Low levels of neuroticism have variably been associated with a better outcome of depressive disorders but the interpretation of this phenomenon is clouded by the knowledge that severity of depression may itself confound ratings of neuroticism. To clarify any relationship between neuroticism and outcome, the authors assessed the predictability of neuroticism (as assessed by questionnaire, psychiatrist rating, and subject self-report) in separate groups of psychiatric patients and symptomatic volunteers with nonmelancholic depression. Depressives' judgments of their own neuroticism correlated with neuroticism scores on the Eysenck Personality Inventory (EPI), but neither of these measures predicted improvement at 6 nor 20 weeks. Judgments of neuroticism made by psychiatrists, which did not correlate with EPI neuroticism scores, did predict improvement. The study suggests that differences in defining neuroticism contribute to its variable association with outcome and further suggests the relevance of clinically assessing neuroticism as a personality variable.