Research Brief Gender Comparisons of Unhealthy Weight-control Behaviors Among Sixth-Graders Deborah Cragun, MS, CGC1; Rheanna N. Ata, BA2; Rita D. DeBate, PhD, MPH, CHES1; J. Kevin Thompson, PhD2 ABSTRACT Objective: To examine gender differences in unhealthy weight-control behaviors (UWCB) and identify key psychosocial and demographic correlates of UWCB among sixth-graders. Methods: A cross-sectional survey was completed by 146 boys and 153 girls at a middle school. Secondary data analyses included bivariate tests and multivariable logistic regression. Results: Forty-seven percent of participants reported 1 or more UWCB, with no differences by gender (P ¼ .75). Factors common to boys and girls included: lower global self-esteem; lower body-esteem; and greater negative parental modeling among participants who engaged in UWCB compared to those who did not. However, multivariable models revealed gender differences. Among boys, body mass index, negative parental modeling, and global self-esteem retained statistically significant associations with UWCB after controlling for other variables in the model, whereas race and weight-related body-esteem remained significant for girls. Conclusions and Implications: This research highlights the need for gender-specific UWCB prevention programs implemented in late childhood and early adolescence. Key Words: body mass index, disordered eating behaviors, body image, self-esteem, parental influence (J Nutr Educ Behav. 2013;45:450-454.)

INTRODUCTION Disordered eating behaviors are deemed significant public health issues because of their prevalence,1 association with other unhealthy behaviors,2 and tendency to progress to long-term physical, mental, and psychosocial health consequences.3,4 Also referred to as unhealthy weight-control behaviors (UWCB), disordered eating behaviors include the following: fasting; eating very little food; consuming meal supplements; and skipping meals.1,4 More extreme disordered eating behaviors, such as self-induced vomiting and laxative, diuretic, and diet pill use/misuse, have been classified as very unhealthful weight control behaviors.4 The importance of developing interventions for decreasing UWCB has 1

been noted in Healthy People 2020, the science-based, 10-year health agenda of the United States (US), with the addition of the new Mental Health Objective: ‘‘Reduce the proportion of adolescents (10-19 years of age) who engage in disordered eating behaviors in an attempt to control their weight.’’5 In order to determine whether separate interventions should be developed to target early adolescent boys and girls, this study employed Social Cognitive Theory to explore potential gender differences and psychosocial correlates of UWCB among sixth-graders.

a state-level policy requiring the reporting of body mass index (BMI) measurements to parents.6 The original cross-sectional study gathered data from students using a self-report questionnaire. The University of South Florida Institutional Review Board, as well as the Hillsborough County Public School Research Board, granted approval for the original study. The current study received institutional review board exemption as it consisted of secondary analysis of deidentified data.

METHODS Study Procedures

Participants

The current research is based on data from a study exploring impacts of

Department of Community and Family Health, College of Public Health, University of South Florida, Tampa, FL 2 Department of Psychology, College of Arts and Sciences, University of South Florida, Tampa, FL Address for correspondence: Rita D. DeBate, PhD, MPH, CHES, Department of Community and Family Health, College of Public Health, University of South Florida, 13201 Bruce B. Downs Blvd, Tampa, FL 33612; Phone: (813) 974-6683; Fax: (813) 974-5172; E-mail: [email protected] Ó2013 SOCIETY FOR NUTRITION EDUCATION AND BEHAVIOR http://dx.doi.org/10.1016/j.jneb.2012.09.002

450

Of 346 sixth-grade students at a middle school in Tampa, Florida, 86% (n ¼ 299) assented to participate. The sample was roughly evenly distributed between boys (48.8%) and girls (51.2%); the mean age was 11.9 years old (SD ¼ 0.54, range ¼ 11-13 years old). Students self-identified into the following combined ethnic/ racial categories: white, nonHispanic/Latino (41.0%); black, nonHispanic/Latino (22.5%); Hispanic (26.4%); Asian (6.6%); and other (3.5%). Patterns of racial/ethnic identification were similar for boys and

Journal of Nutrition Education and Behavior  Volume 45, Number 5, 2013

Journal of Nutrition Education and Behavior  Volume 45, Number 5, 2013 girls. Although socioeconomic status was not assessed, the participating middle school was not designated as a Title 1 school, and it was located within a geographic zip code in which the median household income was more than $30,000 higher than the median household income in the US, according to 2010 census data. Based on the Centers for Disease Control and Prevention's sex-specific BMI-for-age growth charts,7 3.0% of participants were observed to be underweight; 61.5% normal weight; 17.1% overweight; and 18.4% obese. No significant sex differences were observed by proportions classified in each weight category (c2 [3, 296] ¼ 2.92; P ¼ .40).

