Delivering circles of support and accountability to protect communities from sexual harm and sexual reoffending


An Evaluation of 131 Circles of Support and Accountability

This week we present a guest blog by Oli Preston an Associate for Research in Practice. He talks about the evaluation of 131 Circles of Support and Accountability Delivered by Circles South West from 2017-2022.

Circles of Support and Accountability (CoSA) are a community-based approach aimed at preventing sexual offences against children, young people and vulnerable adults. A ‘circle’ is formed of a person who presents a risk of offending (the ‘core member’), trained volunteers from their local community, and a professional coordinator who oversees the circle, offers support and provides a safeguarding role. Providers of circles are accredited by a national body, Circles UK, who ensure continuity of the format across England and Wales.

Circles South West is the provider of circles across South West England. As well as providing the standard CoSA approach, Circles South West has developed three new circle formats through Big Lottery funding secured in 2016. These adaptations are aimed at (1) young people with harmful sexual behaviour; (2) people with intellectual disabilities who present a risk of harm; and (3) people who are leaving prison after serving time for sexual offences.

Typically, a circle lasts for around a year, during which time approximately 3 – 5 volunteers meet weekly with the core member. The circle discusses a variety of topics, primarily addressing areas of ‘dynamic risk’ of offending. Dynamic risks are factors which can change over time, such as hobbies and activities, strategies for managing risky behaviour and positive relationships. Each circle is tailored to the core member; their needs, requirements and offending histories.

In 2017, Research in Practice – a charity providing research and evaluation services – partnered with Circles South West to carry out an extensive evaluation of the circles they provide. Understanding the effectiveness of circles presents a challenge to research given (1) the complexity of sexual offending and the varying underlying causes, and (2) the difficulty of collecting the required quantity of reliable evidence. To address these challenges, the evaluation looked at previous research into circles and other restorative approaches, the existing data available, and worked with Circles South West stakeholders to develop a clear evaluation plan.


Both new and existing data sources were used, including: routine data collection; risk reviews of the core members; bespoke questionnaires designed by Research in Practice; and, a range of validated psychometric tools relating to wellbeing and loneliness. Over 5 years, the evaluation looked at 131 circles, of which 65 circles were ‘complete’ in an evaluation sense (contributing data at the start, middle and end of the circle). The dataset included 1750 completed questionnaires from a variety of stakeholders including coordinators, core members, volunteers and parents/carers (for the young person circles).

To analyse the data collected, a custom computer program was written. This allocated thousands of individual questions and answers against a framework related to dynamic risk developed by the evaluation in partnership with Circles South West coordinators, volunteers and stakeholders.


The research highlighted several key findings, the foremost of which was that in over three quarters of circles, the core member made progress in areas of their dynamic risk. These included improvements in wellbeing, loneliness, social support, careful decision making and managing thoughts and behaviours. For example, in standard circles, 85% reported improved wellbeing: this rose to 92% in circles for those leaving prison. In standard circles, 80% reported improved careful decision making, while 93% of circles for people with intellectual disabilities reported improved management of thoughts and behaviour. The young person circles reported 76% improved management of thoughts and behaviour, with 93% of parents and carers reporting reduction in perceived risk. Qualitative feedback from core members highlighted the differences that circles made to their lives, including giving them a space to talk openly and improving their confidence to engage in pro-social behaviours. The full report and research summaries are available on the Circles South West website.

The data were also analysed to investigate a challenging area of reliability in criminological research: how can we trust that the responses of people at risk of offending are a reliable marker of ongoing risk?

Evaluation data showed that the responses to questionnaires from different stakeholders were broadly in agreement with each other across a range of different categories of risk. Meaning that the responses of core members in their questionnaires were generally corroborated by the volunteers who worked with them and the professional coordinators. We also found a strong positive correlation between an externally validated loneliness questionnaire (the UCLA Loneliness Scale; Hughes et al., 2004) and several other factors of risk, suggesting that self-reported loneliness is a useful indicator of wider dynamic risk. In summary, we would say that Circles is a really important part of the wider system, providing community led, strengths-based, restorative approaches to reducing sexual harm in communities.


This evaluation is one of the largest single evaluations of the CoSA approach, particularly in terms of the volume of data collected and analysed. It demonstrates not only the value of circles in the South West of England, but also the dedication of CSW to evidence informed practice. This research did not carry out randomised trials or quasi-experimental methods, due to limitations in access to the wider offending data required. However, using a mixed methods approach, the evaluation of Circles South West has clearly shown the value of CoSA to core members and local communities, as well as contributing to the evidence base of CoSA more widely, including on suitable evaluation approaches in this space.


Hughes, M.E., Waite, L.J., Hawkley, L.C., & Cacioppo, J.T., (2004). A Short Scale for Measuring Loneliness in Large Surveys: Results from Two Population-Based Studies. Research on Aging, 26 (6): 655-672. doi: 10.1177/0164027504268574.

Our Funders