Applying this 8% difference to Wisconsin's 1981 crash data, had the legal drinking age been 21 instead of 18 since 1972, there would have been 16 more fatal crashes involving 18 through 20 year-old drivers in Wisconsin in 1981 [3].

Further analyses of these data, however, reveal no statistically significant difference in 18 through 20 year-olds' fatal highway crash involvement due to differences in the states' legal drinking ages (t = 1.085 p.29). Thus, one cannot conclude that a higher legal drinking age would have resulted in any difference in Wisconsin in 1981. Indeed, these tests confirm the null hypothesis; that is, one can conclude that higher or lower legal drinking ages do not effect long-term differences in fatal highway crash involvement among 18 through 20 year-old drivers.

CONCLUSIONS: IMPLICATIONS. Despite favorable and unfavorable short-term changes observed in highway crashes involving those affected by raising the legal drinking age in a number of states, higher legal drinking ages have not been demonstrated to effect any long-term differences in underage drivers' highway crash involvement, and can not be considered effective long-term public health countermeasures for morbidity, mortality or disability associated with highway crashes involving underage drivers.

Determination of the legal drinking age involves numerous considerations, including: the social, political, legal and economic costs of denying the legal drinking privilege to eighteen, nineteen or twenty year-old adults; the difficulties of attempting to regulate by formal social policies, widespread individual behaviors which informal social attitudes tolerate or encourage [10]; and, the availability of less costly and more effective mechanisms for

preventing alcohol abuse and alcohol-related highway crashes among adolescents and young adults [11].

Current research clearly indicates the futility of attempting to achieve those objectives by manipulation of the minimum legal drinking age [12].

NOTES

(1.) Wagenaar also examined changes in reported had been-drinking to had-not been drinking crashes involving the same age groups in both states. As explained in his study, results of analyses based on had been-drinking data are subject to distortion due to changing observational biases of original reporters (police officers) induced by changes in the law Consequently these are less reliable indicators of changes in real occurances than driver deaths or three-factor surrogate data Results discussed here are those derived from analysis of less subjective indicators; fatalities, nighttime-to-daytime fatal crash ratios, and three-factor surrogates

(2) Analyses of incomplete data for lowa, Minnesota and Illinois omitted. See Williams [5].

(3.) 18-20 year-old drivers were involved in 202 fatal crashes, 1981 Source WiscONSIN Accident Facts, 1981, Department of Transportation

REFERENCES

[1] Wagenaar AC Effects of the Minimum Drinking Age on Automotive Crashes Involving Young Drivers, Ann Arbor. The University of Michigan Highway Safery Research Institute, 1982.

[2] Dunham NC, Detmer DE: Traffic accidents and the legal drinking age in Wisconsin Wisconsin Medical Journal, 1981 (March)

[3] Wagenaar AC Alcohol Availability and the Raised Drinking Age. Recent Fin dings, Ann Arbor, The University of Michigan Highway Safety Research Institute, 1981

[4] Wagenaar AC. Effects of an Increase in the Legal Minimum Drinking Age, Journal of Public Health Policy, 1981. 2,3 206-224

[5] Williams, AF, Zador PL. Harris 55, Karpf RS. The Effects of Raising the Legal Minimum Drinking Age on Fatal Crash Involvement, Washington DC, Insurance Institute for Highway Safety. 1981

[6] Klein TM: The Effect of Raising the Minimum Legal Drinking Age on Traffic Ac cidents in the State of Maine, Washington, D C., National Highway Traffic Safety Administration, 1981

[7] Maxweli, DM: Impact Analysis of the Raised Drinking Age in lines, Washington, DC, National Highway Traffic Safety Administration, 1982 [8] Motor Vehicle Accident Experience, Boston, Massachusetts Registry of Motor Vehicles, 1978-81

[9] Cook, PJ The Effect of Minimum Drinking Age Legislation on Youthful Auto Fatalities, 1970-1977, Chapel Hill, Duke University Department of Economics, (Unpublished) 1982.

[10] Bonnie, RJ "Discouraging Unhealthy Personal Choices through Government Regulation: Some Thoughts about the Minimum Drinking Age." in H. Wechsler (ed.) Minimum-Drinking- Age Laws, Lexington, Mass. D.C. Heath and Co, 1980. 155-176

[11] No author: Youth, Alcohol and the Law Final Report, Task Force report to the State Council on Alcohol and Other Drug Abuse, Madison, Wis, Wisconsin Department of Health and Social Services, 1978

[12] Barkley, MM "Effects of Controls on Underage Drinking: Research Indications ** Report to the Governor, Bureau of Alcohol and Other Drug Abuse, Madison, Wis. Wisconsin Department of Health and Social Services, 1979.

An additional note to the reader.....

