Six California communities were selected to participate in the study on the basis of comparability and isolation from each other. Four of the communities' police departments implemented programs of sobriety checkpoints: the checkpoint configuration varied in terms of staffing level (three to five officers vs eight to twelve) and mobility of the checkpoints (remaining in one location for the evening vs three frequential locations within the city). The fifth community's police department implemented a program of aggressive roving patrols that focused on DWI enforcement. The sixth community refrained from implementing any special DWI enforcement effort for the duration of the project and served as the experimental comparison site; state-wide totals provided additional comparison. The level of effort devoted to the roving patrols was equal to the officer hours required to operate the highstaffing level checkpoints. The California Office of Travel Safety provided each of the checkpoint departments with a trailer and equipment-set necessary to conduct their programs of frequent checkpoints (18 in a nine-month period). Committees of concerned local citizens were organised to develop and implement vigorous public information and education programs to support the special enforcement efforts. Crash, arrest, and BAC data were obtained from the participating police departments and a state reporting system; and, data regarding public awareness of the programs and perceived risk of arrest were obtained through a survey conducted at local Department of Motor Vehicles offices. There were no significant differences in the decline in alcohol involved crashes among the four configurations of checkpoints tested in this study. Thus, decisions regarding an optimum checkpoint configuration can be made on the basis of other factors, including cost, traffic volume, and demographics. Further, the checkpoint communities experienced declines in the proportions of alcohol-involved crashes of 43, 32, 19, and 16 percent, while the state wide decline for communities was only eight percent; the proportion in the roving patrol community declined by five percent. Paired samples analysis found a statistically significant reduction in alcohol-involved crashes in one of the sobriety checkpoint programs, and for all of the checkpoint programs when data from the four checkpoint communities were combined. Logistic regression analysis indicated alcohol involved crashes declined significantly in the checkpoint sites, and did not change significantly at the comparison site during corresponding periods. Comparing with state-wide data, the checkpoint communities' decline was more than 3 times greater. (Author/publisher)
Samenvatting