- Email: [email protected]
The relation between parental influence, body image, and eating behaviors in a nonclinical female sample
Body Image 9 (2012) 93–100
Contents lists available at SciVerse ScienceDirect
Body Image journal homepage: www.elsevier.com/locate/bodyimage
The relation between parental influence, body image, and eating behaviors in a nonclinical female sample Michelle Abraczinskas a , Brian Fisak Jr. b,∗ , Rachel D. Barnes c a
University of North Carolina at Charlotte, Charlotte, NC, USA University of North Florida, Jacksonville, FL, USA c Yale University School of Medicine, New Haven, CT, USA b
a r t i c l e
i n f o
a b s t r a c t
Article history: Received 10 September 2010 Received in revised form 14 October 2011 Accepted 19 October 2011 Keywords: Body image Parental influence Measurement Modeling Eating disorders
The purpose of the current study is to create a comprehensive composite measure of parental influence based on previously developed measures to clarify the underlying dimensions of parental influence and to determine the degree to which parental influence relates to body image and dysfunctional weight concerns. Previously published literature was reviewed for measures of parental influence, and items from 22 measures were condensed and combined into a single questionnaire, which was completed by 367 female undergraduate psychology students. Two dimensions emerged from a principle components analysis: Direct Influence, which includes weight and eating related comments, and Modeling, which includes parental modeling of dieting and related behavior. Direct Influence and Modeling were significantly related to eating disturbance, such as drive for thinness and bulimic symptomatology. Overall, the results integrate the previous literature and clarify the underlying dimensions of parental influence. Further, this study provides directions for future research related to the development and maintenance of body image and eating disturbance. © 2011 Elsevier Ltd. All rights reserved.
Introduction Several studies have focused on the influence of social factors in the development and maintenance of body image and eating disturbance, and parents are among the most commonly examined sources of influence (Hill & Franklin, 1998; Moreno & Thelen, 1993; Pike & Rodin, 1991; Smolak, Levine, & Schermer, 1999; Thelen & Cormier, 1995). One reason for the focus on parental influence is that parents typically are the first sources of socialization (McCabe & Ricciardelli, 2003). Further, based on the Tripartite Influence Model, parental influence, along with peer and media influence, has been described as a distinct factor in the development of body image and eating disturbance (Keery, van den Berg, & Thompson, 2004; Shroff & Thompson, 2006; van den Berg, Thompson, Obremski-Brandon, & Coovert, 2002). Although considerable variability exists in the operationalization of parental influence in the development of maladaptive eating behaviors and body image, researchers often discuss parental influence as two distinct categories. The first category is direct influence, also referred to as verbal communication or verbal influence (Fulkerson, McGuire, Neumark-Sztainer, Story, French, & Perry, 2002; Smolak et al., 1999; Vincent & McCabe, 2000; Wertheim, Mee,
∗ Corresponding author. E-mail address: b.fi[email protected] (B. Fisak Jr.). 1740-1445/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.bodyim.2011.10.005
& Paxton, 1999). The second category is modeling, also referred to as indirect influence (Ogden & Steward, 2000; Pike & Rodin, 1991; Smolak et al., 1999; Stice, 1998; Vincent & McCabe, 2000; Wertheim et al., 1999). Direct influence typically includes parental behaviors such as discussion and encouragement of child dieting or other attempts to restrict the child’s weight (Fulkerson et al., 2002; Smolak et al., 1999; Vincent & McCabe, 2000; Wertheim et al., 1999). Modeling includes behaviors such as parental dieting, parental expression of weight or body dissatisfaction, and other observable behaviors on the part of the parent to maintain or reduce weight (Ogden & Steward, 2000; Pike & Rodin, 1991; Smolak et al., 1999; Stice, 1998; Vincent & McCabe, 2000; Wertheim et al., 1999). Research has found consistent support for direct influence as a predictor of body dissatisfaction and weight-loss behaviors (Smolak et al., 1999; Wertheim et al., 1999; Young, Clopton, & Bleckley, 2004). For example, Vincent and McCabe (2000) found parental discussion about weight loss predicted disordered eating behavior in a sample of nonclinical adolescent girls. In addition, parental encouragement of daughters to lose weight or diet has been found to be associated with bulimia symptoms, including vomiting and the use of laxatives (Benedikt, Wertheim, & Love, 1998; Dixon, Adair, & O’Connor, 1996; Hanna & Bond, 2006; Moreno & Thelen, 1993; Pike & Rodin, 1991). In addition to direct influence, parental modeling is predictive of body dissatisfaction and maladaptive eating behaviors in children (Edlund, Halvarsson, & Sjödén, 1996; Pike & Rodin, 1991; Rodgers
94
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
& Chabrol, 2009; Smolak et al., 1999; Stice, 1998; Wertheim et al., 1999). Specifically, by engaging in weight loss behaviors, or through the expression of their own body image dissatisfaction, parents may indirectly transmit messages that reinforce the importance of the thin ideal (Wertheim et al., 1999). For example, Stice (1998) found that modeling of family abnormal eating behavior was associated with bulimic symptoms and predicted the onset of bingeing and purging in daughters. Pike and Rodin (1991) found that mothers’ dieting behavior and weight concerns were associated with weight concerns in their daughters. Although the studies cited above found support for modeling, it is noteworthy that other studies failed to find an association between modeling and body image or eating disturbance (Byely, Archibald, Graber, & Brooks-Gunn, 2000; Dixon et al., 1996; Fulkerson et al., 2002; Kanakis & Thelen, 1995; Ruther & Richman, 1993). There are a number of potential explanations for these mixed findings. In particular, Fulkerson et al. (2002) suggested that body mass may play a role in parental influence. Children and adolescents with a higher BMI are more likely to receive direct feedback from parents about weight loss (Thelen & Cormier, 1995). Furthermore, studies that adjust for BMI are less likely to find an association between parental modeling and body image or eating disturbance (Fulkerson et al., 2002). In addition, modeling may serve a particularly important role in early adolescence, since that is a time when social comparison begins to emerge as well as an increase in weight and shape consciousness in girls (Levine & Smolak, 1992). Modeling may become less important as women age and their values are based more on internalized beliefs; therefore, developmental stage may be an additional factor that impacts the transmission of body image concerns and eating disturbance. It is also possible that methodological differences may account for the inconsistent support for modeling. In particular, studies utilizing clinical level samples more often find an association between modeling and eating disturbance (Pike & Rodin, 1991); however, the findings in this area have been inconsistent (Hill & Franklin, 1998; Kanakis & Thelen, 1995). It is also noteworthy that, although several measures of parental influence have been developed, research examining the psychometric properties of these measures has been limited (Wood, Becker, & Thompson, 1996). In particular, few studies have examined the factor structure of measures of parental influence (Benedikt et al., 1998). A related limitation is that there appears to be a general lack of consistency regarding the operationalization of the dimensions or components of parental influence. More specifically, several researchers have conceptualized direct communication and modeling as two distinct categories (Fulkerson et al., 2002; Vincent & McCabe, 2000); however, other researchers have treated parental influence as a single construct (Hill & Franklin, 1998; Pike & Rodin, 1991). In addition to the above limitations, there appear to be inconsistencies in the terminology or labels used to describe similar, and perhaps overlapping, aspects of parental influence. For example, Fulkerson et al. (2002) used the term “direct influence” to describe parental encouragement of dieting and weight-related comments, while Kichler and Crowther (2001) used “negative familial influence.” Overall, researchers concluded that the construct of parental influence needs to be clarified, as the full scope of the construct may not have been addressed in previous studies (Byely et al., 2000; Dixon et al., 1996; Fulkerson et al., 2002; Kanakis & Thelen, 1995; Ruther & Richman, 1993; Shroff & Thompson, 2006). In summary, although several studies have examined the role of parental influence in the development of body image and eating disturbance, there are a number of limitations to the previous literature. These limitations include inconsistent labeling and classification dimensions, and few studies have examined the psychometric properties of these measures, including the factor structure. In response to these limitations, the primary purpose
of the current study was to conduct an integrative analysis of previously developed measures of influence to: (1) elucidate the underlying dimensions of parental influence, and (2) develop a comprehensive and psychometrically sound measure of parental influence on the development of body image and eating disturbance. It was hypothesized that parental influence will be a multi-dimensional construct, which includes a direct influence and a modeling dimension, along with the potential for more. It was also hypothesized that each of the dimensions of parental influence will exhibit concurrent validity, as these dimensions are anticipated to be associated with body image and eating disturbance. Method Participants The sample consisted of 367 female undergraduate students at a large university in the southeast United States. The majority of the participants were White/Caucasian (75.7%). The average participant age was 22.69 years (SD = 5.61). Based on self-report of height and weight, the BMI for a majority of participants (67.6%) was within the normal range, which was operationalized as 18.5–24.9. Further, 26.0% of the participants fell within the overweight/obese range, 25.0 or greater. Finally, 6.3% fell within the underweight range, a BMI of 18.5 or lower. Development of the Parental Influence Questionnaire The Parental Influence Questionnaire (PIQ) is a retrospective measure that was developed for the purpose of this study. In development of this measure, efforts were made to obtain a broad and comprehensive pool of potential items (Clark & Watson, 1995). First, an extensive literature review was conducted using the Psycinfo database, in which all measures of parental influence related to body image and eating disturbance through July 2010 were identified and reviewed. These measures were located using combinations of the following search terms: Body Image, Eating disturbance, Parent influence, Direct influence, Indirect influence, Modeling, Children, Adolescents, and Dieting. Measures were included if the article in which it was found was peerreviewed, written in English, and was related to the influence parents have on the body image or eating disturbance of their own children. Based on this review, the initial item pool was created by extracting all potentially relevant items from measures referenced in published studies. When items were not available from published manuscripts, the authors of the manuscripts were contacted for the specific items. All contacted authors responded to requests. Based on this review, 118 items from 22 measures were included in the initial item pool (see Table 1). When necessary, items were modified to fit a 5-point Likert scale. This format is consistent with a majority of previously published measures of parental influence and the following response options were designated: Strongly Disagree, Somewhat Agree, Neither Agree nor Disagree, Somewhat Agree and Strongly Agree. Further, items were modified where necessary to fit the retrospective nature of this study, and when asked to rate the items, participants were asked to rate their interactions with their parents when a child and adolescent. Validation Measures Eating Disorder Inventory-3. The Eating Disorder Inventory (EDI-3; Garner, 2004) is a self-report inventory for the identification of symptoms of eating pathology. The Drive for Thinness, Bulimia, and Body Dissatisfaction subscales were utilized. The Drive
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
95
Table 1 Summary of measures and populations used to develop parental influence questionnaire. Authors
Measure
Gender
Age range
Ethnicity
Nationality
Source
Baker et al. (2000)
Attitude and Behavior Scale (ABS), Criticism about Eating & Appearance Maternal Encouragement of Daughter to Lose Weight Dieting History and Encouragement Single item measuring parental dieting Mother’s Perceptions of Adolescents Encouragement to Diet Parental Eating and Weight Measures Survey (PEWM) Perceived Frequency of Negative Comments about Appearance Family Teasing & Criticism, Parental Mother Investment in Daughter’s Shape, Mother’s Investment in her Own Thinness Family History Inventory
M&F
College-age
Predominately Caucasian
United States
Self, Mother, Father
F
15–16 yrs
Australia
Self, Mother
F
10–14 yrs
Caucasian Asian-Pacific Caucasian
United States
Self, Mother
F
13–16 yrs
Caucasian
New Zealand
Self
M&F
Middle/High School
Diverse
United States
Self, Mother
F
College-age
United States
Self
F
College-age
Predominantly Caucasian Predominantly Caucasian
United States
Self, Mother
F
10–14 yrs
Predominantly Caucasian
United States
Self
F
Predominantly Caucasian Mostly Anglo-Saxon
United States
Self
Australian
Self
Predominantly Caucasian Spanish
United States Spain
Self, Mother, Father Self
Predominantly Caucasian Predominantly Caucasian Predominantly Caucasian Predominantly Caucasian Predominantly Caucasian Predominantly Caucasian
United States
Self
United States United States
Self, Mother, Father Self
United States
Self
United States
Self
Australia
Self, Mother, Father
Benedikt et al. (1998) Byely et al. (2000) Dixon et al. (1996) Fulkerson et al. (2002)
Gross and Nelson (2000) Kichler and Crowther (2001)
Levine, Smolak, and Hayden (1994)
MacBrayer, Smith, McCarthy, Demos, and Simmons (2001) McCabe and Ricciardelli (2001)
Moreno and Thelen (1993) Senra, Sánchez-Cao, Seoane, and Leung (2007) Shroff and Thompson (2006) Smolak et al. (1999) Stice, Ziemba, Margolis, and Flick (1996) Tantleff-Dunn, Thompson, and Dunn (1995) Thompson, Cattarin, Fowler, and Fisher (1995) Wertheim et al. (1999)
Perceived Sociocultural Influences on Body Image and Body Change Questionnaire Concern with Child’s Weight
M&F
College-age Grades 6–8 12–16 yrs
F
College-age
Family Concern about Appearance Scale (FCAS) Tripartite Influence Scale (Parent Influence Subscale) Parent Questionnaire
F
17–26 yrs
F
10–15 yrs
M&F
Grades 4–5
F
16–40 yrs
M&F
17–35 yrs
F
College-age (17–42 yrs) 14–17 yrs
Perceived Sociocultural Pressures Scale Feedback on Physical Appearance Scale (FCAS) Perception of Teasing Scale (POTS) Parent Encouragement of Daughter to Lose Weight
F
Abbreviations: M, male; F, female; M & F, combined male and female sample; yrs, years old.
