We examined the efficacy of three Driving While Intoxicated (DWI) programs in New Jersey from 1980 through 1985, using covariance analysis of county data. Road blocks, the major component of the Strike Force program, were associated with a drop of 10-15 per cent in the single vehicle nighttime crash rate and showed a relatively stable effect over time. DWI Task Force, an education program, was associated with a 6-10 per cent total decline in the crash rate and declining impact over time. SOBER, another education program, was associated with a small effect in the first year and little or no effect thereafter.