Skip to main content


Rehabilitation programmes – do we know whether they work?

Penelope Gibbs
04 Jun 2019

Anything which helps those who commit crimes change their behaviour has to be a good thing. For many years the rehabilitative efforts of prison and probation have been focused on behaviour changing courses or interventions. Evidence is amassed and analysed as to how one or the other programme seems to be successful in reducing reoffending. Programmes provide an easy way of “measuring” whether those who have committed crime are on the path to turning their lives around and of monitoring the performance of organisations (such as CRCs) tasked with fostering rehabilitation.

Last year Transform Justice published a report on domestic abuse and how we could reduce it. Perpetrator programmes are part of the answer but its not clear they all work equally well. We asked for the evidence that the programmes delivered in prisons and for those on probation worked and were surprised when none was forthcoming. No outcome evaluations had been done on the programmes running now (the main one has been running for five years) – the government is scoping how they might evaluate them. All the main HMPPS programmes are accredited – but all this means is that a panel of experts has said that the programme’s design is sound.

The news that the government does not know whether its own domestic abuse programmes work hardly caused a ripple. But I was disturbed, particularly since excellent out of court programmes are frowned on. I put in a freedom of information request to find out what other prisons and probation programmes were not (outcome) evaluated and the results were even more disturbing – most programmes that are running, and many that no longer run have no outcome evaluations. This means we have no idea whether they work or not. We calculate that 118,096 people have started prison and probation programmes which do not have impact evaluations.

It is really surprising that the government has not sought to find out whether these programmes work. A few years ago they had a bad experience. In 2017 they found out that the programme they had been running since 1992 to help prisoners reduce their likelihood of committing sex offences in fact had the opposite effect – doing the programme increased the chance that someone would offend again on leaving prison. The sex offender treatment programme had in fact been evaluated in 2003 – then it appeared to have no effect in reducing sex offending but did reduce “the overall combined sexual and violent reconviction rate”. But the problem with programmes is that they can work in different ways in different circumstances. Trainers may deviate from the original programme, or it maybe delivered to a different cohort. Anyway, the Sex Offender Treatment Programme was running another decade before they decided to run a new evaluation. When this showed such negative results, they quietly axed the programme and the fiasco was only revealed through Frances Crook of the Howard League and journalist David Rose.

The problem with not evaluating programmes is not just that we don’t know if they work (or worse – if they are making people more likely to re-offend) but that we are forcing people to do these programmes and punishing non-completion. Those in prison applying for parole have to complete certain programmes to be eligible for release and there are many prisoners, particularly those on IPP sentences, who have spent extra years in prison mainly because they have not completed programmes. People are ordered to go on programmes as part of community sentences. If they don’t complete them, they can be prosecuted for breach and punished with imprisonment. It seems grossly unfair that we should deprive people of their liberty for failing to do courses which may or may not work.

Plenty of drugs and interventions used by the health service have a shaky evidence base. If they go wrong they can do people great harm. The same is true of programmes prisoners are pressurised to go on. Those who have committed crimes want support to turn over a new leaf. But we let them down if we make them do courses and never seek to find out whether they help or hinder rehabilitation.

NB In our spreadsheets etc we are referring specifically to outcome/impact evaluations not process evaluations – the latter are not designed to assess whether programmes have an impact