Do Friends Share Similar Body Image and Eating Problems The

ARTICLE IN PRESS
Behaviour Research and Therapy 45 (2007) 15571577 www.elsevier.com/locate/brat
Do friends share similar body image and eating problems? The role of social networks and peer inuences in early adolescence
Delyse M. Hutchinson, Ronald M. Rapee
Department of Psychology, Macquarie University, Sydney, NSW 2109, Australia Received 21 October 2005; received in revised form 15 November 2006; accepted 21 November 2006
Abstract This study examined the role of friendship networks and peer inuences in body image concern, dietary restraint, extreme weight loss behaviours (EWLBs) and binge eating in a large community sample of young adolescent females. Based on girls self-reported friendship groups, social network analysis was used to identify 173 friendship cliques. Results indicated that clique members shared similar scores on measures of dieting, EWLB and binge eating, but not body image concern. Average clique scores for dieting, EWLB and binge eating, were also correlated signicantly with clique averages on measures of perceived peer inuence, body mass index and psychological variables. Multiple regression analyses indicated that perceived peer inuences in weight-related attitudes and behaviours were predictive of individual girls level of body image concern, dieting, EWLB use and binge eating. Notably, an individual girls dieting and EWLB use could be predicted from her friends respective dieting and EWLB scores. Findings highlight the signicance of the peer environment in body image and eating problems during early adolescence. r 2006 Published by Elsevier Ltd.
Keywords: Body image; Eating disturbances; Peer inuence; Social networks; Early adolescence
Introduction A concerning number of people continue to suffer from severe body image and eating disturbances. In the western world, as many as 23% of adolescent and adult females have a clinical eating disorder (Favaro, Ferrara, & Santonastaso, 2003; Kinzl, Traweger, Trefalt, Mangweth, & Biebl, 1999), and sub-clinical eating disorders have been estimated to be at least twice as common as full syndrome disorders (Hoek & Van Hoeken, 2003). High rates of eating disorders are concerning for health professionals as these disorders often require long-term, multidisciplinary interventions that are costly to the community (Agras, 2001). This is especially the case for anorexia nervosa which has one of the highest mortality rates of any mental illness (Keel et al., 2003).
Corresponding author. National Drug and Alcohol Research Centre, University of New South Wales, Sydney, NSW 2052, Australia. Tel.: +61 2 9385 0148; fax: +61 2 9385 0222. E-mail address: [email protected] (D.M. Hutchinson).
0005-7967/$ - see front matter r 2006 Published by Elsevier Ltd. doi:10.1016/j.brat.2006.11.007
ARTICLE IN PRESS
1558 D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577
In light of concerning prevalence rates and the signicant resources involved in treating eating disorders, growing emphasis has been placed on the potential value of prevention and early intervention (Pearson, Goldklang, & Striegel-Moore, 2002). Researchers have argued that to prevent or intervene in the development of eating problems it is necessary to accurately identify and target early risk factors for eating disorders (Jacobi, Hayward, De Zwaan, Kraemer, & Agras, 2004). Targeting early risk factors is also indicated by longitudinal studies that have identied a developmental trajectory from minor body image distortions and dieting behaviours, typically starting during early to mid-adolescence, to more extreme weight loss behaviours (EWLBs) and clinical disorders, developing during mid- to late adolescence and early adulthood (Shisslak, Crago, & Estes, 1995). To identify early risk factors for eating disorders researchers have examined a range of individual attributes and sociocultural factors associated with body image and eating problems. Individual risk factors consistently associated with body image concerns, dieting and binge eating have included body mass index or BMI (Cattarin & Thompson, 1994; Picot & Lilenfeld, 2003), low self-esteem (Button, Sonuga-Barke, Davies, & Thompson, 1996; Ghaderi & Scott, 2001), negative affect (Leon, Fulkerson, Perry, Keel, & Klump, 1999; Stice, Killen, Hayward, & Taylor, 1998) and depression (Graber, Brooks-Gunn, Paikoff, & Warren, 1994; Stice, Burton, & Shaw, 2004). These risk factors have been included in well-known theoretical models of the development of disordered body image and eating behaviours (e.g. Bardone, Vohs, Abramson, Heatherton, & Joiner, 2000; Heatherton & Polivy, 1992; McCarthy, 1990; Stice & Agras, 1998). Sociocultural risk factors have also been implicated both theoretically and empirically in the development of body image and eating disturbances (Levine & Smolak, 1992; Striegel-Moore, Silberstein, & Rodin, 1986; Thompson, Heinberg, Altabe, & Tantleff-Dunn, 1999). Theoretical models have included the family, media and peers as important sociocultural sources of inuence in eating problems (Stice & Agras, 1998; Thompson et al., 1999). While signicant research has focused on family and the media, peer inuence has only received moderate attention in the literature. This is somewhat surprising as it has been established by both developmental theorists and clinical researchers that during adolescence young people attribute considerable importance to the attitudes, beliefs and behaviours of their peers (Coleman, 1980). Theorists have suggested that peers are inuential in the development of individual personality traits, physical characteristics and behavioural tendencies, especially in adolescence (Coleman, 1980; Epstein, 1989). Consistent with this suggestion, empirical research has demonstrated that young people tend to resemble their friends in appearance and social attributes (Berscheid, Dion, Walster, & Walster, 1971), as well as interests, attitudes and behaviours (Kandel, 1978b; Tolson & Urberg, 1993). Peers have been found to share similarities in risk-taking behaviours including smoking (e.g. Tolson & Urberg, 1993), drug use (e.g. Aseltine, 1995; Kandel, 1978b), alcohol consumption (e.g. Curren, Stice, & Chassin, 1997; Jessor, 1987) and delinquent behaviour (e.g. Matsueda & Anderson, 1998). Similarities in peoples attitudes and behaviours are known to be important in determining interpersonal attraction (Kandel, 1978a). Such attraction tends to be greater when individuals share mutually valued attitudes and behaviours (Kandel, 1978b; Tolson & Urberg, 1993). Given that body image concerns and eating disturbances typically develop during adolescence, it has been suggested that adolescent friends might also share similarities in mutually valued body concerns and eating behaviours (Crandall, 1988; Paxton, Schutz, Wertheim, & Muir, 1999; Pike, 1995). To date, research examining the role of peers in body image and eating problems has primarily focused on individual perceptions of peer inuence. Findings suggest that perceived peer inuence is important in adolescent body attitudes and eating behaviours (Lieberman, Gauvin, Bukowski, & White, 2001; Wertheim, Paxton, Schutz, & Muir, 1997). In several studies, perceived peer pressure to be thin was more strongly associated with disordered eating attitudes and behaviours in adolescent girls than perceived pressure from alternate sources such as family, male friends or the media (Matsumoto, Kumano, & Sakano, 1999; Stice, 1998). Moreover, both perceived peer talk about weight loss and perceived peer dieting have been associated with individual dieting attempts by adolescent girls (Levine & Smolak, 1992; Levine, Smolak, & Hayden, 1994; Wertheim et al., 1997). These ndings provide indirect support for the notion that adolescent friends share similarities in their body attitudes and weight loss behaviours. Theorists have suggested that friendship similarities derive from both selection and socialisation processes (Ennett & Bauman, 1994; Epstein, 1989). According to selection, girls chose friends with whom they share similar beliefs and behaviours about the importance of body weight and shape. Via socialisation, peer
ARTICLE IN PRESS
D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 1559
discussion about thinness and dieting then increase preexisting similarities and promote group members to engage in dieting as a means of achieving thinness. An alternate explanation is that girls only perceive their friends to be dieting and that similarities in weight- and eating-related attitudes and behaviours do not exist (Paxton et al., 1999). Only a small number of studies have assessed whether female friends share actual similarities in body image and eating behaviours. These studies used social network methodology (Scott, 2001; Wasserman & Faust, 1994), which involves examining the overlap in individuals self-reported friendship links to identify cohesive peer groups, often referred to as cliques. Based on clique groupings, researchers then examine whether similarities exist between friends in their body attitudes and eating behaviours. Crandall (1988) prospectively examined friendship cliques among young adult females in two college sororities over the course of an academic year. By the end of the year individual binge eating scores could be predicted from clique binge eating levels. Pike (1995) used a cross-sectional design to examine binge eating among adolescent females in grades 9 through 12. Results indicated that grade 9 and 10 students shared similar levels of selfreported binge eating to their friends. However, friendship similarities were not observed in grade 11 and 12 students. Pike suggested that there might be specic time points when peer group similarities are more prominent. Paxton et al. (1999) used social network analysis to examine body image and eating similarities among grade 10 females. Clique members had similar scores on measures of body image concern, dietary restraint and EWLBs. In contrast with previous ndings, however, within-group similarities were not identied for binge eating. More recently, Meyer and Waller (2001) examined change in body image attitudes, dietary restraint and binge eating in a group of women commencing prearranged share living at university. After 14 weeks the study found signicant convergence in self-reported body image attitudes and dietary restraint, which were socially valued by group members, but divergence in attitudes and behaviours relating to bulimia, which were not mutually valued by group members. Although, Meyer and Waller (2001) only examined a small sample N 41, their ndings indicate that socialisation contributed signicantly to changes in body concerns and eating behaviours in young adult women. Taken together, the preceding studies identied similarities in body image attitudes and eating behaviours among female social groups. Results