Objective: To test the efficacy of a high-intensity social norms media intervention to reduce the prevalence of driving after drinking among 21-to-34-year-olds living in western Montana.

Method: The efficacy of a high-intensity social norms media campaign was tested in this quasi-experimental controlled design. A baseline survey collected self-reported data on the target population's behavior with respect to impaired driving, as well as on its perceptions of the behavior of their peers. Normative messages and media were developed from these data. Approximately half of the 21-to-34-year-olds in the State reside in 15 of its western counties. These Montana counties served as a high-dosage paid social norms media intervention area and were exposed to high levels of campaign television, radio, newspaper, billboard, and movie slide advertisements; 26 eastern counties served as a low-dosage control environment and were exposed only to low levels of free media (television and radio public service announcements) and paid newspaper advertisements. Promotional items bearing the campaign message were distributed statewide. In order to make the paid media intervention as powerful as possible, fear-producing, deterrent-based media efforts were eliminated or severely restricted in the treatment counties.

Random samples of the target population were surveyed a total of four times. At Time 1 (November 2001) and Time 2 (November 2002) 1,000 respondents were surveyed. At Time 3 (March 2003) 1,005 respondents were surveyed; at Time 4 (June 2003) a reduced sample of 517 respondents was surveyed. Ten- to twelve-minute telephone interviews were conducted at each time point by trained interviewers through a computer-assisted telephone interviewing laboratory. Each survey gathered information on respondents' exposure to the campaign message, and on their perceptions and reported behaviors regarding driving after drinking.

Results: The campaign successfully reduced the target population's misperceptions of the frequency of impaired driving among their peers. When compared to the control counties, follow-up surveys found a 7.5-percent relative decrease in the percentage who believed the average Montanan their age drove after drinking during the previous month and an 11.0-percent relative increase in the percentage who accurately perceived that the majority of their peers use a non-drinking designated driver. The change in perceptions was associated with a change in reported behavior. In the target area there was a 13.7-percent relative decrease in the percentage that reported personally driving after drinking and a 15.0-percent relative increase in the percentage that reported always using non-drinking designated drivers. The campaign also affected attitudes towards impaired driving enforcement policy. Target county residents reported a 16.5-percent relative increase in the percentage who would support passing a law to decrease the blood alcohol concentration (BAC) legal limit for driving to .08 percent from .10 percent.

Conclusions: A high-intensity paid media social norms intervention can be successful on a statewide scale, across a variety of measures including perceptions, reported behaviors, attitudes, and support for policy. Self-reported surveys are a reliable and widely used method of data collection. However, the results of the data reported by respondents could not be corroborated with changes in BAC of arrested drivers or numbers of alcohol-related fatalities due to accessibility and availability of BAC data of arrested drivers.