Scales and protocols developed by expert panels were used to code study characteristics, including dimensions of methodological quality. Results revealed that better methodological quality was associated with smaller effect size and less variation in effect size. The average effect of remediation on drinking and driving recidivism was an 8- to 9-percent reduction in comparison to no remediation, in the studies that used adequate methods. A similar effect size existed for traffic crashes that involved alcohol. However, licensing actions tended to be associated with a reduction in the occurrence of traffic crashes that were not related to alcohol. Exploratory regression analysis and confirmatory within-study analysis suggested that combinations of modalities that included education, psychotherapy/counseling, and follow-up contact/probation were more effective than other evaluated methods of reducing recidivism. Findings also suggested that the literature overestimates treatment effects due to overemphasis on education for a treatment for all offenders and the use of drinking/driving recidivism as the most frequent outcome measure. Tables and 24 references (Author abstract modified)