Research and analysis

Technical note: prisoner Universal Credit claim preparation test

Published 3 November 2026

This technical note accompanies a slide deck and report evaluating a test of approaches for prison work coaches (PWCs) to support prison leavers in preparing their Universal Credit (UC) claim and enabling them to submit it on the day of their release. 

The test was carried out in 16 prisons across England and Wales, along with 2 in Scotland.  The starting date for each prison was staggered between September 2022 and August 2023, with a formal evaluation period covering 12 months from this date.  

The evaluation consisted of qualitative research carried out by the Department for Work and Pensions (DWP) social researchers with prison work coaches and other staff involved in the test, analysis of tracker data completed by the PWCs and quantitative analysis carried out using linked data from the Ministry of Justice (MoJ) and DWP to analyse the time taken by prison leavers to claim UC following their release. 

The information gathered was used to evaluate how effectively the approach had been adopted and to understand issues faced during its implementation.

1. Qualitative Research 

1.1 “Business As Usual” Activity 

Prior to the start of the test, PWCs within the test prisons were interviewed online and asked to describe the support currently provided as part of their ‘Business as Usual’ processes. 

Three structured focus groups and one interview were also carried out online with PWCs from 11 comparable prisons which were not part of the tests to understand the type of support they provided to prison leavers.  

This information was used to understand which of the test prisons already provided prison leavers with some support in making a UC claim prior to the start of the test, and to what extent any processes existed in comparable prisons who were not part of the test. 

1.2 Focus Groups 

A series of focus group sessions were run with PWCs from 16 out of the 18 test prisons shortly after the test approach had been implemented.  Two prisons were not included following delays in starting the test. 

The purpose of these sessions was to identify any issues faced in implementing the test approach and to understand how closely PWCs were able to follow the test process.  This information was fed back to the project working group on an anonymous basis to inform the continued implementation and management of the test.  The focus groups also provided an opportunity for PWCs from different prisons to share information and learn from each other.  

1.3 Completion Interviews 

Face to face visits took place in 6 of the test prisons following the completion of the test period.  During these visits, semi-structured interviews were held with PWCs and other staff working in HM Prison and Probation Service (HMPPS) who had been involved in the tests (for example, ID and Banking Leads, Prison Employment Leads, Custody Managers and Heads of Reducing Reoffending).   

Separate online interviews were held with PWCs from the 2 prisons in Scotland which had implemented the test approach. 

Two focus groups were also held with some ‘Single Points of Contact’ (SPOCs) from Jobcentre Plus, who were representing job centres which had been involved in the warm handover element of the approach. 

The transcripts from these interviews and focus groups were analysed using NVivo to help draw out themes and provide insight into how effectively different aspects of the approach had been implemented and the issues which were identified and addressed as the test progressed.   

Staff were also given the opportunity to provide feedback on the merits of the test approach and the potential for adopting it across the prison estate, as well as their perceptions of the impact that it had on the prison leavers they engaged with. 

2. Tracker Data 

PWCs were asked to complete a tracker to log key information about prison leavers who had been invited to participate in the tests.  They recorded information at different stages of the process relating to:  

• whether the individual intended to participate in the test  

• the date and outcome(s) of the 3-week pre-release appointment  

• the date and outcome(s) of the day-of-release appointment 

• any information available about whether the individual attended their scheduled claimant commitment appointment following their release

The data was anonymised before being shared with DWP analysts by removing all personal details, so only aggregate information and analysis can be produced for the test.  It was not possible to track individuals or link the tracker data with wider DWP benefit data. 

The information from the tracker was used to monitor implementation and progress of the test, providing information about how many prison leavers were engaging with and any issues encountered at different stages of the process.  Aggregated information across each prison was reported at monthly Working Group meetings and has been presented in the final evaluation report.

3. Quantitative Analysis 

3.1 Data analysed 

A quarterly data-share between MoJ and DWP provides analysts with anonymised information about individuals who have been released from prison and the DWP benefits they claim following their release.  This enabled analysis of the time between release and the start of a UC claim. 

Each quarterly extract of the data contains information on individuals who have left custody in the last 12 months.  Personal data such as names, dates of birth and postcodes of known addresses, are securely shared with DWP who apply a fuzzy matching algorithm to link this information with data from DWP systems.  A match is typically found for around 90% of individuals, which ensures that the data share is broadly representative, if not complete. 

3.2 Measures analysed 

The quantitative analysis carried out was based on all individuals who were released from the test prisons and used to assess whether there had been any difference in times taken to claim UC before and after the start of the test.   

Claim times were determined by comparing the date of release recorded by MoJ with the start date of the UC claim recorded by DWP.  The recorded release dates may differ slightly from the actual date an individual left prison as they are based on the date an individual was no longer included on the roll-call, The start date of the UC claim can also differ from the date the claim was submitted. Any inaccuracy should be consistent across prisons and over time meaning that comparisons are still valid. 

Primary measures: 

• the proportion of UC claims made within 1 day of release 

• the proportion of advance payments approved within 1 day of release 

Secondary measures: 

• the proportion of prison leavers making a UC claim 

• the average (mean) time taken to claim UC 

• the proportion of UC claims made within 3 days of release 

• the proportion of UC claimants receiving an advance payment 

• the average (mean) time taken to approve an advance payment 

• the proportion of advance payments approved within 3 days of release 

• the average (mean) time taken to accept the claimant commitment 

• the proportion of claims leading to a full payment 

• the average (mean) time taken to receive a full payment 

Only claims which were made within 30 days of leaving prison were included in the analysis as a way of limiting this to individuals with an immediate need to claim UC.  Most claims are received within a few days of an individual leaving prison.  Where claims are made much later, this may reflect a change in an individual’s circumstances, rather than reflecting their needs at the time of release. 

3.3 Non-participating comparator prisons 

Comparator prisons with similar characteristics to the test prisons are compiled by the Ministry of Justice (MoJ) based on performance hub data, using an algorithm to match each individual prison with up to 8 ‘most similar’ prisons based on characteristics such as prison type (for instance, closed, open, local), geographical location, operational capacity and types of prisoners held (young offenders / gender / category / average sentence length).  Test prisons were excluded from these comparator groupings (without replacement) meaning that between 4 and 8 suitable non-participating comparator prisons were identified for each of the test prisons. 

3.4 Difference in difference approach 

A difference-in-difference approach was used to analyse changes in the primary measures before and after the test and between test and comparator prisons.  Formal comparisons have been made based on activity in the 6 months before the test commenced and activity in the final 6 months of the test period. 

Additional data were extracted up to 30 months prior to the test, during which time we would ideally see a similar trend in measures for test and comparator prisons – though not necessarily at exactly the same level.  Data from 24 months after the start of the test were also analysed to check that any impact observed during the test had been maintained – many of the prisons continued with the approach after the official test period had finished.