A major benefit of education is the lower risk of unemployment at higher educational levels. In PSID (Panel Study of Income Dynamics) data on the male labor force1 the reduction of the incidence of unemployment is found to be far more important than the reduced duration of unemployment in creating the educational differentials in unemployment rates. In turn, the lesser unemployment incidence of the more educated workers is, in about equal measure, due to their greater attachment to the firms employing them, and to the lesser risk of becoming unemployed when separated from the firm. The lesser frequency of job turnover of more educated workers, which creates fewer episodes of unemployment, is in large part attributable to more on-the-job training. In explaining the lesser conditional unemployment of educated workers and the somewhat shorter duration of their unemployment, indirect evidence is provided that (1) costs of on-the-job search for new employment relative to costs of searching while unemployed are lower for more educated workers; (2) that these workers are also more efficient in acquiring and processing job search information; and (3) that firms and workers search more intensively to fill more skilled vacancies.