were mixed with regard to binge eating, with only two of the four studies identifying clique similarity. Although, these ndings are consistent with sociocultural theories that emphasise the importance of peer inuence in weight concerns and disordered eating, further research is needed to clarify conicting results for binge eating. Moreover, no known studies have assessed whether similarities are identiable earlier in adolescence. This is surprising given that: (a) early adolescence is a high-risk period for the development of body image and eating disturbances, which are known risk factors for the later development of clinical eating disorders; and (b), around this time girls start to place increasing importance on the attitudes, values and behaviours of their friends. Examining females during late adolescence and adulthood, when more severe problems have already developed, provides somewhat limited information for the identication of early risk factors and in turn, the development of prevention and early intervention programs. Thus, the rst aim of this study was to examine the extent of peer similarity in body image and eating problems in early adolescence, using a large community sample. Specically, it was hypothesised that there would be signicantly greater between than within clique variability on measures of body image concern, dietary restraint, EWLBs and binge eating, after controlling for relevant physical and psychological variables. If similarities in body image or eating behaviours were identied, the second aim of this study was to examine whether cliques classied as high and low on these measures could also be characterised by their average scores on physical, psychological, behavioural and peer inuence variables. Measures of perceived peer inuence included friends concern with thinness and dieting, friends as a source of inuence, peer pressure to lose weight and be thin and weight-related teasing (Paxton et al., 1999; Thompson & Heinberg, 1993). Average clique alcohol intake, cigarette use and impulsivity were also examined because: (a) substance use and risk taking have been associated with body image and eating disturbances (e.g. French et al., 1997; Garry, Morrissey, & Whetstone, 2003); and (b), within-group similarities in substance use have been identied in adolescent samples (e.g. Ennett & Bauman, 1994; Jessor, 1987). Signicant correlations were expected between average clique body image and eating behaviours, and average scores for cliques on physical, psychological, behavioural and peer variables.
ARTICLE IN PRESS
The third aim of this study was to examine the characteristics of girls who were not classied part of a friendship clique. Research in the substance use area found that adolescents who are not part of a friendship network are more likely to smoke (Ennett & Bauman, 1994). Based on this nding, it might also be the case that girls who are not members of a school-based clique have different body attitudes or eating behaviours to clique members. To explore whether this is the case, clique members and non-members were compared on body image, eating, physical and psychological measures. In addition, the two groups were compared on measures of perceived peer inuence, number of close friends outside school, impulsivity and substance use. The nal aim of this study was to examine whether girls perceptions of their friends weightand shape-related attitudes and behaviours contributed to the prediction of individual body concerns and eating behaviours. Specically, it was hypothesised that elevated scores on measures of perceived peer inuence would contribute signicantly to the prediction of individual body image concern and eating behaviours, after accounting for the contribution of relevant physical, psychological and behavioural variables. Method Participants Participants were grade 7 female students from 10 girls schools in New South Wales, Australia. Schools were selected from a range of socioeconomic and geographic areas and included ve private schools, three state schools and two Catholic schools. In total, 1094 students participated in the study with a mean age of 12.3 years (SD .52), ranging from 10 to 14 years. This represented an overall response rate of 71.4%. The majority of participants were born in Australia (84.1%), followed by Asia (10.1%), the Middle East (1.2%), New Zealand (.9%) and the United Kingdom (.8%; Oceania, America, Canada, Papua New Guinea, Africa and Russia represented less than .7% each). There was greater variability in parents country of birth: 59.2% of mothers were born in Australia, followed by Asia (20.3%), the United Kingdom (5.7%), the Middle East (4.8%) and New Zealand (2.0%). Likewise, 60.5% of fathers were born in Australia, followed by Asia (17.2%), the United Kingdom (5.5%), the Middle East (4.8%) and Oceania (1.8%). Despite this diversity, most students reported English as the main language spoken at home (83%). The cultural and lingual background of the sample was reective of the population of New South Wales, Australia (Australian Bureau of Statistics (ABS), 1996, 2001). The majority of participants lived with both biological parents (82.0%) and had one (39.7%) or two (32.4%) siblings. Parent occupational status indicated that the sample was predominantly of middle to upper-middle class socioeconomic background. Procedure All grade 7 students at each school were invited to participate in the research and active written consent was obtained from parents before the study commenced. To encourage students to return their consent forms, a small gift was given to students who returned a signed consent form by the due date set at each school. Students received a gift irrespective of their decision to participate. Questionnaires were administered during class and took approximately 90 min to complete. A follow-up day was scheduled for students who were absent on earlier testing days or who required additional reading assistance. Questionnaires were administered under the supervision of the primary researcher. At some schools teaching staff assisted with the supervision of large groups of students. All teachers who assisted were given written information with specic instructions for teachers and students, as well as answers to common student questions. Participants were also given a word glossary to improve consistency in the interpretation of questionnaire items. Words included in the glossary were based on feedback from grade 7 students in a pilot test of the questionnaire. Participants rst completed demographic and clique identication questions, followed by measures of body image, dieting, EWLB use, binge eating, perceived peer inuences, weight-related teasing, self-esteem, depression, negative affect, impulsivity and substance use. Height and weight measurements were assessed privately by a female researcher.
ARTICLE IN PRESS
Measures Demographics Participants were asked to indicate their age, date of birth, country of birth, language spoken at home, relationship to adults living at home, number of siblings, and their mothers and fathers country of birth and occupation. Body image concern and eating behaviours The Body Disparagement, Feeling Fat, Lower Body Fatness and Salience of Weight and Shape subscales of the Body Attitudes Questionnaire (BAQ, Ben-Tovim & Walker, 1991) were summed to an overall measure of body image concern. The possible range of scores was 33165. This scale has been used with Australian adolescents (Paxton et al., 1999) and has demonstrated construct validity and internal reliability (Ben-Tovim & Walker, 1991; Ben-Tovim & Walker, 1994). Cronbachs a for the combined scales in this sample was .95 and the four subscales were also signicantly correlated with each other (from .59 to .76). Dietary restraint was assessed using the Restraint subscale of The Dutch Eating Behaviour Questionnaire (DEBQ-R, van Strien, Frijters, Bergers, & Defares, 1986). This scale has demonstrated construct validity and internal reliability in adolescents (Banasiak, Wertheim, Koerner, & Voudouris, 2001). The possible range of scores was 1050. Cronbachs a for the scale in this sample was .92. The use of more extreme weight loss behaviours was assessed using The EWLB Checklist (Paxton et al., 1991). The frequency of use of six extreme weight loss strategies (appetite suppressants, crash dieting, diuretics, fasting, laxatives and vomiting) was rated on a 4-item scale from 0 (never) to 3 (daily). Ratings on each item were summed to a single score to provide an overall indication of how frequently participants used extreme weight loss strategies. The possible range of scores was 018. This scale has demonstrated reliability in adolescent samples (Banasiak et al., 2001). Cronbachs a for the scale in this sample was .83. The Bulimia subscale of the Eating Disorder Inventory (EDI-B, Garner, Olmsted, & Polivy, 1983) was used to assess binge eating. The 7-item Bulimia subscale has demonstrated validity and reliability in non-clinical samples when scores are based on untransformed responses (Shoemaker, van Strien, & van Der Staak, 1994). The most asymptomatic response received a score of 1 and the most symptomatic response received a score of 6. Untransformed responses were considered appropriate given the non-clinical sample. The possible range of scores was 742, and Cronbachs a for the scale in this sample was .72. Friendship clique assessment measure To identify friendship cliques participants were given a list of the names of all grade 7 students at their school, each with a corresponding number. Students were then asked to identify their friends by writing down the number corresponding to each friends name. Group membership was determined from responses to the following items adapted from Paxton et al. (1999): 1. Write down the group of girls from grade 7 that you usually spend time with at school. If you have more than one group of friends, write down the group that you feel most part of. 2. Write down your friends from grade 7 at your school. That is, the girls you spend time with and enjoy being with. You can include friends that are not in your main group. 3. Write down the numbers of your best friends. That is, the girls you hang around with the most and are closest to. Friendship clique identication Friendship groups generally consist of a core group of tight-knit friends, surrounded by connected yet successively looser links of friends. Individuals can vary in their opinions about: (a) who are part of a core group; and (b), the perceived strength of links between group members (Scott, 2001; Wasserman & Faust, 1994). As this study aimed to examine mechanisms of peer inuence it was considered important to identify cliques in which group processes were likely to be operating. Based on Paxton et al. (1999), the term friendship clique was operationalised in a way that would assist identication of cohesive groups of friends with a moderate to high degree of: (a) membership agreement; and (b), closeness between group members.
