This study reviewed all available published studies that met certain criteria. The studies selected measured recidivism through traffic records; studies that used only self-reports or other subjective measures were excluded. In all the studies reviewed, the number of suspended drivers exceeded 500; and the suspension assessed was full suspension, not a restricted, "hardship," "occupational," or other partial suspension. License suspension is expected to reduce recidivism among drivers convicted for DWI through incapacitation, that is, preventing driving while on suspension, and deterrence, that is, preventing future instances of drinking and driving. When compared with a lesser restriction (including no restriction), license suspension has a greater incapacitation effect but no greater deterrent effect. When license suspension is compared with remediation (education, counseling, treatment), suspension has the greater incapacitation effect, but remediation evidences the greater deterrent effect. Based upon the review of relevant studies, this paper concludes that the effect of suspension upon DWI is primarily to limit driving, and hence exposure to convictions and accidents of all kinds. Remedial programs, coupled with some driving instruction, are apparently more effective in deterring drinking-driving behavior itself. 2 tables, 2 figures, and 11 references