A study examined the nature and the relative incidence of different forms of underemployment and how and under what conditions underemployment of older workers differed from that of younger workers. Data from May 1979, May 1988, and April 1993 Current Population Surveys (CPSs) for workers aged 50-64 were used. Two categories of underemployed workers were identified. Type I underemployment, which could be directly estimated from CPS data, referred to workers who wanted full-time work but could find only part-time employment. Determined by multivariate analysis, Type II underemployment encompassed full- and part-time workers employed at jobs below their education and skill levels. It was defined by comparing each worker's actual wages to the expected wages for a worker with similar education, demographic characteristics, industry and occupation, region, and size of firm. CPS data indicated Type I underemployment had been steadily rising over time. Type I underemployed were found to be an extremely disadvantaged group in terms of pension and health insurance coverage; Type II underemployed were also disadvantaged. Nearly 400,000 workers between the ages of 50-64 were estimated to be Type I underemployed; the estimated number of Type II underemployed was much larger--7.2 million. Single people, females, minorities, and persons with less education tended to have higher underemployment rates. The magnitude of Type II underemployment highlighted the need for further work to refine these estimates. (Contains 10 references.) (YLB)