In the spring and summer of 1981, night-time surveys of drivers' alcohol use were undertaken in the provinces of British Columbia, Saskatchewan and Quebec. The surveys were performed by agencies of the provincial governments, using teams of students as interviewers. Advice and assistance in survey planning, sampling, instrument design, data processing and analysis was provided by the federal Department of Transport, based on experience with a national survey in 1974 and a survey in Ontario in 1979.

A sample of sites was selected using a stratified, multi-stage procedure, and 90-minute survey periods were assigned to each site, within the survey limits of 9 p.m. and 3 a.m. on the nights of Wednesday to Saturday. Cars, light trucks or motorcycles were selected essentially at random from the traffic stream by police officers, taking the next available vehicle whenever interviewers were free. Interviews were conducted with drivers in off-road areas, each driver being asked to volunteer information and to give a breath sample into a J3D ALERT device, which gave an immediate estimate of blood alcohol content. Interviews were completed with samples of over 7300, 3200 and 6300 drivers respectively in British Columbia, Saskatchewan and Quebec, with very low rates of refusal to provide breath samples.

The paper describes the survey and sampling methodologies, focussing on modifications made to improve upon the 1974 survey, particularly in the size and quality of the sample and in the breadth of information obtained. Results of the surveys are presented and compared to the earlier surveys.