Cragun et al 451

Table 1. Gender Comparisons of Weight-control Behaviors Among Sixth-graders (n ¼ 299)

One or more unhealthy weight-control behaviorsb (a) Took diet pills (b) Ate very little food for a day or more (c) Skipped breakfast (d) Skipped meals other than breakfast (e) Ate more fruits and vegetables (f) Ate less high-fat food (g) Exercised more (h) Ate fewer sweets

Girls, Boys, c2 P* n (%)a n (%)a 68 (48) 70 (46) 0.20 .75 3 (2) 0 (0) 3.20 .07 37 (26) 40 (27) 0.02 .91 35 (25) 56 (37) 5.47 .02 26 (18) 26 (17) 0.05 .83 88 (63) 108 (72) 2.76 .10 72 (51) 92 (61) 2.67 .10 106 (74) 112 (74) 0.00 .99 82 (58) 104 (69) 4.24 .04

*Results were considered significant at P # .006 based on Bonferroni correction; Valid percentages are based on the number of respondents who answered each question; bUnhealthy weight-control behavior included: skipping breakfast, skipping meals other than breakfast, taking diet pills, and eating very little food. a

Measures Weight-control behaviors were assessed using the yes/no questions listed in Table 1. A dummy variable was created by classifying participants who answered yes to 1 or more of the 4 items that described UWCB (items ad, Table 1) and assigning them a code of 1. Other participants who reported no UWCB were coded 0. The Body-Esteem Scale (BE) for Adolescents and Adults8 was used to measure self-evaluations of overall appearance and weight using Likert-type items for each subscale, including: (1) BE-Appearance (10 items, Cronbach a ¼ .91); and (2) BE-Weight (8 items; Cronbach a ¼ .91). An average score (ranging from 1 to 5) was obtained for each subscale. Higher scores indicate more positive feelings about one's body. The Rosenberg Self-Esteem Scale includes 10 items for use in assessing global self-esteem (Cronbach a ¼ .82).9 This is a widely used scale that has been determined to be reliable and valid among various different populations.10 An average scale measure (ranging from 1 to 4) was calculated for all students who answered 8 or more of the 10 items. Higher scores indicate higher self-esteem. Child report of negative parental modeling was assessed via the following questions: In the past month, how often have your parents/guardians: Made a comment to you about your weight that made you feel bad; Encouraged you to diet in order to lose weight; Complained about how they

look; Complained about their weight; Talked about wanting to lose weight; Gone on a diet; Made comments about other people's weight (Cronbach a ¼ .82). All items were rated on a Likert-type scale ranging from not at all (1) to every day (4). An average score (ranging from 1 to 4) was calculated for all children who answered 5 or more of the 7 questions. The higher the score, the greater the child report of negative parental modeling.

Data Analyses Data were analyzed using SPSS (version 19.0, SPSS, Inc, Chicago, IL, 2010). Initial analyses included chisquare tests to determine whether the proportion of boys and girls who engaged in weight-control behaviors differed. Separate bivariate analyses for boys and girls consisted of independent samples t tests and chi-square tests to determine which psychosocial variables demonstrated a significant relationship with UWCB. Psychosocial variables from bivariate analyses that were statistically significant at a 2-sided value of P # .05 were entered into 2 logistic regression models, 1 for each sex, in a forward stepwise fashion after first entering variables to control for age, BMI, and race/ethnicity.

RESULTS Nearly half of male (n ¼ 68, 48%) and female participants (n ¼ 70, 46%)

engaged in 1 or more UWCB, with no statistically significant difference by sex (Table 1). After employing a Bonferroni correction to reduce the risk of a type I error, no significant gender differences in specific types of weight-control behaviors were identified (Table 1). Sex-stratified bivariate analyses suggested possible differences in age, BMI, and race/ethnicity based on whether or not participants engaged in UWCB (Supplementary Table 1). Accordingly, these variables were controlled for in the regression models. Bivariate analyses also revealed lower global self-esteem, lower body-esteem, and greater negative parental modeling among both male and female participants who engaged in UWCB compared to those who did not (Supplementary Table 2). Age, parent modeling, and selfesteem differed by sex after controlling for the effects of other variables in the models (Tables 2 and 3). Thirteen-yearold male participants were approximately 5 times more likely to engage in UWCB compared to 11-year-old boys (odds ratio [OR] ¼ 5.30, 95% confidence interval [CI] ¼ 1.10, 25.46). In addition, a 2-fold (95% CI ¼ 1.09, 4.17) increase in the odds of engaging in UWCB was observed for every 1-point increase in negative parental modeling; for every 1-point increase in self-esteem, the odds of engaging in UWCB were halved (OR ¼ 0.49, 95% CI ¼ 0.24, 0.98).

