Abstract
Research is limited on mechanisms of action in restorative justice interventions. This multimethods study delineates the change processes underlying a successful in-prison group treatment program by (a) examining shifts in offenders’ self-schemas and (b) identifying key program components that influence this movement. Researchers assigned to small groups as “co-facilitators” gathered data using participant observation, semi-structured interviews, and psychological assessments at three time points. Mechanisms of action include group norms and behaviors that contrast with prior experiences and uncover offenders’ self-schemas through intrapsychic processes, which prompt them to test and act upon new possible selves through the group process.
High incarceration and recidivism rates continue to overburden the criminal justice system. At the close of 2013, 1.57 million people were incarcerated in the United States (Bureau of Justice Statistics, 2014). In addition, national studies have estimated the percentage of prisoners returning to prison for new crimes or technical violations within 3 years anywhere from 40% (Pew Center on the States, 2011) to 52% (Bureau of Justice Statistics, 2002) to an astounding 68% (Bureau of Justice Statistics, 2014). As a result, there is growing emphasis on desistance-promoting practices most of which target the role of penal professionals. Scant attention has been given to using naturalistic approaches, namely, relationship building with crime victims and community members affected by criminal behavior, as the medium to further desistance in offenders. Bridges to Life (BTL) is one such program, which demonstrates 3-year recidivism rates of 14% and only 3% for violent offenses (BTL, 2015b). This study undertakes an in-depth multimethod exploration of the BTL program to identify the mechanisms of action, which mediate the program’s success and the process by which those mechanisms affect change in offenders’ self-schemas.
The BTL Intervention
BTL is a 12-week manualized faith-based intervention that uses a small group format for offenders and volunteer victims to engage in personal dialogue about the harms caused during crime. Its goals are to reduce recidivism and facilitate healing for all parties. Weekly group sessions focus on topics including accountability, repentance, and reconciliation and include personal storytelling by offenders and victims and writing of unsent accountability letters to victims and family members. Although the program rests philosophically on Christian principles that are congruent with changing offenders’ moral motivations, its focus is ecumenical and non-evangelist. A prior study (n = 102) found that 30% of BTL offender participants reported being Jewish, Muslim, humanist, Wiccan, Native American, believing in God, and having more than one religion (Armour, Windsor, Aguilar, & Taub, 2008). (See Armour et al., 2008, for a detailed description of the BTL intervention, which is excluded here to honor limited space). BTL operates in 63 jails and prisons, including general population, pre-release and substance abuse facilities. The program has completed 649 projects serving more than 23,000 inmates since 2000 (BTL, 2015a). Preliminary analysis of the program indicates a cost savings of US$610,468.72 for every 100 BTL participants, based on annual incarceration costs and mean stays in Texas state prisons and jails (Armour & Rubin, 2006).
Although it can be expected the BTL participants—who are self-selecting based on their motivation and willingness to change—will demonstrate lower recidivism rates than the general population, program evaluations have also revealed significant changes in offenders’ (n = 102) empathy, vengefulness, forgiveness, and relationships post-graduation (Armour et al., 2008). A study of BTL participant responses (n = 1,021) to a nine-item evaluation survey found that BTL fostered a positive response and a capacity for other-oriented empathy and guilt proneness such as “my behavior is bad, but I am worthwhile” (Armour, Sage, Rubin, & Windsor, 2005). Although the results of BTL suggest shifts in offenders’ self-schemas, little is known about the actual mechanisms of action in the group processes or their relationship to these shifts.
Mechanisms of Action and Crime Desistance
The mechanisms of action concept is a relatively new focus in the research literature that examines how interventions for which there is an empirical consensus of efficacy work. Beyond simply identifying mediators, mechanisms of action reflect the sequential, temporal process through which an intervention produces change (Kazdin, 2007). Although methods to uncover mechanisms of action have been applied to psychotherapeutic interventions such as Dialectic Behavior Therapy (Lynch, Chapman, Rosenthal, & Linehan, 2006), the therapeutic alliance (Barber, Connolly, Crits-Christoph, Gladis, & Siqueland, 2000), and Cognitive Therapy (Kwon & Oei, 2003), their use in criminal justice research is severely limited.
