A good deal of research now indicates that the personality dimension of optimism-pessimism plays an important role in a wide range of behavioral and psychological outcomes when people confront adversity (reviewed in Scheier & Carver, 1992). What is less clear is the mechanism (or mechanisms) by which the beneficial effects of optimism take place. One possibility is that optimists do better than pessimists because they cope more effectively.' There is an abundance of evidence that they at least cope differently. Optimists and pessimists differ from one another in reports of their general coping tendencies (Carver, Scheier, & Weintraub, 1989) and in the coping responses they bring to mind when considering hypothetical situations (Scheier, Weintraub, & Carver, 1986), recalling a stressful situation from the recent past (Scheier et al., 1986), dealing with infertility problems (Litt, Tennen, Affleck, & Klock, 1992), managing a life transition (Aspinwall & Taylor, 1992), coping with a serious disease (Friedman et al., 1992), and dealing with worries about specific health threats (Stanton & Snider, 1993;Taylor et al., 1992).Far less information is available, however, concerning the hypothesis that these differences in coping serve as the vehicle by which optimists experience better eventual outcomes. Three studies in the literature are relevant to the question. One of them (Scheier et al., 1989) examined men undergoing coronary artery bypass surgery. These subjects did not complete a full measure of coping but indicated their use of several cognitive-attentional strategies before and after the surgery. Although optimism was related to several of these strategies, there was scant evidence that the strategies mediated the beneficial effect of optimism on subsequent outcomes. The second study (Aspinwall & Taylor, 1992) assessed optimism and coping in a group of students entering college and assessed well-being 3 months later. In this case, the beneficial effects of optimism appeared to operate at least in part through differences in both active coping and avoidance coping.Both of these studies have an important limitation, however. Neither included an initial measure of the variables that served as the later outcome measure. Thus,