The Youth Endowment Fund (YEF) funded Cordis Bright to deliver a Randomised Controlled Trial (RCT) of Cerridwen, a mentoring and case management intervention developed by Media Academy Cymru (MAC). Cerridwen supports young people aged 10 to 17 who are involved in, or at risk of, violent behaviour.

Since autumn 2023, we have worked closely with MAC to design and deliver an RCT that is both methodologically rigorous and feasible in practice. Between April and December 2024, we completed the internal pilot phase. In this blog, we share some lessons about running RCTs in practice - what worked, what we had to adjust, and what may help others embarking on similar trials.

What did the internal pilot find?

The pilot showed that Cerridwen could be implemented as planned. With initial recruitment targets exceeded, the trial is now moving to the full efficacy phase. You can read the detailed internal pilot findings in our published report.

Key lessons from the internal pilot?

1. Be transparent about randomisation, and explain why it matters

Randomisation is key to understanding impact, but it can feel uncomfortable for frontline staff and partners. Some questioned whether it was fair to randomise young people, or whether it aligned with youth-led practice.

We found that honest conversations, transparency about the reasons for randomisation, and strong safeguarding measures helped build trust and confidence in the trial design.

2. Make RCTs feasible in complex frontline settings

The key lesson was not that RCTs “can” work in practice, but rather what conditions make them work effectively in demanding, real-world environments. Traditionally, in youth violence reduction, some stakeholders have suggested that RCTs are hard to implement. Cerridwen shows success depended on:

• Robust and user-friendly processes: We co-designed systems with MAC and partners for referral, randomisation and data collection, making sure they worked well for the people using them.
• Strong early recruitment: Thanks to MAC’s commitment and effective local engagement, we exceeded early recruitment targets.
• Practical support for frontline staff: We developed a practitioner handbook and delivered tailored training so that MAC staff felt confident supporting young people to complete questionnaires consistently.

3. Prioritise collaboration to keep deliver on-track

Collaboration between evaluators, funders and delivery partners was essential to keeping the trial on track:

• Regular problem-solving spaces: Weekly check-ins and monthly cross-partner meetings allowed issues to be identified early and addressed collaboratively.
• Building buy-in: MAC were proactive in engaging referral partners, listening to concerns and building trust across health, education and children’s services. This proved vital for both recruitment and retention of young people.
• Remaining flexible together: This collaborative approach helped us respond quickly when challenges arose, such as seasonal dips in referrals.
• Testing under real conditions: We used the internal pilot to trial referral, randomisation and data processes under real-world pressures.
• Feedback loops: Tools like RAG ratings, data audits and feedback and reflection sessions with delivery staff gave us real-time insight into how the trial was running and what needed to be adjusted.

4. Treat the pilot as a testbed, not a tick box: The pilot was not a formality – it was a genuine opportunity to learn what worked, what didn’t and what needed to be adapted before further scaling up. The pilot incorporated:

5. Keep equity central to design and delivery

Young people from racially minoritised backgrounds repeatedly face structural barriers to support and are disproportionately represented in the youth justice system. Cerridwen was designed with a strong commitment to race equity. Achieving this in practice required deliberate steps:

• Culturally responsive recruitment: We worked with MAC and local partners to make sure referral processes were inclusive and accessible.
• Co-developing research tools: Our team and MAC designed questionnaires and processes together.
• Community-led engagement: MAC used trusted local relationships to reach young people who might otherwise have been missed.

Taking this learning forward

The internal pilot confirmed that Cerridwen could be delivered within the structure of an RCT, but more importantly it showed what it takes to make trials like this work in demanding, real-world contexts. The key ingredients included co-designed systems, strong collaboration, careful attention to equity, and a willingness to adapt when challenges arose.

As we move into the efficacy phase, our focus will remain on maintaining rigour while staying realistic, and on ensuring that the trial produces insights that are useful not only for YEF and MAC, but also for others developing and evaluating violence reduction programmes.

We hope these insights are useful for others designing and delivering violence reduction interventions, and for those considering how best to evaluate their impact.

If you would like to learn more about our evaluation of Cerridwen, please see our project page, and if you have any thoughts of questions about this work, please contact Matt Irani or Madeleine Morrison.