Differential impact analysis revealed that the inmates most likely to recidivate are former residents of the Boston area, black, and unmarried, and they had served indeterminate sentences at the Concord or Framingham prisons. Data indicate that neither the level nor the quality of reintegration programming was associated with the increased recidivism rate. On the contrary, data indicate the reintegration programming may have improved during the year (1979) in which recidivism increased. The final area of investigation involved the construction of extended followup periods to determine whether the detected increase in recidivism persisted over time. Findings indicate that when the followup period was extended to 2 or more years, the previously documented rise in the recidivism rate no longer existed. Data showing the increased recidivism rate during the year after release may be due to a systematic reduction of court base backlog, which may have produced an increased probability that an ex-inmate would be adjudicated a recidivist in a shorter time period, but without increasing the absolute number of recidivists. 11 tables. (Author abstract modified)