Stress generation theory initially posited that depression elevates risk for some stressful events (i.e., dependent events) but not others (i.e., independent events). This preregistered meta-analytic review examined whether stress generation occurs transdiagnostically by examining 95 longitudinal studies with 38,228 participants (537 total effect sizes) from over 30 years of research. Our multilevel meta-analyses found evidence of stress generation across a broad range of psychopathology, as evidenced by significantly larger prospective effects for dependent (overall psychopathology: r = .23) than independent (overall psychopathology: r = .10) stress. We also identified unique patterns of effects across specific types of psychopathology. For example, effects were larger for depression than anxiety. Furthermore, effects were sometimes larger in studies with younger participants, shorter time lags between assessments, checklist measures of stress, and for interpersonal stressors. Finally, a multilevel meta-analytic structural equation model suggested that dependent stress exacerbates psychopathology symptoms over time (β = .04), possibly contributing to chronicity. Interventions targeting the prevention of stress generation may mitigate chronic psychopathology. Conclusions of this study are limited by the predominance of depression effect sizes in the literature and our review of only English language articles. On the other hand, the findings are strengthened by rigorous inclusion criteria, lack of publication bias, and absence of moderating effects by publication year. The latter underscores the replicability of the stress generation effect over the last 30 years. Taken together, the review provides robust evidence that stress generation is a cross-diagnostic phenomenon that contributes to a vicious cycle of increasing stress and psychopathology. (PsycInfo Database Record (c) 2023 APA, all rights reserved).