for Thinness (DT) scale contains seven items that assess preoccupation with restrictive dieting, concern about dieting and fears about weight gain. The Bulimia (BU) scale contains seven items that assess the tendency to think about and engage in uncontrollable overeating and purging, and the Body Dissatisfaction (BD) subscale contains 11 items that assess body shape dissatisfaction. The responses are rated on a 6-point Likert scale, and continuous scoring was used for each of the subscales. The reliability and validity of each of the subscales has been well-established (Garner, 2004), and in the current sample, the Cronbach’s alphas for the Bulimia, Drive for Thinness, and Body Dissatisfaction subscales were .90, .82, and .90, respectively. Tripartite Influence Scale. The Tripartite Influence Scale is a 43-item self-report measure designed to assess common sources of influence of body image disturbance (Keery et al., 2004; Thompson, Heinberg, Altabe, & Tantleff-Dunn, 1999). The factor structure supports three subscales/categories of influence (i.e., Parent, Media, and Peer Influence), and this measure has been found to yield adequate reliability and validity (Keery et al., 2004). The items from the Parent Influence subscale were included in the item pool for the development of the PIQ. In addition, the Peer and Media Influence subscales were used to assess the incremental validity of the PIQ. In the current sample, the Cronbach’s alphas for the Media and Peer Influence subscales were .86 and .94, respectively.
Procedure After receiving institutional ethics approval, survey packets were administered to participants at the completion of psychology classes. Students completed the packets in exchange for extra credit, and alternative extra credit options were available for students who chose not to participate. Participants first reviewed and signed informed consent forms. Following completion of the consent forms, participants then completed the survey packets, were debriefed, and thanked for their participation. Although response rate was not formally assessed, the response rate was estimated to be approximately 85%. Data Analysis Strategies The first step was a principal components analysis of the PIQ items as the authors did not have sufficient evidence to make a priori predictions about the factor structure of the PIQ. Following the establishment of the factor structure, the association between the obtained subscale(s) and body image (as measured by the Body Dissatisfaction subscale of the EDI-3) and eating disturbance (as measured by the DT and BU subscales of the EDI-3) was conducted. In order to examine the incremental validity of the PIQ, hierarchical regressions were planned, in which the Peer Influence and Media
96
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
Influence subscales of the Tripartite Influence Scale, along with BMI, were entered as predictor variables in the first step of the regression equation, and the PIQ subscales were entered in the second step. It was anticipated that PIQ would predict variance in body image disturbance, drive for thinness, and bulimic symptomatology after controlling for media influence, peer influence, and BMI. In addition, an examination of the association between BMI categories and levels of parental influence, as measured by the PIQ, was planned. Significance for all statistical tests was set at .05 unless otherwise noted. Missing data were managed by deleting cases listwise. Results Item Selection Items were simultaneously reviewed by the first and second authors for clarity, redundancy, and relevance to the construct of parental influence. Further, inter-item correlations were considered in the assessment of item redundancy. In particular, items with particularly large inter-item correlations were removed (i.e., greater than or equal to .80). Based on these procedures, the initial item pool was reduced to 33 items. Factor Structure of the PIQ The PIQ was subjected to a principal components analysis with an axis analysis extraction method. An obtained value of .90 on the Kaiser–Meyer–Olkin measure of sampling accuracy suggested that the correlation matrix was adequate for a principal components analysis. In the decision to determine the number of factors (i.e., components) to retain, a scree plot was examined, and an item analysis was conducted for various factor solutions. The scree plot suggested a 2 or 3 factor solution. Consequently, 2 and 3 factor models were forced and subjected to an item analysis. The 2 factor model was retained, as this model provided the most clear and parsimonious item loadings. In particular, based on examination of the 3 factor model, it was apparent that many of the significant item loadings on the third factor also loaded on the first or second factors. The two retained factors were rotated utilizing varimax factor rotation. Items with factor loadings of .4 or greater on only one factor were included in the final version of the measure, and based on this criteria, 28 items were retained. The first and most robust factor, labeled Direct Influence, included 18 items. The eigenvalue for Direct Influence was 9.78, and this factor accounted for 30.55% of item variance. Items that load on Direct Influence appear to assess direct parental communication of weight and/or appearance, and include items described in previous studies as verbal communication and verbal influence. More specifically, this scale includes parental concern about the participant’s weight/appearance, including encouragement to lose weight, negative comments about weight, and appearance-related criticism. The second factor, labeled Modeling, included 10 items that measure parental modeling of body image and weight dissatisfaction and items described in previous studies as indirect influence (Vincent & McCabe, 2000). The eigenvalue for Modeling was 4.32, and this factor accounted for 13.51% of item variance. Items that loaded on Modeling focused on parental modeling of the importance of thinness and parental dieting, and parents’ interactions with each other regarding the importance of weight, shape, and dieting. Cronbach’s alphas for the Direct Influence and Modeling subscales were .91 and .84, respectively, and the factor loadings are provided in Table 2. Finally, the eight items with the most robust loadings on each factor were selected to create parsimonious subscales to assess
Direct Influence (items 4, 12, 14, 18, 19, 23, 26, 32) and Modeling (items 55, 57, 60, 61, 62, 65, 66, 72). Cronbach’s alphas for Direct Influence and Modeling were .93 and .89, respectively. The eightitem version of the Direct Influence and Modeling subscales were used in subsequent analyses. Descriptive Statistics and Correlations A correlation matrix of the study variables is provided in Table 3. The Direct Influence subscale of the PIQ was significantly correlated with all other study variables. The Modeling subscale of the PIQ followed the same pattern; except that it was not significantly correlated with BMI. Although most items were intercorrelated, the correlations were not of the magnitude to raise significant concerns about multicollinearity in subsequent regression analyses (i.e., most correlations were well below .80). Parental Influence and Body Dissatisfaction It was anticipated that the PIQ would be significantly associated with body shape dissatisfaction. To test this hypothesis, a hierarchical regression analysis was conducted to determine the degree to which scores on the PIQ were associated with the BD subscale of the EDI-3, after controlling for Media Influence, Peer Influence, and BMI. Media and Peer Influence subscales of the Tripartite Influence Scale and BMI were entered into the first step of the regression equation, and the Direct Influence and Modeling subscales of the PIQ were entered into the second step. The first step of the model was significant and the addition of the Modeling and Direct Influence subscales led to a significant improvement in the model, F(2, 348) = 4.33, R2 = .016, p < .05. Regarding predictor variables in the final model, Media Influence (ˇ = .26, p < .001), BMI (ˇ = .38, p < .001), Peer Influence (ˇ = .12, p < .05), and Modeling (ˇ = .12, p < .01) were significantly associated with EDI-3 BD scores. Direct Influence was not significantly associated with BD scores. Parental Influence and Drive for Thinness It was also anticipated that the subscales of the PIQ would be significantly associated with drive for thinness as measured by the EDI-DT after controlling for peer influence, media influence, and BMI. Similar to the above analysis, a hierarchical regression analysis was conducted. The first step of the model was significant, and the addition of the PIQ subscales led to a significant improvement in the model, F(2, 348) = 14.58, R2 = .047, p < .001, indicating that the PIQ subscales account for variance in EDI-DT after controlling for Media Influence, Peer Influence, and BMI. Further, based on the final model, the following subscales were found to be significantly associated with EDI-DT: Media Influence (ˇ = .47, p < .001), Direct Influence (ˇ = .15, p < .01), Modeling (ˇ = .15, p < .01), and BMI (ˇ = .092, p < .05). Peer influence was not significantly associated with EDI-DT scores. Parental Influence and Bulimia Symptoms Regarding the relation between parent influence and bulimic symptoms as measured by the EDI-BU, a third hierarchical regression analysis was conducted, again with media influence, peer influence, and BMI as control variables. The first step of the model was significant, with the addition of the PIQ subscales leading to a significant improvement in the model, F(2, 348) = 12.84, R2 = .049, p < .001. This indicates that the PIQ subscales account for variance in EDI-BU after controlling for Media Influence, Peer Influence, and BMI. Further, based on the final model, all subscales were found to be significantly associated with EDI-BU scores: Media Influence (ˇ = .31, p < .001), Direct Influence (ˇ = .19, p < .001),
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
97
Table 2 Factor loadings and source of parent influence questionnaire items. Item
Factor 1 (parental influence)
1. I perceived a strong message from my parents to have a slender figure. 2. My parents wanted me to be thinner. 3. My parents kept me from eating foods that I liked in order to lose weight or keep from gaining weight. 4. My parents watched closely what I ate. 5. My parents asked me how much I weigh. 6. My parents told me that I looked like I gained weight. 7. My parents encouraged me to lose weight. 8. My parents said, “If you do not lose weight, you will never get a date, get a boyfriend/girlfriend, get married, etc.” 9. My parents made negative comments about my physical appearance. 10. My parents said critical things to me about my eating. 11. My parents told me to eat different foods in order to lose weight or keep from gaining weight. 12. I received negative feedback about the size or shape of my body. 13. I received negative feedback from my parents about my eating patterns to change my body size or shape. 14. My parents would say to me, “you do not need to lose weight.”a 15. My parents would say to me, “your health is what is important not your weight.”a 16. My parents would say to me, “what you weigh or how you look like is not what is important.”a 17. My parents would say to me, “you need to make sure you eat enough while you are growing.”a 18. My parents teased me about my appearance. 19. My parents encouraged me to gain weight.a 20. My parents teased me because I was too thin.a 21. My parents commented on each other’s weight. 22. My parent would say to my other parent, “you look heavy.” 23. My parents encouraged each other to lose weight. 24. My parents talked about dieting. 25. My parents complained about their weight. 26. My parents would ask, “Am I gaining weight?” 27. My parents would ask, “Am I as fat as him/her?” 28. My parents worried about their weight. 29. Physical appearance (shape, weight, clothing) was important to my parents. 30. My parent’s weight and shape influenced how they felt about themselves. 31. My parents were satisfied with their body weight.a 32. My parents tried to become more muscular.
