The Alcohol Safety Action Program was introduced in the early 1970s as a comprehensive systems approach for reducing alcohol-related automobile crashes. This dissertation gathers evidence and insights helpful to planners, evaluators, policy-makers, and program implementors. Specifically, A.S.A.P. “Level II” effectiveness in reducing the recidivism rate among program participants was examined. A combination of quantitative and qualitative assessments of the program was performed to gain in-depth insight and to determine which program elements seem associated with its success or failure. Quantitative analysis emphasized A.S.A.P. and non-A.S.A.P. participant two-year recidivism rates. Participants from two Virginia localities in 1977-80 were examined. Independent variables commonly held by both types of participants are age, sex, court delay, and prior offense records. Variables unique to each program were also examined. Qualitative insights were gathered through interviews with current A.S.A.P. participants prior to and following program completion, past program participants, and course instructors. An approximate three-to-one difference in recidivism rate was found between A.S.A.P. and non-A.S.A.P. participants. Variables significantly related to recidivism were prior D.W.I. offenses, court delay, prior reckless driving offenses, and age. Differences based on location were also found. The interviews demonstrate basic satisfaction with the course. Participants cited the overall arrest experience as having the largest impact, with the course providing supportive information. The factor emerging to deter future behavior was the negative experience--the “hassle”--associated with the D.W.I. offense. Overall, it appears that the blend of the educational and punitive approaches makes the A.S.A.P. program more effective than the alternative approaches being used. Specific recommendations emerging from the research are of four general types: administrative mechanisms, laws and policies, the A.S.A.P. course, and evaluation. / Ph. D.