Does perceived behavioural control mediate the relationship between power beliefs and intention?

OBJECTIVES: In order to determine whether the relationship between power beliefs (Sigma(p)) and health-related behavioural intentions is mediated by perceived behavioural control (PBC) we used structural equation modelling of eight cross-sectional data sets. METHOD: Eight studies that examined health-related behaviours and employed representative samples totalling N = 4663 participants were analysed. All studies involved power belief items derived from pilot testing and employed standard multi-item measures of power beliefs, PBC and intention that were highly reliable. RESULTS: Confirmatory factor analysis confirmed the discriminant validity of power beliefs, PBC and intention. Structural equation modelling of relevant paths indicated that PBC only partially mediated the relationship between power beliefs and intention (Z(Sobel) = 5.15, p < .001; Z(Baron&Kenny) = 5.16, p < .001). Power beliefs had a significant direct relationship with intention even after PBC had been taken into account. CONCLUSION: The findings undermine Ajzen's contention that PBC mediates the power beliefs-intention relationship and suggests that it is important to employ measures of power beliefs in addition to measures of PBC in order to enhance the prediction of intentions to perform health-risking, or health-promoting, behaviours.