Media Academy Cymru’s Cerridwen Project
A six-month mentoring programme for children in Cardiff, Swansea and Merthyr Tydfil who are at risk of involvement in violence.
Evaluation type
Funding round
A trusted adultActivity Type
Setting
Evaluator
Completed
July 2025| Project | Funding | Region |
|---|---|---|
| Media Academy Cymru | £2,000,000 | Wales |
What does this project involve?
Cerridwen is a six-month mentoring and case-management programme for 10–17-year-olds at risk of involvement in violence that aims to prevent violence and offending. Delivered by Media Academy Cymru (MAC), children receive 16 weeks of weekly, one-to-one, two-to-three-hour meetings with their case manager. In these sessions, the case manager works with the child on reducing their risk of involvement in violence, focusing on understanding their feelings, improving communication, and building the child’s empathy and identity.
Why did YEF fund this project?
As the YEF Toolkit explains, mentoring is effective in both reducing crime and the behaviours associated with crime and violence. The research suggests that, on average, mentoring reduces violence by 21%, all offending by 14%, and reoffending by 19%. However, we lack rigorous studies of mentoring in a UK context, and so YEF is determined to build the evidence base across England and Wales.
Therefore, we funded this pilot study as the first stage of a large-scale randomised controlled trial (RCT) evaluating the Cerridwen model. The pilot study aimed to answer whether recruitment, randomisation, retention and data collection were effective, establish whether evaluation tools were appropriate, and ascertain what sample size would be required for a future efficacy study. It also aimed to answer whether Cerridwen could recruit and retain enough children for a robust study, in addition to exploring the implementation of the programme.
To address these aims, the evaluation established an efficacy RCT, with an initial, internal pilot. Children were randomised at the individual level. Between April and December 2024, 74 children were randomised to the treatment group to receive Cerridwen, and 77 children were randomised to the control group to receive a lighter-touch intervention (up to eight one-to-one meetings with a case manager over five months). Children were asked to complete a baseline survey and a survey five months into the programme. This included a range of measures, including the Self-Report Delinquency Scale (SRDS) and the Strength and Difficulties Questionnaire (SDQ).
The evaluator also conducted 15 in-depth interviews with children who received Cerridwen, 14 interviews with project staff and 22 interviews with wider programme partners (including referral organisations and participants’ family members). Of the 151 children who were randomised, 90% (136) came from a White ethnic background (slightly lower than the proportion of Wales who identify as White – 94%). 6% identified as being from a mixed ethnic background, 2% from a Black ethnic background and 1% from an Asian ethnic background.
Key Conclusions
| Children were successfully recruited, randomised and retained in the pilot phase of the trial. 56 children were recruited in the first three months (versus a pre-defined aim of 36). 79% of children in treatment and control groups had completed follow-up data collection or were still receiving support in December 2024. |
| Data collection processes were effectively established and embedded; over 79% of participants completed at least 80% of the data collection scales used at baseline and after five months. Evaluation tools were found to be reliable and practical, with good internal consistency and alignment with theoretical expectations. |
| Cerridwen was delivered largely as intended, with adherence to core modules. However, session lengths varied, particularly for some children with neurodivergence, requiring adaptations to sessions. The overall length of the intervention was also longer than anticipated. The programme’s flexibility and youth work approach were valued by participants, with strong engagement from children and positive feedback from parents, carers and staff. |
| The majority of children received the minimum expected number of sessions or are on track to do so. MAC also demonstrated that they have necessary delivery capacity, contacting 98% of 197 referred children within five days. 80% of children also began the programme within 15 days of their referral being accepted. |
| The internal pilot has demonstrated that a full efficacy trial of Cerridwen is feasible. A target sample size for the efficacy study of 592 was originally set to achieve a minimum detectable effect size (MDES) of 0.2; given resource and time constraints, this has been revised down to 367 (which could achieve an MDES of 0.25). |
What will YEF do next?
The internal pilot has demonstrated that a full efficacy trial of Cerridwen is feasible. YEF is, therefore, proceeding with a full trial.