Journal of Abnormal Psychology 1999, Vol. 108, No. 3, 483-489 Copyright 1999 by the American Psychological Association, Inc. 0021-843X/99/$3.00 First Onset Versus Recurrence of Depression: Differential Processes of Psychosocial Risk Peter M. Lewinsohn, Nicholas B. Allen, and John R. Seeley Oregon Research Institute Ian H. Gotlib Stanford University Differential risk factors for the onset of depression were prospectively examined in a community-based sample of adolescents (N = 1,709), some of whom had a history of major depressive disorder (MOD; n = 286) and some of whom did not (n = 1,423). From the theories of J. Teasdale (1983, 1988) and R. Post (1992) concerning the etiology of initial versus recurrent episodes of depression, the authors hypothesized that (a) dysphoric mood and dysfunctional thinking styles would be correlated more highly among those with a previous history of MDD than among those without a history of MDD; (b) dysphoric • moo d or symptoms and dysfunctional thinking would be a stronger predictor of onset of recurrent episodes (n = 43) than of first onsets (n = 70); and (c) major life stress would be a stronger predictor of first onsets of MDD than of recurrent episodes. The results provide support for the 3 hypotheses and suggest that distinct processes are involved in the onset of first and recurrent episodes of MDD. Over the past 2 decades, considerable research has been con- ducted examining psychosocial correlates of depression. The re- sults of these studies are important for our understanding of the phenomenology of depression and have contributed to the devel- opment of theories concerning the cognitive and social functioning of depressed individuals. Fewer investigations have attempted to identify and elucidate risk factors for depression. The identifica- tion of risk factors for depression is critical both for the develop- ment and evaluation of theories of the etiology of depression and for identifying individuals who are at elevated risk for developing this disorder. Of interest is that recent theories of depression highlight the importance of examining risk factors for depression and further- more suggest that different risk factors or processes may be im- plicated in first versus subsequent onsets of this disorder. Given that depression is a highly recurrent disorder (Belsher & Costello, 1988; Lewinsohn, Zeiss, & Duncan, 1989; Mueller et al., 1996), the identification of specific processes and factors associated with risk for first onsets versus recurrent episodes will have important implications not only for understanding the etiology of these states and how depression evolves over the course of recurrent episodes but also for prevention efforts aimed at relapse and recurrence of depression among individuals with a history of this disorder. Teasdale's (1983, 1988) differential activation hypothesis, for example, proposes that vulnerability to severe depressive states is influenced by patterns of information processing that occur during Peter M. Lewinsohn, Nicholas B. Allen, and John R. Seeley, Oregon Research Institute, Eugene, Oregon; Ian H. Gotlib, Department of Psychol- ogy, Stanford University. This research was partially supported by National Institute of Mental Health Grants MH40501 and MH50522. Correspondence concerning this article should be addressed to Peter M. Lewinsohn, Oregon Research Institute, 1715 Franklin Boulevard, Eugene, Oregon 97403-1983. Electronic mail may be sent to pete@ori.org. mild states of dysphoria. Dysphoric mood states are hypothesized to activate negatively biased interpretations of experience, as ex- emplified by Beck's dysfunctional attitudes (Beck, Rush, Shaw, & Emery, 1979) and other negative biases in information processing (e.g., Gotlib & MacLeod, 1997; Williams, Watts, MacLeod, & Mathews, 1988). These dysfunctional cognitive processes, in turn, are hypothesized to maintain and exacerbate the dysphoric mood state, often leading to clinical states of depression. Consistent with these hypotheses and with Bower's (1981) formulation of the effects of mood on memory, a number of empirical investigations have demonstrated that depression is associated with a strong bias for the retrieval of negative material from memory (e.g., Gilboa & Gotlib, 1997; Teasdale & Barnard, 1993). Williams et al. (1988) have similarly postulated that negative self-referential material is extensively rehearsed and elaborated during depressive episodes, strengthening the likelihood that this material might contribute to a recurrence of depression, even though it may not have been implicated in the onset of the initial depressive episode. Teasdale (1983, 1988) has underscored the importance of dif- ferentiating between first onsets and relapse episodes of depression and has proposed that the links between depressogenic information-processing styles and dysphoric affect will be stronger among individuals who have previously experienced a depressive episode than among previously nondepressed persons. Empirical results consistent with this formulation have been reported by Teasdale and Dent (1987) and by Miranda and her colleagues using the Dysfunctional Attitudes Scale (DAS; Weissman & Beck, 1978; e.g., Miranda & Persons, 1988; Miranda, Persons, & Byers, 1990; but see also Rosenbaum, Lewinsohn, & Gotlib, 1996). To date, however, no study has prospectively compared the roles of depressive mood/symptoms and dysfunctional thinking in first onsets and recurrences of depression. Consistent with Teasdale's hypothesis and with Bower's (1981) model, we predicted that the association between depressive symptoms and dysfunctional thinking would be stronger for adolescents with a previous history 483