Data from three samples of adults (Ns = 571, 472, and 989) and a sample of adolescents (N = 1,710) supported the possibility that the prevalence of major depression has been increasing in recent birth cohorts, a phenomenon labeled the age-cohort effect (ACE). A significant ACE for relapse was also found in 1 of the adult samples. In addition, early onset age in the adults (prior to age 25) tended to be associated with relapse. Adults in recent birth cohorts were also found to show an elevated prevalence of other disorders. We examined the power of 4 variables (current mood state, social desirability response bias, labeling, and time interval between the episode and the diagnostic interview) to produce these results without an actual increase in the rate of mental disorder. With 1 exception (labeling), the variables were significantly associated with reports of past episodes of disorder and with birth cohort. Controlling for their influence, however, did not reduce the ACE.