What Works to Reduce Reoffending: update of the evidence on imprisonment and community disposals

Chapter 1: Background and Policy Context

1.1 Introduction

The Scottish Government last published What Works to Reduce Reoffending: A Summary of the Evidence in 2015. This paper provides an update on specific topics included within the original publication: imprisonment, community disposals, suspended sentences and early release schemes and their different impacts on rates of reoffending. This paper also explores the evidence that compares the effectiveness of imprisonment with that of community sentences. . The paper focuses on material drawn from academic papers, grey papers, systematic reviews, and other sources of published evidence. It includes research from Scotland, the rest of the UK and international literature. A second paper, on the topic of the efficacy of rehabilitation programmes, with a specific focus on risk assessment and treatment for sexual offenders, domestic abuse and violent offenders will also be produced. Each paper in the series will cover a specific topic. These topics have been selected according to Scottish Government evidence needs and policy priority as well as availability of other ‘What Works’ reviews at time of drafting.

Key messages

1. Broadly, the findings of the 2025 paper accord with the relevant conclusions of the 2015 review. In particular, where the previous review highlighted that there was a growing evidence base indicating that community sentences can be more effective than short-term prison sentences, it is evident that findings in this area have become more conclusive. Overall, there are no findings in this paper that undermine or call into question the findings of the 2015 paper.

2. When offenders are categorised by risk category in terms of reoffending, there is substantial heterogeneity in outcomes and their desistance patterns will vary according to different levels of need and risk. This suggests that generic or one-size- fits-all interventions are less effective at reducing reoffending than those interventions that target the individual needs of specific cohorts.

3. Overall, the evidence on imprisonment is mixed and generally suggests that having spent time in prison is most likely to have little impact on reoffending. While there is some evidence that prison produces a criminogenic effect, it may also have a limited rehabilitative benefit in specific settings where such programming is emphasised.

4. Overall, the evidence indicates that community sentences are more effective in reducing reoffending than short-term prison sentences (up to 12 months) and may provide greater opportunity for rehabilitation. However, this opportunity for rehabilitation is largely dependent on the offender’s ability to access rehabilitative programmes and the quality of such programmes.

5. Remand can prevent reoffending in the short-term through incapacitation effects. However, it can also impact negatively on needs that have been shown to improve long-term desistance, such as physical and mental health, employment and family relationships.

6. Where evidence is available, the literature suggests that punitive or deterrence-based interventions are less likely to reduce reoffending than interventions based on rehabilitation and reintegration. Further, where punitive or deterrence-based measures are delivered without rehabilitative or wider support, it may instead increase the risk of reoffending. The evidence strongly suggests that the most effective responses to offending typically keep people in their communities and address their underlying criminogenic needs.

7. Evidence suggests that the effectiveness of bail supervision on reducing reoffending is largely dependent on the offender’s willingness to comply with bail conditions and motivation to desist from further offending.

8. Offenders released under electronic monitoring (EM) may be less likely to reoffend than those released without monitoring. However, some studies found no difference in offending behaviour between those released with EM and those without.

9. Evidence on the efficacy of parole on reoffending behaviour is also mixed, with effects differing across jurisdictions and likely dependent on the level of support provided to the offender.

10. Evidence on the impact of age on reoffending suggests that different sentencing options may be more or less effective at different stages of the life course. Therefore, tailoring sentencing options according to the age of the offender is likely to impact on reoffending positively.

11. There is a degree overlap between the criminogenic needs of men and women, although some factors appear to have greater impact on women than men, and female offenders may also have higher non-criminogenic needs. Women may also experience these needs differently, due to structural social differences in, for example, economic independence or trauma related to domestic or gender-based violence. Binge drinking in particular was identified as a major predictor of women’s reoffending but was less strongly linked to reoffending for men. This has implications for the effectiveness of interventions. In terms of physical and mental health, the negative impacts of imprisonment in particular, may be greater for women.

1.2 Aims

This paper provides an update to the previous versions of What Works to Reduce Reoffending (2015 and 2011) series. It focuses on imprisonment and community disposals, and reviews literature published since the last update in 2015. The paper is centred around a number of specific aims to:

• reflect on the different measures of effectiveness, such as desistance or reintegration, that can be used when assessing interventions that may reduce reoffending
• summarise the evidence relating to the impact of different interventions, specifically imprisonment, community disposals, suspended sentences and early release schemes, on rates of reoffending
• where appropriate, compare different interventions and their impact on reoffending rates
• outline the existing gaps in the literature and some of the methodological issues faced by researchers working in this area

1.3 Scope

As discussed above, this paper covers the impact of imprisonment, community disposals, suspended sentences and early release schemes on reoffending. Research aims have been identified above and an appraisal of the relevant evidence has been undertaken. The search methods adopted have been rigorous, however, due to time limitations the review is not comprehensive of all published material on the relevant topics.

