A large scale intervention program conducted by the Virginia Department of Motor Vehicles to increase safety belt use in eight communities (cities, towns, and counties) throughout the Commonwealth of Virginia was evaluated. The police or sheriff's department in each community conducted either one, two, or three repeated applications (waves) of public information/education programs. Each target community was matched with a non-treatment control community with a similar population.

The overall project was very successful. The mean percent safety belt use for the eight target communities began at 52%, and significantly increased to 73% by the end of the project. The mean percent safety belt use for the eight control communities also began at 52%, and dropped insignificantly to 48% by the end of the project.

To extend previous work on the relationship between intervention agents and intervention effectiveness, it was hypothesized that there would be a positive relationship between mean percent safety belt use increase and the agent-to-target ratios (the proportion of police officers or deputies participating in the program per number of citizens in the community), citation-to-target ratios (the proportion of citations or written warnings given during the program per number of citizens in the community), and promotional gimmick-to-target ratios (the proportion of promotional gimmicks given during the program per number of citizens in the community). Each of these ratios were significantly correlated with the percent safety belt use increase for the eight target communities. The terminal model for a stepwise regression procedure, which included the agent-to-target and citation-to-target ratios, yielded an R² of .83.

A plan for institutionalizing safety belt programs that would reduce many of the problems associated with motivating large scale safety belt use is suggested, as well as suggestions for future research.