The study examined official records available to prosecutors in two jurisdictions (Los Angeles County, California, and Middlesex County, Massachusetts). Data were collected from and on 500 male defendants who were ultimately convicted, and official record and self-report data for 452 defendants were actually analyzed. Findings showed prosecutors separately evaluated the three dimensions of a defendant's criminality--rates of committing crimes, dangerous, and persistence. Defendants who were identified as high-rate and dangerous by prosecutors in one jurisdiction were also identified as high-rate and dangerous in the other jurisdiction. Written guidelines concerning selection criteria for career criminals promoted consistency in the judgments of deputy district attorneys about the kinds of defendants who were high-rate dangerous offenders. It was also found that long-term persistent offenders may or may not be high-rate dangerous offenders and that habitual criminality should not be confused with high-rate dangerous criminality. The strongest official record indicators of high-rate offending in the two jurisdictions were found if a defendant had a prior adult conviction for robbery, burglary, arson, forcible rape, sex crime involving a child, kidnapping, or murder; was currently charged for three separate criminal transactions of burglary; was wanted by the authorities for failure to complete a previous sentence; was on parole when arrested; had one or more adult arrests for receiving stolen property; was on pretrial release when arrest; and was known to have a drug problem. Once a group of high-rate offenders was identified, the subset of high-rate dangerous offenders could be identified using a small number of criteria. Several factors commonly perceived as indicative of high-rate dangerousness proved not to be and in some cases were counter-indicators. Examples of factors not associated with high-rate dangerous offending included display or use of a gun to threaten a victim, alcoholism, number of prior arrests for drug distribution or possession, record of previous probation or parole revocations, and record of previous incarceration. While much of the information usually available to prosecutors was useful in identifying high-rate dangerous offenders, the study showed that other commonly used information could be misleading or ineffective for purposes of identification. Recommendations are offered to improve the information available to prosecutors in cases involving high-rate dangerous defendants. 11 notes