Subsequent to completion of the above, an additional study of post-change highway data in the State of Florida revealed no significant change in highway crashes involving 18 year-olds attributable to raising that state's legal drinkiong age to 19 in October 1980. (Morris MF, "Drinking-Driving Behavior In Florida", Pensacola, Pensacola Junior College, March 1983.)

Noting that "drivers in the age group 21-24 account for proportionately more accidents, injuries, fatalities and arrests than any other age group," Morris found reduced involvement "in DUI arrests and alcohol-related accidents for the 18 year olds" but notes also that “a similar decrease was reflected for the 19 and 20 year olds for whom the age change did not apply." Thus, in Florida, raising the legal drinking age generated no significant change in highway crash involvement among those affected by the law.

Adding Florida's post-change experience to those of the seven states cited above, it appears that Dunham and Detmer's prediction that raising the age would reduce alcohol-related highway crash involvement among those affected is at odds with actual occurances in five states (Florida, Maine, Massachusetts, Montana and Illinois) and consistent with the experience of only three of eight states studied, (Michigan, New Hampshire and Tennessee).

Michael M. Birkley June 1983

Applying this 8% difference to Wisconsin's 1981 crash data, had the legal drinking age been 21 instead of 18 since 1972, there would have been 16 more fatal crashes involving 18 through 20 year-old drivers in Wisconsin in 1981 [3].

Further analyses of these data, however, reveal no statistically significant difference in 18 through 20 year-olds' fatal highway crash involvement due to differences in the states' legal drinking ages (t = 1.085 p.29). Thus, one cannot conclude that a higher legal drinking age would have resulted in any difference in Wisconsin in 1981. Indeed, these tests confirm the null hypothesis; that is, one can conclude that higher or lower legal drinking ages do not effect long-term differences in fatal highway crash involvement among 18 through 20 year-old drivers.

CONCLUSIONS: IMPLICATIONS. Despite favorable and unfavorable short-term changes observed in highway crashes involving those affected by raising the legal drinking age in a number of states, higher legal drinking ages have not been demonstrated to effect any long-term differences in underage drivers' highway crash involvement, and can not be considered effective long-term public health countermeasures for morbidity, mortality or disability associated with highway crashes involving underage drivers.

Determination of the legal drinking age involves numerous considerations, including: the social, political, legal and economic costs of denying the legal drinking privilege to eighteen, nineteen or twenty year-old adults; the difficulties of attempting to regulate by formal social policies, widespread individual behaviors which informal social attitudes tolerate or encourage [10]; and, the availability of less costly and more effective mechanisms for

preventing alcohol abuse and alcohol-related highway crashes among adolescents and young adults [11].

Current research clearly indicates the futility of attempting to achieve those objectives by manipulation of the minimum legal drinking age [12].

NOTES

(1.) Wagenaar also examined changes in reported had-been-drinking to had-not-been drinking crashes involving the same age groups in both states. As explained in his study, results of analyses based on had been-drinking data are subject to distortion due to changing observational biases of original reporters (police officers) induced by changes in the law. Consequently these are less reliable indicators of changes in real occurances than driver deaths or three-factor surrogate data Results discussed here are those derived from analysis of less subjective indicators, fatalities, nighttime 10-daytime fatal crash ratios, and three-factor surrogates

(2) Analyses of incomplete data for lowa, Minnesota and Illinois omitted. See Williams [5].

(3) 18-20 year-old drivers were involved in 202 fatal crashes, 1981. Source Wisconsin Accident Facts, 1981, Department of Transportation

REFERENCES

(1) Wagenaar AC: Effects of the Minimum Drinking Age on Automotive Crashes In volving Young Drivers, Ann Arbor, The University of Michigan Highway Safety Research Institute, 1982

[2] Dunham NC, Detmer DE: Traffic accidents and the legal drinking age in Wiscon sin Wisconsin Medical Journal, 1983 (March)

[3] Wagenaar AC Alcohol Availability and the Raised Drinking Age: Recent Findings. Ann Arbor, The University of Michigan Highway Safety Research Institute, 1981

[4] Wagenaar AC Effects of an Increase in the Legai Minimum Drinking Age, Journal of Public Health Policy, 1981. 2.3. 206-224

[5] Williams, AF. Zador PL, Harris SS, Karpf RS. The Effects of Raising the Legal Minimum Drinking Age on Fatal Crash Involvement, Washington D C., Insurance Institute for Highway Safety, 1981

[6] Klein TM: The Effect of Raising the Minimum Legal Drinking Age on Traffic Ac cidents in the State of Maine, Washington, DC, National Highway Traffic Safety Administration, 1981

[7] Maxwell, DM: Impact Analysis of the Raised Drinking Age in Illinois, Washington, D.C., National Highway Traffic Safety Administration, 1982.