ARTICLE IN PRESS
The UCINET VI statistical program (Borgatti, Everett, & Freeman, 2002) was used to analyse social network data for each school. Participant responses to item 3 (best friends) were rst examined to identify the strongest friendship ties. Only reciprocated friendship ties of two or more participants were included. In the next stage, the sociometric data from items 1 (main group) and 2 (friends outside main group) were included in the analyses. Hierarchical clustering based on the clique overlap matrix identied broader social networks that constellated around central friendship ties. The validity of the groupings identied by UCINET VI was then examined by comparing identied cliques to social groupings in the raw data. Discrepancies between the groupings were modied to accurately reect friendship groupings in the raw data. Criteria for excluding data from the clique analyses were based on Paxton et al. (1999). First, dyads N 118 were removed from the sample for the main analyses. This was based on evidence that dyads differ from larger social networks in their intimacy, intensity, degree of conformity and interpersonal style (Ennett & Bauman, 1994; Taylor, de Soto, & Lieb, 1979). Data collected from only two members of a group of three or more girls were also removed from the main analyses. Despite removing dyads from the main data set, we did include the dyadic data in a second set of analyses to examine whether the removal of these more intimate groupings inuenced the overall results. These analyses indicated a very similar pattern of results to the nondyadic data set. Participants were also omitted from the data set if they were not part of a friendship group, or were not closely linked to a particular group (i.e. not identied as a friend by other group members). Likewise, individual participants were not permitted to be a member of more than one group. This was considered important to meet both the statistical assumption of independence of groups, and because the study aimed to examine cohesive, rather than broader friendship groupings. In cases where girls were linked to two or more groups, participants were allocated to the group with which they had the strongest links. Participants who were not clearly linked to one group more strongly than another were excluded from the data set. Finally, to ensure that group members shared moderate to strong links, girls were only classied part of a clique in the nal data set if they shared either level 3 (best friend) or 1 (main group) links with the majority of other group members. In total, 173 cliques were identied N 858. Table 1 summarises the distribution of friendship cliques relative to member size. The majority of groups were triads, followed by groups of four or ve members. Two hundred and thirty-six participants were excluded because they were not part of a friendship clique. An inspection of these participants indicated that half were members of dyadic friendship groups, and that a considerable proportion (87.3%) reported having close friends outside the school environment. As a measure of the attitudes and behaviours characteristic of each friendship clique, group means were calculated on dieting, EWLB and binge eating scores for each individual friendship group. Group means were not calculated for body image concern, as signicant within-group group similarities were not identied. Perceptions of peer attitudes and behaviours A number of measures were used to examine student perceptions of their friends weight-related attitudes and behaviours. Participants completed the 13-item The Friends Concern with Thinness and Dieting Scale
Table 1 Friendship clique distribution relative to member size Clique size 3 4 5 6 7 8 9 10 11 14 No. of participants 150 136 160 144 126 32 45 40 11 14 % of participants 17.5 15.9 18.6 16.8 14.7 3.7 5.2 4.7 1.3 1.6 No. of cliques 50 34 32 24 18 4 5 4 1 1 % of cliques 28.9 19.6 18.5 13.9 10.4 2.3 2.9 2.3 .6 .6
Note: Number of cliques 173.
ARTICLE IN PRESS
(Paxton et al., 1999). Cronbachs a for the scale in this sample was .71. To examine the degree to which friends were considered important in inuencing girls body attitudes and weight loss behaviours participants also completed the 5-item Friends as a Source of Inuence Scale (Paxton et al., 1999; modied from the Magazines as a Source of Inuence Scale, Levine et al., 1994). Cronbachs a for the scale in this sample was .85. A series of single-item scales was also included to examine more direct peer inuence to be thin and diet. Two items assessed the perceived frequency of peer discussion about weight loss and dieting and encouragement by friends to lose weight (see Paxton et al., 1999). Five other items assessed perceived pressure to be thin and diet: Do you think your friends would like you to be thinner than you are now? (responses ranged from never to very often), and perceived pressure to diet: Have dieting attempts by your close friends ever made you feel that you ought to be dieting or watching your weight more?; Have your friends ever made comments about you which have made you feel as if you ought to diet?; Have dieting attempts by your close friends ever made you feel so worried about your weight that you have actually gone on a diet?; Have you ever gone on a diet, not because you really thought you needed to lose weight, but because one or more of your friends was dieting? (responses ranged from denitely not to denitely yes; S. J. Paxton, personal communication, 2000). As these seven items all used conceptually similar 5-point Likert scales to measure peer pressure to be thin and diet, they were combined into a single multi-item measure called The Peer Inuence to be Thin and Diet Scale. Cronbachs a for the scale was .84, indicating that the scale had adequate reliability in this sample. Weight-related teasing Participants completed the 6-item frequency of weight-related teasing subscale from the Perception of Teasing Scale (POTS, Thompson, Cattarin, Fowler, & Fisher, 1995). This scale examines the frequency of weight-related teasing by peers using a 5-point Likert scale. The POTS has demonstrated reliability and validity (Thompson et al., 1995). Cronbachs a for the frequency of weight-related teasing subscale in this sample was .92. Physical measurement Height and weight were measured using a tape measure and electronic weight scales. Measurements were used to calculate body mass index (BMI weight [kg]/height2 [m]). Psychological measures To assess self-esteem participants completed the 10-item Rosenberg Self-Esteem Scale (RSE, Rosenberg, 1965). Cronbachs a for the scale in this sample was .84. Depression was assessed using 17-items from the Center for Epidemiological Studies Depression Scale for Children (CES-DC) which has been validated for use with adolescents (Fendrich, Weissman, & Warner, 1990). Cronbachs a for the scale in this sample was .86. In addition, students completed the 30-item Negative Emotionality (NE) and Constraint (CON) scales of the Multidimensional Personality Questionnaire (MPQ, Tellegen, 1982; Waller, 1990) to examine negative affect and constraint-impulsivity. Cronbachs a for the MPQ-NE and MPQ-CON scales in this sample were .88 and .61, respectively. Substance use Two items asked participants to indicate average weekly alcohol consumption and cigarette intake. Results Data transformations To normalise skewed distributions square root transformations were applied to the following variables: DEBQ-R, EDI-B, EWLB, weekly alcohol and cigarette intake, and weight-related peer teasing. Analyses were conducted using both transformed and untransformed scores. As both methods yielded similar results, the untransformed results have been reported for ease of understanding. Where minor differences were identied in the pattern of results the transformed data have been reported and this has been noted.
ARTICLE IN PRESS
1564 D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 Table 2 Multivariate analysis of covariance (MANCOVA) for friendship cliques on measures of body image concern, dieting, extreme weight loss behaviours and binge eating, after controlling for BMI, self-esteem and negative affect Dependent variable measures Body attitudes questionnaire (BAQ) Dutch eating behaviour questionnaire-Restraint (DEBQ-R) Extreme weight loss behaviors checklist (EWLB) Eating disorder inventory-Bulimia (EDI-B) Note: n 660. Range of clique Ms 41132 1134 04 733 F 1.14 1.45 1.29 1.27 p .145 .001 .020 .026 Effect size .16 .26 .27 .24
Friendship group similarities in body image and eating behaviours Prior to examining friendship similarities in body image and eating behaviours, an analysis of variance was used to ascertain whether clique members were similar on four variables known to be associated with body image and eating variables: BMI, self-esteem, negative affect and depression. These four dependent variables were entered together in MANOVA, as they are also known to be interrelated. The MANOVA was signicant, Wilkss F(4, 672) 1.36, p .000, Z2 .30,1 indicating greater between-group than within-group differences. Univariate F comparisons indicated that clique members were similar on BMI, F(168, 533) 1.69, p .000, Z2 .35; self-esteem, F(168, 533) 1.67, p .000, Z2 .35; and negative affect, F(168, 533) 1.26, p .028, Z2 .28. However, depression scores did not vary signicantly as a function of clique membership F(168, 533) 1.14, p .137. MANCOVA was then used to determine whether friendship clique members were similar on the four key dependent variables: body image concern, dieting, EWLBs and binge eating, after controlling for relevant physical and psychological variables. Participant scores on measures of BMI, self-esteem, and negative affect were previously found to vary as a function of clique membership, therefore these variables were entered together as covariates. The depression measure was not included as a covariate because between-group differences in depression scores were not identied in this sample. Following the covariates, the body image and eating measures were entered in combination as they have consistently been shown to be interrelated. The MANCOVA was signicant, Wilkss F(4, 652) 1.21, p .001, Z2 .26, indicating greater between- than within-group differences. Table 2 summarises the univariate F values for the four dependent variables. The univariate F results indicate signicant within-group similarity in dietary restraint, despite signicant covariation with BMI, F(1, 493) 56.20, p .000, Z2 .06; self-esteem, F(1, 493) 6.56, p .011, Z2 .01; and negative affect F(1, 493) 29.02, p .000, Z2 .03. Clique members also shared similarities in EWLB scores, despite signicant covariation with self-esteem, F(1, 493) 5.86, p .016, Z2 .01; and negative affect, F(1, 493) 9.44, p .002, Z2 .01. Likewise, clique members shared similarities in binge eating, despite signicant covariation with self-esteem, F(1, 493) 5.11, p .024, Z2 .01; and negative affect, F(1, 493) 47.26, p .000, Z2 .05. Levels of body image concern among adolescent females were not signicantly related to clique membership, but covaried signicantly with BMI, F(1, 493) 113.12, p .000, Z2 .10; self-esteem, F(1, 493) 57.06, p .000, Z2 .05; and negative affect, F(1, 493) 60.40, p .000, Z2 .05. It appears that variations in body image concern are better accounted for by physical and psychological factors than peer group membership2.
