The tendency to engage in globally negative self-evaluations (i.e., overgeneralization) is quite prevalent in depression but the precise nature of this tendency is little understood. This research examined whether the self-evaluations of dysphoric individuals differed in ''degree'' or ''kind'' from those of nondysphoric individuals. Dysphoric and nondysphoric students imagined themselves in various hypothetical situations and rated how they would evaluate themselves an 49 trait dimensions that varied in globality. Subjects also rated how depressed or happy each situation would make them feel. Subjects then estimated how they would feel and evaluate themselves 3 days after the original negative events. Results showed that the self-evaluations of dysphoric and nondysphoric subjects differed in both degree and kind. That is, dysphoric subjects rated themselves more negatively than nondysphorics across both global and specific traits, but this negativity was particularly pronounced for global traits. Moreover; supporting the position that globally negative self-evaluations influence the persistence of a depressive episode, globally negative thinking by dysphoric subjects predicted their estimated mood 3 days after the imagined negative events, as well as predicting their actual recovery from depression over the course of a 5-week period. Results are discussed as supporting Teasdale's (1988) differential activation hypothesis regarding the onset and persistence of depression.