An understanding of physical activity through mediators of behaviour change is important to evaluate the efficacy of interventions. The purpose of this review is to update prior reviews with meta-analysis to evaluate the state of physical activity interventions that include proposed mediators of behaviour change. Literature was identified through searching for five key databases. Studies were eligible if they described a published experimental or quasi-experimental trial in English examining the effect of an intervention on physical activity behaviour and mediators in non-clinical adult populations with the necessary statistical information to be included in the meta-analytic structural equation modelling analysis. Fifty-one articles (49 samples) met the eligibility criteria. Small overall effects were identified for mediation paths a (r = .16; 95% CI = .10 to .22), b (r = .21; 95% CI .16 to .27), and c (r = .24; 95% CI .12 to .35), c′ (r = .05 to .19) and ab (r = .02 to .07) that showed similar findings by theory and construct. The effect sizes seen in physical activity interventions are mediated by our current theories, but the effects are very small and no one construct/theory appears to be a critical driver of the mediated effect compared to any other. Innovation and increased fidelity of interventions is needed. © 2020 Informa UK Limited, trading as Taylor & Francis Group.