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

Data Collection

In line with this aim, semi-structured interviews were conducted with three participant groups (total n = 62):

Semi-structured interviews—those guided by a pre-determined list of questions (Chan et al., 2018)—are commonly a feature of qualitative research since they allow the researcher to follows cues given by participants (Chan et al., 2018). The interview schedule was comprised primarily of open-ended questions used to elicit rich, qualitative data from participants. Open-ended questions are designed to elucidate what is of principal importance to the participant (Weller et al., 2018) and to enable participants to make sense of their own experiences without constraints; they are thus integral in phenomenological research (Bevan, 2014; Chan et al., 2018). The interview questions were designed to elucidate how CoSA contribute to core members’ narratives of identity transformation in the context of their desistance from sexual offending. Core members were asked questions about their experiences of CoSA and whether and how these had shaped their sense of identity in the context of desistance. For example, they were asked:

Other participants were asked questions along the same lines, but in relation to the core members with whom they had been involved as a CoSA volunteer or staff member. The interviews with volunteers and staff were vital to the project as they elicited data from these groups about their perspectives and experiences of relational dynamics with core members, thus providing a more fully fleshed out picture.

Sampling and Recruitment

Participants were recruited via existing contacts with six CoSA programs based in and around three locations in Canada and the USA: one major metropolitan centre (Toronto, Ontario, Canada); one large regional town (Fresno, California, USA); and one smaller town (St John’s, Newfoundland and Labrador, Canada). Program staff were provided with Participant Information and Consent Forms and Interview Schedules for each of the participant groups (copies of these are available from the author). Program staff provided this information to prospective participants in-person and/or via email, and either arranged interview times for those willing to take part, or encouraged prospective participants to make contact with the researcher. Program staff were requested by the researcher to inform as diverse a group of potential interviewees as possible so that the project could capture diverse core member experiences. To this end, program staff made deliberate efforts to recruit core members from varying cultural backgrounds, of varying ages, and with varying offending histories and experiences of CoSA (e.g., length of involvement). The approach to sampling can thus best be understood as a combination of convenience and purposive (Gray, 2009).

All interviews except one were conducted in-person by the author (one core member was interviewed via telephone for practical reasons). Interviews were mainly conducted one-on-one, but in a few cases, small groups of staff or volunteers were interviewed together, again for practical reasons. Interviews were predominantly conducted at CoSA program offices or at the facilities used for individual Circles, including church facilities and community-based program offices. They varied considerably in length, from approximately 20 min to over an hour-and-a-half, with a mean length of approximately 1 hr. Interviews were audio-recorded with participants’ consent. In a small number of cases, participants expressed a preference for handwritten notes to be taken in place of an audio-recording.

Data Analysis

All transcripts and notes from interviews were imported into qualitative data analysis software program NVIVO and read through multiple times by the author in order for familiarization to occur (Caulfield & Hill, 2014; Grbich, 2013). A process of coding was then undertaken in NVIVO. The data were primarily coded deductively (i.e., categorized according to pre-determined themes (Hsieh & Shannon, 2005). Excerpts of text were coded according to the ways in which they mirrored elements of the theoretical framework, in particular the relational dynamics of solidarity, reciprocity and subsidiarity. However, as is good research practice in qualitative research (Creswell, 2015), the coding procedure also allowed for themes and concepts to emerge inductively. This step involved categorizing together excerpts of text that emerged from the data (i.e., not according to pre-determined themes [Hsieh & Shannon, 2005]). For example, elements of each of the broader coding categories of solidarity, reciprocity and subsidiarity were identified during this stage.

Following this, a thematic analysis (Braun & Clarke, 2006), which sought to explore the views, perceptions and/or experiences of participants (Caulfield & Hill, 2014), as well as convergences and divergences across these (Caulfield & Hill, 2014; Mason, 2018), was undertaken in light of this theoretical framework. Ethical approval to conduct the research was obtained from Queensland University of Technology’s Human Research Ethics Committee (approval #1600001093).

