Performing and publishing mediator analyses, whether significant or null, provides insight into where research efforts should focus and will assist in developing effective and powerful behaviour change interventions. The present study examined whether self-efficacy, social support and church support mediated changes in leisure-time physical activity (PA) and fruit and vegetable (F&V) consumption in a faith-based intervention.

A 15-month PA and F&V intervention, guided by the structural ecological model, targeted the social, cultural and policy influences within the church. Outcomes and mediators were measured at baseline and follow-up. Data were collected from 2007 to 2011. MacKinnon’s product of coefficients tested for mediation.

Sixty-eight African Methodist Episcopal churches in South Carolina, USA.

Five hundred and eighty-two (PA) and 588 (F&V) church members.

Despite the significant increases in PA and F&V consumption, none of the hypothesized mediators were significant mediators of change in PA or F&V consumption. When examining each path of the mediation model, the intervention did not change any of the hypothesized mediators. However, changes in some mediators were associated with changes in outcomes.

Although there was no significant mediation, the association between changes in mediators and changes in PA and/or F&V consumption suggest that these variables likely play some role in changing these behaviours. Future studies should consider mediation analyses a priori, putting careful thought into the types of measures used and the timing of those measures, while also being cognizant of participant and staff burden. Finding a balance will be fundamental in successfully understanding how interventions exert their effects.