Anti-drunk-driving roadblocks set up by the California Highway Patrol during the Christmas holidays had an apparently significant effect--cutting alcohol-related crashes and increasing arrests--according to figures obtained from the patrol and analyzed by a prominent statistician Thursday at the request of The Times.
But, while the roadblock program seemed to have produced some of the results its advocates claimed it would, they were not nearly as dramatic as many proponents said they expected.
The effects of the program--involving what have been euphemistically called "sobriety checkpoints"--was further clouded by two potentially significant factors: drunk-driving fatalities increased, compared to the same 17-day period in 1983, in the four Highway Patrol areas that used roadblocks. And several experts agreed that if the roadblock program had actually functioned as a true deterrent to drunk driving, arrests should have dropped in the affected areas, but they didn't.
The fatality total involved three deaths in Highway Patrol areas using roadblocks this holiday season versus one death in the four areas last year--figures so small that they are virtually meaningless in a statistical sense.
Opponents of roadblocks point to an increase in drunk-driving fatalities over the holidays. Advocates of the roadblock tactic, most notably the Mothers Against Drunk Driving, immediately contended that the Highway Patrol figures established the effectiveness of the checkpoints. Opponents, including the American Civil Liberties Union, said that the results were ambiguous and that, in any event, they did not bear on the disputed constitutionality of the tactic.
Motorists who were stopped at the roadblocks, moreover, have responded to a Highway Patrol survey assessing the popularity of the technique in numbers far higher than the patrol expected, officials said.
The results of the postcard survey, the patrol disclosed, indicate that 85.5% of motorists responding agree that the roadblocks are a good way to catch drunk drivers and that 83.1% believe that the prospect of encountering roadblocks deters driving under the influence of alcohol or drugs. Another 85.9% said they believe that the roadblocks increase the risk of arrest for drunk driving, and 92.7% said they did not feel that they were unreasonably delayed in their travels by being detained.
A total of 10,150 motorists were stopped in the roadblocks and given survey cards, of which 2,450 were returned, the patrol said--a total significantly greater than commanders were said to have expected. Results of the survey were based on detailed tabulation of only the first 575 responses, but a Highway Patrol spokesman said the proportion of motorists favoring the tactic appeared to be the same for the entire sampling.
The Highway Patrol agreed to supply preliminary statistical totals in response to repeated requests from The Times, cautioning that the agency's own full analysis of the results of the roadblock program will probably not be available for at least two months. The Highway Patrol figures did not take into account roadblock programs set up by other law enforcement agencies around the state. Burlingame authorities used the tactic before the Highway Patrol, and Santa Ana police and the Los Angeles County Sheriff's Department are also experimenting with it.
The checkpoints were set up on surface streets in four Highway Patrol office areas across the state--Glendale (whose territory includes parts of the cities of Los Angeles, Pasadena, Glendale, Burbank and La Canada-Flintridge, as well as portions of unincorporated Los Angeles County), Bakersfield, Redding and North Sacramento.
Checkpoints were conducted over a 17-day period starting Dec. 14. To attempt to measure the effects, The Times obtained figures from the entire territories covered by the four roadblock CHP stations for alcohol-related fatalities, all crashes involving drunk drivers and drunk-driving arrests for the same 17-day periods of 1983 and 1982. The figures were compared against statewide totals for the same statistics for 1983 and 1984.
Comparing the identical 17-day periods of 1983 and 1984, roadblock areas recorded a 19% decline in alcohol-related crashes and a 26% increase in drunk-driving arrests. For the rest of the state--areas where the Highway Patrol has never conducted sobriety roadblocks--crashes went down 5% and arrests rose by 8%.
A standard statistical analysis was conducted for The Times Thursday by William Louv, an expert at the University of Alabama School of Public Health in Birmingham who has taught courses in statistics interpretation for journalists and public health officials across the country. Louv said the numbers of arrests and crashes were large enough to support statistically valid conclusions.