Inmates answered detailed questions about crimes they had committed (especially violent crimes) during the 3 years before their present prison terms. Internal consistency and good rates of response suggest that the accuracy of self-reported survey data was high. Results indicate that most inmates were either occasional offenders or broadly active criminals committing various types of crimes. Younger offenders were most active. While blacks were overrepresented in the survey, their criminal acts were less dangerous and they were less active than whites and Mexican-Americans. Recently active offenders tended to have lengthy criminal records, and respondents who began committing youthful crimes reported higher crime rates as adults than those who began criminal activities later in life. A group of career criminals (25 percent of the sample) committed a large proportion of all reported crimes. These criminals began committing serious crimes as juveniles; they see crime as a safe and enjoyable way to obtain a high standard of living and seem unconcerned about being caught. They expect to return to crime after release. These data lend considerable support to the general objectives of career criminal programs, but suggest that these programs would probably be more effective in reducing property crimes, since only offenders who commit many crimes can be identified with sufficient precision. The limitations and benefits of self-report surveys are discussed. Appendixes include extensive tables, figures, a bibliography of nearly 50 citations, an index, and the survey instrument.