It has been suggested that the low level of effectiveness of youth interventions is due to a lack of knowledge regarding the mechanisms responsible for behaviour change. The identification of behaviour mediators is necessary for the progression of physical activity research, as it allows researchers to determine which components of an intervention are responsible for mediating behaviour change. The purpose of this study was to identify mediators of behaviour change in a physical activity intervention for senior school students. Participants (n=78) were randomly allocated to control or intervention conditions for a period of 10 weeks. Moderate-to-vigorous physical activity (MVPA) and potential mediators were assessed at baseline and post-intervention (10 weeks). Hypothesized mediators were derived from Bandura's Social Cognitive Theory and included: peer support, exercise self-efficacy and outcome expectancy. Mediation was assessed using the product-of-coefficients test described by MacKinnon and colleagues, based on the criteria for mediation identified by Baron and Kenny. While none of the variables satisfied all four criteria for mediation among males or females, self-efficacy was able to satisfy the first three criteria among females in the study. Exercise self-efficacy may be a mediator of physical activity behaviour in adolescent girls.