The sample

All core members (n = 31) were male, and almost all identified as Caucasian (one identified as Mexican-American and another as First Nations Canadian). The mean age of core members was 55 years (range = 24–76). Most (n = 19; 61%) had been incarcerated in relation to contact sexual offences against minors. Smaller numbers (n = 6; 19%) had been incarcerated for non-contact offences (possessing child exploitation material, exhibitionism, or meeting a minor for the purposes of sexual contact), and contact offences against adult women (n = 2; 6%). Three core members chose not to reveal their index offence, and one had not been convicted of a sexual offence, but participated in a CoSA in a preventative capacity as a result of his (self-diagnosed) pedophilia. Some core members discussed having been reincarcerated due to breaching parole conditions; however, none revealed having been reincarcerated due to the commission of a new offence. This is important to note, as the core members in the study are thus likely to be those for whom CoSA have been most beneficial. Core members had served between 14 and 612 months in prison (and/or in civil commitment facilities), with a median time served of 48 months (4 years). Most core members had been incarcerated for the first time for this index offence. Only six had been incarcerated previously for a sexual offence (two of these on multiple occasions). Three had previously been incarcerated in relation to non-sexual offences. This is largely unsurprising, as most individuals charged with sexual offences are previously unknown to criminal justice authorities (Sandler et al., 2008).

Twelve core members were not currently participating in a CoSA, but had effectively completed a CoSA; some of these former core members remained in touch with the program in a variety of ways (e.g., by remaining in contact with volunteers and/or participating in social events with current or former core members, staff and volunteers). For those currently participating in a CoSA, the mean time in the Circle was 58 months (4 years and 8 months; median = 36 months [3 years]). This is far longer than CoSA are typically envisaged to last (Correctional Service Canada, 2002). While Richards (2011) notes that Circles often last far longer than the envisaged 12-month period (cf Duwe, 2013, 2018), in the current study, recruitment may have favored those core members with longer histories of involvement in CoSA. This should be noted when interpreting the findings below.

Seventeen CoSA volunteers (eight males and nine females) were interviewed. Most (n = 11) had participated or were currently participating in multiple Circles. It should be noted that in the St John’s program, volunteer involvement in CoSA is somewhat less structured than elsewhere, and volunteers may engage with core members across multiple Circles at any given time (i.e., the membership of the Circle is more fluid). Volunteer involvement with CoSA ranged from 6 weeks to many years, with some very long-term volunteers unable to recall exactly how long they had been supporting the program.

Staff interviewed for the research (eight males and six females) held a diverse range of roles, from Board Member to Program Coordinator to more specific roles such as Volunteer Coordinator. Staff had been involved with CoSA for an average of approximately 7 years (range = 3 months to approximately 20 years).

Processes of “Saming”: Core Members and CoSA Volunteers as Like Moral Subjects

At the broadest level, in embracing core members back into communities, CoSA involve processes of “saming.” That is, core members and volunteers implicitly and explicitly engage in speech and/or actions that highlight their similarity as moral actors, in a deliberate erasure of difference; “saming” is the set of processes via which we-ness is realized. As set out above, Fox (2015, p. 83; see also Hannem & Petrunik, 2007) persuasively argues that CoSA communicate to core members that the social distance between “us” and “them” is smaller than they might imagine—that core members “share the same moral space as ordinary citizens.” This article proposes, however, that CoSA (attempt to) communicate not just that core members and volunteers inhabit a shared moral space, but that they are like moral subjects. This is an important extension of Fox’s argument, as it asserts that in CoSA, core members are not merely positioned as others occupying the same moral environment, but as comparable others within that moral environment. Consider, for example, this extract from core member Sean’s interview:

One of my volunteers. . . .He just was a Mr MB [Mennonite Brethren] and everybody knew him and I really looked up to him and he was so down to earth and we used to have long talks. I forget now what it was I did, but it wasn’t the whole truth, and so I had come to him. . .well, came to the whole group and said, “I want to apologize for not being totally truthful on that” and he says, “no problem; you’ve asked for forgiveness, I give it freely”. He says, “I understand not being truthful, I have told many lies [in my life]”, and that shocked me because here’s Mr big MB that you really hold up and he says, “yeah, I have struggled with it [dishonesty] in my life”. So that helped that. . .it doesn’t mean that it’s okay for me but it doesn’t mean that, well, I’m weird or I’m the only one; that it is a normal situation that other people have.