1 Based on Levine and Hulletts (2002) recommendations, Z2, rather than partial Z2, was used to interpret effect size (strength of association) in the current sample. Z2 is a more conservative measure of effect size. 2 As a considerable number of participants without BMI scores were excluded from the multivariate analyses, a second set of analyses were conducted to determine whether the exclusion of these participants biased the representativeness of the clique sample. MANOVA was used to compare girls with and without BMI measurements on each of the key variables. Results indicated that girls without BMI scores had signicantly higher BAQ, DEBQ-R and EWLB scores. We therefore decided to take a conservative approach and re-analyse the clique data without BMI as a covariate. The MANCOVA results were in line with the initial ndings, indicating signicant between-group variation in DEBQ-R, EWLB and EDI-B scores, but not BAQ scores. Similar patterns of co-variation with self-esteem and negative affect were also identied.
ARTICLE IN PRESS
D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 Table 3 Correlates of mean friendship group scores on dieting, extreme weight loss behaviours and binge eating Variable/measure Dietary restraint (DEBQ-R) Extreme weight loss behaviours (EWLB) Binge eating (EDI-B) Body image concern (BAQ) Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQ-NE) Depression (CES-DC) Friends concern with thinness and dieting Friends as a source of inuence Peer pressure to lose weight and be thin Peer teasing (POTS) Constraint (MPQ-CON) Alcohol intakeaverage weekly intake Cigarette intakeaverage weekly intake Note: Number of friendship groups 173. ***po.001. DEBQ-R .59*** .53*** .73*** .29*** .35*** .40*** .42*** .55*** .60*** .70*** .34*** .04 .12*** .18*** EWLB .59*** .48*** .57*** .29*** .38*** .33*** .42*** .47*** .45*** .52*** .41*** .01 .13*** .27*** EDI-B .53*** .48*** .66*** .25*** .45*** .52*** .51*** .40*** .49*** .57*** .52*** .00 .26*** .26*** 1565
The Z2 effect sizes shown in Table 2 indicate moderate to large effects on measures of dieting and EWLB use and more moderate effects for binge eating. Around 25% of the variance in dieting, EWLBs and binge eating in this sample can be explained by clique membership. Notably, the effect size for BAQ was moderate even though the univariate F test was not signicant. Intraclass correlations were also calculated to measure the degree of similarity among clique members on measures of dieting, EWLB use and binge eating. Signicant intraclass correlations were found for each of the three dependent variables: dieting, rho (173, 675) .135, p .000; EWLBs, rho (173, 683) .122, p .000; and binge eating, rho (173, 675) .093, p .000. Correlates of friendship group dieting, extreme weight loss behaviours and binge eating To indirectly examine possible mechanisms for friendship similarity we assessed whether other characteristics, including weight-related beliefs and perceived peer inuence, were associated with cliques that reported high levels of eating disturbance. Correlations were calculated between average group DEBQ-R, EWLB and EDI-B scores and group means on a range of physical, psychological and peer inuence variables. In addition, because risk-taking and substance-related impulsivity have been linked to unhealthy eating patterns, average clique constraint scores, and weekly alcohol and cigarette intake were included in the analyses. Body image concern was not included as a dependent variable because previous results indicated that friendship groups did not vary signicantly on this measure. Table 3 summarises the results of these analyses. Table 3 indicates signicant intercorrelations between average clique DEBQ-R, EWLB and EDI-B scores. Group means on the three dependent variables were also correlated signicantly with group BMI, self-esteem, negative emotionality and depression scores. Groups high and low on DEBQ-R, EWLB and EDI-B scores could be characterised by corresponding high and low scores on body image concern and the four measures of perceived peer inuence: friends concern with thinness and dieting, friends as a source of inuence, peer pressure to lose weight and be thin and peer teasing. Small yet signicant correlations were also found between average clique alcohol and cigarette intake and each of the dependent variables. However, constraint scores were not signicantly correlated with average clique dieting, EWLB use or binge eating. Characteristics of Clique Members and Non-Members To examine the characteristics of girls who were not classied part of a friendship clique, MANOVA was used to compare these participants with clique members on the four key dependent variables: body image
ARTICLE IN PRESS
1566 D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 Table 4 Multivariate analysis of variance (MANOVA) and descriptive statistics comparing clique members with non-clique members on body image, eating behaviours, physical and psychological measures Dependent variable measures BAQ DEBQ-R EWLB EDI-B BMI RSE MPQ-NE CES-DC F (1, 971) p Z2 Clique member M (n 760) 72.01 4.24 .59 3.53 20.13 30.18 42.46 13.74 Clique member SD (n 760) 23.89 .93 .71 .67 3.06 4.62 6.60 8.90 Non-member M (n 213) Non-member SD (n 213)
6.73 1.33 4.94 .45 13.32 7.09 .40 2.37
.010 .249 .026 .501 .000 .008 .527 .124
.01 .00 .01 .00 .01 .01 .00 .00
76.99 4.32 .72 3.56 21.05 29.21 42.78 14.82
27.75 1.02 .74 .71 3.84 5.10 6.84 9.65
Note: n 973. Results are based on square root transformations of DEBQ-R, EDI-B and EWLB scores.
concern, dieting, EWLBs and binge eating. The two groups were also compared on physical and psychological variables including BMI, self-esteem, negative emotionality and depression. The MANOVA was signicant, Wilkss F(8, 964) 2.85, p .004, Z2 .02, indicating differences between clique members and excluded participants. Table 4 shows the univariate F values, means and standard deviations for the dependent variables included in the MANOVA. These results indicate that dieting, binge eating, negative emotionality and depression scores did not differ signicantly as a function of clique membership. However, females who were not identied as part of a friendship group had signicantly lower self-esteem and higher BMI, body image concern and EWLB scores than clique members. Examination of effect sizes indicated that clique membership only accounted for 1% of the variance in each of these measures, which are not substantively meaningful effect sizes. These results indicate, however, that the clique analyses were somewhat biased toward individuals with higher self-esteem scores and lower BMI, body image concern, and EWLB scores. A second set of multivariate analyses of variance examined whether clique members and girls who were not classied part of a clique differed signicantly on the peer inuence variables. Measures included friends concern with thinness and dieting, friends as a source of inuence in weight and shape issues, peer pressure to be thin and diet, weight-related teasing and number of close friends outside school. The groups were also compared on measures of constraint-impulsivity and substance use. The MANOVA was not signicant, Wilkss F(7, 948) .91, p .500, indicating that both groups had similar scores on peer variables, constraintimpulsivity and substance use. Perceived peer inuence in the prediction of individual body image and eating behaviours A series of correlations were calculated to independently examine the relationships between individual body image and eating behaviours and scores on peer, physical, psychological and behavioural measures (see Table 5). Hierarchical multiple regression analyses were then conducted to examine whether perceived peer inuence contributed signicantly to the prediction of individual body image concern, dieting, EWLBs and binge eating, after accounting for the predictive contribution of previously identied physical, psychological and behavioural variables. Each regression consisted of four separate blocks of variables, with the independent variables in each block entered simultaneously. This approach reduces the effects of multicollinearity because it measures the degree to which each set of independent variables accounts for variance in the dependent variable, without the inuence of subsequent blocks of variables (Cohen & Cohen, 1983). In each regression model, BMI was entered in the rst block as this variable has been found to contribute to the prediction of body image concern and eating problems. Self-esteem, negative affect and depression were entered in block 2, because these variables have also been associated with body image and eating behaviours. Next, the predictive contribution of individual constraint scores, and average weekly alcohol and cigarette intake were entered as
ARTICLE IN PRESS
D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 1567 Table 5 Correlates of individual body image concern (BAQ), dietary restraint (DEBQ-R), extreme weight loss behaviours (EWLB) and binge eating (EDI-B) Variable/measure Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Constraint (MPQC) Alcohol intake Cigarette intake Adjusted clique mean Friend concern with thinness/dieting Friends as a source of inuence Peer pressure to lose weight/be thin Peer teasing (POTS) BAQ (n 841) .47*** .48*** .47*** .49*** .06 .12*** .08 .40*** .53*** .74*** .47*** DEBQ-R (n 772) .32*** .27*** .34*** .39*** .03 .11** .08 .23*** .44*** .51*** .64*** .32*** EWLB (n 771) .20*** .28*** .26*** .30*** .03 .13*** .08 .22*** .30*** .34*** .40*** .25*** EDI-B (n 770) .18*** .32*** .42*** .37*** .09 .16*** .11** .15*** .24*** .29*** .41*** .37***
Note: **po.01, ***po.001. Results are based on square root transformations of alcohol intake, cigarette intake and EWLB scores.