Because prior research suggests that BTL affects behavior change in offenders through a shift in offenders’ self-schema (Armour et al., 2005), this study utilizes Stein and Markus’s (1996) self-concept and behavioral change framework to examine the mechanisms of action and related shifts in offenders’ self-concept and behavior. Stein and Markus define self-concept or self-schema as a “collection of cognitive schemas” that actively regulate behavior. Behavioral change necessitates the development of new conceptions of “possible selves” that can be used as motivators because they provide goals to approach or avoid (Carver & Scheier, 2000). This framework has been widely utilized in the study of eating disorders (Corte & Stein, 2005; Stein, 1996), as well as by Abrams and Aguilar (2005) to understand the role of self-concept in responses of juvenile offenders to a residential treatment program, though the latter findings were not linked to treatment outcomes.
There is increasing recognition of the importance of self-schema in understanding crime desistance. Paternoster and Bushway (2009) suggest that an offender’s desired possible selves and feared possible selves serve as motivators for behavior change when there is a clear blueprint for what to pursue or avoid. They call for empirical research into causal mechanisms linking identity and social networks with crime desistance. Draycott and Dabbs (1998) further propose that cognitive dissonance generated by inconsistencies within the self-concept produces a strong mechanism by which individuals are motivated to reconcile inconsistencies through changes in behavior or attitude.
This study identifies the mechanisms of action in BTL that generate new possible selves in offenders through the therapeutic group process of surrogate victim–offender dialogue, a restorative justice practice with ample empirical support for its effectiveness in generating empathy, reducing trauma, and increasing offenders’ and victims’ sense of justice. Restorative justice practices are inherently relational, viewing crime as a violation of relationships between people (Zehr, 2002). Restorative processes such as those used in BTL require authentic communication between victims and offenders to identify and respond to the harms caused by crime (Umbreit, Vos, Coates, & Lightfoot, 2005). Restorative justice norms such as authenticity, deep listening, and respect characterize interactions in BTL program sessions. Moreover, because intense engagement between surrogate victims and offenders occur in small groups over 12 weeks, the level of cohesion between group members who honor and actualize restorative justice norms likely influences participant’s individual change processes as well. Group processes such as caring, mutual support, and confrontation help establish norms and contribute to group cohesiveness (Yalom & Leszcz, 2005). Group cohesiveness is particularly relevant in group therapy settings with offenders, who may seek belonging in anti-social groups, which are counteractive to therapy goals (Morgan & Winterowd, 2002). As offenders place significance on the group, they are more able to accept group norms and to hear and accept healthy confrontations.
Analysis of members’ process data indicated that there was a clear relationship between how cohesive the members reported the group to be and the extent to which freedom of action and expressions of feelings were encouraged in groups, as well as treatment outcome as measured by significant reductions in pro-offending attitudes.
Investigation of the influence of the relational context of group-based restorative justice practices, and its impact on the development of new “possible selves” as potential behavioral change mechanisms in male inmates is unique in criminal justice research. This exploratory study proposes a potentially testable theory, which describes the mechanisms of action in the BTL intervention, which influence that change.
Method
This study used a multimethod design, in which qualitative and quantitative data were collected simultaneously, with quantitative data providing a supportive, secondary role (Creswell & Clark, 2007; Morse, 2003). The study primarily relied on in-depth qualitative data including participant observation, BTL participant interviews, and field notes to explore the key components of the BTL intervention that influenced shifts in offender participants’ self-concept, social identity, and motivation to change criminal behaviors. Quantitative data were collected solely to document change in group cohesion and criminal attitudes in congruent directions.
Sample
The sample (n = 18) of offenders (n = 12), victims (n = 4), and facilitators (n = 2) were recruited from a pool of registrants scheduled to take part in a typical BTL prison project at a minimum security substance abuse therapeutic community for male, pre-release offenders. Offender eligibility criteria for BTL include pending release within 9 months. Offenders, victims, and facilitators were evenly divided and assigned to one of two small BTL groups of nine participants each. Two offenders dropped out of the program before completion.
The mean age of offenders was 44 years (SD = 6.96). Inmates were African American (50%), Hispanic (16.7%), and Caucasian (33.3%). Most participants had completed high school or a high school equivalency degree (42%) or some college (25%). Marital status varied between those married (41.7%) and not married (58.3%). Most crimes connected to current incarcerations were drug related (41.7%), followed by property (16.7%), violent (16.7%), and other (25%). Mean time incarcerated upon enrollment was 39.55 months (SD = 55.06), with a range of 4 to 185 months. The majority of inmates identified with a Christian denomination such as Baptist or Protestant (75%), whereas 25% identified as non-denominational Christian or acknowledged a belief in a higher power. All victims were Caucasian, three of whom were female and one of whom male. They had completed some college, and had participated in other BTL projects. Their mean age was 47 years (SD = 19.82 years). Half of the victims were married. Victims identified as either Christian (75%) or non-denominational (25%). The two facilitators were male and female, both Caucasian.