Like Sean, other core members emphasized in their interviews that their volunteers had engaged in processes of saming, in which they had recounted their own immoral and even unlawful acts, and thereby sought to reduce the distance between themselves and the core member. Daniel, for example, seemed grateful and relieved that his volunteers had shared in his Circle their own wayward actions as young people. Terry was particularly fond of two of his volunteers who “weren’t afraid to share their deficiencies or the things that they struggled with.”

Volunteers similarly sought to minimize the moral distance between themselves and core members. For example, volunteer 15 stated in his interview that “everything started to make sense when I said to myself ‘there but by the grace of God go I’. Compared to a lot of these people. . . I had a privileged background.”

In some instances, volunteers’ expressions of saming were deliberately related to sexuality. In other words, volunteers enacted a type of “sexual saming” by seeking to minimize the sexual difference between core members and themselves, positioning themselves as like moral subjects in sexual terms. For example, volunteer 13 stated in his interview that:

Having had some exposure to people having difficulty with their sexual identities. . .and being able to identify with some of these, I sort of say well “by the grace of God” and the whole cliché, I could be one of these.

Similarly, volunteer 3 sought to erase the sexual difference between himself and his core members, stating:

I have my own sexual triggers. I’m a law-abiding citizen. . . .[but]. . . .to say that I don’t ever have anything that I’m not - any thoughts I’m not proud of, is not true but I have some checks and balances. . . .Not all people have those boundaries.

Consider further this excerpt of a group interview with CoSA volunteers 14 and 15:

While there is an immense difference between adolescent sexual experimentation and sexual offending, in these narratives, volunteers deliberately seek to minimize the sexual difference between core members and themselves. Volunteers position themselves as potential offenders, characterizing their own lack of trouble with the law as a manifestation of luck and/or privilege. They thereby cast themselves alongside core members as similar moral subjects. In doing so, they powerfully express and create “we-ness,” or relations of solidarity, reciprocity and subsidiarity in Donati’s (2009) terms. The relations created by saming processes speak directly to the dynamic of solidarity as they reflect the common goods to which all members of a Circle subscribe—that is, the desistance of the core member and the safety of the community. They also indirectly reflect the dynamics of reciprocity and subsidiarity, as one can only enact unilateral, charitable acts for an other who is sufficiently morally distinct as to require such acts. In other words, minimizing the moral distance between volunteers and core members negates opportunities for charitable relations and instead fosters those based on subsidiarity. Processes of saming likewise create opportunities for reciprocity to be enacted. If the moral distance between the core member and volunteers is insurmountable, reciprocity will be difficult to realize.

Of course, having inappropriate sexual thoughts or having committed wayward deeds as a young person is in no way akin to having committed a sexual felony serious enough to result in incarceration. Even attempting to lure a girl into a hut as a young man is hardly morally equivalent to the offences in relation to which core members have been convicted. The point here is not that core members’ and volunteers’ misdeeds are actually similar, or that they reflect a shared morality. Rather, it is that in their attempts to position core members as like moral subjects, volunteers signal to core members an alternative way of conceptualizing and constructing a desistant self. This relational dynamic in CoSA co-produces morphogenesis in the form of identity transformation leading to desistance.

Will the Real Core Member Please Stand Up?