Journal of Nutrition Education and Behavior  Volume 45, Number 5, 2013

452 Cragun et al

Table 2. Estimates and Adjusted ORs for Boys Engaging in Unhealthy Weightcontrol Behavior (n ¼ 122) ORa

95% CI

P

reference 0.91 (0.64) 1.67 (0.80)

2.49 5.30

0.72-8.92 1.10-25.46

.14 .04*

reference 0.15 (0.52) 0.04 (0.54) 0.01 (0.82)

1.15 0.95 0.84

0.42-3.25 0.33-2.76 0.20-4.96

.77 .94 .99

Body mass index category Normal weight At risk for overweight Overweight

reference 1.13 (0.54) 0.45 (0.53)

3.10 1.58

1.07-8.92 0.56-4.45

.04* .35

Negative parent modeling

0.76 (0.34)

2.13

1.09-4.17

.03*

0.72 (0.36)

0.49

0.24-0.98

.05*

Variable Age, y 11 12 13 Race/ethnicity White/non-Hispanic Black/non-Hispanic Hispanic/Latino Other

Global self-esteem

Estimate (SE)

CI indicates confidence interval; OR, odds ratio; SE, standard error. *Statistically significant at P # .05; aORs are adjusted for all other variables listed.

Among female participants (Table 3), the odds of engaging in UWCB were halved (OR ¼ 0.51, 95% CI ¼ 0.32, 0.81) for every 1-point increase on the BE-Weight subscale. Additionally, compared to female participants who identified as white, black girls were significantly more likely to engage in UWCB (OR ¼ 5.72, 95% CI ¼ 1.73, 16.11).

DISCUSSION The high prevalence of UWCB among early adolescent boys (48%) and girls (46%) in the current study underscores the need for interventions, as indicated in the newly developed Healthy People 2020 objective.5 These findings are further supported by several large studies that have confirmed

Table 3. Estimates and Adjusted ORs for Girls Engaging in Unhealthy Weightcontrol Behavior (n ¼ 132) Estimate (SE)

ORa

95% CI

P

reference 0.069 (0.46) 0.347 (1.04)

0.93 1.42

0.38-2.31 0.18-10.93

.88 .07

Race/ethnicity White/non-Hispanic Black/non-Hispanic Hispanic/Latino Other

reference 1.66 (0.57) 0.54 (0.46) 0.32 (0.80)

5.72 1.72 0.72

1.73-16.11 0.70-4.18 0.15-3.47

<.001* .24 .69

Body mass index category Normal weight At risk for overweight Overweight

reference 0.23 (0.55) 0.18 (0.62)

1.25 0.83

0.45-3.70 0.25-2.83

.68 .77

Weight-related body-esteem

0.68 (0.22)

0.51

0.32-0.81

<.001*

Variable Age, y 11 12 13

CI indicates confidence interval; OR, odds ratio; SE, standard error. *Statistically significant at P # .05; aORs are adjusted for all other variables shown.

a high prevalence of UWCB among both male and female adolescents in the US and other countries around the world.11,12 The current study was conducted to identify key factors that may be useful targets for preventing UWCB among a group of early adolescents and to determine whether it may be beneficial to implement sex-specific interventions that target different factors. The following discussion addresses this 2-fold study purpose and the high prevalence of UWCB among black girls who participated in this study. Among adolescents, concern about body weight has previously been identified as an important factor related to UWCB.13,14 The current study also supports the assertion that negative feelings about one's weight may be a key target in prevention efforts among early adolescent girls, given that BE-Weight was the only psychosocial variable to retain a significant relationship with UWCB in the multivariable female model. In contrast, BE-Weight did not remain statistically significant in the male model after controlling for BMI and other demographic and risk factors. Consequently, weight perception does not appear to be the best target for change in interventions to prevent UWCB among this group of young boys. Based on the current results from the multivariable male model, some promising targets for interventions among this group of early adolescent boys include global self-esteem and negative parental modeling of weightrelated behaviors. A causal role of parental influences is supported by longitudinal and cross-sectional studies.13-16 Although parental influence may also be a good target for interventions among girls, findings from the current study in conjunction with findings from 2 earlier studies suggest that familial factors may not be as strongly predictive of disordered eating among adolescent girls.17,18 Interestingly, the relationship between parental influences and weightrelated behaviors may also vary among different cultures, as evidenced by a large cohort study in which father's BMI, weight control, and fear of getting fat appeared to mediate the relationship between skipping breakfast and overweight among Finnish but not Greek adolescent boys.12

