1

1

Running Head: Healthy beliefs and food intake. 2

3

4

Perceived healthiness of food: If it’s healthy, you can eat more! 5

6

Véronique Provencher, Janet Polivy, & C. Peter Herman 7

Department of Psychology, University of Toronto, Canada 8

9

10

Corresponding Author (and requests for reprints): 11

Véronique Provencher, RD, Ph.D. 12

Department of Psychology, University of Toronto at Mississauga 13

3359 Mississauga Rd N., Mississauga, ON L5L 1C6 14

Phone: (905) 828-3957; Fax: (905) 569-4326 15

E-mail: veronique.provencher@utoronto.ca 16

This research was supported by a grant from the Canadian Institute for Health Research to 17

the first author and from the Social Sciences & Humanities Research Council to the second 18

and third author. 19

2

Abstract 1

The main aim of this study was to investigate the effects of food-related beliefs 2

about the healthiness of foods, restrained eating, and weight salience on actual food intake 3

during an ad libitum snack. In a 2 (healthy vs. unhealthy) by 2 (restrained vs. unrestrained 4

eaters) by 2 (weight salient vs. not salient) factorial design, 99 female undergraduate 5

students were invited to taste and rate oatmeal-raisin cookies. Dietary restraint and weight 6

salience did not influence snack intake, but participants ate about 35% more when the 7

snack was regarded as healthy than when it was seen as unhealthy. Ratings of the snack 8

food’s “healthiness,” “capacity to affect weight” and “appropriateness in a healthy menu” 9

also indicated that the “healthy” manipulation was effective. In addition, the “weight 10

salience” manipulation appears to influence perceptions about food differently in restrained 11

versus unrestrained eaters, in that restrained eaters rated the snack food more negatively 12

than unrestrained eaters did when they received weight feedback before eating. Beliefs 13

about the healthiness of foods may thus be of great relevance to both food intake and 14

weight gain. 15

16

Key words: Food perceptions, food intake, restrained eating, weight, eating behavior. 17

18

19

20

21

3

Introduction 1

Calorie-restricted diets are the weight-loss strategy of choice for overweight and 2

obese individuals; unfortunately these diets do not appear to be successful on a long-term 3

basis (Miller, 1999; Polivy & Herman, 2002). Consequences of caloric restriction, such as 4

hunger and disinhibited eating, have been related to weight regain (Elfhag & Rossner, 5

2005). Restrictive diets may also increase appetite (Doucet et al., 2000) and the frequency 6

of obsessive thoughts about food and eating (Hart & Chiovari, 1998). Polivy, Coleman, and 7

Herman (2005) reported a significant effect of deprivation on craving and overeating, 8

mainly in restrained eaters, suggesting that eliminating “forbidden” foods may be 9

counterproductive as a strategy for better managing food intake. Nevertheless, energy 10

intake is obviously a core component in weight management, as is encouraging individuals 11

to make healthier food choices. We still do not fully understand which factors determine, 12

individually or in combination, healthy eating patterns. 13

Perceptions of healthy eating could be considered as one of the many determinants 14

of eating patterns (Paquette, 2005). Previous studies have shown that foods can be (and 15

often are) categorized as healthy or unhealthy (Oakes & Slotterback, 2001; Carels, Harper, 16

& Konrad, 2006; Carels, Konrad, & Harper, 2007). Various factors may influence the 17

healthy/unhealthy categorization of foods, such as their perceived fat content (Carels et al., 18

2006) as well as some stereotypical beliefs related to their names (Oakes, 2006). 19

Furthermore, it has been demonstrated that perceptions about healthiness or “fatteningness” 20

of foods may bias estimations of caloric content of foods (Carels et al., 2006; Carels et al., 21

2007). More specifically, when compared to the actual caloric content of the foods, 22

“healthy” food choices were perceived as having a lower caloric content (underestimation) 23

4

whereas “unhealthy” food choices were considered as having a higher caloric content 1

