Early research on the effects of prison concluded that prisoner rehabilitation programmes do not work (Martinson 1974). This research was influential in rehabilitation gradually taking a back seat in favour of prison policies emphasising punishment and incapacitation. Subsequent scholars questioned the evidence base for this conclusion (Cullen 2005), but as Nagin et at. (2009) summarise, “Remarkably little is known about the effects of imprisonment on reoffending. The existing research is limited in size, in quality, [and] in its insights into why a prison term might be criminogenic or preventative.”
Understanding the effects of imprisonment has proven challenging for two reasons. The first is data availability, since in many countries the required panel datasets on crime and labour market outcomes either cannot be accessed or linked together. Second, which individuals are sent to prison is not random. While simple correlations reveal that ex-convicts have relatively high rates of criminal activity and weak labour market attachment, this could be driven by their unobserved characteristics as opposed to the experience of being in prison. An emerging literature in the past decade has recognised the limitations of correlational data and uses new approaches and new datasets to sort out causal effects.