How confident are we that social support mediates changes in walking behaviour? Well, it depends ...

Abstract

To enhance the effectiveness of physical activity interventions, it is important to identify the mechanisms (mediators) through which changes in physical activity occur. Exercise-related social support from family and friends is one of the hypothetical mechanisms of physical-activity change. Data from a randomized controlled trial (N=52) were used to examine the evidence of a mediating effect of social support on changes in walking. Inactive adults were randomized into either a print or a print-plus-telephone intervention. Walking and exercise-related social support were assessed at baseline, after the intervention and 4 weeks later. Four methods of mediation analysis were used to establish whether social support explained the effects of the intervention on initial behavior change and on maintenance of change. These were the Baron-Kenny method, the FreedmanSchatzkin method, MacKinnon’s product-of-coefficients test based on the empirical distribution of estimates, and the bootstrap product-of-coefficients test. Sufficient evidence for a mediating effect of social support on initial changes in walking was provided by all approaches but the Baron-Kenny method. There was insufficient support for a mediating effect on maintenance of walking. The strength of evidence that a theoretical construct is a mediator of physical activity change depends in part on the power of the statistical approach to detect a mediating effect of a certain size. The choice of method is especially important when analyzing distal mediating effects, effects of small-to-moderate size, and data from small-scale trials