We examined the prevalence, population attributable risk (PAR), and clinical characteristics of occupational asthma (OA) in a randomly selected population in six communities in Canada. Our study followed the European Community Respiratory Health Survey protocol. A randomly selected population of 18,701 (87% response rate) persons from the study communities, ranging in age from 20 to 44 yr, completed an initial questionnaire, of whom 2,974 (39% response rate) attended the laboratory and completed supplementary questionnaires. Of these latter individuals, 383 had asthma. Asthma was defined as physician-diagnosed asthma, and adult-onset asthma was defined as a first attack at age 15 yr or older. We used several methods for estimating OA as follows: (1) reporting of a high-risk job (occupation and industry) for OA at the time of asthma onset (Probable OA); (2) reporting of exposure to a substance that may cause OA (Possible OA) while not in a high-risk job at the time of asthma onset; and (3) combination of the PAR for high-risk jobs and exposures. The prevalence (95% confidence interval [CI]) of Probable OA and Possible OA combined was 36.1% (31.3 to 41.0%) among subjects with adult-onset asthma. The occupations most commonly reported in association with OA were nursing in the Probable OA group and clerical and food preparation in the Possible OA group. The clinical characteristics and exposures reported by both groups were similar. The PAR for adult-onset asthma in high-risk jobs and exposures was 18.2%. The assessment of occupation and industry alone, rather than of exposures, may underestimate the contribution of occupational exposures to asthma prevalence.