Factor 2 (modeling)
Source
.61
.32
3
.85 .70
.13 .18
1, 2, 4 5
.42 .57 .75 .83 .74
.25 .28 .18 .12 .06
6 6, 7 7 7, 9, 10 7
.71
.17
10
.79 .76
.24 .23
11 5
.77
.23
12
.63
.31
12
.45
−.15
7
.53
−.08
7
.54
−.04
7
.54
−.09
7
.50 −.33 −.18 .12 .09
.27 .22 .28 .67 .37
1 7 1 1 7
.08 .17 .04 .13 .14 .04 .21
.64 .72 .80 .74 .66 .80 .68
1 1 7 7 7 12 8
.16
.71
11
.05 .03
.38 .41
12 7
Note. Source (original author of item): 1 = Shroff and Thompson (2006); 2 = Levine et al. (1994); 3 = Stice et al. (1996); 4 = Benedikt et al. (1998); 5 = Moreno and Thelen (1993); 6 = Tantleff-Dunn et al. (1995); 7 = Smolak et al. (1999); 8 = Senra et al. (2007); 9 = Wertheim et al. (1999); 10 = Kichler and Crowther (2001); 11 = Baker et al. (2000); 12 = McCabe and Ricciardelli (2001). Principal axis extraction method was used with a Quartimax factor rotation. Factor loading of .4 or greater on one factor were significant. Items 33–37 were reverse-scored. Items 19, 20, 22, and 31 were dropped from subsequent analyses due to nonsignificant factor loadings. a Reverse-scored items.
Table 3 Correlation matrix of study variables. Variable
PIQ-Direct Influence
PIQ-Modeling
EDI-DT
EDI-BU
TI-Media
PIQ-Direct Influence PIQ-Modeling EDI-DT EDI-BU TI-Media TI-Peers BMI EDI-BD
– .34** .39** .41** .27** .31** .41** .34**
– .37** .32** .32** .28** -.01 .24**
– .59** .59** .34** .19** .65**
– .47** .35** .24** .49**
– .40** .07 .39**
TI-Peer
BMI
EDI-BD
– .42**
–
– .09 .29**
Note. PIQ-Direct Influence: Parent Influence Questionnaire–Direct Influence Subscale; PIQ-Modeling: Parent Influence Questionnaire–Modeling Subscale; EDI-DT: Eating Disorder Inventory–Drive for Thinness subscale; EDI-BU: Eating Disorder Inventory–Bulimia Subscale; TI-Peer: Tripartite Influence Scale–Peer Influence Subscale; TI-Media: Tripartite Influence Scale–Media Influence Subscale; BMI: Body Mass Index; EDI-BD: Eating Disorder Inventory–Body Dissatisfaction subscale. ** p < .01.
98
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
BMI (ˇ = .14, p < .01), Peer Influence (ˇ = .13, p < .01), and Modeling (ˇ = .12, p < .05). Parental Influence and BMI Independent samples t-tests were conducted to determine if participants with BMIs in the overweight to obese range reported greater levels of parental influence and modeling when compared to participants with BMIs in the normal range. Underweight participants (with BMIs of less than 18.5) were excluded from the analysis because few participants fell into this category (6.3%). A significant difference between normal and overweight/obese participants was found for Direct Influence, t(334) = 6.28, p < .001, 2 = .015, with overweight/obese participants (M = 18.59, SD = 9.38) reporting higher levels of Direct Influence when compared to normal weight participants (M = 12.88, SD = 6.59). In contrast, the normal weight participants (M = 22.37, SD = 8.70) did not differ significantly from the obese/overweight participants (M = 20.99, SD = 8.59) on levels of Modeling, t(335) = 1.31, p = .19, 2 = .0004. Overall, these results indicate that obese/overweight participants’ recall more direct influence, but no difference based on BMI category was found for Modeling. Discussion The primary purpose of the current study was to conduct an integrative analysis of previously developed measures of parental influence to: (1) clarify the underlying dimensions of parental influence, and (2) develop a comprehensive composite measure that can be utilized consistently in future research to clarify the construct of parental influence. Based on analyses, two clear dimensions emerged, direct influence and modeling. These dimensions supplement the factor analysis conducted by Benedikt et al. (1998) and are consistent with a number of previous studies, which have proposed that direct influence and modeling are distinct dimensions of parental influence (Annus, Smith, Fischer, Hendricks, & Williams, 2007; Baker, Whisman, & Brownell, 2000; Smolak et al., 1999). The direct influence dimension consisted of direct parental communication, including feedback/comments about weight, eating, and exercise. More specifically, this dimension includes parental concern about the participant’s weight, parental messages about the importance of being thin and losing weight, parental pressure for the participant to diet or restrict food intake, parental encouragement of the participant to exercise to lose weight, parental criticism or negative comments about the participant’s weight (along with appearance and eating behaviors), and parental catastrophizing about the participant’s weight. The parental modeling dimension included participant recall of parental dieting and other strategies to manage weight, parental weight dissatisfaction, parental self-emphasis on weight, and parental appearance comments to his or her spouse. The direct influence and modeling dimensions were found to be associated with measures of eating disturbance, including drive for thinness and bulimic symptomatology. It is noteworthy that these associations were significant even after controlling for media and peer influence, suggesting that parental influence may be a distinct pathway in the development of eating disturbance. The association between parental influence and body shape dissatisfaction was less consistent. In particular, after controlling for BMI, modeling was found to be associated with scores on the BD subscale of the EDI-3; however, direct influence was not found to be associated with the BD subscale. These findings indicate that parental modeling, rather than direct parental influence, may impact levels of body shape dissatisfaction. However, in the regression equation, shared variance between direct influence and BMI
may have prevented direct influence from being a significant predictor of body shape dissatisfaction. Based on this observation, a direction for future research is to examine the degree to which direct influence and BMI may interact to influence levels of body shape dissatisfaction. Overall, it appears that direct influence has more of a negative impact on participants’ bulimic symptomatology relative to parental modeling. As an interpretation of this finding, it is reasonable that participants report being more influenced by direct negative messages when compared to more subtle messages that may not be directed towards the participant (i.e., modeling). It is also noteworthy that the weaker association between modeling and measures of eating disturbance is generally consistent with previous literature, as studies that have examined the effect of both direct influence and modeling typically find stronger and more consistent support for direct influence (Baker et al., 2000; Fulkerson et al., 2002; Moreno & Thelen, 1993; Rodgers & Chabrol, 2009; Vincent & McCabe, 2000). Interestingly, consistent with research by Fulkerson et al. (2002) and Thelen and Cormier (1995), direct influence was found to be related to BMI, as overweight/obese participants reported receiving greater levels of direct feedback from parents; however, parental modeling was not found to be associated with BMI. It is noteworthy that examination of BMI as a continuous variable yielded the same results, as BMI was found to be significantly correlated with direct influence but not modeling (see Table 3). Although the direction of the association between these variables is uncertain, the findings suggest that parents of overweight/obese participants may be more likely to engage in direct feedback as an attempt to manage their child’s weight. Another hypothesis is that overweight mothers are more likely to encourage their daughters to diet (Fulkerson et al., 2002). This can send a conflicting and confusing message to daughters. In contrast, modeling behavior, which includes attempts by parents to manage their own weight, is not a strategy to control or manage the weight of overweight/obese participants. However, if the parents of the overweight participants are overweight themselves; they may be modeling unhealthy overeating habits or poor food choices. More research in this area is needed. Despite the strengths of this study, there are a number of noteworthy limitations and directions for future research. First, the findings may not generalize to populations that are either not represented or underrepresented in the current study. In particular, the participants were college females in their late teens/early twenties. The study’s results therefore may not be generalizable to children, young adolescents, males, or older adults. In addition, the current study, along with the items derived from previous studies, have been based on predominantly Caucasian samples. Consequently, the findings from this study may not generalize to other ethnic groups. Furthermore, a majority of participants were of normal weight. As a result, the factor structure of the PIQ may not generalize to those who are either underweight or overweight. In response to these limitations, a direction for future research would be to examine the consistency of the factor structure of the PIQ for those who do not fall in the normal weight range, in males, and also over a variety of age ranges. In addition, it is noteworthy that the current results were based on a nonclinical sample. Consequently, it is recommended that the psychometric properties of the PIQ are examined in clinical samples. Finally, the PIQ and the BD subscale are limited in that both measures focus only on weight and shape and do not focus on other dimensions of body image disturbance. Consequently, a recommended direction for future research is to examine the role of parental influence on other dimensions of body image, including general appearance and muscularity. An additional limitation relates to the self-report and retrospective nature of this study. Specifically, participants with body
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