Because all offenders were male and the majority of volunteer victims were female, there is a possibility, particularly in a same sex prison population, that gender dynamics influence offender outcomes. Such a dynamic was difficult to attend to in the study of this particular program, as BTL is offered almost exclusively at male facilities. Furthermore, BTL’s victim volunteers are frequently female, as are most victims who participate in in-prison restorative processes in the United States (see Umbreit, Vos, Coates, & Brown, 2003). Although the gender makeup of this sample is indicative of the “norm” in BTL programming, it should be noted that female offender populations are understudied and are likely to exhibit different dynamics in treatment than their male counterparts (Belknap, 2014).
Participation in the BTL program is voluntary for offenders, victims, and facilitators, a practice that potentially confounds treatment effects because of a self-selection bias. Latimer, Dowden, and Muise (2005) suggest assessing treatment effects relative to participants’ motivation prior to program participation to control for this bias. In this study, all inmates fell above the midpoint in assessing their readiness for the program and most fell above the midpoint in rating the importance of their participation (83.4%) and expected helpfulness of BTL (75%).
Data Collection
Data were collected from offenders, victims, and participant observers assigned for the duration of the 12-week program to two small BTL breakout groups. Each group consisted of six offenders, two victims, and a volunteer community facilitator. Because of the prison prohibition against recording, as well as the potential disruption of note taking on group process, a research investigator was assigned to each group to be a co-facilitator-in-training and to gather data about group interaction and change processes using participant observation, a research tool widely used for naturalistic observation and holistic understanding of observation data from both an insider’s and an outsider’s perspective (Jorgensen, 2003). Research investigators recorded their observations in field notes and voice recordings immediately following each session, and the principal investigator conducted 3-hr weekly audiotaped debriefing meetings with the research investigators. During this time, interactions between the researchers and groups were explored to monitor for the introduction of bias into the process.
Each of the research investigators also conducted in-person semi-structured interviews using an interview guide about offender change and administered quantitative measures with victims and offenders of the group to which they were not assigned. Administration of the interview guide and measures occurred at three time points: before, during, and after the intervention, with the “during” administration staggered for each offender to provide a sense of temporal processes over the course of 12 weeks. The interview guide asked participants what they admired and disliked about themselves, the impact and influence of BTL on self-knowledge, their assessment of behavior change in self and other offender participants, their behavior in helping other participants, and their physical and emotional pain while participating in the group.
Quantitative measures were used to assess group cohesion and the direction in criminal involvement change processes and readiness to change. Group cohesion was measured by the Harvard Community Health Plan Group Cohesiveness Scale (HCHP GCS; Budman et al., 1987). The Group Cohesiveness Scale (GCS) is an observation checklist-coding scheme designed to assess cohesiveness in psychotherapy groups and consists of five subscales (Trust, Focus, Interest, Facilitative Behavior, Bonding) and a global scale (Global Cohesiveness). Each scale ranges from (1) very slight to (9) very strong with explicit descriptions for odd-numbered levels. Inter-rater correlations range from .68 to .85.
Criminal involvement change processes were assessed by the Criminal Involvement Change Process Scale (CICPS). The CICPS is a 16-item scale developed for this study that measures change in eight processes: consciousness raising, dramatic relief, environmental reevaluation, self-reevaluation, self-liberation, helping relationships, reinforcement management, and stimulus control. Each item is rated in a 5-point Likert-type scale (1 = never to 5 = repeatedly). It is adapted from the original 40-item Processes of Change Questionnaire (PCQ) developed by Prochaska, Velicer, DiClemente, and Fava (1988). Importance of, confidence in, and readiness to change criminal behavior were assessed by the Rulers for Criminal Behavior (RCB). The RCB is a three-item scale developed for this study that measures change in centimeters on a 10-inch ruler. This scale is adapted from Miller and Rollnick’s (1991) scaling techniques for behavior change.
