Of the sample of 1,632 men and 128 women, the mean participation rate was about 90 percent. A self-administered survey instrument containing 65 pages was used to elicit information on the offender's childhood and family, offenses, and criminal history variables. Official data were also obtained from Colorado prison records, and qualitative data were collected from interviews with selected respondents. Data analysis revealed that crime rates were sensitive to cohort differences in 1986 and 1989 and to measurement effects. Crime frequencies were higher in 1986 than in 1989. Blacks tended to report higher crime participation rates using an anonymous version of the survey instrument. Crime frequency estimates remained fairly stable across different data collection methods, and no differences were found across race, age, or education. Skewed offending rates were found among female prisoners; most reported very few crimes, while a small proportion reported extremely high criminal activity rates. Race differences between males and females on employment and educational dimensions suggested differential socioeconomic marginalization that might be correlated with criminal activity. Crime participation rates changed when the definition of a crime changed. Also, asking about criminal behavior during smaller versus larger increments of time affected responses. In general, self- report and official data were complementary. An appendix contains additional survey data, along with data collection forms. 98 references, 75 tables, and 27 figures