Journal of Nutrition Education and Behavior  Volume 45, Number 5, 2013 Additional differences between boys and girls in the current study suggest a relationship between UWCB and demographic factors. In contrast to earlier findings,17,19 the current study, as well as a recent study of Massachusetts middle school students,18 found that UWCB were more prevalent among black compared to white early adolescent girls. Although not directly comparable to the current study of sixth-graders, evidence that a relationship between UWCB and race may be moderated by socioeconomic status (SES) comes from a study that revealed that higher-SES African American girls in third grade reported more weight concerns than lower-SES African American girls, but the opposite trend was observed for white girls.20 Alternatively, general findings of earlier onset of puberty among African American girls21 may also help to explain the relationship, as onset of puberty is related to disordered eating behaviors.22 These findings highlight the need for caution when making generalizations about the prevalence of disordered eating based on race/ethnicity. The results of the current study should be interpreted with the following limitations in mind. First, the cross-sectional nature of the study does not allow for causation to be determined. Second, although the participation rate was high, generalizability is limited because participants represented a single grade in 1 middle school. Third, those who did and did not engage in UWCB may have differed on other key characteristics (ie, onset of puberty, media and peer influences) that could affect dieting behaviors. Fourth, questions to assess UWCB were not previously validated, did not ask about the frequency of the behaviors, and did not include questions to assess very unhealthful weight control behaviors (ie, vomiting and laxative or diet pill use). Finally, several confidence intervals in the current study are wide, which may be a result of the small size of subgroups within the sample. Despite these limitations, findings from the current study offer some information regarding possible leverage points for intervention development and support the need for additional research among young adolescent boys and girls.

IMPLICATIONS FOR RESEARCH AND PRACTICE This study raises important implications pertinent to the development of interventions to prevent UWCB among various groups of early adolescents. Current findings support a need to focus interventions on both girls and boys, as well as on all ethnic/racial groups. Targeting different interventions to early adolescent boys and girls may be appropriate given differences in the relative strength of psychosocial correlates of UWCB. The strong association of negative modeling behaviors by parents and UWCB in boys supports the need for a holistic environmental approach to preventing eating problems among boys. Furthermore, the association between global selfesteem with UWCB among boys is of practical significance for intervention planning and provides support that ecological approaches (eg, positive youth development [PYD] programs) that do not focus specifically on weight management, bodyesteem, or healthy eating may be successful in preventing or reducing UWCB through improvements in internal assets such as general self-esteem.23 Although a PYD approach may also be beneficial for girls, improvement in weight satisfaction is a critical target for reducing UWCB among girls.3 Weight satisfaction should therefore be assessed among early adolescent girls, even if it is not the focus of PYD programs. Relationships between psychosocial and demographic variables with UWCB are complex and may vary among boys and girls and among different subpopulations, which suggests the need for additional assessment so that interventions can be targeted to specific groups. Further research is also necessary to evaluate the effectiveness of targeted interventions to reduce UWCB among early adolescents.

SUPPLEMENTARY DATA Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/ j.jneb.2012.09.002.