Another expression of the relational dynamic of saming emerged from exploring the meanings given to the term “core member” in the interviews. This line of questioning had been included in the interview schedule in order to elucidate whether and how participants understood the value of avoiding stigmatizing labels like “sex offender,” and the relevance of this to their identity transformation and desistance in the context of CoSA. As Hannem and Petrunik (2007, p. 161) note, in CoSA, the term “core member” is used so that terms pointing to the criminal history or clinical diagnosis of the perpetrator can be avoided (see also Blagden et al., 2018). In the current study, some interviewees’ responses reflected this understanding. Daniel, for example, stated that in a CoSA, “you don’t say ‘sex offender’, you just say ‘core member’.” Far more striking, however, was the apparent confusion among interviewees, especially core members themselves, about exactly who the core member is in the context of a given CoSA. Numerous instances of this confusion emerged during the interviews. Morgan, for example, stated in his interview that “I know one of my core members went through an abusive relationship that I was able to provide some light on,” inadvertently confusing his own status as core member with that of his volunteers. Adam similarly understood the term “core member” to apply to his volunteers:

Okay. I look at my Circle, [NAME] being the leader. The other two that are in it or three that are in it, they’re core members, because they’ve been here for quite a while and gone through whatever. Me, I’m just a member, but I’m not a core member. Someday maybe I will be, but at this point, no.

Other core members’ apparent misapprehension as to whom the term “core member” applies similarly denotes their desire for—and perhaps realization of—a sense of belonging in the Circle in the same way as the other (volunteer) members. Responding to a question about whether his CoSA was coming to a close, core member Damien stated: “Yeah, gradually, but I still meet with them. I’m a member.” Kevin described one of his volunteers as “a core member. He’s one of my support workers, my support members.”

Importantly, volunteers and staff almost without exception understood the term “core member” and to whom it applies. However, they frequently downplayed or even actively resisted the unique status it bestows on core members, instead highlighting the equal standing of all members of Circles. In the following examples, staff and volunteers alluded to having a shared journey and purpose with the core member, and were thereby somewhat dismissive of the term “core member:”

These people - to have a core member is - that’s just a term. They do become part of family. They do become a brother. . . .If we do feel that way, that we’re on that journey with them now, so it’s just as important to us that no more harm comes (Volunteer 1).

We use the term core member as a - when we’re doing paperwork and things. . . .[We say] “we’re going to meet [Dean]”, not “[we’re going to meet] core member [Dean]”. . . .Once we get to know individuals, there’s a - there still are boundaries but we don’t think like core member, non-core member. . . .It’s just a friend that we’re helping out (Staff member 9).

Others conceded that while the core member occupies a special position in the CoSA initially, this diminishes over the life of the Circle until there is little differentiation among the members:

…this person knows that we’re there for them. I think that as most of these CoSA groups go on, I don’t think that we think of anyone as the core member. Eventually, the core member just becomes a part of the Circle (Volunteer 2).

It goes away. . . .you forget about who the core member is after a while – not completely but it really becomes secondary. . .after a year. . .it’s even hard for me to talk about the core member anymore because that person really isn’t defined that way (Volunteer 3).

This again highlights the relational dynamic of saming, attempts toward the erasure of difference between core members and volunteers, and the creation of we-ness. This is, of course, aspirational rather than “real.” It would be neither possible nor desirable for no distinction to exist between a core member and his volunteers. Indeed, a CoSA would not exist without the unique identity of the core member, and a complete erasure of difference could raise issues about the accountability of the core member. The point here is not that such as erasure of difference is ideal. Rather, it is that volunteers’ efforts toward this end demonstrate the co-production of core member desistance through this relational dynamic. In their attempts to minimize the moral distance between core members and themselves, volunteers communicate to core members the possibility of a different, desistant, identity.