(overestimation) (Carels et al., 2006; Carels et al., 2007). Moreover, restaurants claiming 2

to serve “healthy” foods may lead consumers to underestimate the caloric density of the 3

foods they offer (Chandon & Wansink, 2007). The “health halo” effect refers to the fact 4

that individuals are more likely to underestimate the caloric content of main dishes and to 5

choose high-caloric side dishes in restaurants claiming to offer “healthy” food choices (e.g., 6

Subway) than in restaurants that do not make that claim (e.g., McDonalds) (Chandon & 7

Wansink, 2007). Low-fat nutrition claims could also contribute to overeating by increasing 8

consumer’s serving-size estimate and reducing guilt associated with eating (Wansink & 9

Chandon, 2006). These results suggest that food intake may be influenced by beliefs about 10

the healthiness of foods. The present study is thus designed to test the associations between 11

such food beliefs and actual food intake. 12

Some limited previous research has suggested that individual differences could have 13

an impact on food-related beliefs and accordingly may influence caloric estimation (Carels 14

et al., 2007). For example, even though both dieters and non-dieters were inaccurate in 15

estimating the caloric content of foods, current dieters were less inaccurate, showing a 16

smaller discrepancy between their caloric estimations of foods and the actual caloric 17

content of these foods than did non-dieters (Carels et al., 2007). However, some 18

differences in the behaviors of chronic dieters (restrained eaters) and current dieters (people 19

who simply report being on a diet at this moment) have been reported, with chronic dieters 20

often being more likely to overeat in response to situational provocations (e.g., Polivy, 21

1996). In addition, restrained eaters are usually engaged in a long-term struggle to achieve 22

(or at least maintain) a body shape that conforms to society’s thin-ideal female physique 23

5

(e.g., Polivy & Herman, 2007) whereas the restrictive eating behaviors of current dieters 1

might reflect a relatively short-term restriction (e.g., until they reach a limited weight-loss 2

target). Restrained eating or chronic dieting, then, may differ from current dieting in how it 3

influences the relations between food-related beliefs about healthiness of foods and actual 4

food intake. 5

Food intake in restrained eaters also appears to be strongly affected by weight 6

feedback. Restrained eaters who were led to believe that they were heavier than they 7

previously thought ate significantly more during a subsequent “taste test” than did 8

restrained eaters who did not receive such weight feedback; they also ate more than did 9

unrestrained eaters (regardless of whether or not they received false weight feedback) 10

(McFarlane, Polivy, & Herman, 1998). This finding suggests that simply being weighed 11

(i.e., having one’s weight made salient) could alter the effects of food-related beliefs on 12

actual food intake in restrained eaters. 13

The main objective of the current study was to investigate the effects of food-related 14

beliefs about the healthiness of foods on actual food intake during an ad libitum snack. In 15

addition, we were interested in examining the extent to which restrained eating and having 16

one’s weight made salient might moderate the effects of food-related beliefs on food intake 17

(during an ad libitum snack as well as the day’s total energy intake). As underestimations 18

of caloric content of healthy food have been previously observed (Carels et al., 2006; 19

Carels et al., 2007; Chandon & Wansink, 2007), we hypothesized that food intake would be 20

greater if the ad lib snack was described as “healthy” rather than “unhealthy.” Given that 21

perceived caloric content is affected by perceptions of healthiness, we also predicted that 22

the “healthy” snack will be estimated as having lower caloric content that the “unhealthy” 23

6

snack. Further, we predicted that food intake during a “healthy” ad lib snack would be 1

higher in restrained eaters than in unrestrained eaters, and that the opposite would be 2

observed with the “unhealthy” ad lib snack. In fact, chronic dieters might be more 3

susceptible to overeat healthy food because beliefs about the food’s healthiness could 4

authorize them to eat; by the same token, perceiving the food to be unhealthy might have a 5

more powerful inhibitory effect on the intake of restrained eaters. Finally, because 6

restrained eaters were sensitive to weight feedback (McFarlane et al., 1998), and because 7

simply being weighed could remind them that they should restrain their eating, we 8

