Objective: The objective of this study was to determine the relative efficacy of primary and secondary enforced motor vehicle occupant restraint laws on the outcomes of restraint use, crash-related mortality, and crash-related injuries. Search strategy: We used the Cochrane Collaboration search strategy to search the following electronic databases: MEDLINE, EMBASE, National Technical Information Service (NTIS), PsycINFO, ERIC, Nursing and Allied Health (CINAHL), Transportation Research Information Service (TRIS), and EI Compendex. The reference lists from each potentially eligible study were checked, and knowledgeable people in the field were contacted, for additional leads to published reports. Selection criteria: Studies had to include a comparison of primary enforcement law to no law, secondary enforcement law to no law, or a primary to a secondary law. Any study design was acceptable. Acceptable outcome measures included observed restraint use, and counts or rates of deaths or serious injuries. Data collection: Data were collected using a standard abstract reporting form. Relative differences in outcomes and absolute differences were calculated when possible. Main results: We identified 48 studies for the review. When places or time periods with primary enforcement laws were compared to those without such laws, the relative prevalence of seat belt use ranged from 1.5 to 4.5; the prevalence differences ranged from 10 to 50 per 100 observed drivers. Secondary laws had smaller effects. Two studies evaluated a change in law from secondary to primary enforcement; this was associated with an increase in belt use 6 months later of 5.3 per 100 observed drivers in Louisiana and 18 per 100 drivers in California. Primary enforcement laws were associated with a relative risk of death in MV crashes of .54 to .97. The reduction in mortality associated with secondary enforcement laws was much more modest, with relative risks estimates of .81 to 1.025. Primary enforcement laws were associated with a relative risk of severe injuries of .20 to .89; the association of secondary enforcement laws with severe injuries was smaller. Conclusions: Our review of existing studies suggests that primary enforcement laws are likely to be more effective than secondary laws. However, few studies are of good quality, and quantitative estimates of the relative effect of primary compared with secondary laws are limited.