Measures for assessing daytime sleepiness can be categorized into four general types-behavioral observation, laboratory performance, introspection, and physiological techniques. Each approach has its advantages and disadvantages. Thus, for example, observational techniques may provide a minimally 'contaminated' sample, but inferences may be quite difficult; laboratory performance measures may indicate vulnerabilities associated with sleepiness, but may be limited because of practice effects; introspection provides a personal perspective, but that point of view may be influenced by past history or ability to articulate; physiological measures may provide the most 'objective measures', but are difficult to apply to field studies. The Multiple Sleep Latency Test (MSLT) is the most thoroughly studied physiological measure of sleepiness relying upon serial assessments of the speed of falling asleep in standard conditions intended to optimize the sleep onset process. Procedures for administering, analyzing, and interpreting the MSLT are presented.