« AnteriorContinuar »
Should eighteen, nineteen, or twenty year-old adults be denied the legal right to sell, purchase, possess, or consume alcoholic beverages? Responding to scattered evidence that lowering the legal drinking age in 1971-73 may have resulted in increased alcohol-related highway crashes among newly enfranchised legal drinkers (1), a number of states have subsequently rescinded the drinking priveleges of 18, 19, and 20 yearold adults. Federal policy-makers and legislators in lower legal drinking age states look to the experiences of the age-change states to assist in determining whether there is, or is not, a compelling public safety interest in denying the right to drink alcoholic beverages to any group of legal adults, solely on the basis of their age. Despite the fact that, since 1975, twenty-one states have raised their legal drinking ages, scientifically reliable studies of the highway safety impact of those changes are few and generally inconsistent. in Michigan, Wagenaar found decreases in underage highway crash involvement ranging from 17.7% to 30.7% (2), and Williams reported a 41% reduction among 18-20 year-olds after the age was raised to 21 (3). Scholarly observers disagree on the size and significance of post-change crash involvement in Maine, where Wagenaar found no significant shift in serious crashes (2), klein reported decreases of 18.6% for 18 year-olds and 13.9% for 19 year-olds (4); and Williams found a 14% aggregate reduction for 18-19 year-olds after the age was raised to 20 (3). Where Williams found a 10% decrease in night-time to day-time fatal crash ratios (3), Massachusetts authorities report a 23% increase in fatal crashes involving 18-19 year-old drivers subsequent to raising the minimum legal drinking age to 20 (5). Williams also reported a 17% increase in fatal crash ratios for 18 year-olds in Montana after the age was raised to 19 (3), and Maxwell found an 8.8% decrease in 19-20 yearold Illinois drivers' highway crash involvement after that state raised the legal drinking age from 19 to 20 (6). In the absence of conclusively consistent findings among the available drinking age impact studies, it was decided to examine and compare recent crash experience among 18-20 year-old drivers in Wisconsin, which has retained 18 as the legal drinking age for all alcoholic beverages since 1972, with those observed and attributed by others to raising the legal drinking age in Michigan and Illinois, to determine whether evidence could be found to support the highway safety rationale for restricting or denying the drinking priveleges of young adults.
The data analyzed consisted of the number of driver deaths tested in Wisconsin's fatal accident blood testing program which were found to have blood alcohol levels less than .05% (Non-Impaired Driver Deaths, and .05% or greater (Impaired Driver Deaths), the number of drivers involved in fatal crashes (Fatal Crashes), and the number of all highway crashes in which police reported that the driver had been drinking and was imparied by alcohol (Had-Been-Drinking) for 18 through 24 year-old drivers in the years 1976-1981. Crash rates per thousand licensed drivers were calculated for each of the age groups, 18-20 and 21-24, for each of the six years studied.
The resulting data sets were subjected to partitioned chi-square analyses (7) to identify significance levels of annual changes in values of each variable for each driver age group. Linear least-squares regressions (8) were used to analyze and predict rates and frequencies for item/years identified as significant by partition analyses. Chi-squares were calculated for predicted-versus-actual data values in each significant change year. In brief, these analyses assessed the degree to which actual crash experience differed from that expected on the basis of previous trends.
Findings Table 1 shows the results of various analyses of the four crash statistics (Fatal Crashes, Had-Been-Drinking Crashes, Alcohol - Impaired Driver Deaths, Non-Impaired Driver Deaths) for drivers in the two age groups. Percentages in the last column indicate the difference between actual rates per thousand licensed drivers for the most significant change year, 1981, and the rates expected based on analyses of the 1976-80 data.
As Table I shows, alcohol-related crashes involving the younger driver age group decreased dramatically in 1981, while non-alocohol related crashes (Non-Impaired Driver Deaths) were only slightly lower than predicted. Overall reductions in driver deaths and fatal crashes were too small to account for the significant decrease in alcohol-related crashes among 18-20 year-old drivers. The contrast between these sizeable reductions, the much smaller decrease in Had-Been-Drinking Crashes and the insignificant increase in Alcohol-Impaired Driver Deaths among 21-24 year-old drivers supports the conclusion that there was a significant change in younger drivers' drinking and driving behaviors in Wisconsin in 1980-81.
Table 2 shows the crash totals for the two age groups of drivers during the six-year period, 1976-81, along with the 1981 figures projected by linear regression analyses. Note in the line for 1981 how the alcoholrelated crashis decreased for drivers aged 18-20, in contrast to those involving the next older driver age group. Figure 1 illustrates the 1980-81 reduction in fatal crashes among 18-20 year-olds and increased fatal crashes involving 21-24 year-old drivers in the same year. Note the significant difference in predicted and actual crash levels for younger drivers and the insignificant difference for the older driver age group in 1981.
Table 3 compares changes in underage drivers' crash statistics observed after raising the legal drinking age in Michigan by Wagenaar (9) and in Illinois by Maxwell (6), with those involving drivers of the same age in Wisconsin in 1980-81. The values shown in Table 3 are the difference in actual versus predicted crash levels in significant change years. Figures for Michigan and Illinois reflect changes in three-factor surrogate crashes (night-time, single vehicle crashes involving male drivers ) resulting from Box-Jenkins Time Series analyses. (See Wagenaar (2) for detailed discussion of data and methodology.) Wisconsin figures are for changes in fatal crashes and driver deaths. The data selected for
18-20 21-24 Predicted * 18-20 **21-24