S1 Dataset -

A growing body of research suggests that obesity can be understood as a complex and biobehavioral condition influenced by social relationships ─social networks. Social network analysis allows us to examine how an individual’s network characteristics (e.g., popularity) are associated with obesity and obesity-related behaviors. The objectives of the study were to (a) examine whether network members in African American churches are similar in body mass index (BMI) and obesity-related behaviors (physical activity, eating, alcohol consumption) and (b) examine whether an individual’s network characteristics, such as popularity (i.e., receiving nominations from peers) and expansiveness (i.e., sending nominations to peers) are associated with BMI and obesity-related behaviors. We used a cross-sectional study design and conducted social network analysis using Exponential random graph models with three African American church-based social networks (network A, B, and C, n = 281). There were no significant network members’ similarities on BMI in the three church-based networks. One out of three networks showed similarities in fruit and vegetable consumption (network B), fast food consumption (network C), physical activity, sedentary behaviors, and alcohol consumption (network A). African Americans with a high BMI were more popular, as were individuals with greater fat intake and alcohol consumption. Our findings support the perspective that we need to improve obesity-related behaviors by targeting influential individuals and existing ties and to develop obesity interventions using social networks. The degree to which our findings varied across churches also suggests that the relationship among an individual’s obesity-related behaviors and network characteristics should be understood in the unique social context.