Random samples of men born in 1956, 1958, 1960, and 1962 were taken from the Adult Criminal Justice Statistical System, a longitudinal file which permits the tracking of criminal careers of large numbers of offenders over long time periods. The samples were matched with the Wage History and Base Wage Files of the California Employment Development Department, which contains information on individuals' employment and earnings for 1980 through 1986. Matches were made for 60 percent of the records. The merged analysis file contains information on a broad class of arrestees and is more representative of the general population of offenders than the samples of parolees typically studied in the past. The study found that the typical member of the arrestee sample earned less in 1986 on the conventional labor market than the typical young male in the general population. Correlation analysis showed a strong connection between earnings and arrests. The study discusses alternative explanations for these findings. 1 table, 2 figures, 2 notes, 5 references.