The literature on ‘what works’ to reduce reoffending is largely made up of one-off quantitative evaluations comparing reoffending rates of study participants with control groups, often quasi-experimental in design or consisting of randomised control trials. Where possible, this review has prioritised systematic reviews and meta-analyses that bring together these stand-alone studies, in order to leverage the greater analytic power of aggregating these results. The review includes research from Scotland, the rest of the UK and international literature.

Evidence has been selected for review on the basis of its relevance. Factors that were considered included:

• Whether the evidence is Scotland-specific or whether the research took place in a jurisdiction similar enough to Scotland that the findings are relevant.
• Whether the paper has direct relevance to reducing reoffending.
• Whether the evidence has been published since the last What Works to Reduce Reoffending update (2015).
• Whether the research is considered a ‘key study’ either by Scottish Government analysts, Scottish Government policy colleagues or within the academic field.

The review will highlight any areas where the evidence on the efficacy of an intervention is particularly strong or areas where significant development has taken place since 2015.

1.4 Search strategy

This paper provides an update to the evidence base collated in the 2015 review. This project employs a literature review methodology. This is deemed the most appropriate approach as the research questions are focused and the timescale in which the evidence is required is short. The intention is not to review the entirety of the material on the subject, rather this literature review synthesises and evaluates what the researcher was able to read in the required timeframe.

Initial literature searches took their starting point from bibliographies of existing literature reviews on reoffending, including using previous iterations of the ‘What Works’ series to identify key academics in the field. Supplementing this initial evidence gathering, key academic and grey literature was explored via SG library searches and augmented with Google Scholar searches. Searches were conducted using key search terms such as ‘reoffending’, ‘recidivism’, ‘desistance’, ‘deterrence’, etc., and further ‘snowballing’ from relevant reference lists, particularly those included in systematic reviews. Key terms as identified in these initial sources of evidence were then used to refine the search terms for subsequent searches.

Inclusion criteria:

• Material stemming from academic papers, grey papers, systematic reviews, and other sources of published evidence.
• Material published in English between 2015 and 2024.
• ‘What Works’ publications in both the UK and internationally, similar to those produced as part of the Ministry of Justice (MoJ) analytical series.

Exclusion criteria:

• Material in languages other than English and published pre-2015.
• Material from opinion articles and editorials.

1.5 Background

1. In recent years there has been a rapid rise in the prison population in Scotland. Whilst there are a number of factors that have led to a high prison population, the relatively high reconviction rate in Scotland when compared to other European countries is an important consideration (although caution must always be taken when comparing different justice systems serving different communities) (ScotPHO 2024).

2. Ensuring that prisons function effectively, focusing on those who pose the greatest risk of harm, and providing programmes that help reduce the risk of reoffending, is a priority for Scottish Government (Scottish Government 2024a). There is therefore strong interest in understanding which approaches and interventions can impact on reoffending and, in turn, rates of imprisonment.

3. In Scotland, the reconviction rate is a key measure for estimating how many people who have been convicted of a crime go on to commit crimes again. However, reconviction rates are a proxy measure for recidivism, as not all offending is reported to the police, and not all offences committed or recorded by the police will result in a conviction. Therefore, reoffending rates cannot be considered the same as reconviction rates. Reconviction rates and trends over time, however, can be one indicator of the effectiveness of the criminal justice system in the rehabilitation of offenders. Moreover, measuring reconvictions does allow us to track the pressure reoffending is putting on the criminal justice system over time (Scottish Government 2024b). However, It is important to remember that many other factors may influence reconviction rates, such as policing budgets and priorities, prosecutorial priorities and practice, changes in court processing, and wider social factors like levels of economic deprivation and evolving cultural norms.

4. The reconviction rate is presented in national statistics as the percentage of offenders in the cohort who were reconvicted one or more times by a court within a specified follow up period from the date of the index conviction^[1]. For most reconviction analyses in national statistics, the follow-up period is one year.

5. In 2020/21, the most recent year for which data is available, the reconviction rate increased by 2.6 percentage points over the previous year from 24.3% in 2019-20 to 26.9% in 2020-21. The average number of reconvictions per offender increased by 8% in the same period from 0.41 to 0.44. Males are reconvicted more often, on average, than females. In 2020-21, the average number of reconvictions per offender for males was 0.45, which was 16% higher than the value of 0.39 for females (Scottish Government 2024b).

6. However, the most recent reconviction figures are still lower than those ten years ago. Between 2011-12 and 2020-21, the reconviction rate fell by 2.8 percentage points from 29.6% to 26.9%, and the average number of reconvictions per offender decreased by 20% from 0.55 to 0.44 (Scottish Government 2024b)

7. Our latest reconvictions data 2020-21 shows the highest reconviction rate was for offenders whose index conviction was given in courts in Ayrshire, East, North and South (30.3%). Dumfries and Galloway had the highest number of reconvictions per offender on average (0.48). Excluding convictions in the High Court, the lowest reconviction rate (10.6%) was given in courts in Argyll and Bute, which also had the lowest average number of reconvictions per offender (0.15).

