The authors tested whether dispositional pessimism would predict withdrawal from social activities among women treated for breast cancer. In a cross-sectional sample 3-12 months postsurgery, disruption of social and recreational activities (measured by the Sickness Impact Profile) correlated with concurrently assessed pessimism. This association appeared mediated by emotional distress and fatigue. A longitudinal sample was studied shortly postsurgery and over the next year. Initial pessimism predicted disruption of social activities concurrently and prospectively (3, 6, and 12 months later) but predicted change in disruption from one time to the next only at final follow-up. These associations appeared partially mediated by distress. The authors conclude that pessimism places patients at risk for adverse outcomes in several respects rather than solely with regard to emotional distress.