Data Analysis
Although data were collected at the individual level, the unit of analysis was the group. Data were analyzed from transcribed interviews, session observations, debriefing meetings, and field notes using template analysis and a modified form of grounded theory. Template analysis is appropriate for data collected using interview questions or categories derived from a specific theory (Crabtree & Miller, 1992). In this study, categories derived from self-schema concepts (Stein & Markus, 1996), group process indicators (Yalom & Leszcz, 2005), and restorative justice functions (Zehr, 2002). Data were entered into NVivo (Qualitative Research Software, International) and coded based on these a priori codes. In addition, we used a grounded theory approach (Strauss & Corbin, 1990) to conduct open coding, employing constant comparison of indicators to illuminate new codes and their properties and dimensions.
Data were analyzed in sequential order: All data collected for Session 1 were analyzed before moving to Session 2. Memoing was used to document emerging themes from each data piece as well as meanings derived from combined understandings of current and prior sessions (Strauss & Corbin, 1990). The two authors held weekly meetings during the analysis to discuss emerging findings and explore alternative explanations. Final themes were collapsed into axial codes, which were analyzed for the relationships between them. Conceptual memoing and visual diagrams were used to formulate a temporal storyline and theoretical framework about mechanisms of action in BTL. The BTL executive director and four other restorative justice practitioners familiar with the program reviewed results and gave feedback on goodness-of-fit.
Demographic and scale data were entered into SPSS, where descriptive data and frequency tables were reviewed. Due to the limited sample size, t tests were conducted on data from the Group Cohesion Scale, Criminal Involvement Change Processes, and Criminal Rulers Scale only to assess change in the expected direction, not to determine statistical significance.
Findings
Results from the study include offender participant changes, changes in group cohesion, and the key program components found to influence those changes. Scales measuring group cohesion, criminal involvement change processes, and criminal rulers each showed positive change in the expected direction over time, supporting qualitative findings (see Table 1).
Mean Offender Scores on Scales Before, During, and After the Group Process (n = 10).
Note. Scores excluded for two offenders who dropped out.
Directional improvements in both inter- and intra-personal processes were contextualized by interview transcripts and participant observation. We identified four sequential categories of mechanisms acting on program participants (see Figure 1). Community-established group norms predicated group behaviors, which triggered intrapsychic processes that resulted in varying levels of self-system change, in which offender participants tested and acted out behavioral changes suggested by the group process. Throughout the program, a mysterious element labeled synchronicity refers to the shared perception of a divine influence overseeing the process. See the appendix for a case vignette describing the movement of one program participant through the model.

Diagram of mechanisms of action in Bridges to Life.
Group Norms and Behaviors
Group norms refer to values and beliefs associated with the BTL philosophy, which is rooted in restorative justice. Restorative justice values are directly articulated and transferred by program facilitators and volunteers during the program as well as by BTL graduates still in prison. However, some of the norms emerge through shared leadership and seem to unfold organically through modeling or the ethos of the program itself.
Group norms established a sense of trust and community in which high quality sharing and other supportive behaviors were possible. When group norms were violated by a change in structure or behavior, the group process stalled and meetings became less effective. The group norms and accompanying behaviors created a culture, as exemplified by the results from the group cohesion scales, showing that group cohesiveness, interest, trust, cooperation, focus, and caring increased as time passed, reaching nearly maximum average scores on each subscale by the end of the group.
Four norms were definitional: caring connection, expectation to stay, leveling of barriers, and demand for authenticity. These norms were manifested by specific behaviors that imbued them with meaning adding power to the group’s process. A description of the norms and related group behaviors is provided below.
Caring connection
One of the inmates’ first impressions is that volunteer victims and facilitators receive nothing in exchange for their commitment to the program. This creates trustworthiness and a strong sense of mutual caring early on. One offender participant reported,
The volunteers seem to have genuine concern. They are doing it from their hearts—not counselors trying to get into our brains. It is a human process. You are able to be yourself. It is amazing that people want to come in and spend time with alcoholics, criminals and addicts, but they do.
The volunteers themselves echo a similar sense of love and caring toward the offender participants. One stated,
I just hope that they see that I really truly care about all of them. And by helping them, you know, learn to let go of some of that and see that there is forgiveness and love in this world and there are people that don’t want to just take from them. I want them to really feel that.
