Adult diabetics (N = 94) completed a questionnaire containing measures of diabetes-related self-efficacy, outcome expectancies, social support and diabetes self-care. Self-efficacy was consistently associated with self-care in the areas of diet, exercise and blood glucose testing. Outcome expectancies were correlated with exercise and blood glucose testing. The relationship between self-efficacy and blood glucose testing was moderated by outcome expectancies, such that self-efficacy had a greater effect when combined with strong beliefs in outcomes. At low levels of self-efficacy, strong outcome beliefs were associated with poorer self-care. Social support was associated with exercise self-care, and diet-specific positive family interactions were associated with better diet adherence. The relationship between social support and exercise self-care was mediated by self-efficacy. When the effects of self-efficacy were controlled, social support was no longer a significant independent predictor of self-care. The results of this study suggest that programmes designed to increase confidence in self-care abilities are likely to be effective. However, no definitive comment can be made concerning causal order from the present results. It is equally possible that a history of successful diabetes self-care builds patients' confidence in their self-care abilities. The issue of causality is a matter for future research.