Previous studies of the impact of drinking age laws on traffic accidents have often used statistical techniques or short data sets that can lead to misleading conclusions. Studies generally have focused on fatalities or surrogate variables or both. Using autoregressive integrated moving average techniques and total accidents by month over a 20-year period, the experience in illinois is reexamined. Thelower drinking age in illinois from october 1973 to january 1980 was responsible for an increase of more than 5, 000 accidents per month, a 14% increase. When the age was raised back to 21 in 1980, the figures reversed a similar amount. More than 1, 000 additional 18- and 19-year-old drivers were involved in accidents each month with the lower age limit, a 20% increase. It was also found that total monthlyproperty-damage-only accidents were highly positively correlated with wet or snowy pavement and negatively correlated with dry pavement; fatal accidents behaved in the opposite way. The techniques used can be used to analyze the impact of other public policies on accidents. This paper appears in transportation research record no. 1325, Highway safety: older drivers, seat belts, alcohol, motorcycles, and pedestrians 1991
Abstract