predicted that the intake of a “healthy” or “unhealthy” ad lib snack would be lower in 9

restrained eaters weighed before eating than in restrained eaters who were weighed after 10

eating or in unrestrained eaters (weighed before or after eating). 11

12

Methods 13

Participants and study design 14

Participants were 99 female undergraduate students (mean age = 19.4 ± 2.8 yrs; 15

mean body mass index [BMI] = 23.2 ± 4.2 kg/m2) recruited online from the introductory 16

psychology course at the University of Toronto at Mississauga. They were invited to 17

participate in a market-research study investigating various dimensions of new snack 18

products. In compensation for their participation in the 2 sessions of the study, participants 19

received experimental course credit or were paid a small fee ($20). The study was 20

approved by the Office of Research Ethics at the University of Toronto, and all participants 21

provided informed consent. 22

7

Participant were randomly assigned to one of the experimental conditions in a 2 1

(healthy vs. unhealthy) by 2 (restrained vs. unrestrained eater) by 2 (weight salient or not) 2

factorial design. Each participant was tested on an individual basis twice, in separate 45-3

minute experimental sessions (between 11:00 a.m. and 5:00 p.m. for the first session and 4

between 9:00 a.m. and noon for the second session). 5

Measurements and procedure 6

Taste-rating task. Participant arrived at the laboratory for their first appointment in a 7

pre-meal state (i.e., at least 2 hours without food prior to the experiment). At that time, 8

participants were informed that this market-research study involved a taste-rating task in 9

which they would taste and rate a new snack food. More specifically, a pre-weighed plate 10

of approximately 500g of freshly baked bite-sized English Bay oatmeal-and-raisin cookies 11

was presented to the participant, with a taste-rating form and a glass of water. On average, 12

one bite-size cookie is about 9g, which represents approximately 40 kcal per cookie. 13

Oatmeal-raisin cookies were then described by the experimenter, but the description given 14

to the participant differed according to the condition to which she was randomly assigned 15

(“healthy snack” vs. “unhealthy snack”). In the “healthy snack” condition, the cookies 16

were described by the experimenter as follows: “The snack product that you have to taste 17

today is a new high-fibre oatmeal snack made with healthy ingredients. You have certainly 18

heard that whole oatmeal is good for your health because it contains soluble fibres. So, this 19

new oatmeal snack is high in soluble fibres, as well as low in saturated fat and free from 20

trans fat.” These are all characteristics of healthy food choices so this description was 21

designed to make the cookies appear to be a healthy snack. In the “unhealthy snack” 22

condition, the cookies were described as “new gourmet cookies made with fresh butter and 23

8

old-fashioned brown sugar. So, these new cookies are a great treat with a pleasant, sweet 1

taste.” The experimenter asked the participant to taste and rate the snack food on the 2

dimensions listed on the taste-rating form during a subsequent 10-minute period. With 3

visual analogue scales (VAS) (150mm), the taste-rating form measured the perceived 4

palatability of the snack tested (i.e., salty, sweet, crunchy, bitter, sour, and good-tasting). 5

Participants were instructed to eat as many cookies as were needed to achieve accurate 6

ratings and to feel free to help themselves to cookies after they completed the taste-rating 7

task, as long as they did not change their initial ratings. After participants completed the 8

ratings, the plate of cookies was removed and weighed to measure grams of cookies eaten 9

by each participant. 10

Weight, height and BMI. Described as normative data collection, weight and height 11

were measured in all participants, and BMI was then calculated (kg/m2). According to 12

weight-salience condition randomization (i.e., weight salient or not), measurements were 13

performed for half of the participants before the taste-rating task whereas the remaining 14

participants were weighed immediately after the taste-rating task but before the completion 15

of the questionnaires. The participants were made aware of their weight by being told 16

directly what their weight was on the scale. 17

Questionnaires. After the taste-rating task, participants were asked to complete 18

some questionnaires. The Restraint Scale (Herman & Polivy, 1980) was used to assess 19

whether participants exhibit behavioral and attitudinal concerns about dieting and weight 20

control. Participants were categorized as restrained eaters (scores of 15 or higher) or 21

unrestrained eaters (scores below 15), as performed previously (Goldman, Herman & 22

