Differences in exposure -- the extent to which different people are exposed to the risk of an automobile crash -- is a variable which has been difficult to measure and frequently ignored in epidemiologic studies of traffic accidents. To study exposure, we interviewed 3000 drivers living in Metropolitan Toronto by telephone during 1978, to assess the amounts and types of driving they had done. Drivers had been selected at random from provincial records to include drivers from six different groups:​ males with 0-2 points, 3-5 points, 6-8 points or 9+ demerit points and females with 0-2 points or 3-5 demerit points. (Demerit points are assigned when a driver is convicted of one of 26 traffic offences.)

Women had driven fewer miles than men at the same level of demerit points, for each type of driving (city, highway, rushhour, night). Within each sex, drivers with more demerit points had driven significantly more than drivers with fewer demerit points.

The actual number of convictions and collisions experienced by the sample of drivers was obtained from provincial records. Rates of collision and conviction were calculated using persons and person-miles in the denominator:​ when account was taken of the mileage, women had higher collision rates than men with comparable demerit points. Furthermore, drivers who, typically, drove small amounts each week (under 120 miles/week) had a much increased risk of collision per mile, which was even greater if the driver had acquired demerit points. Drivers with greater mileages per week had a relatively low risk of accident per mile which was not altered by the number of points.

Obtaining valid and reliable data on driving behaviour can present difficulties. The results of a validation study were reassuring, but our inability to reach transient individuals may have introduced some underestimation in these results.