Sleepiness is a commonplace experience for all individuals at some point in their daily lives. Whether such a ubiquitous experience has utility in discriminating a diseased population from the general population has yet to be determined. The purpose of this study was to determine the accuracy of reports of sleepiness across different situations by identifying individuals with sleep apnea. We compared a clinical population of 630 patients aged 50-65 with polysomnographically diagnosed sleep apnea with a randomly sampled population of 1,877 individuals of similar age from a nearby community. Sleepiness while watching television was the most prevalent behavior (about 36% of the reference population and 90% of the apnea patients), but potentially better discrimination occurred with items inquiring about sleepiness in other situations. The data suggest that discrimination between diseased and nondiseased populations may be more successful with sleepiness items involving relatively low, rather than high, rates of endorsement.