Polivy, 1991). The validity of the Restraint Scale has been previously reported (Allison, 23

9

Kalinsky, & Gorman, 1992; Laessle, Tuschl, Kotthaus, & Pirke, 1989; Polivy, Herman, & 1

Howard, 1988; van Strien, Herman, Engels, Larsen, & van Leeuwe, 2007). Participants 2

were also asked to report their opinions (on 8-point scales) about the new snack food tested. 3

More specifically, they were asked 1) “How healthy is this snack for you?” (from very 4

unhealthy [1] to very healthy [8]), 2) “If you were eating this snack regularly, how would it 5

affect your weight?” (from weight loss [1] to weight gain [8]), and 3) “Do you consider this 6

snack as appropriate in a healthy menu?” (from very inappropriate [1] to very appropriate 7

[8]). In addition, they had to estimate the caloric content of a serving size of 6 pieces of the 8

new snack food. 9

Dietary intake. After having completed the first experimental session, participants 10

were told to return to the laboratory the following day to taste and rate a second new snack 11

food. Participants therefore arrived the next morning believing that they would have to 12

taste and rate another snack, but the purpose of the second session was actually to assess 13

food consumption during the previous day (i.e., the day of the experiment). Food intake 14

was assessed by 24-h recall, administrated by the experimenter following standardized 15

procedures. Note that participants were also asked to report the amount of snack they ate 16

during the experiment (self-reported amount of snack intake eaten). A dietician reviewed 17

the food-recall report and calculated nutrient intakes using the Food Processor SQL 18

software (ESHA Research, 2008). At the end of the second session, the participant was 19

informed that there was no additional taste-rating test. The experimenter then explained to 20

her the true purpose of the study and asked her not to discuss any of the details of the study 21

with other people who might participate in the study. 22

23

10

Statistical analyses 1

A 2 (healthy vs. unhealthy) by 2 (restrained vs. unrestrained eater) by 2 (weight salient or 2

not) analysis of variance (ANOVA) was conducted to assess the effects of these three 3

variables on actual food intake during an ad libitum snack as well as on total energy intake 4

during the day of the experiment. Opinions about the new snack food (i.e., “healthy,” 5

“weight-gain” and “appropriateness” ratings, and perceived caloric content) were also 6

analyzed using a 3-way ANOVA. Pearson correlation coefficients were also calculated to 7

assess associations between actual and estimated amounts of snack food eaten, as well as 8

between participants’ ratings of the snack food’s “healthiness,” “capacity to affect weight” 9

and “appropriateness in a healthy menu.” The probability level for significance used for 10

the interpretation of all statistical analyses was set at an alpha level of p<0.05, and all data 11

were analyzed using SPSS statistical software (version 15.0 for Windows). 12

13

Results 14

Table 1 presents descriptive characteristics of the sample in each experimental 15

condition. No significant differences were observed between experimental groups for age 16

and BMI, while as expected, a difference in restraint scores was observed between 17

restrained and unrestrained eaters, F(7,91) = 38.48; p<0.0001. 18

19

20

21

11

Food consumption 1

A 3-way ANOVA revealed only a main effect of beliefs about healthiness of food 2

on food intake during the taste test, F(1,91) = 4.92; p<0.05. Participants ate more (14.4g or 3

about 56 kcal or 35% more) when the food was described as a healthy snack (M=56.0, 4

SD=34.6 g; healthy condition) than when it was presented as cookies (M=41.6, SD=27.8 g; 5

unhealthy condition). No differences were found as a function of restraint status or weight 6

salience. When the magnitude of the significant difference observed was assessed using 7

Cohen’s d effect-size (ES) estimates (d = standardized difference, i.e. difference between 8

means divided by their pooled standard deviation [strength of ES defined as small (d=0.20 9

to 0.49), moderate (d=0.50 to 0.79) and large (d ≥ 0.80)]) (Cohen, 1992), we obtained an 10