risk-taking and substance use have been associated with both unhealthy eating patterns and peer inuence. Measures of perceived peer inuence were entered in block 4 and included: friends concern with thinness and dieting, friends as a source of inuence, peer pressure to lose weight and be thin and peer teasing. Finally, as a measure of the self-reported attitudes and behaviours characteristic of each girls peer environment, clique averages on DEBQ-R, EWLB and EDI-B measures were recalculated for each girl after removing her own contribution to the group mean (Paxton et al., 1999). Adjusted clique means for each measure were then entered in block 4 of the corresponding regression model (i.e. adjusted clique DEBQ-R means were only entered in the regression equation predicting individual DEBQ-R scores), along with the other peer variables. Adjusted clique means were not calculated for body image concern in either the univariate or multivariate analyses as previous results indicated that body image concern was not a variable that characterised young adolescent friendship groups. To control for type I errors, alpha levels were set at .0125 (.05 divided by 4 blocks). Tables 69 summarise the results of the regression analyses. Correlates of individual body image concern, dieting, extreme weight loss behaviours and binge eating Signicant positive correlations were identied between individual BAQ, DEBQ-R, EWLB and EDI-B scores and measures of perceived peer inuence and relevant adjusted clique averages. BAQ, DEBQ-R, EWLB and EDI-B scores were also signicantly correlated with measures of BMI, self-esteem, negative emotionality, depression and alcohol intake, but not constraint. A small positive relationship was identied between individual EDI-B scores and tobacco intake. Multivariate prediction of individual body image concern BMI contributed signicantly to the prediction of individual body image concern, accounting for 20% of the variation in BAQ scores (see Table 6). The addition of self-esteem, depression and negative affect variables in block 2 contributed a signicant increment to the variation in individual body image concern scores, DR2 .28, po.001, however, the addition of constraint and substance use measures in block 3 did not improve prediction. Peer inuence variables were entered in block 4, accounting for a further 20% increase in the prediction of individual levels of body image concern, DR2 .20, po.001. Perceived peer pressure to be thin and diet contributed the largest amount of explained variance, followed by self-esteem and BMI. Friends as a source of inuence in weight and shape concerns and depression scores also contributed to the variation in individual girls body image concern. Taken together, these variables accounted for 68% of explained variation in body image concern scores.
ARTICLE IN PRESS
1568 D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 Table 6 Hierarchical multiple regression analyses for individual prediction of body image concern (BAQ) from physical, psychological, behavioural and peer variables Variables predicting individual body image concern (BAQ) Step 1 Body mass index (BMI) Step 2 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Step 3 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Constraint (MPQC) Alcohol intake Cigarette intake Step 4 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Constraint (MPQC) Alcohol intake Cigarette intake Friend concern with thinness/dieting Friends as a source of inuence Peer pressure to lose weight/be thin Peer teasing (POTS) Note: n 841. **po.01, ***po.001. B SE b T R R2 DR2
.45 3.29 2.71 1.37 .59 .59 2.71 1.38 .58 .59 .02 3.34 .68 1.46 1.13 .14 .29 .10 2.46 .59 .01 .43 2.31 .18 .23 .19 .16 .13 .10 .19 .16 .13 .10 .16 2.17 .70 .16 .12 .11 .08 .13 1.74 .55 .17 .13 .16 .17 .45 .37 .26 .16 .22 .37 .26 .15 .21 .00 .05 .03 .20 .22 .04 .11 .02 .04 .03 .00 .09 .48 .03 14.64*** .69 14.68*** 8.86*** 4.45*** 5.99*** .69 14.57*** 8.90*** 4.37*** 5.91*** .14 1.54 .97 .82 9.02*** 9.20*** 1.29 3.69*** .78 1.41 1.06 .07 3.28** 14.91*** 1.01
.20 .48
.20*** .28***
.48
.00
.68
.20***
Multivariate prediction of individual dietary restraint Table 7 summarises the results of regression analyses predicting individual DEBQ-R scores. In block 1, BMI accounted for 12% of explained variation in dieting scores. The addition of self-esteem, negative affect and depression in block 2 contributed a signicant increment to the variation in individual dieting scores, DR2 .15, po.001, although BMI and depression were the only statistically signicant predictors. In block 3, constraint and substance use variables did not improve prediction signicantly, however, the subsequent addition of peer inuence variables and adjusted clique mean DEBQ-R scores in block 4 resulted in a 21% increase in explained variance, DR2 .21, po.001. Notably, with the exception of weight-related teasing, the remaining four peer inuence variables were all signicant predictors. Perceived peer pressure to be thin and diet contributed the largest amount of explained variance in dieting scores, and to a lesser extent, so too did friends as a source of inuence, friends concern with thinness and dieting, and the adjusted clique average for dieting. In the nal model, BMI, depression and constraint also contributed to the variation in girls dieting scores. Taken together, these variables accounted for 48% of explained variance in dieting scores. Multivariate prediction of individual use of extreme weight loss behaviours Table 8 summarises the results of regression analyses predicting individual EWLB scores. In block 1, BMI was a signicant predictor of individual use of EWLBs, DR2 .05, po.001. The addition of self-esteem, depression and negative emotionality variables in block 2 contributed a signicant increment in the variation
ARTICLE IN PRESS
D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 1569 Table 7 Hierarchical multiple regression analyses for individual prediction of dietary restraint (DEBQ-R) from physical, psychological, behavioural and peer variables Variables predicting individual dieting (DEBQR) Step 1 Body mass index (BMI) Step 2 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Step 3 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Constraint (MPQC) Alcohol intake Cigarette intake Step 4 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Constraint (MPQC) Alcohol intake Cigarette intake Adjusted clique mean Friend concern with thinness/dieting Friends as a source of inuence Peer pressure to lose weight/be thin Peer teasing (POTS) Note: n 772. **po.01, ***po.001. B SE b T R R2 DR2
.34 .95 .80 .15 .10 .28 .80 .17 .09 .28 .18 1.46 .21 .39 .08 .07 .18 .18 .62 .20 .13 .29 .29 .57 .01 .09 .09 .07 .06 .04 .09 .07 .06 .04 .07 .96 .31 .08 .06 .05 .04 .06 .83 .26 .05 .08 .06 .07 .08 .35 .29 .08 .07 .28 .29 .09 .06 .29 .08 .06 .03 .14 .04 .05 .19 .08 .03 .03 .08 .11 .17 .33 .01 10.33*** .52 9.35*** 2.18 1.67 6.39*** .52 9.31*** 2.41 1.45 6.49*** 2.45 1.52 .68 .70 4.85*** 1.43 1.37 4.85*** 3.02** .76 .77 2.78** 3.49** 4.68*** 7.75*** .16
.12 .27
.12*** .15***
.27
.00
.48
.21***
of EWLB scores, DR2 .12, po.001, although negative emotionality was not a statistically signicant predictor. In block 3 constraint and substance use variables did not improve predictive efcacy. However, in block 4, the addition of peer inuence variables resulted in a signicant improvement in prediction, DR2 .08, po.001. In the nal model, self-esteem scores contributed the largest amount of explained variance, followed by peer inuence to be thin and diet and the adjusted clique mean for EWLB use. To a lesser extent, friends as a source of inuence and depression scores also contributed to the explained variance. However, when BMI and depression scores were entered together with the peer variables they no longer emerged as signicant predictors. Overall, the nal regression model explained 26% of the variability in individual use of EWLBs. Multivariate prediction of individual binge eating Table 9 summarises the results of regression analyses predicting individual EDI-B scores. In block 1, BMI was a signicant predictor of binge eating, DR2 .04, po.001. Block 2 variables contributed a signicant improvement in explained variance in individual binge eating, DR2 .21, po.001, with negative emotionality accounting for the largest proportion of variance, followed by self-esteem, BMI and depression. In block 3 constraint and substance use variables did not improve predictive efcacy. The subsequent addition of peer inuence variables in block 4 resulted in a small yet signicant increase in prediction, DR2 .06, po.001. In the nal model, negative emotionality contributed the largest amount of explained variance in individual binge eating scores, followed by weight-related peer teasing, self-esteem and perceived peer pressure to be thin and
ARTICLE IN PRESS
1570 D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 Table 8 Hierarchical multiple regression analyses for individual prediction of extreme weight loss behaviours (EWLB) from physical, psychological, behavioural and peer variables Variables predicting individual use of extreme weight loss behaviours (EWLB) Step 1 Body mass index (BMI) Step 2 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Step 3 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Constraint (MPQC) Alcohol intake Cigarette intake Step 4 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Constraint (MPQC) Alcohol intake Cigarette intake Adjusted clique mean Friend concern with thinness/dieting Friends as a source of inuence Peer pressure to lose weight/be thin Peer teasing (POTS) B SE b T R R2 DR2
.21 .05 .03 .03 .01 .01 .03 .03 .01 .02 .01 .17 .02 .01 .02 .00 .01 .01 .05 .00 .11 .01 .02 .03 .01 .01 .01 .01 .01 .00 .01 .01 .01 .01 .01 .08 .03 .01 .01 .01 .00 .01 .08 .03 .03 .01 .01 .01 .01 .21 .15 .17 .06 .18 .15 .18 .06 .18 .06 .08 .03 .04 .14 .01 .12 .06 .02 .00 .11 .04 .11 .18 .04 5.99*** .41 4.61*** 4.41*** 1.39 3.82*** .42 4.51*** 4.62*** 1.19 3.85*** 1.78 2.00 .81 .51 1.16 3.83*** .24 2.62** 1.90 .57 .05 3.40** 1.06 2.71** 3.52*** .89
.05 .17
.05*** .12***
.18
.01
.26
.08***
Note: n 771. **po.01, ***po.001. Results are based on square root transformations of EWLB scores.
diet. However, after the peer variables were entered in block 4, BMI and depression were no longer signicant predictors. The nal regression model accounted for 32% of explained variation in individual binge eating scores. Discussion The present research focused on expanding the understanding of peer factors in body image and eating problems during early adolescence. This is the rst known study to investigate whether friendship similarities exist in body image attitudes and eating behaviours in a large cohort of girls entering adolescence. The results of this study extend on previous research in a number of ways. First, clique analyses indicated that during early adolescence friendship group members shared similarities in dietary restraint, EWLBs and binge eating, but not body image concern. Second, cliques characterised by high and low levels of dieting, use of EWLBs and binge eating, could also be distinguished by their average scores on physical, psychological, behavioural and peer inuence measures. Third, compared with clique members, girls who were not identied as part of a friendship group had lower self-esteem, and higher BMI, body image concern, and use of EWLBs. Finally, consistent with past research, perceived friendship inuence in body image attitudes and eating behaviours contributed signicantly to the prediction of individual girls body image concern, dieting, use of EWLBs and
