Used on individuals who break the law to punish them, protect the public and prevent future offending through reform and rehabilitation.
Focus of the intervention
Custodial sanctions in the review included:
- prison
- young offender institutions
- other secure residential facilities
They were compared to non-custodial sanctions, which included:
- probation
- community service or fines
- suspended sentences or dismissals
- electronic tagging
- treatment-focused interventions not based in custodial settings
Most samples were exclusively made up of adults. Most samples were either exclusively males, or males made up 80% of the sample.
This narrative is based on one meta-analytic review covering 116 studies. Nearly 80% (n=93) of studies were from the USA, 5% (n=6) were from the UK and 8% (n=9) were from Europe. The rest (n = 8) were from other countries.
Effect – how effective is it?
Overall, the evidence suggests that custodial sanctions have increased reoffending, but there is no evidence of significant changes for individual studies. Effect sizes for individual studies are not provided in the systematic review.
On average, evidence suggests that custodial sanctions increase reoffending compared to non-custodial sanctions. The specific reoffending outcomes covered by the review included:
- new convictions
- arrests or charges
- reimprisonment
- not complying with conditions set
- other types of reoffending, for example, self-reported
The meta-analysis found that on average, being sentenced to custody had a statistically significant increase on reoffending. This translated to an estimated 54% of the group of offenders given a custodial sanction reoffending, compared to 46% of the comparison group (which did not receive custodial sanctions).
Primary studies with stronger research designs tended to produce smaller effect sizes than weaker designs, but still produced either no effect or an increase in reoffending.
The review found no conditions where custody reduced reoffending.
How strong is the evidence?
The review was sufficiently systematic that many forms of bias that could influence the study conclusions can be ruled out.
This evidence is taken from a meta-analytic review covering 116 studies. This demonstrated a high-quality design. It had a well-designed search strategy, calculated appropriate effect sizes with the attention to the issue of dependency, had high statistical conclusion validity and assessed the potential of publication bias.
However, the review did not sufficiently consider inter-rater reliability, statistical outliers, assessment of the risk of bias, or the influence of unanticipated outcomes on the size of the effects.
Mechanism – how does it work?
Underlying the support for custodial sanctions is an assumption that individuals are both:
- rational actors conducting a cost-benefit analysis of committing a criminal act
- deterred away from criminal behaviour and activity by the increasing severity of punishments
That is, the risk of committing a criminal act that might result in imprisonment with a loss of freedoms outweighs the reward you get from the offending behaviour.
This assumption is based on theory rather than empirical evidence. The authors present three logical and empirical reasons why this may not hold true in practice.
- Reoffending rates for those sentenced to imprisonment are high.
- The deterrence effect of a potential sanction is linked to its certainty as well as severity.
- The deterrence effect assumes offenders always perceive imprisonment as a more severe sentence to a non-custodial sanction. Research suggests this is not always the case.
More recent research suggests imprisonment can increase the risk of reoffending. This is because individual’s are exposed to risk factors during their time in prison. For example, associating with other offenders and other negative events – including exposure to violence and increased victimisation – which are hypothesised to increase psychological strain.
Protective factors such as family and pro-social contacts are at a distance to the individual. This also increases the risk of reoffending. The experience of being in prison will also create barriers to accommodation and employment.
However, the review did not test the mechanisms for change.
Moderators – in which contexts does it work best?
Moderator analyses found that the increase in reoffending following custody remained true regardless of offender characteristics, type of custodial sanction, country or time spent imprisoned.
Initial moderator analyses suggested that there was a stronger relationship between custody and reoffending for juveniles than for adults. However, further analysis found that when other factors are accounted for, the outcome of imprisonment for adults and young offenders are the same.
Implementation – what can be said about implementing this initiative?
The review gave no account of how custodial or non-custodial sanctions were implemented, nor of any implementation challenges encountered by the primary studies.
Economic considerations – how much might it cost?
The review did not mention the costs or benefits of custodial sanctions and reoffending. No formal economic analysis was provided.
General considerations
- Almost 80% of the studies included in the review are from the USA. Caution should be taken when applying the findings to other geographical contexts. However, the findings are in line with a 2019 Ministry of Justice study. This found that custodial sentences of under 12 months with supervision on release are associated with higher levels of reoffending than sentences served in the community via court orders (community orders and suspended sentence orders).
- Most of the participants in the primary studies were exclusively male or made up 80% of the sample. There should be caution if generalising the findings across the population.
- The length of follow-up in terms of reoffending across the primary studies varied. This ranged from less than one year to up to four years.
- The most common form of custodial sanctions from the primary studies was prison.
Summary
This review found that on average, custodial sanctions increased reoffending compared to noncustodial sanctions.
It is likely that individuals who are in custodial settings are more exposed to risk factors associated with criminal activity and behaviour, and have less access to protective factors to protect them from this behaviour.
The moderator analyses found no conditions where custodial sanctions reduced reoffending.
Reviews
Review one
Reference
Petrich DM, Pratt TC, Jonson CL and Cullen FT. (2021). 'Custodial Sanctions and Reoffending: A Meta-Analytic Review'. Crime and Justice, 50(1).
Additional resources
Ministry of Justice. (2019). 'The impact of short custodial sentences, community orders and suspended sentence orders on reoffending'.
Summary prepared by
This narrative was prepared by UCL Jill Dando Institute and was co-funded by the College of Policing and the Economic and Social Research Council (ESRC). ESRC grant title: 'University Consortium for Evidence-Based Crime Reduction'. Grant reference: ES/L007223/1.