Offenders experience this caring connection as an unexpected and deeply touching force: “The victims come in to be with you in a group . . . The victims are sincere, show compassion and share themselves with me—a criminal. This gives me courage.” Another observed, “Feeling the genuine caring . . . that changes a person.” This connection is manifested behaviorally by offenders facing the harm done and victims’ visioning for them of a possible future.
Recognition of harm
By seeing harms as lived by group members who they care about, offenders came to recognize how they had caused damage to others. A victim of domestic violence described how offenders reacted to her story:
From being the macho man, it just shifted to . . . , “How could someone do that to you?” Or, “Wow. Why would you let anyone do that?” Or, “You tell us where he lives.” They switched to that protective thing so they got to see . . . what happens when someone that you care about is being hurt like that . . . It was like a light switch.
Offenders supported this observation, saying, “Before, I thought these things didn’t happen to me or affect me. Then I had to face them. It could have been me. This was a scary thought. I had to feel it.” Feelings of compassion turned to repentance for harm caused:
I hadn’t been able to see how the victims felt before. When I took my aggression out on others, I never knew how they felt . . . Now I can perceive . . . hurt and anger. It’s making me feel a lot of regret.
Visioning
The genuine caring and respect in the group also allowed members to challenge offenders’ view of self by sharing their own visions of the offenders’ strengths, possibilities, and opportunities. This visioning through the eyes of others who truly cared helped participants to see themselves in a new light. For example, a research investigator challenged offenders’ views of themselves as unworthy or unimportant.
I said, “It really angers me that you think that little of yourself that you think, ‘Well if I die, I die.’” I said, “That’s not at all how I see you.” And he started to cry. I stopped . . . And there was silence for a good 10 minutes.
Expressions of visioning were bi-directional. Offenders also helped victims to see hidden aspects of their own victimization or new possibilities within themselves. Researchers observed the contribution this made on the offenders’ personal vision of self. “They [felt] like they too could make a contribution. That they could make a difference in the lives of others . . . That people could see beyond their criminality.”
Expectation to Stay
All members were expected to be physically present and fulfill their commitment to the group until the end. This is modeled by the victim volunteers: “I’m there every Tuesday unless I’m sick. They know that they can count on me. Where they probably haven’t had a lot of people they can count on in the past.” Members confronted each other for not complying with the established norm. When the two facilitators were absent, the following exchange occurred:
Devon confronted Jessica and Tony about not coming . . . Jessica flippantly said, “We went dancing.” [H]e goes, “[The rest of us] haven’t missed any. And when Denise had to miss, she let us all know why she wasn’t here.”
Behavioral manifestations of staying in the program, no matter what, included negotiating disruptions, accepting conflict, and having faith in the process.
Negotiating disruptions
Committed to stay, group members found a way to tolerate otherwise disruptive events and to move past them. One offender commented on a fellow inmate:
Someone spoke too much. I had to listen to him at work every day so it was hard to listen to him again in the group. There was tension. I wanted to tell him to shut up. But, I saw him grow a lot.
Indeed, as the group progressed, members were able to not only negotiate disruptions but also assign positive meaning to them. For example, an offender remarked about a disruptive volunteer: “I think it went just the way it needed to go. They have issues too and we get to see that they have issues.”
Accepting conflict
The expectation to stay supports the healthy conflict that often accompanies holding one another accountable. A volunteer observed, “If that commitment’s not there, I don’t think you can reach that deep, deep place where you want to go. They don’t trust each other and then they don’t know if you’re going to get mad.” By staying and accepting conflict rather than running from it, group members made more powerful connections with one another. One volunteer spoke about facing a conflict after she walked out on a session:
[T]here was this bond made . . . Even though I was mad, I still came back to face it. “Okay, I’ll just put it on the table for you. Here it is . . . why I got upset. You have to know.” [A]fter that it was kind of like a different, I felt very close to the group.
During this conflict and others, offenders demonstrated an extraordinary ability to tolerate ambiguity and uncertainty, holding diverging points of view without making judgment.
Having faith in the process
Very early on, researchers notice that volunteers have an unwavering faith in the ultimate positive outcomes of the program, regardless of individual’s progress or lack thereof. Although BTL is designed as a faith-based program, participants far more frequently refer to their faith in the program than any particular faith-based elements. One offender said, “The whole thing has to do with faith. It takes faith to dump your guts on the table.” A volunteer observed,
It’s for anybody and everybody and it’s beautiful and it just seems to work no matter what religion you are. So by faith I think it’s just faith in yourself, faith in the program and faith in your higher power.
