This paper discusses empirical observations that indicate problems in the classification of personality disorders (e.g., the high comorbidity among personality disorders), draws conclusions from these observations regarding the nature and assessment of personality disorders, and raises further questions that concern the construct validity of the current taxonomic system. In particular, the need for clarification of trait dimensions that comprise the personality disorders is indicated. The paper explores the possibility of pooling existing data sets to investigate these questions, discusses several examples of the types of analyses that might be done, and provides some illustrative data. Finally, methodological problems in pooling data sets are considered.