It also speaks to Donati’s (2015) notion of super-functionality, in that relationships that characterize CoSA extend beyond the functions of a CoSA: “They’ve become part of our lives and we’ve become part of his or her lives” (Volunteer 3). This speaks to the (re)integration of the core member to which CoSA aspire (Hannem & Petrunik, 2007). CoSA have been criticized on the grounds that volunteers can relate to core members in a way that fosters their dependence on the Circle rather than their (re)integration into the community more broadly (Richards et al., 2020). Differentiating functional from super-functional social relations provides a useful framework for understanding how and why this occurs and how CoSA might more effectively create the conditions under which genuine social (re)integration can take place.

Role-modeling or Reciprocity?

As outlined above, it is widely accepted that CoSA provide prosocial modeling to core members. This claim was, however, largely contested and even rejected outright by core members in the current study. In seeking to explore core members’ understandings of the role of CoSA in their desistance, participants in the study were asked how they would characterize the relationship they have with their volunteers, and more specifically, whether they viewed their volunteers as role models. In the main, core members did not consider volunteers to be role models, but instead viewed them as, in Keith’s words, “friends and support.” The significance of this distinction cannot be overstated; for core members, volunteers cannot be role models because this disavows the mutuality that characterizes and defines their view of their relationships with CoSA volunteers. Consider, for example, Zack’s response to the question of whether he sees his volunteers as role models:

Here, Zack implies that his volunteers cannot be role models because this would repudiate the reciprocal nature of the relationships he has with them. While his volunteers are there to listen and provide advice, he seeks to do the same for them. Leon’s response to the same question, while dissimilar to Zack’s, also (implicitly this time) resists the characterization of his volunteers as role models, instead emphasizing the mutuality inherent in these relationships. Leon first responded by agreeing: “they had qualities that were examples, yes,” but immediately qualified his response with the following anecdote about one of his CoSA volunteers:

I remember the retired doctor, he just lived across the - about five minutes away from where I am in that senior’s high rise apartment or condo[minium]. So we spent time together, we’d go for a coffee. Or they have a games room down in the basement level. . . .So he had never played pool in his life so I taught him how to play pool. I’ll tell you, you should have seen him strutting around the table when he’d have a few balls go down!

Leon’s immediate reaction to the question of whether his volunteers are role models is thus an account of him mentoring one of his volunteers rather than the reverse, again underscoring the reciprocity that he sees as characterizing his relationships with his volunteers.

Volunteers were likewise resolute that their relationships with core members should not predominantly be understood as an expression of prosocial modeling. In fact, they conveyed considerable resistance to this suggestion. For Volunteer 3, for example, to be a role model suggests hierarchical relations—a “pedestal of some sort.” In his experience, however, this “is not what happens if relationships develop to the point where they’re mutual, which is the goal.” In his understanding, core members do not and ought not relate to volunteers as role models, but as “fellow citizens.” Volunteer 12 conceded that in some instances, volunteers may be perceived by the core member as role models, but viewed this as something of a “side effect” of CoSA:

If they’re trying to be role models, if they see that as their role, I think they’re missing the point. . . .inevitably, some people become role models. But the important thing is that they [core members] know that there is somebody there for them.

It has been documented previously that core members often view their volunteers as friends and/or supporters (Richards, 2011; Thompson et al., 2017). The current study, however, extends beyond this to demonstrate the reciprocity with which core member-volunteer relationships are imbued, and the level of resistance among CoSA participants to asymmetrical notions of role-modeling. In a number of instances, volunteers and staff even implicitly described their efforts to provide core members with opportunities to practise reciprocity. Given the life circumstances of many core members, and the often profound difference between their own life circumstances and that of their volunteers, core members are not often easily able to practise reciprocity. Nonetheless, they described taking to any opportunity to do so with considerable enthusiasm. Morgan, for example, described helping one of his volunteers with home renovations, by ripping up the carpet in her home, and providing advice to another volunteer who had experienced an abusive relationship. Frank practised reciprocity by supporting two of his volunteers in their aim to become police officers, and Ewan explained that in his CoSA, “We look out for each other.” Dean had put his bookkeeping skills to use to reciprocate the support he received from his CoSA, an opportunity facilitated by Staff Member 8 (who also acts as a volunteer in a number of CoSA). Consider the following extract from their joint interview:

This exchange powerfully demonstrates the reciprocity that (ideally) characterizes core member-volunteer relationships by highlighting both the volunteer’s efforts to provide an avenue for the core member to practise reciprocity, and the core member’s eager uptake of this opportunity in order to do so. In numerous other instances, core members and volunteers described their CoSA meetings as settings in which all members, not just the core member, are both givers and receivers of support and advice. This is often achieved through a “checking in” process at the commencement of each meeting, in which each participant is given a chance to share personal news, such as recent life struggles and/or achievements. Many interviewees emphasized that this practice has the effect of fostering relationships of reciprocity, and core members embraced this understanding of the purpose of their CoSA:

We all shared. We would always go around. I’d hear about their lives as well as my own. Some days I sat there and it seemed like we’d talk more about them than me. That’s okay because we all need some sort of a Circle to talk in (Rory, core member).

Everybody can use support. It doesn’t matter who you are. . . .it’s a two-way street. I’ve been told that I’ve inspired them [volunteers] to do some things too, which is really cool (Joshua, core member).

That’s the purpose of the group – we’re here to support each other as well, not to receive support from the - like the support people [volunteers]. We have to support each other. . . .It’s a communist group (Leon, core member).

It thus appeared vital to both core members and volunteers that core members not be positioned as passive or dependent recipients of support, especially as the CoSA matured over time. When Rory claimed that “we all need some sort of a Circle,” he implied not only that he is essentially no different from his volunteers, but that the support he receives from his volunteers is characterized by subsidiarity—it is at least theoretically mutual and does not cast him as a passive dependent but a co-creator of the common good of his own desistance. As Volunteer 1 argued, this practice of sharing among all participants “really makes them [core members] feel much more than just a work in progress that needs help and a hand-out.” In the words of Volunteer 8, “I’m not meeting with [Dean] because I have some charity concern. He’s a friend of ours.” While the relationships formed among CoSA participants may shift over the life of a Circle and beyond—encompassing friend, ally, mentor, and so on—they thus clearly embody the interrelated elements of solidarity, reciprocity, and subsidiarity, as well as super-functionality.

It follows that Donati’s (2015) concept of “relational reflexivity” ought to be key to understanding the capacity of CoSA to produce the common good of desistance, or at least to understanding how the support function of CoSA produces desistance. Whether and how “relational reflexivity” can expand our understanding of the accountability function of CoSA was not the focus of the current study but ought to be considered in future research. For Donati, behavioral change is shaped by our relations with others. The above discussion has sought to demonstrate that the behavioral change that CoSA produce among core members occurs in relational context with volunteers. The emphasis on with volunteers here is central to this theorization. In the hypotheses outlined earlier (e.g., that CoSA provide social support or role-modeling to core members), the social relations envisaged are largely unilateral—that is, CoSA are done to or for core members. According to this existing understanding, core members relate to their volunteers by, for example, “living up to” their expectations. The above discussion suggests, however, that social relations in CoSA are (ideally) reciprocal—that is, created with core members. For example, rather than living up to an other’s expectations, CoSA co-create new expectations for core member behavior, and collectively navigate these. CoSA thus provide an opportunity for core members to “re-establish the circuit of reciprocity” (Donati, 2009, p. 226), and in doing so, co-produce desistant identities among core members. Weaver demonstrates how important this step is to desistance, highlighting the importance that many desisting offenders place on taking up the role of breadwinner (Weaver & McNeill, 2015, p. 102), peer support mentor or volunteer (Weaver, 2012, pp. 408–409). Such roles are, however, a virtual impossibility for those who have sexually offended in the current punitive and risk-averse sex offender management landscape (see e.g., Harris, 2017; Wacquant, 2009). CoSA are one of very few avenues via which the circuit of reciprocity might be established. Crucially, understanding CoSA in this way reinforces Weaver and McNeill’s (2015, p. 104) assertion that desistance should be understood “not as an end in itself—as some studies tend to imply—but as a means to realizing and maintaining. . .individual and relational concerns.” In other words, CoSA do not seek to directly produce desistance but to produce a series of social relations via which desistance will be (co-)produced. This occurs because the relational concerns formed in and by CoSA are (ideally) incompatible with continued offending (see Weaver, 2012, p. 406).