ES of 0.46, which is considered to be a small ES of perceived healthiness of food on intake. 11

There was a marginal main effect of beliefs about healthiness of food on the amount of 12

snack intake reported in the 24-h food recall, F(1,91) = 3.46; p<0.08. Self-reported amount 13

of snack eaten (in kcal) tended to be higher in the healthy condition (M=195.2, SD=112.9 14

kcal) than in the unhealthy condition (M=155.8, SD=100.9 kcal). No significant effects 15

were observed for the difference between the actual versus self-reported snack intakes, and 16

the amounts of snack food eaten and reported were significantly correlated, r=0.77; 17

p<0.0001. 18

Dietary intake during the entire day of the experiment was also assessed by a 24-h 19

food recall and a 3-way ANOVA was performed on total energy intake. While no main 20

effects of beliefs about healthiness of food and weight salience were observed, a main 21

effect of restraint status was noted on total energy intake, F(1,91) = 5.96; p<0.05. 22

12

Restrained eaters consumed about 16% (271) fewer kcal during the day of the experiment 1

than did unrestrained eaters. 2

Opinions about the new snack food 3

Participant’s ratings of the snack food’s “healthiness,” “capacity to affect weight” 4

and “appropriateness in a healthy menu” were all separately analyzed with 3-way 5

ANOVAs. For the “healthy” rating, there was a main effect of beliefs about healthiness of 6

food, F(1,91) = 32.08; p<0.0001. The snack food tested was perceived as healthier in the 7

healthy condition (M=5.7, SD=1.3) than in the unhealthy condition (M=4.2, SD=1.4), 8

which confirmed the effectiveness of the “healthy” manipulation. Furthermore, a 9

significant interaction between restraint status and weight salience was observed, F(1,91) = 10

4.02; p<0.05. Making weight salient in restrained eaters decrease their “healthy” rating of 11

the snack when compared to unrestrained eaters, whereas no differences between the two 12

groups were noted when weight was not made salient (see Figure 1). 13

Similar differences were also observed for the “weight gain” rating of the snack 14

food tested. There was a main effect of beliefs about healthiness of food (F(1,91) = 8.19; 15

p<0.01), with the snack food perceived as having a lower capacity for weight gain in the 16

healthy condition (M=5.2, SD=1.2) than in the unhealthy condition (M=5.8, SD=1.0). A 17

significant interaction between restraint status and weight salience was also observed 18

(F(1,91) = 4.96; p<0.05), indicating that the capacity of the snack to induce weight gain 19

was perceived as higher by restrained eaters than by unrestrained eaters when their weight 20

was made salient (M=5.9, SD= 1.1 vs. M=5.0, SD=1.0, respectively) but not when weight 21

was not made salient (M=5.5, SD= 1.3 vs. M=5.5, SD=1.1, respectively). 22

13

Regarding the appropriateness of the snack food in a healthy menu, similar findings 1

were again noted. A main effect of beliefs about healthiness of food was noted (F(1,90) = 2

18.34; p<0.0001), with the snack being perceived as more appropriate in a healthy menu in 3

the healthy condition (M=5.3, SD=1.2) than in the unhealthy condition (M=4.2, SD=1.4). 4

In addition, a significant interaction between restraint status and weight salience was 5

observed, F(1,90) =11.27; p<0.001. When weight was made salient, restrained eaters rated 6

the “appropriateness” of the snack lower than did unrestrained eaters (M=4.2, SD= 1.4 vs. 7

M=5.0, SD=1.0, respectively), whereas when weight was not made salient, restrained eaters 8

rated the snack higher on “appropriateness” than did unrestrained eaters (M=5.4, SD= 1.6 9

vs. M=4.6, SD=1.3, respectively). 10

A 3-way ANOVA showed no main effect of beliefs about healthiness of food, 11

restraint status or weight salience on the perceived caloric content of a serving of six 12

cookies (healthy condition: M=201.8, SD=234.9 kcal vs. unhealthy condition: M=245.4, 13

