Driving at night is an additional behavioral risk factor for MVCs. Williams (1985) analyzed data from the 1977 National Personal Transportation Study, a nationally representative sample of more than 24,000 households, and the FARS data, described above, for 1976-1978. For drivers of all ages, nighttime rates of fatal crashes per miles driven were higher than the corresponding daytime rates, suggesting that driving at night is a risk factor for fatal crashes among all drivers, including adolescents. This study also indicated that adolescents do more of their driving at night than do older drivers, and, at every age, males do more of their driving at night than do females. Notably, one cross-sectional study of 706 licensed drivers (65% European American, 25% Latino, 4% African American, 4% Asian American, and 2% other; 52% male) between the ages of 16 and 20 years suggests that driving at night may be related to driving after drinking (Grube & Voas, 1996). In this sample, those who drove most often at night were more likely than were their peers to report drinking and driving.

A nighttime restriction on adolescent drivers is one component, common to many graduated licensing systems, which has been evaluated separately. In one study regarding curfew laws, Preusser, Williams, Zador, and Bloomberg (1984) assessed the patterns of police-recorded MVCs among 16 year-old drivers in several matched states with and without curfew laws. New York and Pennsylvania, each with curfew laws, were compared with Ohio, a state without a curfew law. Similarly, Louisiana was compared with Mississippi. The pattern of MVCs in Maryland was compared before and after implementation of a curfew law. Primarily in New York and Pennsylvania as compared with Ohio, the presence of a curfew law was associated with a decreased risk of MVCs during the curfew hours. This decreased crash involvement was observed among 16-year-old drivers but not among older drivers. These findings suggest that curfew laws may be variably associated with decreased crash risk, but limitations in the study design should inspire careful interpretation of the findings. The analyses were not multivariate and the effects of the curfew laws in New York and in Pennsylvania were compared only with a single comparison state (Ohio).

In a second study of curfew laws, Preusser, Zador, and Williams (1993) compared data on fatal MVCs from the 1984-1990 FARS, described above, for cities with and without curfew ordinances for adolescent drivers ages 17 and younger. All of the 124 cities included in the study had a population of 100,000 or larger, and none was covered by a statewide curfew law. Cities with curfew ordinances had fewer nighttime highway fatalities among adolescents ages 13 to 17 years than did cities without such ordinances. During the daytime, there were no significant differences in adolescent highway fatalities across cities with and without curfew laws.