Predicting Offender-Patients' Reconvictions
CLIVE PAYNE 1,
SARAH McCABE 2, and
NIGEL WALKER 3
1 Senior Research Officer, Research Services Unit, Nuffield College, Oxford, OX1 1NF
2 Senior Research Officer, Oxford University Penal Research Unit, St. Cross Building, Manor Road, Oxford
3 Wolfson Professor of Criminology and Director of the Institute of Criminology, 7 West Road, Cambridge, CB3 9DT
A logistic analysis of Walker and McCabe's follow-up data for hospital order patients who are allowed to leave within a year of admission suggests a simple prediction model which is promising enough to call for validation on similar samples. As with ordinary penal samples, the number of previous convictions and type of current offence are powerful predictors of reconviction, but a puzzling difference from ordinary penal samples is the negligible predictive power of sex and age. The diagnosis is a useful predictor only if those patients who are re-admitted without being reconvicted are included in the sample. The probable explanation of these findings is: (i) that hospital orders are used to deal with patients whose `way of life' involves frequent petty dishonesties, either because they have been reduced to it by their handicaps or because they belong to a subculture in which dishonest acquisition is not strongly disapproved of; and (ii) that the schizophrenic hospital order case has a greater chance than others of being re-admitted to hospital either before he gets into further trouble with the law or without prosecution when he does. The analysis confirms the importance of efforts to arrange after-care, expecially for high-risk groups.
Submitted on August 28, 1973