image disturbance, a drive for thinness or bulimic symptomatology may have been biased in their recall of parental comments and behavior, and it is possible that participants with body shape and eating disturbance may have overestimated the association between parental behaviors and eating pathology (Jansen, Smeets, Boon, Nedrekoorn, Roefs, & Mulkens, 2007). An additional limitation is that current BMI was assessed, not the retrospective BMI at the time of parental influence. As a result of the above limitations, additional research is recommended, including studies that examine the consistency between parent and child-report. Finally, in addition to the above-mentioned directions for future research, it is recommended that follow-up psychometric studies are conducted with the PIQ to confirm the factor structure and examine test–retest reliability of this measure. Another limitation and potential direction for future research is that participants were instructed to consider the influence of both parents when completing the questionnaire, instead of distinguishing between the mother and father. Providing the distinction between mother and father feedback may have yielded different results. For example, mothers and fathers may provide different types of feedback or the impact of the feedback may differ based on the gender of the parent. Consequently, comparison of mother’s and father’s influence is a possible direction for future research. A further limitation and direction for future research relates to the fact that the authors used a somewhat narrow operational definition of body image, which was measured by body shape dissatisfaction. Consequently, it may beneficial for future studies to consider other aspects of body image, including muscularity. In addition, some items did not load on either factor, and appeared to be related to pressure to gain weight. In order to examine this category more thoroughly, a direction for future research would be to examine these items in a sample of underweight individuals, possibly including individuals with anorexia nervosa. A final limitation and future direction would be to include moderators, which was beyond the scope of the current study. For example, modeling may be moderated by direct parental influence. Kichler and Crowther (2001) found that under high levels of negative, direct parental influence, modeling was significantly associated with body image; however, no association was found between parental modeling and body image under low levels of negative parental influence. Another example is that modeling has been found to be significant when more extreme and restrictive weight loss behaviors are utilized by parents, such as fasting, skipping meals, and crash dieting versus the child observing more general dietary restraint (Benedikt et al., 1998; Wertheim et al., 1999). Taking moderators into account may further elucidate and clarify the pathway from parental influence to body image and eating disturbance. Overall, the current study provides a number of significant contributions to the understanding of parental influence as a risk factor in the development of a drive for thinness, body shape dissatisfaction, and bulimic symptomatology. For example, in addition to the development of a promising measure, this study provides insight into the dimensions or facets of parental influence. More specifically, this study is among the first to provide strong psychometric support for the premise that parental influence forms two distinct dimensions, direct influence and modeling. Further, because the items of the PIQ are derived from previously published measures of parental influence, this study also assists in the clarification of inconsistent terminology utilized to discuss various components of parental influence. For example, it appears that terms direct influence and parental criticism about appearance refer to overlapping constructs. This measure has the potential to serve as an impetus for future research, and the PIQ may serve as a much needed tool to assess the degree to which parents directly and indirectly communicate appearance-related messages to their children or
99
adolescents. Finally, we hope that additional research is conducted to confirm the factor structure of the PIQ and to refine this measure.
Acknowledgement This manuscript is based on the first author’s undergraduate honors thesis, and the support for the purchase of study-related measures was provided by University of North Florida Undergraduate Enrichment Program.
References Annus, A. M., Smith, G. T., Fischer, S., Hendricks, M. & Williams, S. F. (2007). Associations among family-of-origin food-related experiences, expectancies, and disordered eating. International Journal of Eating Disorders, 40, 179–186. Baker, C., Whisman, M. & Brownell, K. (2000). Studying intergenerational transmission of eating attitudes and behaviors: Methodological and conceptual questions. Health Psychology, 19, 376–381. Benedikt, R., Wertheim, E. & Love, A. (1998). Eating attitudes and weight-loss attempts in female adolescents and their mothers. Journal of Youth and Adolescence, 27, 43–57. Byely, L., Archibald, A., Graber, J. & Brooks-Gunn, J. (2000). A prospective study of familial and social influences on girls’ body image and dieting. International Journal of Eating Disorders, 28, 155–164. Clark, L. & Watson, D. (1995). Constructing validity: Basic issues in objective scale development. Psychological Assessment, 7, 309–319. Dixon, R., Adair, V. & O’Connor, S. (1996). Parental influences on the dieting beliefs and behaviors of adolescent females in New Zealand. Journal of Adolescent Health, 19, 303–307. Edlund, B., Halvarsson, K. & Sjödén, P. (1996). Eating behaviours and attitudes to eating, dieting and body image in 7-year-old Swedish girls. European Eating Disorders Review, 4, 40–53. Fulkerson, J., McGuire, M., Neumark-Sztainer, D., Story, M., French, S. & Perry, C. (2002). Weight-related attitudes and behaviors of adolescent boys and girls who are encouraged to diet by their mothers. International Journal of Obesity, 26, 1579–1587. Garner, D. M. (2004). Eating Disorders Inventory-3: Professional manual. United States: Psychological Assessment Resources. Gross, R. & Nelson, E. (2000). Perceptions of parental messages regarding eating and weight and their impact on disordered eating. Journal of College Student Psychotherapy, 15, 57–78. Hanna, A. & Bond, M. (2006). Relationships between family conflict, perceived maternal verbal messages, and daughters’ disturbed eating symptomatology. Appetite, 47, 205–211. Hill, A. & Franklin, J. (1998). Mothers, daughters and dieting: Investigating the transmission of weight control. The British Journal of Clinical Psychology, 37, 3–13. Jansen, A., Smeets, T., Boon, B., Nedrekoorn, C., Roefs, A. & Mulkens, S. (2007). Vulnerability to interpretation bias in overweight children. Psychology and Health, 22, 561–574. Kanakis, D. & Thelen, M. (1995). Parental variables associated with bulimia nervosa. Addictive Behaviors, 20, 491–500. Keery, H., van den Berg, P. & Thompson, J. (2004). An evaluation of the Tripartite Influence Model of body dissatisfaction and eating disturbance with adolescent girls. Body Image, 1, 237–251. Kichler, J. & Crowther, J. (2001). The effects of maternal modeling and negative familial communication on women’s eating attitudes and body image. Behavior Therapy, 32, 443–457. 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. Stephens, J. H. Crowther, D. L. Tennenbaum & M. Stephens (Eds.), The etiology of bulimia nervosa: The individual and familial context (pp. 59–80). Washington, DC, USA: Hemisphere Publishing Corp. Levine, M., Smolak, L. & Hayden, H. (1994). The relation of sociocultural factors to eating attitudes and behaviors among middle school girls. The Journal of Early Adolescence, 14, 471–490. MacBrayer, E., Smith, G., McCarthy, D., Demos, S. & Simmons, J. (2001). The role of family of origin food-related experiences in bulimic symptomatology. International Journal of Eating Disorders, 30, 149–160. McCabe, M. & Ricciardelli, L. (2001). The structure of the Perceived Sociocultural Influences on Body Image and Body Change Questionnaire. International Journal of Behavioral Medicine, 8, 19–41. McCabe, M. & Ricciardelli, L. (2003). Sociocultural influences on body image and body changes among adolescent boys and girls. The Journal of Social Psychology, 143, 5–26. Moreno, A. & Thelen, M. (1993). Parental factors related to bulimia nervosa. Addictive Behaviors, 18, 681–689. Ogden, J. & Steward, J. (2000). The role of the mother–daughter relationship in explaining weight concern. International Journal of Eating Disorders, 28, 78–83. Pike, K. & Rodin, J. (1991). Mothers, daughters, and disordered eating. Journal of Abnormal Psychology, 100, 198–204.
100
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
Rodgers, R. & Chabrol, H. (2009). Parental attitudes, body image disturbance and disordered eating amongst adolescents and young adults: A review. European Eating Disorders Review, 17, 137–151. Ruther, N. & Richman, C. (1993). The relationship between mothers’ eating restraint and their children’s attitudes and behaviors. Bulletin of the Psychonomic Society, 31, 217–220. Senra, C., Sánchez-Cao, E., Seoane, G. & Leung, F. (2007). Evolution of self-concept deficits in patients with eating disorders: The role of family concern about weight and appearance. European Eating Disorders Review, 15, 131–138. Shroff, H. & Thompson, J. (2006). The tripartite influence model of body image and eating disturbance: A replication with adolescent girls. Body Image, 3, 17–23. Smolak, L., Levine, M. & Schermer, F. (1999). Parental input and weight concerns among elementary school children. International Journal of Eating Disorders, 25, 263–271. Stice, E. (1998). Modeling of eating pathology and social reinforcement of the thinideal predict onset of bulimic symptoms. Behaviour Research and Therapy, 36, 931–944. Stice, E., Ziemba, C., Margolis, J. & Flick, P. (1996). The dual pathway model differentiates bulimics, subclinical bulimics, and controls: Testing the continuity hypothesis. Behavior Therapy, 27, 531–549. Tantleff-Dunn, S., Thompson, J. & Dunn, M. (1995). The Feedback on Physical Appearance Scale (FOPAS): Questionnaire development and psychometric evaluation. Eating Disorders: The Journal of Treatment & Prevention, 3, 332–341.
Thelen, M. & Cormier, J. (1995). Desire to be thinner and weight control among children and their parents. Behavior Therapy, 26, 85–99. 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, 146–157. Thompson, J., Heinberg, L., Altabe, M. & Tantleff-Dunn, S. (1999). Exacting beauty: Theory, assessment, and treatment of body image disturbance. Washington, DC, USA: American Psychological Association. van den Berg, P., Thompson, J., Obremski-Brandon, K. & Coovert, M. (2002). The tripartite influence model of body image and eating disturbance: A covariance structure modeling investigation testing the meditational role of appearance comparison. Journal of Psychosomatic Research, 53, 1007–1020. Vincent, M. & McCabe, M. (2000). Gender differences among adolescents in family, and peer influences on body dissatisfaction, weight loss, and binge eating behaviors. Journal of Youth and Adolescence, 29, 205–221. Wertheim, E., Mee, V. & Paxton, S. (1999). Relationships among adolescent girls’ eating behaviors and their parents’ weight-related attitudes and behaviors. Sex Roles, 41, 169–187. Wood, K., Becker, J. & Thompson, J. (1996). Body image dissatisfaction in preadolescent children. Journal of Applied Developmental Psychology, 17, 85–100. Young, E., Clopton, J. & Bleckley, M. (2004). Perfectionism, low self-esteem, and family factors as predictors of bulimic behavior. Eating Behaviors, 5, 273–283.
