University of Cambridge
How Research Design can Impact the Evaluation of Policing Interventions: The development of an integrity model as part of a randomised controlled trial to evaluate a police-led diversion scheme
Abstract
dc:description.abstractThe trend in England & Wales is to prosecute and incarcerate more offenders, and projected figures for the prison population suggest that the prison estate will need to be significantly developed to accommodate them. Yet, a consensus has developed derived from research studies and meta-analysis that custodial sanctions, overall, do not reduce reoffending. Whist it is a key responsibility for the police in England & Wales to bring offenders to justice, the police also retain the option to deal with less serious offending by way of out-of-court disposals, known as OOCDs. These are predominantly in the form of cautions and community resolutions, but such disposals are mainly ‘administrative’ in terms of what is expected from the offender. Offender Desistance Policing is a criminological construct and process designed to divert offenders predominantly from prosecution but also from the criminal justice system altogether. Although still representing an OOCD, it offers offenders the opportunity to reflect on their behaviour and to bring about sustainable desistance from offending. Checkpoint is a police-led diversion scheme based on Offender Desistance Policing and is an offender management programme developed and implemented by a police force in the Northeast of England. The hypothesis for Checkpoint was that it could reduce reoffending post-intervention in comparison to normal criminal justice processes and was tested using a randomised controlled trial (RCT). Whilst the integrity of the Checkpoint intervention or ‘treatment’ was found to have reached the required level, the integrity of the random assignment into the experiment was less so. An in-depth investigation was conducted to establish what impact the degradation in the integrity of the experiment would have on the analytical approach. This led to the development of an ‘integrity model’ for experiments to guide the measurement of key statistical parameters that underpin the analysis of RCTs. As only a proportion of the Checkpoint treatment cohort were ‘treated as assigned’, high experiment integrity sub-samples were established in order to provide a sensitivity analysis in addition to the ‘intention to treat’ approach. The prevalence of reoffending did not differ between Checkpoint and the controls, but there was clear statistically significant evidence that the frequency of offending reduced to a greater extent with the Checkpoint treatment cohort in the follow-up period. This reduction correlated to improved experiment integrity, that is, the reduction in the frequency of reoffending was greater in the Checkpoint treatment cohort as more of that cohort were treated as assigned. How harm of reoffending varied from pre- to post-intervention was also considered and found to be seriously impacted by outlier crime severity score values. However, evidence of a greater reduction in the Checkpoint treatment cohort could be found by applying different statistical techniques to account for these complexities in the dataset. An economic analysis using the results from the trial found that diverting offenders from prosecution results in substantial non-cashable savings for the criminal justice system, even when there is no difference in reoffending compared to the controls.
Degree
thesis:*- Name dc:type.qualificationname
- Doctor of Philosophy (PhD)
- Level dc:type.qualificationlevel
- Doctoral
- Grantor dc:publisher.institution
- University of Cambridge
- Year dc:date.issued
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Cooper, John
- Advisor dc:contributor.advisor
-
- Sherman, Lawrence
Subjects
dc:subject × 5Rights
dc:rightsIdentifiers
dc:identifier.*- DOI dc:identifier.doi
- https://doi.org/10.17863/CAM.113111
- OAI identifier oai:identifier
- oai:www.repository.cam.ac.uk:1810/375397