The National Job Corps Study has been conducted since 1993 to provide Congress and program managers with the information they need to assess how well Job Corps attains its goal of helping students become more responsible, employable, and productive citizens. The cornerstone of the study was the random assignment of all youths found eligible for Job Corps to either a program group (who could enroll in Job Corps) or a control group (who could not). Previous reports presented impact and benefit-cost estimates by comparing the experiences, in particular, the earnings, of program and control groups using data from follow-up interviews conducted during the four years after random assignment. The current report presents findings from an analysis of administrative earnings records using: (1) annual social security earnings (SER) data reported by employers to the Internal Revenue Service (IRS); and (2) quarterly wage records reported by employers to state unemployment insurance (UI) agencies. According to both the survey and administrative records data, the estimated earnings impacts are negative in the first and second years after random assignment and positive and statistically significant in the third and fourth years after random assignment. No impacts were found for the full sample on employment or earnings after the four-year period covered by the survey. Revised benefit-cost estimates suggest that the benefits to society of Job Corps are smaller than the substantial program costs, particularly for younger participants. Job Corps, however, may be cost-effective for the 20- to 24-year-olds at program application whose earnings impacts persisted during the post-survey period. Older students remain in Job Corps longer than younger ones, receive more hours of vocational training while enrolled, and are more highly motivated and well-behaved (as reported by program staff). However, many of the youngest sample members in the control group returned to high school after being rejected from Job Corps, whereas fewer older control group members enrolled in alternative education and training programs. These findings suggest that to improve overall program effectiveness, Job Corps needs to fully address differences by age in program structure and experience, and perhaps, to reassess the target population served by the program. Further long-term follow-up is required to eliminate uncertainties about the effectiveness of the program, which may indicate that Job Corps may increase the long-term earnings of those who were 16 to 19 years old at program application as they mature, find stable jobs, and experience the full benefits of program participation, both from increased vocational and academic skills gained in the program, as well as from improved social skills and attitudes toward work. The persistent earnings impacts among students who were 20 to 24 at program application raise this as a possibility. The following are appended: (1) Selecting States and Sample Member for the UI Study and Obtaining IU-Based Impact Estimates Using the Full Sample; (2) Supplementary Tables to Chapter III; (3) Supplementary Tables to Chapter IV; and (4) Supplementary Tables to Chapter V. (Contains 1 figure and 26 tables.)