The methodological changes included a more detailed calendar system for cueing recall of past events and month- by-month reporting of rates of offending. The present study also used individual interviews rather than self- administered surveys. Data were gathered from 700 convicted male offenders, who were randomly assigned to two groups, one using the RAND method and the other using the modified method. Findings revealed that only a small proportion of frequent offenders are active every month in a particular crime category. The majority of offenders reported more than one rate of offending during the active periods, with activity patterns varying by crime types. Finally, the lambda estimates from the revised methodology did not differ from those obtained with the RAND questions. Figures, tables, and 24 references (Author abstract modified)