A meta-analysis of 17 studies that have evaluated the effects on traffic safety of using daytime running lights (DRL) on cars is presented. A distinction is made between studies that have evaluated the effects of DRL on the accident rates of each car using it and studies that have evaluated changes in the total number of accidents in a country following the introduction of mandatory use of DRL. Three different definitions of the measure of safety effects are compared and their validity discussed. It is concluded that the use of DRL on cars reduces the number of multi-party daytime accidents by about 10-15% for cars using DRL. The estimated effects on the total number of accidents of introducing DRL laws are somewhat smaller, 3-12% reduction in multi-party daytime accidents, and are likely to contain uncontrolled confounding effects. There is no evidence to indicate that DRL affects types of accident other than multi-party daytime accidents.