Does treatment, education and work during the prison term have any significance for the risk of recidivism?

By Lena Roxell

Photo of a teal-coloured door ajar, through which you can see a green garden
Photo by Jan Tinneberg on Unsplash

Since the turn of the millennium, levels of recidivism following a prison term have declined in Sweden. One explanation that has been posited for this trend is that the composition of the inmate population has changed during the period in question. Sex, age, prior criminal histories, the offences leading to imprisonment and the length of prison sentences are all factors that might explain the decline. However, it was noted that the decline in recidivism between 2011 and 2014 could not be completely explained by these factors.

According to this background, my research has focused on whether prison service measures, such as the provision of treatment, education and work, might have contributed to the decline in recidivism. Several international meta-analyses have examined the effects of various treatment and work-focused programmes on the risk for recidivism. The reported results vary. It is unusual to examine different measures, such as treatment, education and work programmes, in the same study.

To investigate whether various interventions are important in reducing the risk of recidivism, I conducted a new study. The sample comprises women and men who were released from Swedish prisons between 2012 and 2015. The analyses are based on a total of 14,805 individuals. Data have been collected from the Prison and Probation Service registers, Statistics Sweden and the Convictions Register.

Analysis of the data took place through propensity score matching (PSM). All those who have completed treatment measures, have received education or have engaged in work have been matched to individuals who have not. The matching has been based on the background covariates sex, age, immigrant/native background, criminal history, offence type, sentence length and substance abuse. The matched datasets have then been used to estimate the effect of measures experienced during the time in prison on recidivism subsequent to release. The analyses have employed linear regression, with recidivism as the dependent variable.

The results show that treatment, education and work, and combinations of these measures, reduce the risk for recidivism. The reduction is primarily found in relation to the risk for reconviction following release from prison, rather than in relation to receiving a new prison sentence. At the same time, the explanatory power of these variables was low throughout. In other words, the measures provided to inmates by the Prison and Probation Service only constitute a partial explanation for the reduced levels of recidivism that have been noted during the period examined in the study.

About the author

Lena Roxell works as a senior lecturer at the Department of Criminology at Stockholm University. Her research has mainly been about the prison and the inmates.