LAUSR

OBJECTIVE Disordered eating attitudes and behaviors are prevalent among college women, and peers appear to influence current and future eating pathology. Social network analysis (SNA) is an innovative quantitative method to examine relationships (i.e., ties) among people based on their various attributes. In this study, the social network of one sorority was modeled using exponential random graph model (ERGM) to explore if homophily, or the tendency for relationship ties to exist based on shared attributes, was present according to sorority members' disordered eating behaviors/attitudes and their body mass index (BMI). METHOD Participants included members of one sorority at a large Southeastern university. All members were included on a roster unless they elected to opt out during the consent process, and 41 (19%) of the members completed the study measures. Participants completed the Social Network Questionnaire developed for this study (degree of "liking" of every member on the roster), the Eating Disorder Examination-Questionnaire (EDE-Q), and a demographics questionnaire in exchange for one hour of community service credit. RESULTS The final sample consisted of mostly White women with an average age of 20. Homophily across liking ties was examined based on the EDE-Q Global scale, episodes of binge eating, and BMI. The greater the difference in EDE-Q Global scores, the more likely the participants were to like one another. The greater the difference in BMI, the less likely the participants were to like one another. Binge eating was unrelated to homophily. DISCUSSION College sorority women appear to prefer other women with dissimilar levels of disordered eating attitudes, suggesting complex interactions between stigmatized or valued disordered eating attributes. Women with similar BMI were more likely to like one another, confirming past findings.