ARTICLE IN PRESS
D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 1571 Table 9 Hierarchical multiple regression analyses for individual prediction of binge eating (EDI-B) from physical, psychological, behavioural and peer variables Variables predicting individual binge eating (EDI-B) Step 1 Body mass index (BMI) Step 2 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Step 3 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Constraint (MPQC) Alcohol intake Cigarette intake Step 4 Body mass index (BMI) Self-esteem (RSE) Negative emotionality (MPQN) Depression (CESD-C) Constraint (MPQC) Alcohol intake Cigarette intake Adjusted clique mean Friend concern with thinness/dieting Friends as a source of inuence Peer pressure to lose weight/be thin Peer teasing (POTS) B se b T R R2 D R2
.20 .04 .03 .02 .03 .01 .03 .02 .03 .01 .01 .14 .00 .00 .02 .02 .00 .01 .03 .02 .02 .01 .00 .02 .03 .01 .01 .01 .01 .00 .01 .01 .01 .00 .01 .07 .02 .01 .01 .00 .00 .01 .07 .02 .01 .01 .01 .01 .01 .20 .13 .16 .27 .12 .12 .16 .27 .11 .06 .07 .01 .01 .13 .20 .06 .06 .02 .03 .07 .06 .03 .15 .13 5.77*** .50 4.08*** 4.23*** 6.14*** 2.70** .51 3.76*** 4.24*** 6.21*** 2.47 1.92 1.83 .14 .57 .18 3.57*** 4.68*** 1.25 2.00 .43 .89 2.14 1.53 .78 3.14** 3.59***
.04 .25
.04*** .21***
.26
.01
.32
.06***
Note: n 770. **po.01, ***po.001. Results are based on square root transformations of EDI-B scores.
binge eating. Taken together, these results provide support for theoretical models that emphasise the signicance of sociocultural factors in the development of body image disturbance and disordered eating patterns (Levine & Smolak, 1992; Stice & Agras, 1998; Striegel-Moore et al., 1986; Thompson et al., 1999). Consistent with previous research in older samples (Meyer & Waller, 2001; Paxton et al., 1999), greater between- than within-group variation in dietary restraint and use of EWLBs were identied, after controlling for signicant covariation with BMI, self-esteem and negative affect. This suggests that girls share similarities in weight loss behaviours considerably earlier in adolescence than previously identied. Regression analyses also indicated that average clique scores on DEBQ-R and EWLB measures, adjusted to remove the contribution of each individual girls own score, signicantly contributed to the respective prediction of individual dietary restraint and EWLBs. Put simply, girls with friends who dieted were also likely to diet. These results are consistent with the notion that girls who exhibit high levels of dieting and EWLBs occupy peer environments that foster dieting as a means of achieving a socioculturally valued thin-ideal (Levine et al., 1994; Paxton et al., 1999; Stice, 1994). Adolescent friends also reported similar levels of binge eating, after controlling for signicant covariation with self-esteem and negative affect. Effect sizes indicated that as much as 24% of the variation in binge eating in this sample was explained by clique membership. While this is consistent with two earlier studies that identied friendship similarities in binge eating among older samples (Crandall, 1988; Pike, 1995), several
ARTICLE IN PRESS
other studies found conicting results (Meyer & Waller, 2001; Paxton et al., 1999; Pike, 1995). One explanation for these discrepancies is that similarities in binge eating only occur at specic time points (Paxton et al., 1999; Pike, 1995). As adolescent development involves increasing individuation from both family and peers (Douvan & Adelson, 1966), it may be the case that girls are more susceptible to social inuence in the early stages of adolescence. Research in the area of substance use has also indicated that social inuence is stronger in the initiation than in the maintenance of drug use behaviours (Kandel, 1974). As girls in the current sample were in their rst year of high school, when new friendships are established, peer pressure to initiate binge eating may have been especially strong, contributing to friendship similarity in binge eating. An alternate explanation for mixed ndings is that similarities are only observed in behaviours that are socially valued by clique members (Meyer & Waller, 2001). As binging is often considered shameful and secretive, similarities are less likely to be observed in binge eating than in dieting behaviours, which tend to be more socially accepted. Consistent with this, ndings indicated that clique membership explained somewhat more of the variance in dieting and EWLBs than in binge eating. Mixed ndings in the literature may therefore reect the extent to which binge eating was socially valued among cliques in different samples (Meyer & Waller, 2001). Longitudinal research is needed to clarify whether similarities in binge eating only occur at specic developmental time points in adolescence, or whether discrepancies between cross-sectional studies reect cohort effects. In contrast to previous research (Meyer & Waller, 2001; Paxton et al., 1999), body image concern did not vary signicantly as a function of clique membership. However, girls with elevated levels of body image concern did report signicantly lower self-esteem and higher BMI and negative affect scores. This suggests that variations in body image concern in early adolescence are better accounted for by psychological and physical variables than clique membership. One explanation for why friendship similarities were not identied is that people tend to share greater similarity on observable attributes and behaviours, than on covert attitudes and beliefs (Kandel, 1978b; Tolson & Urberg, 1993). It has been suggested that being similar to ones friends in observable ways minimises conict between differing attitudes and beliefs, and increases the likelihood of engaging in mutually valued activities (Kandel, 1978a). Consistent with this explanation, female friends shared similarities in dieting, EWLB use and binge eating, but not covert body attitudes. Moderate inter-correlations did indicate, however, that cliques with high and low DEBQ-R, EWLB and EDI-B scores could be characterised by corresponding high and low levels of body image concern. Thus, there was a trend for cliques that engaged in binge eating and weight loss behaviours to report high levels of body image concern, even though signicant between-group differences were not identied. It may be the case that this trend strengthens over time, with female friendship cliques becoming more differentiated in their levels of body image concern. This would account for within-group similarities that have been identied in older samples (Paxton et al., 1999). An alternate explanation for why friendship similarities in BAQ scores were not identied is that body image concerns are normative in young adolescent girls. In this study, body image concerns were clearly more common than dieting, EWLBs and binge eating. This ts with the notion that females from many societies develop within a sociocultural milieu that espouses thinness and physical beauty. Not only do adolescent girls compare their developing bodies to unrealistic physical ideals, but they are frequently exposed to information about how to diet and exercise to attain this thin-ideal (Levine et al., 1994; Striegel-Moore et al., 1986). Of course, not all weight and shape concerns are pathological. It has been argued that some level of body dissatisfaction is actually healthy, promoting normal weight and obese individuals to engage in healthy eating and exercise (Heinberg, Thompson, & Matzon, 2001). That higher levels of individual body image concern are more pathological, however, was supported by the nding that the majority of variance in both clique and individual body image concern was explained by psychological and physical factors. Further research is needed to elucidate why adolescent friendship groups are more readily distinguished by eating behaviours than weight-related attitudes and whether this pattern changes over the course of adolescence and adulthood. After examining peer similarity in body image and eating behaviours, this study assessed whether cliques that had elevated levels of dieting, EWLB use and binge eating could also be characterised by their scores on a range of physical, psychological, behavioural and peer variables. In line with previous research among older adolescents (Paxton et al., 1999), cliques characterised by elevated levels of dieting, EWLB use and binge eating tended to have similarly high scores on peer-related variables including: friends concern with thinness
ARTICLE IN PRESS
and dieting, friends as a source of inuence, peer pressure to lose weight and be thin and weight-related teasing by peers. Clique members that regularly dieted, used EWLBs or engaged in binge eating also reported signicantly lower self-esteem, and higher negative emotionality, depression and BMI scores. These ndings indicate that even in early adolescence, female friends resemble each other in more complex ways than merely their eating behaviours. Not only does this highlight the need for prevention and early intervention at the level of the peer group, but also the importance of targeting emotional and physical problems alongside body image and eating disturbances. The current study extended on past research by examining the role of impulsivity and substance use in body image and eating behaviours. Results indicated that average clique DEBQ-R, EWLB and EDI-B scores were not associated with clique impulsivity. Although impulsivity did contribute to the prediction of individual dieting in the multivariate model, the proportion of explained variance was relatively small. Thus, impulsivity did not emerge as a signicant factor in early adolescence. In regard to substance use, neither alcohol nor cigarette use predicted an individual girls body image or eating behaviours in the multivariate models. However, small positive relationships were identied between average clique alcohol and cigarette use, and clique averages for dieting, EWLBs and binge eating. The fact that these relationships were weak is consistent with the low frequency of smoking (1.5%) and alcohol use (4.3%) in the sample. Stronger relationships may exist among older adolescents who typically engage in higher levels of substance use. The third aim of this study was to examine the characteristics of girls who were not classied part of a friendship clique. Comparisons between these girls and clique members indicated that the two groups did not differ signicantly on measures of impulsivity and substance use, nor on any of the peer variables of interest, including friends concern with thinness and dieting, friends as a source of inuence, peer pressure to lose weight and be