[8] Motor Vehicle Accident Experience, Boston, Massachusetts Registry of Motor Vehicles, 1978-81.

191 Cook, PJ The Effect of Minimum Drinking Age Legislation on Youthful Auto Fatalities, 1970-1977," Chapel Hill, Duke University Department of Economics, (Unpublished) 1982.

[10] Bonnie, RJ Discouraging Unhealthy Personal Choices through Government Regulation. Some Thoughts about the Minimum Drinking Age," in H. Wechsler (ed.) Minimum Drinking-Age Laws, Lexington, Mass DC Heath and Co., 1980, 155-176

[11] No author: Youth, Alcohol and the Law: Final Report, Task Force report to the State Council on Alcohol and Other Drug Abuse, Madison, Wis Wisconsin Department of Health and Social Services, 1978

[12] Barkley, MM "Effects of Controls on Underage Drinking: Research Indications”*. Report to the Governor, Bureau of Alcohol and Other Drug Abuse, Madison, Wis., Wisconsin Department of Health and Social Services, 1979

An additional note to the reader.....

Subsequent to completion of the above, an additional study of post-change highway data in the State of Florida revealed no significant change in highway crashes involving 18 year-olds attributable to raising that state's legal drinkiong age to 19 in October 1980. (Morris MF, "Drinking-Driving Behavior In Florida", Pensacola, Pensacola Junior College, March 1983.)

Noting that "drivers in the age group 21-24 account for proportionately more accidents, injuries, fatalities and arrests than any other age group." Morris found reduced involvement "in DUI arrests and alcohol-related accidents for the 18 year olds" but notes also that “a similar decrease was reflected for the 19 and 20 year olds for whom the age change did not apply." Thus, in Florida, raising the legal drinking age generated no significant change in highway crash involvement among those affected by the law.

Adding Florida's post-change experience to those of the seven states cited above, it appears that Dunham and Detmer's prediction that raising the age would reduce alcohol-related highway crash involvement among those affected is at odds with actual occurances in five states (Florida, Maine, Massachusetts, Montana and Illinois) and consistent with the experience of only three of eight states studied, (Michigan, New Hampshire and Tennessee).

Michael M. Birkley June 1983

Table 3 compares changes in underage drivers' crash statistics observed after raising the legal drinking age in Michigan by Wagenaar and in Illinois by Maxwell', 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 for detailed discussion of data and methodology.) Wisconsin figures are for changes in fatal crashes and driver deaths. The data selected for comparison in Table 3 are those least likely to be distorted by influence other than changes in real occurances. Each item bears a direct relationship to alcohol. involvement and reflects the degree of change in alcoholrelated crashes among the populations studied.

As shown in Table 3, alcohol-related crash involvement decreased relative to the level predicted on the basis of

CONCLUSIONS

The study results show that raising the legal drinking age had no greater impact on 18-20 year-old drivers' highway crash involvement in Michigan and Illinois than not changing the legal drinking age had on crashes involving 18-20 year-old drivers in Wisconsin. These results are entirely consistent with those of earlier research conducted by the National Highway Traffic Safety Administration and more recent analysis of data for twenty-two states which retained different drinking ages over time", which found no significant difference in 18-20 year-old

(1) Douglas RL, Filkins LD, and Clark FA The Effect of Lower Legal Drinking Age on Youth Crash Involvement, Final Report, Washington, DC, National Highway Traffic Administra fron, 1974

(2) Wagenaar AC Effects of Raising the Legal Drinking Age on Traffic Accident involvement of Young Drivers, Ann Arbor, The University of Michigan Highway Safety Research institute. 1981

(3) Wiliams AF. Zador PL Hams 55. Karp! RS. The Effects of Raising the Legal Minimum Drinking Age on Fatal Crash Involvement, Washington. D. C. Insurance Institute for Highway Safety, 1981

(4) Klein TM The Effect of Raising the Minimum Legal Drinking Age on Traffic Accidents in the State of Maine, Washington, DC. National Highway Traffic Safety Administration, 1981

(5) No Author Motor Vehicle Accident Experience, Boston Mass Massachusetts State Registry of Motor Vehicles 1978-81

(6) Maxwell DM impact Analysis of the Raised Drinking Age in Illinois, Washington, DC, Na tional Highway Traffic Safety Administration, 1982

previous trends for 18-20 year-olds in each of the states studied. Reductions among younger drivers in Michigan (17.7%) and Wisconsin (20.5% -28.6%) were substantially greater than the 8.8% decrease observed among young IIlinois drivers subsequent to raising the legal drinking age. Reductions among 18-20 year-old drivers attributed to intervention of the drinking age change in Michigan in 1978-79 are well within the range of decreases in alcoholrelated highway crashes involving drivers of the same age in Wisconsin in 1980-81, where there was no evident intervening legal variable.