Objectives: To examine 9-year trends and relationships regarding misperceptions of body size and dieting for weight loss among adolescents from 24 countries, and explore the influence of country-level overweight prevalence.
Methods: Sociodemographic characteristics, body size perception and dieting for weight loss were assessed in the Health Behaviour in School-aged Children survey conducted in 24 countries cross-sectionally at three time points (2001/2002, 2005/2006 and 2009/2010). Logistic regression models examined change over time in overestimation of body size in non-overweight adolescents, underestimation of body size in overweight adolescents, dieting for weight loss in non-overweight and overweight adolescents and relationships between body size perception and dieting. Analyses were stratified by weight status and sex. Covariates included country-level overweight prevalence, family affluence and country level of development. Body mass index was only included in models examining dieting for weight loss.
Results: Country-level overweight prevalence increased over time (11.6-14.7%). Compared with Time 1, overweight adolescents had greater odds of body size underestimation at Time 3 (odds ratio (OR)=1.68 for girls; OR=1.10 for boys), whereas non-overweight adolescents had lower odds of body size overestimation at Time 3 (OR=0.87 for girls; OR=0.89 for boys). Controlling for country-level overweight prevalence attenuated these relationships. Compared with Time 1, overweight and non-overweight boys were 10% more likely to diet at Time 3, whereas overweight and non-overweight girls were 19% and 16%, respectively, less likely to diet at Time 3. Controlling for country-level overweight prevalence did not impact trends in dieting for weight loss. Additionally, the association of self-perceived overweight with increased odds of dieting diminished over time.
Conclusions: Body size perceptions among adolescents may have changed over time concurrent with shifts in country-level body weight. However, controlling for country-level overweight prevalence did not impact trends in dieting for weight loss, suggesting a potentially stronger impact of social comparison on weight-related perceptions than on behavior.