Contents lists available at SciVerse ScienceDirect
Body Image journal homepage: www.elsevier.com/locate/bodyimage
The relation between parental influence, body image, and eating behaviors in a nonclinical female sample Michelle Abraczinskas a , Brian Fisak Jr. b,∗ , Rachel D. Barnes c a
University of North Carolina at Charlotte, Charlotte, NC, USA University of North Florida, Jacksonville, FL, USA c Yale University School of Medicine, New Haven, CT, USA b
a r t i c l e
i n f o
a b s t r a c t
Article history: Received 10 September 2010 Received in revised form 14 October 2011 Accepted 19 October 2011 Keywords: Body image Parental influence Measurement Modeling Eating disorders
The purpose of the current study is to create a comprehensive composite measure of parental influence based on previously developed measures to clarify the underlying dimensions of parental influence and to determine the degree to which parental influence relates to body image and dysfunctional weight concerns. Previously published literature was reviewed for measures of parental influence, and items from 22 measures were condensed and combined into a single questionnaire, which was completed by 367 female undergraduate psychology students. Two dimensions emerged from a principle components analysis: Direct Influence, which includes weight and eating related comments, and Modeling, which includes parental modeling of dieting and related behavior. Direct Influence and Modeling were significantly related to eating disturbance, such as drive for thinness and bulimic symptomatology. Overall, the results integrate the previous literature and clarify the underlying dimensions of parental influence. Further, this study provides directions for future research related to the development and maintenance of body image and eating disturbance. © 2011 Elsevier Ltd. All rights reserved.
Introduction Several studies have focused on the influence of social factors in the development and maintenance of body image and eating disturbance, and parents are among the most commonly examined sources of influence (Hill & Franklin, 1998; Moreno & Thelen, 1993; Pike & Rodin, 1991; Smolak, Levine, & Schermer, 1999; Thelen & Cormier, 1995). One reason for the focus on parental influence is that parents typically are the first sources of socialization (McCabe & Ricciardelli, 2003). Further, based on the Tripartite Influence Model, parental influence, along with peer and media influence, has been described as a distinct factor in the development of body image and eating disturbance (Keery, van den Berg, & Thompson, 2004; Shroff & Thompson, 2006; van den Berg, Thompson, Obremski-Brandon, & Coovert, 2002). Although considerable variability exists in the operationalization of parental influence in the development of maladaptive eating behaviors and body image, researchers often discuss parental influence as two distinct categories. The first category is direct influence, also referred to as verbal communication or verbal influence (Fulkerson, McGuire, Neumark-Sztainer, Story, French, & Perry, 2002; Smolak et al., 1999; Vincent & McCabe, 2000; Wertheim, Mee,
∗ Corresponding author. E-mail address: b.fi[email protected] (B. Fisak Jr.). 1740-1445/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.bodyim.2011.10.005
& Paxton, 1999). The second category is modeling, also referred to as indirect influence (Ogden & Steward, 2000; Pike & Rodin, 1991; Smolak et al., 1999; Stice, 1998; Vincent & McCabe, 2000; Wertheim et al., 1999). Direct influence typically includes parental behaviors such as discussion and encouragement of child dieting or other attempts to restrict the child’s weight (Fulkerson et al., 2002; Smolak et al., 1999; Vincent & McCabe, 2000; Wertheim et al., 1999). Modeling includes behaviors such as parental dieting, parental expression of weight or body dissatisfaction, and other observable behaviors on the part of the parent to maintain or reduce weight (Ogden & Steward, 2000; Pike & Rodin, 1991; Smolak et al., 1999; Stice, 1998; Vincent & McCabe, 2000; Wertheim et al., 1999). Research has found consistent support for direct influence as a predictor of body dissatisfaction and weight-loss behaviors (Smolak et al., 1999; Wertheim et al., 1999; Young, Clopton, & Bleckley, 2004). For example, Vincent and McCabe (2000) found parental discussion about weight loss predicted disordered eating behavior in a sample of nonclinical adolescent girls. In addition, parental encouragement of daughters to lose weight or diet has been found to be associated with bulimia symptoms, including vomiting and the use of laxatives (Benedikt, Wertheim, & Love, 1998; Dixon, Adair, & O’Connor, 1996; Hanna & Bond, 2006; Moreno & Thelen, 1993; Pike & Rodin, 1991). In addition to direct influence, parental modeling is predictive of body dissatisfaction and maladaptive eating behaviors in children (Edlund, Halvarsson, & Sjödén, 1996; Pike & Rodin, 1991; Rodgers
94
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
& Chabrol, 2009; Smolak et al., 1999; Stice, 1998; Wertheim et al., 1999). Specifically, by engaging in weight loss behaviors, or through the expression of their own body image dissatisfaction, parents may indirectly transmit messages that reinforce the importance of the thin ideal (Wertheim et al., 1999). For example, Stice (1998) found that modeling of family abnormal eating behavior was associated with bulimic symptoms and predicted the onset of bingeing and purging in daughters. Pike and Rodin (1991) found that mothers’ dieting behavior and weight concerns were associated with weight concerns in their daughters. Although the studies cited above found support for modeling, it is noteworthy that other studies failed to find an association between modeling and body image or eating disturbance (Byely, Archibald, Graber, & Brooks-Gunn, 2000; Dixon et al., 1996; Fulkerson et al., 2002; Kanakis & Thelen, 1995; Ruther & Richman, 1993). There are a number of potential explanations for these mixed findings. In particular, Fulkerson et al. (2002) suggested that body mass may play a role in parental influence. Children and adolescents with a higher BMI are more likely to receive direct feedback from parents about weight loss (Thelen & Cormier, 1995). Furthermore, studies that adjust for BMI are less likely to find an association between parental modeling and body image or eating disturbance (Fulkerson et al., 2002). In addition, modeling may serve a particularly important role in early adolescence, since that is a time when social comparison begins to emerge as well as an increase in weight and shape consciousness in girls (Levine & Smolak, 1992). Modeling may become less important as women age and their values are based more on internalized beliefs; therefore, developmental stage may be an additional factor that impacts the transmission of body image concerns and eating disturbance. It is also possible that methodological differences may account for the inconsistent support for modeling. In particular, studies utilizing clinical level samples more often find an association between modeling and eating disturbance (Pike & Rodin, 1991); however, the findings in this area have been inconsistent (Hill & Franklin, 1998; Kanakis & Thelen, 1995). It is also noteworthy that, although several measures of parental influence have been developed, research examining the psychometric properties of these measures has been limited (Wood, Becker, & Thompson, 1996). In particular, few studies have examined the factor structure of measures of parental influence (Benedikt et al., 1998). A related limitation is that there appears to be a general lack of consistency regarding the operationalization of the dimensions or components of parental influence. More specifically, several researchers have conceptualized direct communication and modeling as two distinct categories (Fulkerson et al., 2002; Vincent & McCabe, 2000); however, other researchers have treated parental influence as a single construct (Hill & Franklin, 1998; Pike & Rodin, 1991). In addition to the above limitations, there appear to be inconsistencies in the terminology or labels used to describe similar, and perhaps overlapping, aspects of parental influence. For example, Fulkerson et al. (2002) used the term “direct influence” to describe parental encouragement of dieting and weight-related comments, while Kichler and Crowther (2001) used “negative familial influence.” Overall, researchers concluded that the construct of parental influence needs to be clarified, as the full scope of the construct may not have been addressed in previous studies (Byely et al., 2000; Dixon et al., 1996; Fulkerson et al., 2002; Kanakis & Thelen, 1995; Ruther & Richman, 1993; Shroff & Thompson, 2006). In summary, although several studies have examined the role of parental influence in the development of body image and eating disturbance, there are a number of limitations to the previous literature. These limitations include inconsistent labeling and classification dimensions, and few studies have examined the psychometric properties of these measures, including the factor structure. In response to these limitations, the primary purpose
of the current study was to conduct an integrative analysis of previously developed measures of influence to: (1) elucidate the underlying dimensions of parental influence, and (2) develop a comprehensive and psychometrically sound measure of parental influence on the development of body image and eating disturbance. It was hypothesized that parental influence will be a multi-dimensional construct, which includes a direct influence and a modeling dimension, along with the potential for more. It was also hypothesized that each of the dimensions of parental influence will exhibit concurrent validity, as these dimensions are anticipated to be associated with body image and eating disturbance. Method Participants The sample consisted of 367 female undergraduate students at a large university in the southeast United States. The majority of the participants were White/Caucasian (75.7%). The average participant age was 22.69 years (SD = 5.61). Based on self-report of height and weight, the BMI for a majority of participants (67.6%) was within the normal range, which was operationalized as 18.5–24.9. Further, 26.0% of the participants fell within the overweight/obese range, 25.0 or greater. Finally, 6.3% fell within the underweight range, a BMI of 18.5 or lower. Development of the Parental Influence Questionnaire The Parental Influence Questionnaire (PIQ) is a retrospective measure that was developed for the purpose of this study. In development of this measure, efforts were made to obtain a broad and comprehensive pool of potential items (Clark & Watson, 1995). First, an extensive literature review was conducted using the Psycinfo database, in which all measures of parental influence related to body image and eating disturbance through July 2010 were identified and reviewed. These measures were located using combinations of the following search terms: Body Image, Eating disturbance, Parent influence, Direct influence, Indirect influence, Modeling, Children, Adolescents, and Dieting. Measures were included if the article in which it was found was peerreviewed, written in English, and was related to the influence parents have on the body image or eating disturbance of their own children. Based on this review, the initial item pool was created by extracting all potentially relevant items from measures referenced in published studies. When items were not available from published manuscripts, the authors of the manuscripts were contacted for the specific items. All contacted authors responded to requests. Based on this review, 118 items from 22 measures were included in the initial item pool (see Table 1). When necessary, items were modified to fit a 5-point Likert scale. This format is consistent with a majority of previously published measures of parental influence and the following response options were designated: Strongly Disagree, Somewhat Agree, Neither Agree nor Disagree, Somewhat Agree and Strongly Agree. Further, items were modified where necessary to fit the retrospective nature of this study, and when asked to rate the items, participants were asked to rate their interactions with their parents when a child and adolescent. Validation Measures Eating Disorder Inventory-3. The Eating Disorder Inventory (EDI-3; Garner, 2004) is a self-report inventory for the identification of symptoms of eating pathology. The Drive for Thinness, Bulimia, and Body Dissatisfaction subscales were utilized. The Drive
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
95
Table 1 Summary of measures and populations used to develop parental influence questionnaire. Authors
Measure
Gender
Age range
Ethnicity
Nationality
Source
Baker et al. (2000)
Attitude and Behavior Scale (ABS), Criticism about Eating & Appearance Maternal Encouragement of Daughter to Lose Weight Dieting History and Encouragement Single item measuring parental dieting Mother’s Perceptions of Adolescents Encouragement to Diet Parental Eating and Weight Measures Survey (PEWM) Perceived Frequency of Negative Comments about Appearance Family Teasing & Criticism, Parental Mother Investment in Daughter’s Shape, Mother’s Investment in her Own Thinness Family History Inventory
M&F
College-age
Predominately Caucasian
United States
Self, Mother, Father
F
15–16 yrs
Australia
Self, Mother
F
10–14 yrs
Caucasian Asian-Pacific Caucasian
United States
Self, Mother
F
13–16 yrs
Caucasian
New Zealand
Self
M&F
Middle/High School
Diverse
United States
Self, Mother
F
College-age
United States
Self
F
College-age
Predominantly Caucasian Predominantly Caucasian
United States
Self, Mother
F
10–14 yrs
Predominantly Caucasian
United States
Self
F
Predominantly Caucasian Mostly Anglo-Saxon
United States
Self
Australian
Self
Predominantly Caucasian Spanish
United States Spain
Self, Mother, Father Self
Predominantly Caucasian Predominantly Caucasian Predominantly Caucasian Predominantly Caucasian Predominantly Caucasian Predominantly Caucasian
United States
Self
United States United States
Self, Mother, Father Self
United States
Self
United States
Self
Australia
Self, Mother, Father
Benedikt et al. (1998) Byely et al. (2000) Dixon et al. (1996) Fulkerson et al. (2002)
Gross and Nelson (2000) Kichler and Crowther (2001)
Levine, Smolak, and Hayden (1994)
MacBrayer, Smith, McCarthy, Demos, and Simmons (2001) McCabe and Ricciardelli (2001)
Moreno and Thelen (1993) Senra, Sánchez-Cao, Seoane, and Leung (2007) Shroff and Thompson (2006) Smolak et al. (1999) Stice, Ziemba, Margolis, and Flick (1996) Tantleff-Dunn, Thompson, and Dunn (1995) Thompson, Cattarin, Fowler, and Fisher (1995) Wertheim et al. (1999)
Perceived Sociocultural Influences on Body Image and Body Change Questionnaire Concern with Child’s Weight
M&F
College-age Grades 6–8 12–16 yrs
F
College-age
Family Concern about Appearance Scale (FCAS) Tripartite Influence Scale (Parent Influence Subscale) Parent Questionnaire
F
17–26 yrs
F
10–15 yrs
M&F
Grades 4–5
F
16–40 yrs
M&F
17–35 yrs
F
College-age (17–42 yrs) 14–17 yrs
Perceived Sociocultural Pressures Scale Feedback on Physical Appearance Scale (FCAS) Perception of Teasing Scale (POTS) Parent Encouragement of Daughter to Lose Weight
F
Abbreviations: M, male; F, female; M & F, combined male and female sample; yrs, years old.
