Nonparticipants in general population health surveys have been found to be less healthy than participants, but data on nonparticipants in work-site health surveys have been more scarce. We compared cause-specific mortality among 11,156 male employees of The Dow Chemical Company who participated in at least one work-site health examination between 1967 and 1978 with 6915 employees who did not participate. The nonparticipants experienced higher mortality rates for nearly every cause of death examined but particularly from smoking and alcohol-abuse related diseases. This was especially true during the first 5 years of follow-up, suggesting that some employees do not participate because they are already ill. These findings have important implications for the use of examination data for both primary and secondary disease prevention purposes, and these are discussed.