The first of two statistical analyses employed for the study estimated how many among two representative samples of offenders would have been prevented from committing their current offenses if an 18-month mandatory prison sentence had been imposed at previous conviction. The second analysis investigated the consequences of shorter prison sentences upon the size of the prison population. Findings from the two analyses suggest that the overall number of convictions could be reduced by imposing more severe sentences, but only at the cost of substantial increases in the length of imprisonment. Modest reductions in prison sentences would not lead to a large number of additional convictions, but would significantly decrease the size of the prison population. For the second study, statistical analysis was conducted in which future criminal behavior of 811 male of offenders was predicted. Results showed that of the 77 acts of violence which were subsequently recorded against the total sample, only 13 were committed by the 52 offenders classified as dangerous who were free to reoffend and for whom records were available. It is concluded that predictive assessments of dangerousness are too unreliable to use as the basis for a policy of reducing serious crime by selective incapacitation of offenders. Seventy references and nine tables are provided. A separate listing of 63 correctional studies published by the British Home Office is also included.