for Thinness (DT) scale contains seven items that assess preoccupation with restrictive dieting, concern about dieting and fears about weight gain. The Bulimia (BU) scale contains seven items that assess the tendency to think about and engage in uncontrollable overeating and purging, and the Body Dissatisfaction (BD) subscale contains 11 items that assess body shape dissatisfaction. The responses are rated on a 6-point Likert scale, and continuous scoring was used for each of the subscales. The reliability and validity of each of the subscales has been well-established (Garner, 2004), and in the current sample, the Cronbach’s alphas for the Bulimia, Drive for Thinness, and Body Dissatisfaction subscales were .90, .82, and .90, respectively. Tripartite Influence Scale. The Tripartite Influence Scale is a 43-item self-report measure designed to assess common sources of influence of body image disturbance (Keery et al., 2004; Thompson, Heinberg, Altabe, & Tantleff-Dunn, 1999). The factor structure supports three subscales/categories of influence (i.e., Parent, Media, and Peer Influence), and this measure has been found to yield adequate reliability and validity (Keery et al., 2004). The items from the Parent Influence subscale were included in the item pool for the development of the PIQ. In addition, the Peer and Media Influence subscales were used to assess the incremental validity of the PIQ. In the current sample, the Cronbach’s alphas for the Media and Peer Influence subscales were .86 and .94, respectively.
Procedure After receiving institutional ethics approval, survey packets were administered to participants at the completion of psychology classes. Students completed the packets in exchange for extra credit, and alternative extra credit options were available for students who chose not to participate. Participants first reviewed and signed informed consent forms. Following completion of the consent forms, participants then completed the survey packets, were debriefed, and thanked for their participation. Although response rate was not formally assessed, the response rate was estimated to be approximately 85%. Data Analysis Strategies The first step was a principal components analysis of the PIQ items as the authors did not have sufficient evidence to make a priori predictions about the factor structure of the PIQ. Following the establishment of the factor structure, the association between the obtained subscale(s) and body image (as measured by the Body Dissatisfaction subscale of the EDI-3) and eating disturbance (as measured by the DT and BU subscales of the EDI-3) was conducted. In order to examine the incremental validity of the PIQ, hierarchical regressions were planned, in which the Peer Influence and Media
96
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
Influence subscales of the Tripartite Influence Scale, along with BMI, were entered as predictor variables in the first step of the regression equation, and the PIQ subscales were entered in the second step. It was anticipated that PIQ would predict variance in body image disturbance, drive for thinness, and bulimic symptomatology after controlling for media influence, peer influence, and BMI. In addition, an examination of the association between BMI categories and levels of parental influence, as measured by the PIQ, was planned. Significance for all statistical tests was set at .05 unless otherwise noted. Missing data were managed by deleting cases listwise. Results Item Selection Items were simultaneously reviewed by the first and second authors for clarity, redundancy, and relevance to the construct of parental influence. Further, inter-item correlations were considered in the assessment of item redundancy. In particular, items with particularly large inter-item correlations were removed (i.e., greater than or equal to .80). Based on these procedures, the initial item pool was reduced to 33 items. Factor Structure of the PIQ The PIQ was subjected to a principal components analysis with an axis analysis extraction method. An obtained value of .90 on the Kaiser–Meyer–Olkin measure of sampling accuracy suggested that the correlation matrix was adequate for a principal components analysis. In the decision to determine the number of factors (i.e., components) to retain, a scree plot was examined, and an item analysis was conducted for various factor solutions. The scree plot suggested a 2 or 3 factor solution. Consequently, 2 and 3 factor models were forced and subjected to an item analysis. The 2 factor model was retained, as this model provided the most clear and parsimonious item loadings. In particular, based on examination of the 3 factor model, it was apparent that many of the significant item loadings on the third factor also loaded on the first or second factors. The two retained factors were rotated utilizing varimax factor rotation. Items with factor loadings of .4 or greater on only one factor were included in the final version of the measure, and based on this criteria, 28 items were retained. The first and most robust factor, labeled Direct Influence, included 18 items. The eigenvalue for Direct Influence was 9.78, and this factor accounted for 30.55% of item variance. Items that load on Direct Influence appear to assess direct parental communication of weight and/or appearance, and include items described in previous studies as verbal communication and verbal influence. More specifically, this scale includes parental concern about the participant’s weight/appearance, including encouragement to lose weight, negative comments about weight, and appearance-related criticism. The second factor, labeled Modeling, included 10 items that measure parental modeling of body image and weight dissatisfaction and items described in previous studies as indirect influence (Vincent & McCabe, 2000). The eigenvalue for Modeling was 4.32, and this factor accounted for 13.51% of item variance. Items that loaded on Modeling focused on parental modeling of the importance of thinness and parental dieting, and parents’ interactions with each other regarding the importance of weight, shape, and dieting. Cronbach’s alphas for the Direct Influence and Modeling subscales were .91 and .84, respectively, and the factor loadings are provided in Table 2. Finally, the eight items with the most robust loadings on each factor were selected to create parsimonious subscales to assess
Direct Influence (items 4, 12, 14, 18, 19, 23, 26, 32) and Modeling (items 55, 57, 60, 61, 62, 65, 66, 72). Cronbach’s alphas for Direct Influence and Modeling were .93 and .89, respectively. The eightitem version of the Direct Influence and Modeling subscales were used in subsequent analyses. Descriptive Statistics and Correlations A correlation matrix of the study variables is provided in Table 3. The Direct Influence subscale of the PIQ was significantly correlated with all other study variables. The Modeling subscale of the PIQ followed the same pattern; except that it was not significantly correlated with BMI. Although most items were intercorrelated, the correlations were not of the magnitude to raise significant concerns about multicollinearity in subsequent regression analyses (i.e., most correlations were well below .80). Parental Influence and Body Dissatisfaction It was anticipated that the PIQ would be significantly associated with body shape dissatisfaction. To test this hypothesis, a hierarchical regression analysis was conducted to determine the degree to which scores on the PIQ were associated with the BD subscale of the EDI-3, after controlling for Media Influence, Peer Influence, and BMI. Media and Peer Influence subscales of the Tripartite Influence Scale and BMI were entered into the first step of the regression equation, and the Direct Influence and Modeling subscales of the PIQ were entered into the second step. The first step of the model was significant and the addition of the Modeling and Direct Influence subscales led to a significant improvement in the model, F(2, 348) = 4.33, R2 = .016, p < .05. Regarding predictor variables in the final model, Media Influence (ˇ = .26, p < .001), BMI (ˇ = .38, p < .001), Peer Influence (ˇ = .12, p < .05), and Modeling (ˇ = .12, p < .01) were significantly associated with EDI-3 BD scores. Direct Influence was not significantly associated with BD scores. Parental Influence and Drive for Thinness It was also anticipated that the subscales of the PIQ would be significantly associated with drive for thinness as measured by the EDI-DT after controlling for peer influence, media influence, and BMI. Similar to the above analysis, a hierarchical regression analysis was conducted. The first step of the model was significant, and the addition of the PIQ subscales led to a significant improvement in the model, F(2, 348) = 14.58, R2 = .047, p < .001, indicating that the PIQ subscales account for variance in EDI-DT after controlling for Media Influence, Peer Influence, and BMI. Further, based on the final model, the following subscales were found to be significantly associated with EDI-DT: Media Influence (ˇ = .47, p < .001), Direct Influence (ˇ = .15, p < .01), Modeling (ˇ = .15, p < .01), and BMI (ˇ = .092, p < .05). Peer influence was not significantly associated with EDI-DT scores. Parental Influence and Bulimia Symptoms Regarding the relation between parent influence and bulimic symptoms as measured by the EDI-BU, a third hierarchical regression analysis was conducted, again with media influence, peer influence, and BMI as control variables. The first step of the model was significant, with the addition of the PIQ subscales leading to a significant improvement in the model, F(2, 348) = 12.84, R2 = .049, p < .001. This indicates that the PIQ subscales account for variance in EDI-BU after controlling for Media Influence, Peer Influence, and BMI. Further, based on the final model, all subscales were found to be significantly associated with EDI-BU scores: Media Influence (ˇ = .31, p < .001), Direct Influence (ˇ = .19, p < .001),
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
97
Table 2 Factor loadings and source of parent influence questionnaire items. Item
Factor 1 (parental influence)
1. I perceived a strong message from my parents to have a slender figure. 2. My parents wanted me to be thinner. 3. My parents kept me from eating foods that I liked in order to lose weight or keep from gaining weight. 4. My parents watched closely what I ate. 5. My parents asked me how much I weigh. 6. My parents told me that I looked like I gained weight. 7. My parents encouraged me to lose weight. 8. My parents said, “If you do not lose weight, you will never get a date, get a boyfriend/girlfriend, get married, etc.” 9. My parents made negative comments about my physical appearance. 10. My parents said critical things to me about my eating. 11. My parents told me to eat different foods in order to lose weight or keep from gaining weight. 12. I received negative feedback about the size or shape of my body. 13. I received negative feedback from my parents about my eating patterns to change my body size or shape. 14. My parents would say to me, “you do not need to lose weight.”a 15. My parents would say to me, “your health is what is important not your weight.”a 16. My parents would say to me, “what you weigh or how you look like is not what is important.”a 17. My parents would say to me, “you need to make sure you eat enough while you are growing.”a 18. My parents teased me about my appearance. 19. My parents encouraged me to gain weight.a 20. My parents teased me because I was too thin.a 21. My parents commented on each other’s weight. 22. My parent would say to my other parent, “you look heavy.” 23. My parents encouraged each other to lose weight. 24. My parents talked about dieting. 25. My parents complained about their weight. 26. My parents would ask, “Am I gaining weight?” 27. My parents would ask, “Am I as fat as him/her?” 28. My parents worried about their weight. 29. Physical appearance (shape, weight, clothing) was important to my parents. 30. My parent’s weight and shape influenced how they felt about themselves. 31. My parents were satisfied with their body weight.a 32. My parents tried to become more muscular.
