9 Analytic Plan

Aim 1: Confirm the feasibility and acceptability of a fatphobia and weight stigma intervention. We will confirm the feasibility and acceptability of our intervention structure and content via (a) study attrition at all timepoints, (b) participant attendance to the 4 sessions, (c) completion rates of study activities both prior to and following sessions, and (d) intervention adherence and facilitator competence. To assess intervention adherence and facilitator competence, sessions will be audio recorded and reviewed by two members of research staff using a 12-item leader competence assessment. Over the course of the study, we expect study attrition to be <20%, attendance at all 4 sessions to be >80%, and completion rates of study activities to be >75%. Feasibility will be benchmarked via two primary outcomes: (a) we expect high levels of intervention adherence in the intervention, with high interrater reliability and average treatment fidelity and facilitator competence ratings at or above 8.0 on a 10-point scale, and (b) we expect to meet recruitment goals (N = 30).

Aim 2: Estimating intervention effects. Reductions in anti-fat bias will be assessed via changes in self-report fatphobia and anti-fat bias as well as shifts in IAT and AAT task performance from baseline to post-intervention and baseline to 4-week follow-up. To evaluate changes over time for the singular group, we would use a paired t-test to assess if the intervention changed participants’ cognitions about fatphobia and weight stigma.

Aim 3: Eating Disorder Prevention. Reductions in eating disorder symptomology will be assessed via changes in self-report eating disorder behaviors and cognitions from baseline to post-intervention and baseline to 4-week follow-up. We will specifically assess how the intervention has affected participants that have been identified as demonstrating disordered eating (ED) at baseline. To evaluate changes over time between the ED and healthy group, we will use mixed effects modeling, estimating fixed effects of Condition, Time, and a Time x Condition interaction, with a random intercept of participant to account for baseline differences.