Research on the impact of managed care on access, quality, outcomes, and costs of mental health and substance abuse services has been limited because of the lack of generalizability of studies, the "black box" focus of much of the research, the dynamic changes in mental health and substance abuse managed care, and the proprietary nature of the health care industry. This article provides a framework for understanding the organizational, financial, and procedural features of health plans and the effect of these features on the characteristics and flow of patients through health plans and the selection and utilization of treatments. The diverse research priorities of key stakeholders, i.e., public and private purchasers, managed care organizations, providers, and patients and their families, are described along with a broader societal agenda for delineating the outcomes of health care plans. Critical research and methodologic issues in studying the effects of managed care are outlined, including issues related to identifying and selecting appropriate outcome measures and developing appropriate methods for risk adjustment to adequately control for patient selection bias. This article asserts that government, purchasers, health care plans, providers, consumers, and researchers must collaboratively develop resources and research approaches to fully evaluate the effects of managed care. To realize this objective, auspices with reasonable objectivity are needed along with access to necessary data within the black box of health care systems, a cadre of trained investigators, and sufficient research funding, including the development of an all payers fund to support clinical and health services research.