Factor 2 (modeling)
Source
.61
.32
3
.85 .70
.13 .18
1, 2, 4 5
.42 .57 .75 .83 .74
.25 .28 .18 .12 .06
6 6, 7 7 7, 9, 10 7
.71
.17
10
.79 .76
.24 .23
11 5
.77
.23
12
.63
.31
12
.45
−.15
7
.53
−.08
7
.54
−.04
7
.54
−.09
7
.50 −.33 −.18 .12 .09
.27 .22 .28 .67 .37
1 7 1 1 7
.08 .17 .04 .13 .14 .04 .21
.64 .72 .80 .74 .66 .80 .68
1 1 7 7 7 12 8
.16
.71
11
.05 .03
.38 .41
12 7
Note. Source (original author of item): 1 = Shroff and Thompson (2006); 2 = Levine et al. (1994); 3 = Stice et al. (1996); 4 = Benedikt et al. (1998); 5 = Moreno and Thelen (1993); 6 = Tantleff-Dunn et al. (1995); 7 = Smolak et al. (1999); 8 = Senra et al. (2007); 9 = Wertheim et al. (1999); 10 = Kichler and Crowther (2001); 11 = Baker et al. (2000); 12 = McCabe and Ricciardelli (2001). Principal axis extraction method was used with a Quartimax factor rotation. Factor loading of .4 or greater on one factor were significant. Items 33–37 were reverse-scored. Items 19, 20, 22, and 31 were dropped from subsequent analyses due to nonsignificant factor loadings. a Reverse-scored items.
Table 3 Correlation matrix of study variables. Variable
PIQ-Direct Influence
PIQ-Modeling
EDI-DT
EDI-BU
TI-Media
PIQ-Direct Influence PIQ-Modeling EDI-DT EDI-BU TI-Media TI-Peers BMI EDI-BD
– .34** .39** .41** .27** .31** .41** .34**
– .37** .32** .32** .28** -.01 .24**
– .59** .59** .34** .19** .65**
– .47** .35** .24** .49**
– .40** .07 .39**
TI-Peer
BMI
EDI-BD
– .42**
–
– .09 .29**
Note. PIQ-Direct Influence: Parent Influence Questionnaire–Direct Influence Subscale; PIQ-Modeling: Parent Influence Questionnaire–Modeling Subscale; EDI-DT: Eating Disorder Inventory–Drive for Thinness subscale; EDI-BU: Eating Disorder Inventory–Bulimia Subscale; TI-Peer: Tripartite Influence Scale–Peer Influence Subscale; TI-Media: Tripartite Influence Scale–Media Influence Subscale; BMI: Body Mass Index; EDI-BD: Eating Disorder Inventory–Body Dissatisfaction subscale. ** p < .01.
98
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
BMI (ˇ = .14, p < .01), Peer Influence (ˇ = .13, p < .01), and Modeling (ˇ = .12, p < .05). Parental Influence and BMI Independent samples t-tests were conducted to determine if participants with BMIs in the overweight to obese range reported greater levels of parental influence and modeling when compared to participants with BMIs in the normal range. Underweight participants (with BMIs of less than 18.5) were excluded from the analysis because few participants fell into this category (6.3%). A significant difference between normal and overweight/obese participants was found for Direct Influence, t(334) = 6.28, p < .001, 2 = .015, with overweight/obese participants (M = 18.59, SD = 9.38) reporting higher levels of Direct Influence when compared to normal weight participants (M = 12.88, SD = 6.59). In contrast, the normal weight participants (M = 22.37, SD = 8.70) did not differ significantly from the obese/overweight participants (M = 20.99, SD = 8.59) on levels of Modeling, t(335) = 1.31, p = .19, 2 = .0004. Overall, these results indicate that obese/overweight participants’ recall more direct influence, but no difference based on BMI category was found for Modeling. Discussion The primary purpose of the current study was to conduct an integrative analysis of previously developed measures of parental influence to: (1) clarify the underlying dimensions of parental influence, and (2) develop a comprehensive composite measure that can be utilized consistently in future research to clarify the construct of parental influence. Based on analyses, two clear dimensions emerged, direct influence and modeling. These dimensions supplement the factor analysis conducted by Benedikt et al. (1998) and are consistent with a number of previous studies, which have proposed that direct influence and modeling are distinct dimensions of parental influence (Annus, Smith, Fischer, Hendricks, & Williams, 2007; Baker, Whisman, & Brownell, 2000; Smolak et al., 1999). The direct influence dimension consisted of direct parental communication, including feedback/comments about weight, eating, and exercise. More specifically, this dimension includes parental concern about the participant’s weight, parental messages about the importance of being thin and losing weight, parental pressure for the participant to diet or restrict food intake, parental encouragement of the participant to exercise to lose weight, parental criticism or negative comments about the participant’s weight (along with appearance and eating behaviors), and parental catastrophizing about the participant’s weight. The parental modeling dimension included participant recall of parental dieting and other strategies to manage weight, parental weight dissatisfaction, parental self-emphasis on weight, and parental appearance comments to his or her spouse. The direct influence and modeling dimensions were found to be associated with measures of eating disturbance, including drive for thinness and bulimic symptomatology. It is noteworthy that these associations were significant even after controlling for media and peer influence, suggesting that parental influence may be a distinct pathway in the development of eating disturbance. The association between parental influence and body shape dissatisfaction was less consistent. In particular, after controlling for BMI, modeling was found to be associated with scores on the BD subscale of the EDI-3; however, direct influence was not found to be associated with the BD subscale. These findings indicate that parental modeling, rather than direct parental influence, may impact levels of body shape dissatisfaction. However, in the regression equation, shared variance between direct influence and BMI
may have prevented direct influence from being a significant predictor of body shape dissatisfaction. Based on this observation, a direction for future research is to examine the degree to which direct influence and BMI may interact to influence levels of body shape dissatisfaction. Overall, it appears that direct influence has more of a negative impact on participants’ bulimic symptomatology relative to parental modeling. As an interpretation of this finding, it is reasonable that participants report being more influenced by direct negative messages when compared to more subtle messages that may not be directed towards the participant (i.e., modeling). It is also noteworthy that the weaker association between modeling and measures of eating disturbance is generally consistent with previous literature, as studies that have examined the effect of both direct influence and modeling typically find stronger and more consistent support for direct influence (Baker et al., 2000; Fulkerson et al., 2002; Moreno & Thelen, 1993; Rodgers & Chabrol, 2009; Vincent & McCabe, 2000). Interestingly, consistent with research by Fulkerson et al. (2002) and Thelen and Cormier (1995), direct influence was found to be related to BMI, as overweight/obese participants reported receiving greater levels of direct feedback from parents; however, parental modeling was not found to be associated with BMI. It is noteworthy that examination of BMI as a continuous variable yielded the same results, as BMI was found to be significantly correlated with direct influence but not modeling (see Table 3). Although the direction of the association between these variables is uncertain, the findings suggest that parents of overweight/obese participants may be more likely to engage in direct feedback as an attempt to manage their child’s weight. Another hypothesis is that overweight mothers are more likely to encourage their daughters to diet (Fulkerson et al., 2002). This can send a conflicting and confusing message to daughters. In contrast, modeling behavior, which includes attempts by parents to manage their own weight, is not a strategy to control or manage the weight of overweight/obese participants. However, if the parents of the overweight participants are overweight themselves; they may be modeling unhealthy overeating habits or poor food choices. More research in this area is needed. Despite the strengths of this study, there are a number of noteworthy limitations and directions for future research. First, the findings may not generalize to populations that are either not represented or underrepresented in the current study. In particular, the participants were college females in their late teens/early twenties. The study’s results therefore may not be generalizable to children, young adolescents, males, or older adults. In addition, the current study, along with the items derived from previous studies, have been based on predominantly Caucasian samples. Consequently, the findings from this study may not generalize to other ethnic groups. Furthermore, a majority of participants were of normal weight. As a result, the factor structure of the PIQ may not generalize to those who are either underweight or overweight. In response to these limitations, a direction for future research would be to examine the consistency of the factor structure of the PIQ for those who do not fall in the normal weight range, in males, and also over a variety of age ranges. In addition, it is noteworthy that the current results were based on a nonclinical sample. Consequently, it is recommended that the psychometric properties of the PIQ are examined in clinical samples. Finally, the PIQ and the BD subscale are limited in that both measures focus only on weight and shape and do not focus on other dimensions of body image disturbance. Consequently, a recommended direction for future research is to examine the role of parental influence on other dimensions of body image, including general appearance and muscularity. An additional limitation relates to the self-report and retrospective nature of this study. Specifically, participants with body
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
image disturbance, a drive for thinness or bulimic symptomatology may have been biased in their recall of parental comments and behavior, and it is possible that participants with body shape and eating disturbance may have overestimated the association between parental behaviors and eating pathology (Jansen, Smeets, Boon, Nedrekoorn, Roefs, & Mulkens, 2007). An additional limitation is that current BMI was assessed, not the retrospective BMI at the time of parental influence. As a result of the above limitations, additional research is recommended, including studies that examine the consistency between parent and child-report. Finally, in addition to the above-mentioned directions for future research, it is recommended that follow-up psychometric studies are conducted with the PIQ to confirm the factor structure and examine test–retest reliability of this measure. Another limitation and potential direction for future research is that participants were instructed to consider the influence of both parents when completing the questionnaire, instead of distinguishing between the mother and father. Providing the distinction between mother and father feedback may have yielded different results. For example, mothers and fathers may provide different types of feedback or the impact of the feedback may differ based on the gender of the parent. Consequently, comparison of mother’s and father’s influence is a possible direction for future research. A further limitation and direction for future research relates to the fact that the authors used a somewhat narrow operational definition of body image, which was measured by body shape dissatisfaction. Consequently, it may beneficial for future studies to consider other aspects of body image, including muscularity. In addition, some items did not load on either factor, and appeared to be related to pressure to gain weight. In order to examine this category more thoroughly, a direction for future research would be to examine these items in a sample of underweight individuals, possibly including individuals with anorexia nervosa. A final limitation and future direction would be to include moderators, which was beyond the scope of the current study. For example, modeling may be moderated by direct parental influence. Kichler and Crowther (2001) found that under high levels of negative, direct parental influence, modeling was significantly associated with body image; however, no association was found between parental modeling and body image under low levels of negative parental influence. Another example is that modeling has been found to be significant when more extreme and restrictive weight loss behaviors are utilized by parents, such as fasting, skipping meals, and crash dieting versus the child observing more general dietary restraint (Benedikt et al., 1998; Wertheim et al., 1999). Taking moderators into account may further elucidate and clarify the pathway from parental influence to body image and eating disturbance. Overall, the current study provides a number of significant contributions to the understanding of parental influence as a risk factor in the development of a drive for thinness, body shape dissatisfaction, and bulimic symptomatology. For example, in addition to the development of a promising measure, this study provides insight into the dimensions or facets of parental influence. More specifically, this study is among the first to provide strong psychometric support for the premise that parental influence forms two distinct dimensions, direct influence and modeling. Further, because the items of the PIQ are derived from previously published measures of parental influence, this study also assists in the clarification of inconsistent terminology utilized to discuss various components of parental influence. For example, it appears that terms direct influence and parental criticism about appearance refer to overlapping constructs. This measure has the potential to serve as an impetus for future research, and the PIQ may serve as a much needed tool to assess the degree to which parents directly and indirectly communicate appearance-related messages to their children or
99
adolescents. Finally, we hope that additional research is conducted to confirm the factor structure of the PIQ and to refine this measure.
