fedoroff, I. C., Polivy, J., & Herman, C. P. (1997). The effect of pre-exposure to food cues on the eating behavior of restrained and unrestrained eaters. Appetite, 28(1), 33–47. Finucane, M. L., Slovic, P., & Mertz, C. K. (2000). Public perception of the risk of blood transfusion. Transfusion, 40(8), 1017–1022. http://dx.doi.org/10.1046/j.15372995.2000.40081017.x. Guinot, C., Latreille, J., Malvy, D., Preziosi, P., Galan, P., Hercberg, S., & Tenenhaus, M. (2001). Use of multiple correspondence analysis and cluster analysis to study dietary behaviour. Food consumption questionnaire in the SU.VI.MAX. cohort. European Journal of Epidemiology, 17(6), 505–516. http://dx.doi.org/10.1023/ a:1014586129113. Hartmann, C., Siegrist, M., & van der Horst, K. (2012). Snack frequency. Associations with healthy and unhealthy food choices. Public Health Nutrition. http:// dx.doi.org/10.1017/S1368980012003771. Herman, C. P., & Polivy, J. (1984). A boundary model for the regulation of eating. In J. A. Stundkard & E. Stellar (Eds.), Eating and its disorders (pp. 141–156). New York: Raven Press. Hoefling, A., & Strack, F. (2008). The tempting effect of forbidden foods. High calorie content evokes conflicting implicit and explicit evaluations in restrained eaters. Appetite, 51(3), 681–689. http://dx.doi.org/10.1016/j.appet.2008.06.004. Hofmann, W., Friese, M., & Strack, F. (2009). Impulse and self-control from a dualsystems perspective. Perspectives on Psychological Science, 4(2), 162–176. http:// dx.doi.org/10.1111/j.1745-6924.2009.01116.x. Hofmann, W., van Koningsbruggen, G. M., Stroebe, W., Ramanathan, S., & Aarts, H. (2010). As pleasure unfolds. Hedonic responses to tempting food. Psychological Science, 21(12), 1863–1870. Jensen, K. O., & Holm, L. (1999). Preferences, quantities and concerns. Socio-cultural perspectives on the gendered consumption of foods. European Journal of Clinical Nutrition, 53(5), 351–359. Keller, C., Visschers, V. H. M., & Siegrist, M. (2012a). Affective imagery and acceptance of replacing nuclear power plants. Risk Analysis, 32(3), 464–477. Keller, C., Visschers, V. H. M., & Siegrist, M. (2012b). The content and not only the valence of spontaneous associations should be analyzed. Risk Analysis, 32(3), 481–482. Koningsbruggen, M. G., Stroebe, W., & Aarts, H. (2011). Through the eyes of dieters. Biased size perceptions of food following tempting food primes. Journal of Experimental Social Psychology, 47, 293–299. Le Roux, B., & Rouanet, H. (2010). Multiple correspondence analysis. Thousand Oaks: Sage. Leiserowitz, A. A. (2005). American risk perceptions. Is climate change dangerous? Risk Analysis, 25(6), 1433–1442. http://dx.doi.org/10.1111/j.15406261.2005.00690.x. Mann, T., & Ward, A. (2007). Attention, self-control, and health behaviors. Current Directions in Psychological Science, 16(5), 280–283. http://dx.doi.org/10.1111/ j.1467-8721.2007.00520.x. Ouwehand, C., & Papies, E. K. (2010). Eat it or beat it. The differential effects of food temptations on overweight and normal-weight restrained eaters. Appetite, 55(1), 56–60. Papies, E., Stroebe, W., & Aarts, H. (2007). Pleasure in the mind. Restrained eating and spontaneous hedonic thoughts about food. Journal of Experimental Social Psychology, 43(5), 810–817. http://dx.doi.org/10.1016/ j.jesp. 2006.08.001. Papies, E. K., Stroebe, W., & Aarts, H. (2009). Who likes it more? Restrained eaters’ implicit attitudes towards food. Appetite, 53(3), 279–287. http://dx.doi.org/ 10.1016/j.appet.2009.07.001. Polivy, J., Herman, C. P., & Warsh, S. (1978). Internal and external components of emotionality in restrained and unrestrained eaters. Journal of Abnormal Psychology, 87(5), 497–504. Renner, B., Sproesser, G., Strohbach, S., & Schupp, H. T. (2012). Why we eat what we eat. The eating motivation survey (TEMS). Appetite, 59(1), 117–128. http:// dx.doi.org/10.1016/j.appet.2012.04.004. Roefs, A., Herman, C. P., MacLeod, C. M., Smulders, F. T. Y., & Jansen, A. (2005). At first sight. How do restrained eaters evaluate high-fat palatable foods? Appetite, 44(1), 103–114. http://dx.doi.org/10.1016/j.appet.2004.08.001. Schifferstein, H. N. J., & Ophuis, P. (1998). Health-related determinants of organic food consumption in the Netherlands. Food Quality and Preference, 9(3), 119–133. http://dx.doi.org/10.1016/s0950-3293(97)00044-x.

C. Keller, K. van der Horst / Appetite 62 (2013) 150–159 Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D. G. (2007). The affect heuristic. European Journal of Operational Research, 177(3), 1333–1352. http://dx.doi.org/ 10.1016/j.ejor.2005.04.006. Sparks, P., Conner, M., James, R., Shepherd, R., & Povey, R. (2001). Ambivalence about health-related behaviours. An exploration in the domain of food choice. British Journal of Health Psychology, 6, 53–68. Stroebe, W., Mensink, W., Aarts, H., Schut, H., & Kruglanski, A. W. (2008). Why dieters fail. Testing the goal-conflict model of eating. Journal of Experimental Social Psychology, 44(1), 26–36. http://dx.doi.org/10.1016/j.jesp. 2007.01.005. Swiss Federal Statistical Office (2010). Statistisches lexikon der Schweiz [Statistical lexicon of Switzerland]. . Szalay, L. B., & Deese, J. (1978). Subjective meaning and culture. An assessment through word associations. Hillsdale, NJ: Lawrence Erlbaum Associates.

159

Van Strien, T., Frijters, J. E. R., Bergers, G. P. A., & Defares, P. B. (1986). The Dutch eating behavior questionnaire (DEBQ) for assessment of restrained, emotional, and external eating behavior. International Journal of Eating Disorders, 5(2), 295–315. van Strien, T., Herman, C. P., Engels, R., Larsen, J. K., & van Leeuwe, J. F. J. (2007). Construct validation of the Restraint Scale in normal-weight and overweight females. Appetite, 49(1), 109–121. http://dx.doi.org/10.1016/j.appet. 2007.01.003. van Strien, T., Herman, C. P., & Verheijden, M. W. (2009). Eating style, overeating, and overweight in a representative Dutch sample. Does external eating play a role? Appetite, 52(2), 380–387. http://dx.doi.org/10.1016/j.appet.2008.11.010. Wardle, J., Haase, A. M., Steptoe, A., Nillapun, M., Jonwutiwes, K., & Bellisle, F. (2004). Gender differences in food choice. The contribution of health beliefs and dieting. Annals of Behavioral Medicine, 27(2), 107–116. http://dx.doi.org/ 10.1207/s15324796abm2702_5.