We report a subgroup analysis of 24 out of 42 subjects who were hospitalized for non-psychotic major depressive disorder and who agreed to participate in interviews at admission and 2 years afterwards (as reported previously by Domken, Scott, & Kelly, 1994; Bothwell & Scott, 1997). At 2 year follow-up, these 24 subjects were categorized according to established criteria into clients meeting criteria for full remission (FR; n=9) and those meeting criteria for partial remission (PR; n=15). The most striking findings were that, over time, PR subjects showed significant loss of self-esteem and showed greater divergence in self-ratings compared to observer ratings of their depressive symptoms, whilst the same ratings in the FR group changed in the opposite direction. We suggest that the persistence of depression in PR subjects may provide evidence to support Teasdale’s (1988) hypothesis that some individuals “get depressed about being depressed”. The research and clinical implications of the results are noted.