Samenvatting
Data were obtained from roadside surveys of the breath alcohol levels of non-accident-involved drivers in Australia, Canada and the USA. Only surveys using some form of random selection were used and the comparison was based on data for the same time of day and days of week. Only a small number of studies satisfied the random selection criterion; among those which did there were few meaningful differences in the percentages of drinking drivers but some differences in the percentages at higher BACs.