SD=423.5 kcal). Pearson correlation analyses showed that a higher “healthy” rating of the 14

snack was associated with (a) a lower “weight-gain” rating (r=-0.32; p<0.002), (b) a higher 15

“appropriateness” rating (r=0.63; p<0.0002) and (c) a higher price for a bag of the snack 16

food tested (r=0.24; p<0.05). 17

18

Discussion 19

The main aim of the present study was to investigate the effects of the perceived 20

healthiness of foods, restrained eating and weight salience on food intake during an ad 21

libitum snack. Although restrained eating and weight salience did not influence snack 22

14

intake, participants ate about 35% more in the “healthy” condition than in the “unhealthy” 1

condition. The “healthy/unhealthy” manipulation also affected participants’ ratings of the 2

snack food’s “healthiness,” “capacity to affect weight” and “appropriateness in a healthy 3

menu.” Furthermore, the “weight salience” manipulation appears to influence perceptions 4

of food differently in restrained and unrestrained eaters. Restrained eaters had a more 5

negative evaluation of the snack food in general when their weight was made salient before 6

eating, whereas unrestrained eaters had more positive attitudes when their weight was made 7

salient. 8

As we hypothesized, beliefs about the healthiness of foods significantly affected 9

eating: perceiving a food as healthy increased intake of that food. This finding is in 10

accordance with previous literature; Chandon & Wansink (2007) reported that caloric 11

contents of familiar main dishes from restaurants claiming to offer healthy food choices 12

were estimated by consumers as up to 35% lower in calories than when the dish was from a 13

restaurant not making such health claims. Categorization of foods as healthy, then, may 14

mean that a particular food will be eaten in greater amounts because it is assumed to 15

conduce to health (Ross & Murphy, 1999). Although the difference in perceived caloric 16

content for the healthy versus the unhealthy food in the present study was not significant 17

(due to the extreme variability of the caloric estimates), the healthy food was seen as being 18

not only healthier but more appropriate to eat and less likely to lead to weight gain. 19

The current study also clearly accords with previous research showing that norms 20

can influence food intake (Herman & Polivy, 2005; Herman & Polivy, 2008). More 21

specifically, beliefs about the healthiness of foods could be described as normative, because 22

such beliefs can serve as an indicator of appropriate intake. Accordingly, our finding that 23

15

there was no under-reporting for the amount of snack food eaten (in any condition) suggests 1

that participants generally regarded it as normal to have a higher intake of healthy than of 2

unhealthy foods. Furthermore, believing that cookies were healthy significantly increased 3

food intake among all participants; there was no differential responsiveness between 4

restrained and unrestrained eaters. This finding is in line with Herman and Polivy’s (2008) 5

sensory-normative distinction theory, according to which normative cues affect everyone 6

whereas sensory cues have a more powerful effect in obese and/or restrained individuals. 7

Even if our participants thought they were successfully restricting their food intake, 8

normative beliefs about the healthiness of a given food could lead to overeating (Herman & 9

Polivy, 2007). 10

While weight salience and restraint status did not influence eating behavior, making 11

weight salient affected restrained and unrestrained eaters’ perceptions of foods differently. 12

Restrained eaters had a more negative evaluation of both healthy and unhealthy snack foods 13

when they received weight feedback before eating, whereas unrestrained eaters gave more 14

positive food evaluations in the feedback-before-eating condition. The strong relation 15

between body dissatisfaction and dietary restraint (van Strien et al., 2007), and the fact that 16

restrained eaters had a higher BMI than did unrestrained eaters (p<0.01), might explain 17

why making weight salient adversely influences food perceptions only in restrained eaters. 18

Because of their greater weight dissatisfaction and greater perceived need for weight 19

control, restrained eaters may feel more negatively about foods in general when their 20

weight is made salient to them. Food-related goals are likely to influence perceptions of 21

food (Carels et al., 2007). Our results also showed that restrained eaters reported 22

consuming about 16% fewer kcal than did unrestrained eaters in the 24-hour recall, 23

