There is some evidence to suggest that graduated licensing may have a desirable influence on MVC risk among the new adolescent drivers affected by the law. For example, Ulmer, Preusser, Williams, Ferguson, and Farmer (2000) followed the MVC experience in Florida from 1995 to 1997, to assess the effect of a graduated licensing program, implemented in Florida in 1996. The authors compared changes in Florida crash rates, from pre- to post-implementation, with changes over the same time period in the crash experience in Alabama, a state with no graduated licensing program. The Florida graduated licensing program required 15- to 17-year-old drivers to have a learner's permit for six months. The law disallowed driving between 7 p.m. and 6 a.m. during the first three months of this learner's permit period and disallowed driving between 10 p.m. and 6 a.m. for the second three months. Driving between 11 p.m. and 6 a.m. was prohibited for 16-year-old drivers and unsupervised driving between 1 a.m. and 5 a.m. was prohibited for 17-year-old drivers. Penalty points for moving violations had an enhanced effect for these young drivers, who could lose their licenses after many fewer infractions than under the former rules. Finally, a "zero tolerance policy", meaning a maximum allowed blood alcohol concentration of zero, was put in place for drivers under the age of 21 years. In Florida, between 1995 and 1997, there was a decline in the ratio of the MVC rate for 15- to 17-year-olds to that for drivers ages 25 to 54. Over the same time span, there were no significant changes in the relative MVC rate among Florida 18-year-olds, who were unaffected by graduated licensing, or for adolescents of any age between 15 and 18 years in Alabama. The observed declines in MVC risk among adolescents ages 15 to 17 years in Florida were more pronounced for white than for nonwhite adolescents and were greatest for crashes in urban rather than rural areas, and for crashes occurring at night rather than during the day. These changes were not associated with a decline in licensure, suggesting that some component(s) of the graduated driver licensing program itself may be responsible for the desirable change in MVC risk. However, it is unknown which of the many components of the system are effective.