Circles of Support and Accountability: The Role of Social Relations in Core Member Desistance

• Author: Kelly Richards
• DOI: http://doi.org/10.1177/0306624X20964094
• Published date: 01 August 2022
• Date: 01 August 2022
• Subject Matter: Articles

https://doi.org/10.1177/0306624X20964094

International Journal of

Offender Therapy and

Comparative Criminology

2022, Vol. 66(10-11) 1071 –1092

© The Author(s) 2020

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0306624X20964094

journals.sagepub.com/home/ijo

Article

Circles of Support and

Accountability: The Role

of Social Relations in Core

Member Desistance

Kelly Richards1

Abstract

Circles of Support and Accountability (CoSA) appear to reduce the sexual recidivism

of core members (i.e., individuals convicted of sexual offending). It remains unclear,

however, how they do so. While much previous scholarship has hypothesized that

the relations between core members and CoSA volunteers promote desistance from

sexual offending, there has been no theoretically-informed research that specifically

interrogates these relations. This article begins to address this gap by examining the

relations formed in and by CoSA through the lens of Donati’s theory of relational

reflexivity. Drawing on semi-structured interviews with 62 CoSA participants across

six CoSA programs located in the USA and Canada, it proffers a new theorization of

the role of social relations in core members’ desistance. Findings from the study will

enable CoSA practitioners around the globe to explicate and deepen their practice

around more rigorous theoretical precepts.

Keywords

circles of support and accountability, desistance, social relations, sexual offending

Introduction

Since their emergence in Canada in the mid-1990s, Circles of Support and Accountability

(CoSA) have gained traction across the west. CoSA are now a feature of the criminal

justice landscape across Canada (Chouinard & Riddick, 2014; Wilson et al., 2005,

2009), parts of the United States of America (Duwe, 2018; Fox, 2013), the United

1Queensland University of Technology, Brisbane, QLD, Australia

Corresponding Author:

Kelly Richards, Queensland University of Technology, Faculty of Law, School of Justice, GPO Box 2434,

Brisbane, QLD 4001, Australia.

Email: K1.richards@qut.edu.au

964094IJOXXX10.1177/0306624X20964094International Journal of Offender Therapy and Comparative CriminologyRichards

research-article2020

1072 International Journal of Offender Therapy and Comparative Criminology 66(10-11)

Kingdom (Bates et al., 2012; McCartan, 2016; Nellis, 2009), Western Europe (Höing

et al., 2013, 2015), and Australia (Richards et al, 2020; Richards & McCartan, 2018).

As has been set out in the burgeoning literature, CoSA are groups of trained volunteers

from the community who support people convicted of sexual offences (often against

children) to reintegrate into the community after a period of incarceration.

The emerging evidence about the effectiveness of CoSA in reducing sexual recidi-

vism is promising. Duwe’s (2018) randomized controlled trial of CoSA in Minnesota

compared 50 people who had been incarcerated in relation to sexual offending and

who had participated in a CoSA (known as “core members”) with 50 individuals who

had been incarcerated in relation to sexual offending but who were not assigned a

CoSA, over an average of 6 years. Duwe (2018) found statistically significant differ-

ences between core members and controls, with the rate of rearrest for a new sexual

offence being 88 percent lower for core members (see also Duwe, 2013). Quasi-

experimental research (involving comparing core members to a matched comparison

group rather than random allocation) has been conducted in Canada (Wilson et al.,

2005, 2007, 2009) and the United Kingdom (Bates et al., 2014). While such research

can involve selection bias due to the lack of randomization, these studies have pro-

duced positive results similar to Duwe’s (2013, 2018).

Research also suggests that CoSA may enhance outcomes in terms of core mem-

bers’ relationships (Bates et al., 2012; Elliott et al., 2017), employment and education

(Bates et al., 2012; Clarke et al., 2017; McCartan, Kemshall, Westwood, Cattel, et al.,

2014), housing (Bates et al., 2007; Clarke et al., 2017), health (Bates et al., 2012),

prosocial attitudes (Bates et al., 2012; Höing et al., 2015), participation in prosocial

activities (McCartan, Kemshall, Westwood, Cattel, et al., 2014), emotional regulation

and self-esteem (Höing et al., 2015), and sense of hope (Elliott et al., 2017). While

these studies did not use comparison groups, and it is therefore not possible to con-

clude that CoSA caused these changes, taken together, these findings are likewise

promising.

We should nonetheless interpret the evidence about CoSA critically (Clarke et al.,

2015; Elliott & Zajac, 2015; Thompson et al., 2017). As Duwe (2018) concedes, even

his research - the “strongest evidence to date” (p. 480)—is not conclusive. Nonetheless,

collectively, the research on CoSA highlights a promising practice.

Much less is understood about the mechanisms via which CoSA produce behav-

ioral change among core members. In short, while it appears that CoSA reduce sexual

recidivism, we do not yet adequately understand how they do so. Scholars typically

hypothesize, however, that the relational dynamics between the core member and vol-

unteers in CoSA are central to this behavioural change. A number of interrelated sug-

gestions have been put forward in this regard. First, scholars argue that CoSA reduce

social isolation experienced by people who have sexually offended (Cesaroni, 2001;

Höing et al., 2013). In providing core members with a social outlet and supporting

them to develop a social network, it is envisaged that CoSA will reduce recidivism, as

strong social bonds are associated with reduced reoffending (Bailey & Klein, 2018;

van den Berg, et al. (2018). Second, scholars argue that CoSA volunteers model pro-

social behavior to core members (Chouinard & Riddick, 2014; Höing et al., 2013,