16

suggesting that they were actually trying to control their weight. Nevertheless, despite the 1

fact that dieters should be better able to estimate the caloric content of foods (Carels et al., 2

2007) and they were less convinced that the snack food was good for them (at least when 3

weight was made salient), restrained eaters nevertheless ended up eating the same amount 4

of food in the laboratory as did unrestrained eaters in both the healthy and unhealthy 5

conditions. Their restrictive attitudes and behaviors clearly did not successfully prevent 6

them from eating more of the “healthy” snack. 7

8

Conclusion 9

Although this study has some limitations (e.g., female undergraduate students are not 10

necessarily representative of the general population), these findings contribute to a better 11

understanding of how perceptions of food may influence food intake. Future studies should 12

address this issue in males and older adults of both sexes, as well as in overweight/obese 13

individuals; do beliefs that a given food is healthy make everyone eat more? Beliefs about 14

the healthiness of foods need to be understood in the context of making healthy eating 15

recommendations and claiming health benefits for food products in a society facing an 16

increased prevalence of overeating and obesity. 17

Acknowledgment 18

The first author is recipient of fellowships from the Canadian Institutes of Health Research 19

and Canadian Diabetes Association. The authors would like to express their gratitude to 20

Mary Grace Lao, Maria Krisselle Galvez and Mickey Kaur for their help in the data 21

collection. 22

17

References 1

2

Allison, D. B., Kalinsky, L. B., & Gorman, B. S. (1992). A comparison of the 3

psychometric properties of three measures of dietary restraint. Psychological Assessment, 4, 4

391-398. 5

6

Carels, R. A., Harper, J., & Konrad, K. (2006). Qualitative perceptions and caloric 7

estimations of healthy and unhealthy foods by behavioral weight loss participants. Appetite, 8

46, 199-206. 9

Carels, R. A., Konrad, K., & Harper, J. (2007). Individual differences in food 10

perceptions and calorie estimation: An examination of dieting status, weight, and gender. 11

Appetite, 49, 450-458. 12

Chandon, P. & Wansink, B. (2007). The biasing health halos of fast-food restaurant 13

health claims: lower calorie estimates and higher side-dish consumption intentions. Journal 14

of Consumer Research, 34, 301-314. 15

Cohen, J. (1992). A Power Primer. Psychological Bulletin, 112, 155-159. 16

Doucet, E., Imbeault, P., St Pierre, S., Almeras, N., Mauriege, P., Richard, D. et al. 17

(2000). Appetite after weight loss by energy restriction and a low-fat diet-exercise follow-18

up. International Journal of Obesity and Related Metabolic Disorders, 24, 906-914. 19

18

Elfhag, K. & Rossner, S. (2005). Who succeeds in maintaining weight loss? A 1

conceptual review of factors associated with weight loss maintenance and weight regain. 2

Obesity Review, 6, 67-85. 3

Goldman, S.J., Herman, C.P., & Polivy, J. (1991). Is the effect of a social model 4

attenuated by hunger? Appetite, 17, 129-140. 5

Hart, K. E. & Chiovari, P. (1998). Inhibition of eating behavior: negative cognitive 6

effects of dieting. Journal of Clinical Psychology, 54, 427-430. 7

Herman, C. P. & Polivy, J. (1980). Restrained eating. In A.Stunkard (Ed.), Obesity 8

(pp. 208-225). Philadelphia: W.B. Saunders. 9

Herman, C. P. & Polivy, J. (2005). Normative influences on food intake. Physiology 10

& Behavior, 86, 762-772. 11

Herman, C. P. & Polivy, J. (2007). Norm-violation, norm-adherence, and 12

overeating. Collegium antropologicum, 31, 55-62. 13

Herman, C. P. & Polivy, J. (2008). External cues in the control of food intake in 14

humans: The sensory-normative distinction. Physiology & Behavior, 94, 722-728. 15

