OBJECTIVE: To describe an innovation in performance contracting for substance abuse services in the State of Maine and examine data on measured performance by providers before and after the innovation. DATA SOURCES AND COLLECTION: From the Maine Addiction Treatment System (MATS), an admission and discharge data set collected by the Maine Office of Substance Abuse (OSA). The MATS data for this study include information on clients of programs receiving public funding from October 1, 1989 through June 30, 1994. Additional data are drawn from the contracts between the state and providers, and from service delivery reports submitted to OSA. STUDY DESIGN: Client-level performance measures were calculated directly from MATS using OSA's formulas and standards, and then aggregated to the treatment program level. Multivariate regression analysis was done for each performance indicator as a dependent variable with performance contracting, time, extent of state funding, and provider characteristics as independent variables. PRINCIPAL FINDINGS: Performance contracting is positively related to better performance for effectiveness indicators overall. Individual effectiveness indicators that showed improvement include drug use indicators (abstinence and reduction in use) and social functioning indicators. In addition, performance contracting is associated with an increase in efficiency performance, defined as delivery of the contracted amount of service, for agencies that depend heavily on OSA for funding. Finally, performance contracting appears unrelated to the special populations indicators that measure services to target populations that OSA considers harder to treat. CONCLUSIONS: There is tentative evidence of a relationship between provider performance and the introduction of performance contracting. More definite conclusions await more detailed analyses of client-level data.