Erin Moore and Dr. Renata Forste, Sociology

My ORCA grant was used in a statistical examination using data from the 2001-2002 Health Behaviors in School-age Children, a nationally representative survey measuring children and adolescents on a variety of health-related behaviors and attitudes. With these data, I explored the effects of a discrepancy between actual and perceived body size on various types of weight control behaviors. I also took into account the influences of the sociocultural pressures of family, peers, and media, and also the influence of pubertal development. This examination focused on adolescent girls in 9th and 10th grade in the United States; it yielded a number of interesting findings which hold implications both for policy and future research.

An extensive body of literature chronicles the prevalence and consequences of negative body image on adolescent girls. However, we know very little about the accuracy of that negative image in terms of how closely it mirrors the individual’s actual body. Whether or not an individual possesses an accurate mental image of their body raises an interesting question: we know that a majority of adolescent girls engage in some kind of weight control behavior, but should we not question the mental health of an individual who engages in such behavior needlessly, who treat her body as if it were unhealthy when she is technically at a correct weight? I wanted to explore this issue quantitatively. For each person in my sample of 1,979 high school females, I calculated a score which represents the degree to which her perceived body size deviates from her actual body size, a variable I call the “gap.” This score is created by subtracting the individual’s BMI (based on self-reported height and weight measures, on a scale of 1 to 4) from her perceived body size (also scaled from 1 to 4, where 1=too thin, 2=about the right size, 3=a bit overweight, and 4=very overweight.) Respondents who perceive themselves to be smaller than they actually are were excluded from this analysis, such that the gap is scaled from 0 to 3. A gap of 0 represents no discrepancy between actual and perceived body size, and a gap of 3 represents an individual who is, by government standards, underweight, but perceives herself to be overweight. Scores of 1 or 2 represent less extreme values of this discrepancy. Once every score was calculated, I then examined how this gap affects the likelihood that a person engages in healthy and unhealthy weight control behaviors. Potential healthy weight control behaviors include things like exercising, eating more fruits and vegetables, and eating fewer sweets. Potential unhealthy weight controls include taking pills, smoking, and vomiting. Data were analyzed using Poisson regression, a statistical modeling technique which tests for relationships between variables measuring counts of something, in this case the total number of weight control behaviors the respondent reported using in the previous year. The results of that analysis yielded some very interesting findings.

The influence of the gap was statistically significant for both healthy and unhealthy weight control, such that an increase in gap score was associated with higher expected rates for all types of behavior. This first series of models purposefully excluded all other variables to discern if a relationship existed between the gap and weight control alone. Once established, the resiliency of this relationship was tested in a second succession of models, this time including the other controls for such effects as puberty, family relationships, peer relationships, media intake, and race. When these controls were included in the model, the significance of the gap for healthy weight control behavior disappeared, while one’s pubertal stage and race seemed to have strong effects. For unhealthy weight control, however, the effect of the gap remained constant throughout all models. Even controlling for the influence of family, peers, media, and puberty, possessing a distorted mental image of one’s body was consistently associated with engagement in unhealthy weight control.

Another interesting finding involves the influence of parents: poor relationships with mothers and fathers were associated with unhealthy weight control at a statistically significant level. However, this association was not present in respect to healthy weight control. A wide body of literature connects the influence of parents to girls’ body image, yet a relationship between parents and a specific category of behavior offers a unique insight. Particularly because these unhealthy behaviors can be extremely harmful, understanding the role of parents is a valuable contribution to the body image debate. It may also have implications for the prevention of eating disorders. It has been an interdisciplinary norm in recent years to focus on the harmful influence of media for body image, but these data suggest that intervention at the family level may be a more effective approach.