8. However, comparing reconviction rates across geographical areas should be interpreted with caution, especially if deploying reconviction rates as a performance outcome. The reconviction rate does not take account of (or control for) key factors that could explain variation in reconviction rates across different areas or courts. These factors include differences in offender characteristics, offence mix and population size. Local authority reconviction rates tend to fluctuate year to year and smaller local authorities tend to have larger fluctuations as they have a lower number of offenders. Slight between-year fluctuations in the numbers of offenders reconvicted may lead to larger changes in the reconvictions in percentage terms compared to local authorities with larger numbers of offenders. All of these factors make it difficult to meaningfully compare different areas.

9. At the time of writing, May 2025, Scotland is facing an unprecedented prison population crisis. The average daily prison population in 2023-24 was 7,860. This is an increase of 5.8% from the previous year (7,428) and the highest level since 2019-20 (8,198) (Scottish Government 2024c). The latest prison population projection modelling indicates that it is likely that the overall prison population will increase between 1^st April 2025 and 30^th September 2025 (Scottish Government 2025a). Following a rapid growth in the remand population to September 2020, the prison population stabilised between 7,300 and 7,600 across 2020-22. However, growth over 2023 led to a higher prison population. The prison population has consistently exceeded 8,000 since the end of March 2024, reaching 8,379 on 4^th March 2025 (Scottish Government 2025b).

10. The number of custodial sentences increased by 9% from 8,169 in 2021-22 to 8,923 in 2022-23, Between 2013-14 and 2022-23 all custodial sentences decreased by 37% (Scottish Government 2024d).

11. Previous action to address the increasing prison population in Scotland has included a presumption against short sentences (PASS) (3 months or less), introduced by the Criminal Justice and Licensing (Scotland) Act 2010 which commenced in February 2011. In June 2019, the extension of this presumption to 12 months or less was approved by the Scottish Parliament in the Presumption Against Short Periods of Imprisonment (Scotland) Order 2019. This presumption stated that a court must not pass a sentence of imprisonment for a term of 12 months or less unless it considers that no other method of dealing with the person is appropriate (Scottish Government 2022).

12. Following the extension of the presumption, the proportion of custodial sentences equal to 12 months or less fell by 4% between 2018-19 and 2019-20. Sentences of 12 months or less made up 75% of all custodial sentence lengths in 2020-21, 7% lower than 2011/2012, when they represented 82% of all custodial sentences (Scottish Government 2022).

13. In February 2025, the Prisoners (Early Release) (Scotland) Act 2025 changed the point of release for most short-term prisoners - those serving sentences of less than four years. This change will apply to eligible prisoners currently serving short-term sentences and those who are sentenced in future. Currently most short-term prisoners are released from prison custody once they have served half (50%) of their sentence, the Act changes this release point to following two fifths (40%) of their sentence (Scottish Government 2024e).

14. Recent Scottish Government analysis investigated how the number of short sentences (up to 12 months) issued by the courts related to the number of the overall sentences being served by people in prison. The paper found that a significant number of people appear to be moving through the prison system with short custodial sentences several times over a year and that these multiple short disposals in effect create a longer custodial sentence. In turn, this effect extends overall custodial sentence lengths. For example, figures indicate that 80% of short sentences in 2021-2022 were given to people who were already on remand or were serving previous custodial sentences (Scottish Government 2024f)

15. In response to the growing prison crisis, in February 2024, Cabinet Secretary for Justice and Home Affairs announced in their Parliamentary statement on the prison population that an independent review of sentencing and penal policy would be established. Although this paper will primarily aim to provide an update to the What Works to Reduce Reoffending series, it is hoped that this will be informative to the upcoming Review.

1.6 Limitations

There are several limitations that should be noted in relation to the evidence synthesised for this review. One notable issue is that it is difficult to establish the causal effects of a sentence on an offender. For example, while a sentence may be important in reducing reoffending, there are a number of other factors that affect the likelihood of recidivism. For example, the characteristics of the offender, including the nature and severity of the offending, their age and gender, and the strength of their familial and social bonds. For these reasons, difficulties in comparability between studies also arise, where cohorts are not like for like and a reduction in reoffending may be due to a myriad of reasons not considered by the research.

Additionally, there are several limitations related to the literature review approach utilised in this paper. Whilst every effort has been made to produce a rigorous review within the timescales, the search strategy has been necessarily limited and this means that the evidence reviewed and included is not comprehensive of all published research on the topic. Furthermore, whilst the research reviewed has been subject to methodological scrutiny, a full critical appraisal of the sources has not taken place, as would be required in a Rapid Evidence Assessment.