Over the past three decades, certain clinical factors have been identified as either positive or negative predictors of response to treatment with the tricyclic antidepressants (TCAs), and have been well reviewed in the psychopharmacology literature. The current report is an effort to focus attention on methodologic and clinical issues that have recently been given attention in this area and that may be useful to clinicians and researchers. The specific issues addressed are (a) the implications of the definitions of "response" in studies of predictive factors, (b) the importance of compliance with treatment, (c) the potentially confounding issues of severity of illness and psychoticism in depression, (d) the status of chronicity of illness as a factor in prediction of treatment response, and (e) the possible value of a definition of "atypical depression" for differential selection of treatment.