thin, weight-related teasing by peers and number of close friends outside school. Thus, girls who were not clearly linked to a friendship group perceived themselves to be inuenced by similar weight-related peer attitudes and behaviours as girls who were part of a group. Closer examination of this group indicated that they were not without friends. In fact, half were members of dyadic friendship groups and the majority reported having close friends outside the school environment. It appears that many of these girls may have been exposed to weight-related peer inuences and behaviours from alternate sources, possibly explaining why differences were not observed in perceived peer inuence. Further comparisons between the two groups on physical, psychological, body image and eating variables found that girls who were not part of a friendship clique had signicantly lower self-esteem, and higher BMI, body image concern, and EWLB use compared with clique members. However, friendship status only explained 1% of the variance in each of these variables, which are not substantively meaningful effect sizes. Finally, this study examined the role of perceived peer inuence in the prediction of individual body image concern, dieting, EWLBs and binge eating, after evaluating the contribution of previously identied physical, psychological and behavioural variables. Perceived peer pressure to be thin and diet was a consistent predictor of individual scores on each of the four dependent variables. Thus, girls perceptions that their friends frequently discuss and engage in dieting were important in the prediction of individual body image and eating disturbance. Girls who perceived their friends as an important source of inuence also had higher levels of body image concern, dieting and EWLBs. In addition, perceived friend concern with thinness and dieting predicted individual dieting, and perceived weight-related teasing was predictive of individual binge eating. These ndings are consistent with past research that identied associations between perceived peer inuence and early adolescent weight and shape attitudes (Levine et al., 1994; Lieberman et al., 2001; Thompson & Heinberg, 1993). One explanation for these ndings is that girls who engage in disordered eating merely project their own weight and shape concerns onto their friends. An alternate explanation, consistent with the correlational data, is that adolescent girls who exhibit high levels of weight loss or binge eating behaviour occupy peer environments in which body weight and shape are valued (Paxton et al., 1999). Although this study conrmed that peer similarities in eating behaviours exist earlier than previously identied, several limitations were evident. First, due to the statistical assumption of independence of groups, girls were allocated to discreet friendship cliques. Segregating friends in this way fails to capture the complex overlap that typically occurs among friends involved in different social groups. Future research could examine different types of friendship links, including dual and multiple group membership. A second methodological limitation was that some of the peer inuence measures had not been subject to extensive psychometric
ARTICLE IN PRESS
assessment. However, these measures demonstrated good reliability in the current sample and have been associated with similar trends in past research (e.g. Paxton et al., 1999). Some recent validity studies also suggest that the DEBQ-R scale measures a tendency toward overeating that is frequently associated with dietary restriction, rather than dietary restriction per se (e.g. van Strien, Engels, van Staveren, & Herman, 2006). The validity of the DEBQ-R needs to be re-examined to ascertain whether it is caloric restriction or other factors, such as disinhibition, that are important in the peer context. It is also notable that signicant differences were identied between clique members and non-members. Although non-members were inappropriate for the clique analyses, it is possible that their exclusion impacted on the overall representativeness of the nal clique data set. Finally, although this study adjusted for relevant covariates, there may have been other factors, not assessed in this study, that covaried with body image and eating behaviours. Future research could control for additional factors likely to contribute to body image and eating problems (e.g. family and media inuences, biological factors). Given that early body image and eating disturbances are known precursors to the development of more severe eating problems, the ndings of this cross-sectional study suggest that members of adolescent cliques that exhibit elevated levels of dieting, EWLBs or binge eating, may be at risk for more serious eating problems. To determine whether this is the case, prospective research is needed to map the development of body image and eating problems among friendship cliques across time. This would also clarify whether mixed ndings regarding friendship similarities in body image concern and binge eating reect cohort effects, or whether there are specic time points in adolescence and early adulthood when friendship similarities occur. If girls who inhabit peer environments that value thinness, weight loss or binge eating are demonstrated to be at risk for developing more severe eating disturbances, then prevention and early intervention programs are indicated at the peer level. It follows, however, that to develop effective programs, longitudinal research needs to assess the contribution of selection and socialisation in the development of peer similarity. From a methodological standpoint, some researchers have argued that the use of self-selected friendship groups makes it difcult to determine which mechanisms account for friendship similarity (Meyer and Waller, 2001). To control for pre-existing similarities, it is recommended that baseline assessments in longitudinal research be conducted prior to the formation of strong friendship links (i.e. during the rst few weeks of high school). In summary, the results of this study indicate that members of female friendship cliques shared similarities in dieting, use of EWLBs and binge eating, after controlling for BMI, self-esteem and negative affect. Perceptions of peer attitudes and behaviours also contributed to the prediction of individual body image and eating disturbances after physical, psychological and behavioural factors had been considered. Notably, individual dieting and use of EWLBs could be predicted from friends scores on these respective measures. These ndings extend on previous research, underscoring the importance of the peer environment in body image and eating problems in early adolescence. Acknowledgements The authors would like to thank Professor Susan Paxton for her suggestions regarding the planning and implementation of the study and Dr. Alan Taylor for his helpful statistical advice. Thanks also go to the students who participated in the research and the school staff who kindly assisted with the implementation of the study. Dr. Hutchinson was supported by a Research Training Fellowship from the New South Wales Institute of Psychiatry and an Australian Postgraduate Award Scholarship. References
Agras, W. (2001). The consequences and costs of the eating disorders. Psychiatric Clinics of North America, 24(2), 371379. Aseltine, R. H. (1995). A reconsideration of parental and peer inuences on adolescent deviance. Journal of Health & Social Behavior, 36(2), 103121. Australian Bureau of Statistics (ABS). (1996). New South Wales young people. Canberra: ABS. Australian Bureau of Statistics (ABS). (2001). New South Wales census data. Canberra: ABS.
ARTICLE IN PRESS
D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 1575 Banasiak, S. J., Wertheim, E. H., Koerner, J., & Voudouris, N. J. (2001). Testretest reliability and internal consistency of a variety of measures of dietary restraint and body concerns in a sample of adolescent girls. International Journal of Eating Disorders, 29(1), 8589. Bardone, A. M., Vohs, K. D., Abramson, L. Y., Heatherton, T. F., & Joiner, T. E. J. (2000). The conuence of perfectionism, body dissatisfaction, and low self-esteem predicts bulimic symptoms: Clinical implications. Behavior Therapy, 31(2), 265280. Ben-Tovim, D. I., & Walker, K. (1991). The development of the BenTovim Walker Body Attitudes Questionnaire (BAQ), a new measure of womens attitudes towards their bodies. Psychological Medicine, 21, 775784. Ben-Tovim, D. I., & Walker, M. (1994). The inuence of age and weight on womens body attitudes as measured by the Body Attitudes Questionnaire (BAQ). Journal of Psychosomatic Research, 38(5), 477481. Berscheid, E., Dion, K., Walster, E., & Walster, G. W. (1971). Physical attractiveness and dating choice: A test of the matching hypothesis. Journal of Experimental Social Psychology, 7, 173189. Borgatti, S. P., Everett, M. G., & Freeman, L. C. (2002). Ucinet for Windows: Software for social network analysis. Harvard: Analytic Technologies. Button, E., Sonuga-Barke, E., Davies, J., & Thompson, M. (1996). A prospective study of self-esteem in the prediction of eating problems in adolescent schoolgirls: Questionnaire ndings. British Journal of Clinical Psychology, 35(2), 193203. Cattarin, J. A., & Thompson, J. K. (1994). A three-year longitudinal study of body image, eating disturbance, and general psychological functioning in adolescent females. Eating Disorders: Journal of Treatment & Prevention, 2, 114125. Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Coleman, J. C. (1980). Friendship and the peer group in adolescence. In J. Adelson (Ed.), Handbook of adolescent psychology (pp. 408431). New York: Wiley. Crandall, C. S. (1988). Social contagion of binge eating. Journal of Personality and Social Psychology, 55, 588598. Curren, P. J., Stice, E., & Chassin, L. (1997). The relation between adolescent and peer alcohol use: A longitudinal random coefcients model. Journal of Consulting and Clinical Psychology, 65, 130140. Douvan, E., & Adelson, J. (1966). The adolescent experience. New York: Wiley. Ennett, S. T., & Bauman, K. E. (1994). The contribution of inuence and selection to adolescent peer group homogeneity: The case of adolescent cigarette smoking. Journal of Personality & Social Psychology, 67(4), 653663. Epstein, J. L. (1989). The selection of friends: Changes across the grades and in different school environments. In T. J. Bernt, & G. W. Ladd (Eds.), Peer relationships in child development (pp. 158187). New York: Wiley. Favaro, A., Ferrara, S., & Santonastaso, P. (2003). The spectrum of eating disorders in young women: A prevalence study in a general population sample. Psychosomatic Medicine, 65(4), 701708. Fendrich, M., Weissman, M. M., & Warner, V. (1990). Screening for depressive disorder in children and adolescents: Validating the center for epidemiologic studies depression scale for children. American Journal of Epidemiology, 131(3), 538551. French, S. A., Story, M., Neumark-Sztainer, D., Downes, B., Resnick, M., & Blum, R. (1997). Ethnic differences in psychosocial and health behavior correlates of dieting, purging, and binge eating in a population-based sample of adolescent females. International Journal of Eating Disorders, 22(3), 315322. Garner, D. M., Olmsted, M. P., & Polivy, J. (1983). Development and validation of a multidimensional eating disorder inventory for anorexia nervosa and bulimia. International Journal of Eating Disorders, 2, 1534. Garry, J. P., Morrissey, S. L., & Whetstone, L. M. (2003). Substance use and weight loss tactics among middle school youth. International Journal of Eating Disorders, 33(1), 5563. Ghaderi, A., & Scott, B. (2001). Prevalence, incidence and prospective risk factors for eating disorders. Acta Psychiatrica Scandinavica, 104, 122130. Graber, J. A., Brooks-Gunn, J., Paikoff, R. L., & Warren, M. P. (1994). Prediction of eating problems: An 8-year study of adolescent girls. Developmental Psychology, 30(6), 823834. Heatherton, T. F., & Polivy, J. (1992). Chronic dieting and eating disorders: A spiral model. In J. H. Crowther, D. L. Tennenbaum, S. E. Hobfoll, & M. A. P. Stephens (Eds.), The etiology of bulimia nervosa: The individual and familial context (pp. 133155). Washington: Hemisphere Publishing Corporation. Heinberg, L. J., Thompson, J. K., & Matzon, J. L. (2001). Body image dissatisfaction as a motivator for healthy lifestyle change: Is some distress benecial? In R. H. Striegel-Moore, & L. Smolak (Eds.), Eating disorders: Innovative directions in research and practice (pp. 215232). Washington, DC: American Psychological Association. Hoek, H. W., & Van Hoeken, D. (2003). Review of the prevalence and incidence of eating disorders. International Journal of Eating Disorders, 34(4), 383396. Jacobi, C., Hayward, C., De Zwaan, M., Kraemer, H. C., & Agras, W. (2004). Coming to terms with risk factors for eating disorders: Application of risk terminology and suggestions for a general taxonomy. Psychological Bulletin, 130(1), 1965. Jessor, R. (1987). Problem-behavior theory, psychosocial development, and adolescent problem drinking. British Journal of Addiction, 82(4), 331342. Kandel, D. (1974). Inter- and intragenerational inuences on adolescent marijuana use. Journal of Social Issues, 50, 107135. Kandel, D. B. (1978a). Homophily, selection, and socialization in adolescent friendships. American Journal of Sociology, 84(2), 427436. Kandel, D. B. (1978b). Similarity in real-life adolescent friendship pairs. Journal of Personality & Social Psychology, 36(3), 306312. Keel, P. K., Dorer, D. J., Eddy, K. T., Franko, D., Charatan, D. L., & Herzog, D. B. (2003). Predictors of mortality in eating disorders. Archives of General Psychiatry, 60(2), 179183. Kinzl, J. F., Traweger, C., Trefalt, E., Mangweth, B., & Biebl, W. (1999). Binge eating disorder in females: A population-based investigation. International Journal of Eating Disorders, 25(3), 287292.
ARTICLE IN PRESS
1576 D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 Leon, G. R., Fulkerson, J. A., Perry, C. L., Keel, P. K., & Klump, K. L. (1999). Three to four year prospective evaluation of personality and behavioural risk factors for later disordered eating in adolescent girls and boys. Journal of Youth and Adolescence, 28(2), 181196. Levine, M. P., & Smolak, L. (1992). Toward a model of the developmental psychopathology of eating disorders: The example of early adolescence. In J. H. Crowther, D. L. Tennenbaum, S. E. Hobfoll, & M. A. P. Stephens (Eds.), The etiology of bulimia nervosa: The individual and familial context (pp. 5980). Washington: Hemisphere Publishing Corporation. Levine, M. P., Smolak, L., & Hayden, H. (1994). The relation of sociocultural factors to eating attitudes and behaviors among middle school girls. Journal of Early Adolescence, 14(4), 471490. Levine, T. R., & Hullett, C. R. (2002). Eta squared, partial eta squared, and misreporting of effect sizes in communication research. Human Communication Research, 28(4), 612625. Lieberman, M., Gauvin, L., Bukowski, W. M., & White, D. R. (2001). Interpersonal inuence and disordered eating behaviors in adolescent girls: The role of peer modeling, social reinforcement, and body-related teasing. Eating Behaviors, 2(3), 215236. Matsueda, R. L., & Anderson, K. (1998). The dynamics of delinquent peers and delinquent behavior. Criminology, 36, 269308. Matsumoto, S., Kumano, H., & Sakano, Y. (1999). Investigation of social inuence on eating tendencies and dieting behavior among female students. Japanese Journal of Behavior Therapy, 25(2), 1123. McCarthy, M. (1990). The thin ideal, depression and eating disorders in women. Behavioral Research and Therapy, 28, 205218. Meyer, C., & Waller, G. (2001). Social convergence of disturbed eating attitudes in young adult women. The Journal of Nervous and Mental Disease, 189(2), 114119. Paxton, S. J., Schutz, H. K., Wertheim, E. H., & Muir, S. L. (1999). Friendship clique and peer inuences on body image concerns, dietary restraint, extreme weight-loss behaviors, and binge eating in adolescent girls. Journal of Abnormal Psychology, 108(2), 255266. Paxton, S. J., Wertheim, E. H., Gibbons, K., Szmukler, G. I., Hillier, L., & Petrovich, J. L. (1991). Body image satisfaction, dieting beliefs, and weight loss behaviors in adolescent girls and boys. Journal of Youth & Adolescence, 20, 361379. Pearson, J., Goldklang, D., & Striegel-Moore, R. H. (2002). Prevention of eating disorders: Challenges and opportunities. International Journal of Eating Disorders, 31(3), 233239. Picot, A. K., & Lilenfeld, L. R. (2003). The relationship among binge severity, personality psychopathology, and body mass index. International Journal of Eating Disorders, 34(1), 98107. Pike, K. M. (1995). Bulimic symptomatology in high school girls. Psychology of Women Quarterly, 19, 373396. Rosenberg, M. (1965). Measurement of self-esteem. In M. Rosenberg (Ed.), Society and the adolescent self-image (pp. 297307). New York: Princeton University. Scott, J. (2001). Social network analysis: A handbook (2nd ed.). London: Sage Publications. Shisslak, C. M., Crago, M., & Estes, L. S. (1995). The spectrum of eating disturbances. International Journal of Eating Disorders, 18(3), 209219. Shoemaker, C., van Strien, T., & van Der Staak, C. (1994). Validation of the Eating Disorder Inventory in a non-clinical population using transformed and untransformed responses. International Journal of Eating Disorders, 15, 387393. Stice, E. (1994). Review of the evidence for a sociocultural model of bulimia nervosa and an exploration of the mechanisms of action. Clinical Psychology Review, 14(7), 633661. Stice, E. (1998). Modeling of eating pathology and social reinforcement of the thin-ideal predict onset of bulimic symptoms. Behaviour Research and Therapy, 36, 931944. Stice, E., & Agras, W. (1998). Predicting onset and cessation of bulimic behaviors during adolescence: A longitudinal grouping analysis. Behavior Therapy, 29(2), 257276. Stice, E., Burton, E. M., & Shaw, H. (2004). Prospective relations between bulimic pathology, depression, and substance abuse: Unpacking comorbidity in adolescent girls. Journal of Consulting and Clinical Psychology, 72(1), 6271. Stice, E., Killen, J. D., Hayward, C., & Taylor, C. (1998). Age of onset for binge eating and purging during late adolescence: A 4-year survival analysis. Journal of Abnormal Psychology, 107(4), 671675. Striegel-Moore, R. H., Silberstein, L. R., & Rodin, J. (1986). Toward an understanding of risk factors for bulimia. American Psychologist, 14, 246263. Taylor, R. B., de Soto, C. B., & Lieb, R. (1979). Sharing secrets: Disclosure and discretion in dyads and triads. Journal of Personality and Social Psychology, 37(7), 11961203. Tellegen, A. (1982). Brief manual for the Differential Personality Questionnaire. Unpublished manuscript, University of Minnesota, Minneapolis. Thompson, J., Cattarin, J., Fowler, B., & Fisher, E. (1995). The perception of teasing scale (POTS): A revision and extension of the Physical Appearance Related Teasing Scale (PARTS). Journal of Personality Assessment, 65(1), 146157. Thompson, J. K., & Heinberg, L. J. (1993). Preliminary test of two hypotheses of body image disturbance. International Journal of Eating Disorders, 14(1), 5963. Thompson, J. K., Heinberg, L. J., Altabe, M., & Tantleff-Dunn, S. (1999). Future directions: Integrative theories, multidimensional assessments, and multicomponent interventions. In Exacting beauty: Theory, assessment, and treatment of body image disturbance (pp. 311326). Washington, DC: American Psychological Association. Tolson, J. M., & Urberg, K. A. (1993). Similarity between adolescent best friends. Journal of Adolescent Research, 8(3), 274288. van Strien, T., Engels, R. C., van Staveren, W., & Herman, C. (2006). The validity of dietary restraint scales: Comment on Stice et al. (2004). Psychological Assessment, 18(1), 8994. van Strien, T., Frijters, J. E. R., Bergers, G. P. A., & Defares, P. B. (1986). The Dutch Eating Behaviour Questionnaire (DEBQ) for assessment or restrained, emotional, and external eating behaviour. International Journal of Eating Disorders, 5, 295315.
ARTICLE IN PRESS
D.M. Hutchinson, R.M. Rapee / Behaviour Research and Therapy 45 (2007) 15571577 1577 Waller, N. (1990). [Development of short forms for the negative emotionality and constraint high order factor scales of the Multidimensional Personality Questionnaire]. Unpublished raw data. Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications. Cambridge, New York: Cambridge University Press. Wertheim, E. H., Paxton, S. J., Schutz, H. K., & Muir, S. L. (1997). Why do adolescent girls watch their weight? An interview study examining sociocultural pressures to be thin. Journal of Psychosomatic Research, 42, 345355.

Do Friends Share Similar Body Image and Eating Problems The

Uploaded by

Copyright:

Available Formats

Do Friends Share Similar Body Image and Eating Problems The

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Do Friends Share Similar Body Image and Eating Problems The

Uploaded by

Copyright:

Available Formats

How does the study examine the role of friendship networks and peer influences on body image and eating behaviors?

What individual attributes and environmental factors have researchers examined as potential early risk factors for eating disorders?

ARTICLE IN PRESS

Behaviour Research and Therapy 45 (2007) 15571577 www.elsevier.com/locate/brat

0005-7967/$ - see front matter r 2006 Published by Elsevier Ltd. doi:10.1016/j.brat.2006.11.007

Note: Number of cliques 173.

6.73 1.33 4.94 .45 13.32 7.09 .40 2.37

.010 .249 .026 .501 .000 .008 .527 .124

.01 .00 .01 .00 .01 .01 .00 .00

76.99 4.32 .72 3.56 21.05 29.21 42.78 14.82

27.75 1.02 .74 .71 3.84 5.10 6.84 9.65

You might also like