This study compared the relative efficacy of four imagery techniques in increasing tolerance to cold pressor pain. Relevant pleasant, relevant unpleasant, irrelevant pleasant, and irrelevant unpleasant imagery strategies were compared in a two-way factorial design. Prior research suggested that pleasantness and relevance both affect imagery potency. This study attempted to assess the relative contribution of these two variables to increases in pain tolerance. Also investigated were the roles of several hypothesized mediating variables; namely, contextual valence, self-efficacy, treatment credibility, and involvement in imagery. The subjects were 60 female undergraduates who were randomly assigned to the four imagery groups. Two-way analysis of covariance were performed on all dependent variables, using pain threshold as the covariate. Pearons r.'s were used to test correlational hypotheses.