A significant amount of information about perinatal drug abusers was obtained using CPA's initial test protocol. In particular, the need for routine cognitive assessment was identified. Although the initial program curriculum was too advanced for many program participants, knowledge about the level of individual functioning allowed for appropriate adjustments to be made. At the same time, most participants tested in the normal range of intellectual ability and in-depth intellectual assessment would have been an unnecessary expense. Objective, self-report personality inventories seemed more useful than structured clinical interviews in assessing perinatal drug abusers. The almost exclusive focus of the CPA on measuring psychiatric domains, however, was a mistake. This emphasis neglected such important constructs as learning disabilities, social supports, violence, and parenting experiences. In addition, the initial test protocol overlooked alcohol or other drug responses and cognitions about alcohol or other drug use. General measurement issues relevant to the psychosocial evaluation of perinatal drug abusers are reviewed, including reliability, validity, standardization, and test administration issues. The development of the CPA's addiction testing protocol is described with respect to cognitive functioning, psychiatric diagnosis, personality functioning, addiction and psychiatric severity, global psychiatric functioning, and family functioning. Future directions in the assessment of perinatal drug abuse are noted that focus on neuropsychological functioning, violence, social supports, parenting experiences, and drug use. 50 references