Cragun et al 453

REFERENCES 1. Neumark-Sztainer D, Wall M, Larson NI, Eisenberg ME, Loth K. Dieting and disordered eating behaviors from adolescence to young adulthood: findings from a 10-year longitudinal study. J Am Diet Assoc. 2011;111:1004-1011. 2. Johnson JL, Eaton DK, Pederson LL, Lowry R. Associations of trying to lose weight, weight control behaviors, and current cigarette use among US high school students. J Sch Health. 2009;79:355-360. 3. Neumark-Sztainer DR, Wall MM, Haines JI, et al. Shared risk and protective factors for overweight and disordered eating in adolescents. Am J Prev Med. 2007;33:359-369. 4. Neumark-Sztainer D. Preventing the broad spectrum of weight-related problems: working with parents to help teens achieve a healthy weight and a positive body image. J Nutr Educ Behav. 2005;37(suppl 2):S133-S140. 5. Mental Health and Mental Disorders: Objectives. Healthy People 2020 Web site. http://healthypeople.gov/2020/ topicsobjectives2020/objectiveslist.aspx? topicid¼28. Updated October 30, 2012. Accessed November 20, 2012. 6. Kaczmarski JM, DeBate RD, Marhefka SL, Daley EM. State-mandated school-based BMI screening and parent notification: a descriptive case study. Health Promot Pract. 2011;12: 797-801. 7. Division of Nutrition, Physical Activity, and Obesity, National Center for Chronic Disease Prevention and Health Promotion. About BMI for Children and Teens. Centers for Disease Control and Prevention Web site. http://www .cdc.gov/healthyweight/assessing/bmi/ childrens_bmi/about_childrens_bmi .html. Updated September 13, 2011. Accessed November 20, 2012. 8. Mendelson BK, Mendelson MJ, White DR. Body-esteem scale for adolescents and adults. J Pers Assess. 2001; 76:90-106. 9. Rosenberg M. Society and the Adolescent Self-Image. Princeton, NJ: Princeton University Press; 1965. 10. Blascovich J, Tomaka J. Measures of selfesteem. In: Robinson JP, Shaver PR, Wrightsman LS, eds. Measures of Personality and Social Psychological Attitudes. 3rd ed. Ann Arbor, MI: Institute for Social Research; 1993:115-160. 11. Lai Yeung WL. Gender perspectives on adolescent eating behaviors: a study on

454 Cragun et al

12.

13.

14.

15.

the eating attitudes and behaviors of junior secondary students in Hong Kong. J Nutr Educ Behav. 2010;42:250-258. Veltsista A, Laitinen J, Sovio U, et al. Relationship between eating behavior, breakfast consumption, and obesity among Finnish and Greek adolescents. J Nutr Educ Behav. 2010; 42:417-421. Vander Wal JS. Unhealthy weight control behaviors among adolescents. J Health Psychol. 2012;17:110-120. Linde JA, Wall MM, Haines J, Neumark-Sztainer D. Predictors of initiation and persistence of unhealthy weight control behaviours in adolescents. Int J Behav Nutr Phys Act. 2009; 6:72. www.ijbnpa.org/content/6/1/72. Accessed November 27, 2012. Gardner RM, Stark K, Friedman BN, Jackson NA. Predictors of eating disorder scores in children ages 6 through 14: a longitudinal study. J Psychosom Res. 2000;49:199-205.

Journal of Nutrition Education and Behavior  Volume 45, Number 5, 2013 16. Meesters C, Muris P, Hoefnagels C, van Gemert M. Social and family correlates of eating problems and muscle preoccupation in young adolescents. Eat Behav. 2007;8:83-90. 17. Chao YM, Pisetsky EM, Dierker LC, et al. Ethnic differences in weight control practices among US adolescents from 1995 to 2005. Int J Eat Disord. 2008;41:124-133. 18. Austin SB, Spadano-Gasbarro J, Greaney ML, et al. Disordered weight control behaviors in early adolescent boys and girls of color: an underrecognized factor in the epidemic of childhood overweight. J Adolesc Health. 2011;48:109-112. 19. Neumark-Sztainer D, Croll J, Story M, et al. Ethnic/racial differences in weight-related concerns and behaviors among adolescent girls and boys: findings from Project EAT. J Psychosom Res. 2002;53: 963-974.

20. Robinson TN, Chang JY, Haydel KF, Killen JD. Overweight concerns and body dissatisfaction among thirdgrade children: the impacts of ethnicity and socioeconomic status. J Pediatr. 2001;138:181-187. 21. Styne DM. Puberty, obesity and ethnicity. Trends Endocrinol Metab. 2004; 15:472-478. 22. Tremblay L, Lariviere M. The influence of puberty onset, body mass index, and pressure to be thin on disordered eating behaviors in children and adolescents. Eat Behav. 2009;10:75-83. 23. Levine MP, Smolak L. Recent developments and promising directions in the prevention of negative body image and disordered eating in children and adolescents. In: Smolak L, Thompson JK, eds. Body Image, Eating Disorders, and Obesity in Youth: Assessment, Prevention, and Treatment. 2nd ed. Washington, DC: American Psychological Association; 2009:215-240.