The accident reports of a company were analyzed for 10 years before and after change from 8- to 12-hour shifts. Age-sex standardized ratios were calculated for each year for on- and off-the-job accidents by severity. Times of occurrence of on-the-job accidents were also examined. Overall accident rates were reduced on the 12-hour shift schedule, but statistical significance was reached only for lower accident categories. Off-the-job injuries increased on the 12-hour shift. There were significant gender differences in accident rates on 8-hour but not 12-hour shifts. A distinct circadian pattern of accident frequency was observed. We concluded that the introduction of a 12-hour shift system did not result in increased accidents.