Objectives. The present study examined the extent to which self-help interventions change specific cognitions and the extent to which changes in such cognitions are related to behavioural changes.

Design. A randomized field experiment with follow-ups after 2 weeks and 3 months.

Method. Smokers (N = 1546) were randomly assigned to one of four conditions offering smoking cessation self-help materials containing (1) outcome information, (2) self-efficacy enhancing information, (3) both sorts of information, or (4) no information.

Results. First, with regard to behavioural effects, only self-help interventions that included self-efficacy enhancing information were more effective than no information. Second, with regard to the cognitive changes, the data showed that outcome information led to increase in expected positive outcomes but also to increases in self-efficacy expectations. Further, self-efficacy-enhancing information led to increases in self-efficacy. Third, different cognitive changes between T1 and T2 were related to different types of quitting activity at T3 in different groups of smokers.

Conclusions. Some types of information lead to specific cognitive changes, while other types have more generalized cognitive effects. Further, cognitive changes produced by a self-help intervention predict future quitting activity.