Six-month (first offenders only) and 1-year and 2-year subsequent accidents and convictions were compared for various sanctions for each of the three offender groups. Aggregated county DUI recidivism and accident data were evaluated in relation to the proportions of various sanctions given by each county, the average county fine, and a sanction-stringency score derived from several variables included in a DUI court survey. First offenders receiving license restriction or suspension, either alone or in conjunction with counseling and education, had significantly fewer posttreatment accidents than did groups receiving no license control. License suspension among second offenders was significantly associated with reduced subsequent total accidents, and counseling and education plus license restriction was significantly associated with reduced subsequent alcohol-related accidents and major convictions. Longer jail terms for repeat offenders were associated with fewer subsequent alcohol-related accidents, total accidents, and major convictions. Counties imposing higher fines for repeat offenders had reduced subsequent total accidents. 21 tables and appended supplementary data