Laessle, R. G., Tuschl, R. J., Kotthaus, B. C., & Pirke, K. M. (1989). A comparison 16

of the validity of three scales for the assessment of dietary restraint. Journal of Abnormal 17

Psychology, 98, 504-507. 18

19

McFarlane, T., Polivy, J., & Herman, C. P. (1998). Effects of false weight feedback 1

on mood, self-evaluation, and food intake in restrained and unrestrained eaters. Journal of 2

Abnormal Psychology, 107, 312-318. 3

Miller, W. C. (1999). How effective are traditional dietary and exercise 4

interventions for weight loss? Medicine and Science in Sports and Exercise, 31, 1129-1134. 5

Oakes, M. E. (2006). Filling yet fattening: stereotypical beliefs about the weight 6

gain potential and satiation of foods. Appetite, 46, 224-233. 7

Oakes, M. E. & Slotterback, C. S. (2001). Judgements of food healthfulness: food 8

name stereotypes in adults over age 25. Appetite, 37, 1-8. 9

Paquette, M. C. (2005). Perceptions of healthy eating: state of knowledge and 10

research gaps. Canadian Journal of Public Health, 96 Suppl 3, S15-S21. 11

Polivy, J. (1996). Psychological consequences of food restriction. Journal of the 12

American Dietetic Association, 96, 589-592. 13

Polivy, J., Coleman, J., & Herman, C. P. (2005). The effect of deprivation on food 14

cravings and eating behavior in restrained and unrestrained eaters. International Journal of 15

Eating Disorders, 38, 301-309. 16

Polivy, J. & Herman, C. P. (2002). If at first you don't succeed. False hopes of self-17

change. American Psychologist, 57, 677-689. 18

20

Polivy, J., Herman, C. P., & Howard, K. I. (1988). The Restraint Scale: Assessment 1

of dieting. In M.Hersen & A. S. Bellack (Eds.), Dictionary of behavioral assessment 2

techniques. (pp. 377-380). New York: Pergamon Press. 3

Polivy, J. & Herman, P. (2007). Is the body the self? Women and body image. 4

Collegium antropologicum, 31, 63-67. 5

Ross, B. H. & Murphy, G. L. (1999). Food for thought: cross-classification and 6

category organization in a complex real-world domain. Cognitive Psychology, 38, 495-553. 7

van Strien, T., Herman, C. P., Engels, R. C., Larsen, J. K., & van Leeuwe, J. F. 8

(2007). Construct validation of the Restraint Scale in normal-weight and overweight 9

females. Appetite, 49, 109-121. 10

Wansink, B., & Chandon, P. (2006) Can “low-fat” nutrition labels lead to obesity?. 11

Journal of Marketing Research, 43, 607-617. 12

13

14

15

16

17

21

Table 1: Descriptive characteristics of the sample in each experimental condition (Mean ± SD) 1

Healthy snack description Unhealthy snack description

Restrained eaters Unrestrained eaters Restrained eaters Unrestrained eaters

Weight

salient

(N=12)

Weight not

salient

(N=12)

Weight

salient

(N=12)

Weight not

salient

(N=15)

Weight

salient

(N=11)

Weight not

salient

(N=12)

Weight

salient

(N=13)

Weight not

salient

(N=12)

Age (yrs) 19.1±1.6 19.3±1.2 19.3±1.2 19.1±1.3 21.4±7.2 18.9±1.2 18.8±1.6 19.3±2.3

BMI (kg/m2) 23.5±3.6 25.3±6.7 22.8±3.7 21.7±4.6 24.4±3.1 24.1±3.9 22.2±4.3 21.8±1.6

Restraint score 17.1±1.7 21.0±3.0 8.6±3.1 9.2±3.4 18.8±3.5 18.3±3.1 7.2±3.7 9.2±3.4

2

3

4

22

Title for figure 1

2

Figure 1 3

Mean “healthy” rating of the snack food tested (± SD) (unitless scores) for unrestrained and 4

restrained eaters, exposed to the salient weight condition or no salient weight condition. 5

1

2

3

4

5

6

7

8

9

10

11

12