This article reports two studies assessing the influence of self-efficacy, outcome expectancies and aversive feedback on different aspects of adherence. Study 1 employed a computer simulation of physiotherapy to test experimentally the effects of aversive feedback (i.e., loud noise) experienced during simulated therapy on adherence behaviour in a student population. Study 2 examined whether similar effects of aversive feedback (i.e., pain) experienced during physiotherapy in a clinical setting would be observed in a longitudinal questionnaire study of predictors of adherence. In both studies, self-efficacy and outcome expectancies were assessed at baseline and after experience of the task (performing simulated or actual physiotherapy). Study 1 found that self-efficacy and outcome expectancies predicted persistence with simulated physiotherapy (i.e., completing the experimental session), whereas aversive feedback influenced adherence during sessions (i.e., correct response rate). Study 2 found that self-efficacy and outcome expectancies predicted persistence with actual physiotherapy (i.e., completing the prescribed number of sessions). Aversive feedback and outcome expectancies influenced adherence during sessions. We conclude that different factors predict different aspects of adherence behaviour. It is therefore important to measure both persistence over time and adherence during sessions, and to investigate the predictors of each dimension of adherence.