Offending Behaviour Programmes

Key sources

  • REVIEWS BY GOVERNMENT RESEARCHERS[1] – ‘What works’ reviews (including summaries of international systematic reviews and meta-analyses[2]) are available in Goldblatt and Lewis (1998)[3], Harper and Chitty (2005)[4], Hollis (2007)[5] and Chitty (2009)[6], and MOJ (2010c)[7] – Green Paper Evidence Report for Breaking the Cycle
  • UK FINDINGS[8] – Friendship et al. (2002), Falshaw et al. (2003), Cann et al. (2003) McDougall (2007), Sadlier (2010)[9], and Travers (2011)[10] report prisons’ findings. Reports on probation evaluations[11] are in Ong (2003)[12], Stewart-Ong (2004)[13], Hollin et al. (2002)[14], Hollin et al. (2004)[15], Palmer et al. (2007)[16], and Hollis (2007)[17].
  • OTHER REVIEWS AND DISCUSSION PAPERS – Davis (2008) summarised research for the National Audit Office[18]. Recent debates about ‘What Works’ evidence can be found in: Raynor (2008)[19] and (2004)[20], Sherman (2009)[21], Tilley (2009)[22], Hope (2009)[23], Stanley (2009)[24], and Hollin (2008)[25].
  • NPS PERFORMANCE REPORTS – NPS (2008)[26] [27], NPS (2007)[28] and NPS (2006)[29].
  • NOMS ANNUAL REPORT AND ACCOUNTS (2011)[30]

Key points

1)    What works?

  • The most recent international systematic reviews of evidence strongly support the use of cognitive skills programmes to reduce offending[31].
  • Estimates of how much overall difference they can make have varied between 4 and 15 percentage points reduction[32] [33]. The two most recent analyses of England and Wales prison (not community) data – Sadlier (2010)[34], and Travers (2011)[35] – finding, respectively,  6 percent and 6.4 percentage points reduction in reoffending less than comparisons.
  • Programmes that work address thinking and attitudes and provide skills training, including interpersonal problem solving and anger control[36]. They need to be consistently implemented and be delivered by specially trained staff using pro-social modelling, and as part of an appropriate package of intervention.
  • Cognitive skills programmes appear to work best for higher risk offenders[37].
  • UK evaluations[38]  have shown mixed results. There are promising findings, but these have mostly been limited to studies which suffered from possible selection effects or weak comparisons[39]. Palmer et al. (2007) reported positive findings from the probation pathfinders, using statistical techniques to control for some possible selection effects. The most recent prison results for ETS show effects of 6 percentage points (Sadler, 2010) and 6.4 percentage points (Travers, 2011).
  • Evaluations have consistently pointed to the challenges of large-scale implementation.[40]

2)    What are the implementation problems?

  • The scale and pace of change affects quality of delivery – Lipsey’s (2001)[41] review noted that small pilots were more effective than routinely rolled-out programmes. In the UK, Blud et al. (2001)[42] and (2003)[43] concluded that, for prisons, quality of implementation may be as important for effectiveness as details of the programme itself (Impact was highest when tutor turnover was low, and early positive findings were not replicated when prison implementation was quickly scaled up). Probation evaluations have made similar observations[44].
  • Referring the right offenders Issues with unclear targeting were noted in the probation evaluations[45]. It has been suggested that this was influenced by challenging national targets set for completions. Effects are weakened if programmes are offered to low risk offenders.
  • Drop out reduces effectiveness, and may lead to higher reconviction – Probation evaluations of the pathfinder programme found that only a third of offenders completed.[46]  Several studies have suggested that those who start but failing to complete programmes have higher rates of reoffending. McMurran’s (2007)[47] systematic review of 16 studies supports this view, particularly for community samples.

3)    OBPs in the Community

  • Completions – Targets for accredited programme completions were met in 2007-8 for the fourth successive year[48].  Across all programmes (not including sex offenders and domestic violence), there were 14,531 completions (104% of the target). More recently numbers of completions may have fallen[49]
  • Tackling drop-out – Local areas have analysed reasons for drop out[50] and embarked on retention initiatives[51]. Nationally, the Probation Service has reported on levels of attrition by programme and at different stages. In 2006/07, 68% of offenders who started programmes completed. However, 20% of offenders referred were low risk (below the OGRS risk of reconviction level recommended)[52].
  • The new accredited Thinking Skills Programme (TSP) has been rolled out nationally.
  • Improvement in targeting NOMS (2010)[53] report that probation exceeded its target of 85% of programme starts meeting criteria in 2009/10 (89%) and 2010-11 (94%).

In the pipeline

1)    Large cohort studies

Government researchers in OMSAS[54] are working on a large cohort study of offenders under community supervision. This study is expected to show how the effects of interventions interact with different offender needs and characteristics: that is, what works with whom.

2)    Benefit-Costs

OMSAS unit cost study for interventions delivered to adults in prisons and for YOT practitioners. NOMS Specification, Benchmarking and Costings programme (SBC) is developing cost data to support understanding of cost-effectiveness.

The select committee report Cutting Crime: The Case for Justice Reinvestment[55] pointed to government response to the high costs of imprisonment. In this context, the ability to show the value for money of community interventions is likely to prove critical.