Acknowledgement This manuscript is based on the first author’s undergraduate honors thesis, and the support for the purchase of study-related measures was provided by University of North Florida Undergraduate Enrichment Program.
References Annus, A. M., Smith, G. T., Fischer, S., Hendricks, M. & Williams, S. F. (2007). Associations among family-of-origin food-related experiences, expectancies, and disordered eating. International Journal of Eating Disorders, 40, 179–186. Baker, C., Whisman, M. & Brownell, K. (2000). Studying intergenerational transmission of eating attitudes and behaviors: Methodological and conceptual questions. Health Psychology, 19, 376–381. Benedikt, R., Wertheim, E. & Love, A. (1998). Eating attitudes and weight-loss attempts in female adolescents and their mothers. Journal of Youth and Adolescence, 27, 43–57. Byely, L., Archibald, A., Graber, J. & Brooks-Gunn, J. (2000). A prospective study of familial and social influences on girls’ body image and dieting. International Journal of Eating Disorders, 28, 155–164. Clark, L. & Watson, D. (1995). Constructing validity: Basic issues in objective scale development. Psychological Assessment, 7, 309–319. Dixon, R., Adair, V. & O’Connor, S. (1996). Parental influences on the dieting beliefs and behaviors of adolescent females in New Zealand. Journal of Adolescent Health, 19, 303–307. Edlund, B., Halvarsson, K. & Sjödén, P. (1996). Eating behaviours and attitudes to eating, dieting and body image in 7-year-old Swedish girls. European Eating Disorders Review, 4, 40–53. Fulkerson, J., McGuire, M., Neumark-Sztainer, D., Story, M., French, S. & Perry, C. (2002). Weight-related attitudes and behaviors of adolescent boys and girls who are encouraged to diet by their mothers. International Journal of Obesity, 26, 1579–1587. Garner, D. M. (2004). Eating Disorders Inventory-3: Professional manual. United States: Psychological Assessment Resources. Gross, R. & Nelson, E. (2000). Perceptions of parental messages regarding eating and weight and their impact on disordered eating. Journal of College Student Psychotherapy, 15, 57–78. Hanna, A. & Bond, M. (2006). Relationships between family conflict, perceived maternal verbal messages, and daughters’ disturbed eating symptomatology. Appetite, 47, 205–211. Hill, A. & Franklin, J. (1998). Mothers, daughters and dieting: Investigating the transmission of weight control. The British Journal of Clinical Psychology, 37, 3–13. Jansen, A., Smeets, T., Boon, B., Nedrekoorn, C., Roefs, A. & Mulkens, S. (2007). Vulnerability to interpretation bias in overweight children. Psychology and Health, 22, 561–574. Kanakis, D. & Thelen, M. (1995). Parental variables associated with bulimia nervosa. Addictive Behaviors, 20, 491–500. Keery, H., van den Berg, P. & Thompson, J. (2004). An evaluation of the Tripartite Influence Model of body dissatisfaction and eating disturbance with adolescent girls. Body Image, 1, 237–251. Kichler, J. & Crowther, J. (2001). The effects of maternal modeling and negative familial communication on women’s eating attitudes and body image. Behavior Therapy, 32, 443–457. 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. Stephens, J. H. Crowther, D. L. Tennenbaum & M. Stephens (Eds.), The etiology of bulimia nervosa: The individual and familial context (pp. 59–80). Washington, DC, USA: Hemisphere Publishing Corp. Levine, M., Smolak, L. & Hayden, H. (1994). The relation of sociocultural factors to eating attitudes and behaviors among middle school girls. The Journal of Early Adolescence, 14, 471–490. MacBrayer, E., Smith, G., McCarthy, D., Demos, S. & Simmons, J. (2001). The role of family of origin food-related experiences in bulimic symptomatology. International Journal of Eating Disorders, 30, 149–160. McCabe, M. & Ricciardelli, L. (2001). The structure of the Perceived Sociocultural Influences on Body Image and Body Change Questionnaire. International Journal of Behavioral Medicine, 8, 19–41. McCabe, M. & Ricciardelli, L. (2003). Sociocultural influences on body image and body changes among adolescent boys and girls. The Journal of Social Psychology, 143, 5–26. Moreno, A. & Thelen, M. (1993). Parental factors related to bulimia nervosa. Addictive Behaviors, 18, 681–689. Ogden, J. & Steward, J. (2000). The role of the mother–daughter relationship in explaining weight concern. International Journal of Eating Disorders, 28, 78–83. Pike, K. & Rodin, J. (1991). Mothers, daughters, and disordered eating. Journal of Abnormal Psychology, 100, 198–204.
100
M. Abraczinskas et al. / Body Image 9 (2012) 93–100
Rodgers, R. & Chabrol, H. (2009). Parental attitudes, body image disturbance and disordered eating amongst adolescents and young adults: A review. European Eating Disorders Review, 17, 137–151. Ruther, N. & Richman, C. (1993). The relationship between mothers’ eating restraint and their children’s attitudes and behaviors. Bulletin of the Psychonomic Society, 31, 217–220. Senra, C., Sánchez-Cao, E., Seoane, G. & Leung, F. (2007). Evolution of self-concept deficits in patients with eating disorders: The role of family concern about weight and appearance. European Eating Disorders Review, 15, 131–138. Shroff, H. & Thompson, J. (2006). The tripartite influence model of body image and eating disturbance: A replication with adolescent girls. Body Image, 3, 17–23. Smolak, L., Levine, M. & Schermer, F. (1999). Parental input and weight concerns among elementary school children. International Journal of Eating Disorders, 25, 263–271. Stice, E. (1998). Modeling of eating pathology and social reinforcement of the thinideal predict onset of bulimic symptoms. Behaviour Research and Therapy, 36, 931–944. Stice, E., Ziemba, C., Margolis, J. & Flick, P. (1996). The dual pathway model differentiates bulimics, subclinical bulimics, and controls: Testing the continuity hypothesis. Behavior Therapy, 27, 531–549. Tantleff-Dunn, S., Thompson, J. & Dunn, M. (1995). The Feedback on Physical Appearance Scale (FOPAS): Questionnaire development and psychometric evaluation. Eating Disorders: The Journal of Treatment & Prevention, 3, 332–341.
Thelen, M. & Cormier, J. (1995). Desire to be thinner and weight control among children and their parents. Behavior Therapy, 26, 85–99. 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, 146–157. Thompson, J., Heinberg, L., Altabe, M. & Tantleff-Dunn, S. (1999). Exacting beauty: Theory, assessment, and treatment of body image disturbance. Washington, DC, USA: American Psychological Association. van den Berg, P., Thompson, J., Obremski-Brandon, K. & Coovert, M. (2002). The tripartite influence model of body image and eating disturbance: A covariance structure modeling investigation testing the meditational role of appearance comparison. Journal of Psychosomatic Research, 53, 1007–1020. Vincent, M. & McCabe, M. (2000). Gender differences among adolescents in family, and peer influences on body dissatisfaction, weight loss, and binge eating behaviors. Journal of Youth and Adolescence, 29, 205–221. Wertheim, E., Mee, V. & Paxton, S. (1999). Relationships among adolescent girls’ eating behaviors and their parents’ weight-related attitudes and behaviors. Sex Roles, 41, 169–187. Wood, K., Becker, J. & Thompson, J. (1996). Body image dissatisfaction in preadolescent children. Journal of Applied Developmental Psychology, 17, 85–100. Young, E., Clopton, J. & Bleckley, M. (2004). Perfectionism, low self-esteem, and family factors as predictors of bulimic behavior. Eating Behaviors, 5, 273–283.