The purpose of this study was to evaluate Widiger, Trull, Clarkin, Sanderson, and Costa's (1994) facet-level predictions about depressive personality disorder, thereby evaluating the validity of the depressive personality disorder construct. A sample of 67 patients was collected from two treatment facilities. Participants completed three different measures of depressive personality disorder and the NEO-PI-R (Costa & McCrae, 1992). All measures of depressive personality disorder were significantly correlated with three of the four predicted facets: Anxiety, Depression, and Self-Consciousness. All three measures were significantly correlated with the Vulnerability, Warmth, Gregariousness, Assertiveness, Positive Emotions, Actions, Trusts, and Achievement Striving facets. A series of regression analyses was performed. Two of the predicted facets--Self-Consciousness and Tendermindedness--were able to uniquely account for variance in all three measures of depressive personality across most analyses. However, there was a lack of specificity across other facets in predicting the depressive personality disorder. A facet-level understanding of depressive personality disorder appears to be empirically justified, and the validity of the depressive personality disorder further supported. Additional work is needed on improving the measurement of depressive personality disorder.