| Summary: | <h4>Background</h4>There is ongoing debate about whether friends' greater similarity in Body Mass Index (BMI) than non-friends is due to friend selection, shared environments, or peer influence.<h4>Methods</h4>First-year college students (n = 104) from a southwestern U.S. university were randomly assigned roommates during the university's housing process, effectively removing friend selection effects. Participant BMI was measured up to four times (T1-T4) across 2015-2016. The influence of roommate baseline BMI (T1) on change in participant BMI over time (T2-T4) was analyzed using a linear mixed effects model adjusted for individual socio-demographics, linear time trends, baseline BMI, and physical clustering of students. A sensitivity analysis examining floormates was also conducted.<h4>Results</h4>Consistent with roommate influence, participants randomized to roommates with a higher BMI gained more weight between times T2 and T4 (β = 0.06; 95% CI = 0.02, 0.10). No shared environment effects (shared campus or floor) were found.<h4>Conclusions</h4>Randomly assigned roommates influenced each other's weight trajectories. This clarifies that BMI convergence can occur outside of friend selection or shared environments mechanisms.
|
|---|
