Identifying Mechanisms of Peer Influence on Youth Weight-Related Behaviors

点击翻译按钮获取中文摘要

Summary The United States has experienced a two to three fold increase in pediatric obesity since the 1970’s. To date, school-based interventions to prevent and treat overweight and obesity have realized only limited success. Many of these interventions are guided by health behavior theories and change strategies that address the issue from multiple levels of influence. There is, however, limited information regarding peer influence on youth weight status and weight-related behaviors, such as physical activity, screen time, and dietary patterns. A growing body of literature suggests that such weight-related behaviors are similar among friends, but the mechanisms underlying this clustering of behaviors remain unclear. Friends may influence each other, but also similar students may become friends, or friends may be exposed to similar outside influences. A better understanding of these phenomena would facilitate design of more effective interventions that can leverage the power of peer influence. Therefore, the purpose of this proposed study is to identify these mechanisms of action by collecting and analyzing social network and weight-related behavior data in a cohort of diverse young adolescents during their middle school years (6th to 8th grade). We will distinctly measure networks of interaction (whom the respondent spends his/her time with), sentiments (whom the respondent likes), and organized activities (classes, clubs, and teams). Data will be collected several times each academic year allowing us to analyze these processes in fine time grain and to use external changes in the organized activities as natural experiments and quasi-experiments on social network structures and weight-related behaviors. Stochastic Actor-Oriented Models will be used to rigorously analyze the co-evolution of the network structure and weight-related behaviors. Using findings from those statistical analyses, Agent-Based (simulation) Models will be developed to incorporate direct causal relationships and feedbacks as well as the shapes of these effects over time. Such models will be used to simulate potential intervention scenarios on the behaviors and ultimately, weight status. The proposed research will identify unique leverage points for targeting and timing of WRB interventions. We anticipate that next generation WRB interventions will be able to use the information obtained from this study to improve their ability to prevent excess weight gain in youth thereby reducing the current and future prevalence of related health risk factors and co-morbidities. The project will produce an empirically calibrated test bed for developing, testing, and evaluating intervention strategies, which can be shared with the general public along with privacy-protected study data.