The study’s findings do not discredit existing understandings of CoSA (e.g., that CoSA provide social support to, and reduce the social isolation of core members). Rather, they encourage a reframing of these understandings of CoSA, instead positioning relations of solidarity, reciprocity and subsidiarity as central. They do, however, suggest the notion that CoSA provide prosocial modeling to core members ought to be reconsidered. As the above discussion suggests, CoSA are not and should not be centrally concerned with role-modeling appropriate behaviors to core members, as much of the existing literature asserts. Rather, CoSA (ideally) create relations of solidarity and subsidiarity—relations that are reciprocal rather than charitable.

Furthermore, the findings suggest a new understanding of the dynamics of core member-volunteer relations as compared with the dynamics that commonly characterize relations between core members and statutory agents such as parole officers and mandated therapists. Prior research (Richards, 2011; Thomas et al., 2014; Thompson et al., 2017) has argued that Circle relationships are superior to state (e.g., parole) relationships because core members are grateful that members of the community would volunteer to work with them, unlike state agents, who are mandated to do so. The current study does not challenge this assertion as such. It does, however, significantly build on this understanding by clarifying that the superiority of Circle relations is their super-functionality (Donati, 2015). In contrast, relations between core members and agents of the state can only ever be functional and unilateral—something done to rather than with core members. Moreover, CoSA, unlike correctional or therapeutic supervision, are specifically designed to produce positive social relations, and thereby create, in Donati’s (2014a, p. 161) terms, “added social value.” Non-government services such as CoSA produce added social value precisely because they use “a greater quantity and a better quality of social relations as compared to the social relations used by the market or by public administrative bureaucracies” (Donati, 2014a, p. 161). This underscores McCartan, Kemshall, Westwood, Solle, et al.’s (2014) point that CoSA do not duplicate statutory supervision arrangements, as well as suggesting how and why this is the case.

Limitations

Perhaps the chief limitation of the current study is that it was conducted with CoSA programs in specific local contexts. It must be acknowledged that CoSA programs around the globe vary considerably from one another (Höing et al., 2013), and that the particular locations at which data were collected do not necessarily represent all CoSA programs. In particular, the research was conducted with CoSA programs in the USA and Canada, and it has been acknowledged that CoSA practice in these locations differ from CoSA in the United Kingdom (Richards, 2011). As such, the findings reported here are likely to be of most relevance to CoSA in the North American context; processes of “saming” may be more challenging to achieve in the UK, where CoSA are more closely aligned with the state criminal justice apparatus and rely more heavily on volunteers from the criminal justice sector (Thompson et al., 2017).

As noted above, the average age of core members was 55 years. It is not possible to know whether younger core members might relate to volunteers in different ways, and this could form the subject of future study. However, no differences were noted across the core members of varying ages in the current research. The sampling procedure may also have favored core members with longer histories of involvement with CoSA. The findings may thus better reflect the dynamics of more stable, longer-term Circles. The relational dynamics described in this article may therefore reflect CoSA that have been especially successful. This is not a weakness of the research, but rather underscores that these relational dynamics are aspirational in supporting core members on their desistance journeys. Whether and how these dynamics would assist core members who are less well-suited to CoSA remains to be seen, and should form the focus of future research. They cannot, however, be expected to form rapidly, but rather might be supported to evolve over time. This highlights the need for appropriate resourcing and support for CoSA programs over the longer-term.