Most of the predictive research has beencross-sectional. In one of the fewlongitudinal studies on behaviour and self-efficacy that has been conducted, the issueof direction of causality was raised:behaviour caused self-efficacy rather thanthe reverse. This obviously calls for morelongitudinalresearch,especiallytoestablish the direction of causality of thevariables. Furthermore, causality can bestbe established by using an experimentalparadigm with random assignment ofindividuals to groups and attempts atmanipulation of the underlying constructs.The literature in this area would also benefitfrom experimental basic research