Leveling of Barriers and Releasing Expectations
Through interaction, the lines between victims and offenders become blurred, as does the hierarchy of helping or fixing the person who has something “wrong” with him or her. In BTL, all participants share their struggles to heal and to deal with pain and shame in their past. One offender participant stated,
I related to Ellen’s story. We were in sort of reverse roles. My wife was a dope fiend and I was a workaholic. Ellen’s ex used the dope in her relationship. I got things from her story and I think she got some things from mine.
As participants share equally, the humanness of the participants becomes more essential than their roles. By graduation, researchers noted that offenders were referred to by colloquialisms such as “my guys.” The research team transcriber interjected that she herself had difficulty in labeling the participant types as the study progressed and settled on “participants” rather than “offenders.” Researchers, too, struggled with boundaries in their roles:
When they were telling the story, I fully expected to be moved but not to the point where I would have tears running down my face . . . I’ve been a clinician for many years and know that that’s a way that you connect and show your humanness and be authentic, but not to that degree.
The norm of leveling barriers was expressed behaviorally in participants’ releasing expectations about what could take place.
BTL tears away many assumptions of a traditional treatment program. One offender participant remarked, “This is an awesome program. It is heartbreaking, funny and exciting. I never knew what was going to happen.” It also challenges beliefs of group members about who they are and what they can expect when interacting with others. A research investigator shared,
There are these moments where I feel kind of flipped into this place of uncertainty about how to be. “What, (laughs) what am I? Am I to watch? Am I to intervene? Am I to be a human being here? Or am I supposed to be all those things? And how do you do that?” It’s not just so clear.
Frequently, group sessions diverge from the established curriculum such that responding to each other as human beings took precedence over fidelity to the rules. In one instance, a volunteer approached a crying offender to offer a hug and stated “This is totally against the rules. I don’t care.” Atypical and even seemingly inappropriate expressions of emotions and behavior make interactions not only unpredictable but also interrupt systemic reactions or repetitive patterns that reinforce homeostasis or the status quo. One of the research investigators commented, “It all merges. It crosses all over the place. That’s what BTL does is that it blows assumptions, which then makes chaos. But it also makes room. It makes openings for other things to happen.”
Demand for Authenticity
Authenticity refers to a level of honesty expected from group members, which is—in the words of one offender—“as real as real can get.” Sharing goes beyond a strictly cognitive accounting. An offender described what his buddy went through in preparing to tell his story: “He talked about being very scared . . . he knows now that it just can’t come from the head. He knows he has to live it, feel it, breathe it, have it come from his heart.” Victim volunteers may take first steps in modeling authenticity by sharing completely of themselves. “Melvin just plowed into it, with all his rough edges to sort of model how it’s done . . . showing everybody it’s okay you can tell the most gruesome things.” This demand for authenticity is manifested behaviorally in storytelling and critical feedback.
Storytelling
The relating of each person’s crime story is expected of all participants. It carries significant meaning because of the demand for truth telling. A volunteer observed, “When they have to tell their story in group, they have to learn to trust people again . . . you’ll see grown men break down and cry because it’s the first time they’ve ever told their story.” A researcher’s notes from an interview revealed the significance of this event for the storyteller:
He had enough self-awareness to make it through the program and still remain hidden to others, but he was willing to go deeper within himself. He waited until the very end to share his story so he had much time to psychologically prepare. He took this event very seriously and considered it a powerful part of his inner work.
When a group member “holds back” or does not share honestly, the group notices and expects more. After finally sharing, an offender related, “[S]omebody came up to me and they said, ‘I’m so proud of you.’ And I was like, ‘Why?’ And they said, ‘You finally told your story . . . And you’re finally real.’”
Feedback/interaction
Participants give each other critical feedback. Volunteers take a lead role: “I’m the first to call out somebody if I think they’re lying or . . . I just tell them that I don’t think they’re telling the truth.” As trust grows in the group, offenders offer the same level of feedback to volunteers. One offender challenged the victim’s denial in a domestic violence relationship. “She said, ‘No, I left him.’ He said, ‘No you haven’t.’” Another victim described how group members “poked me because they knew there was stuff I was still downplaying.” In her story of receiving feedback and gaining greater understanding, she explains,
I was driving home and the guys had just told me that night, “You haven’t forgiven yourself.” I was like, “I don’t know why, I’m not mad at myself!” And they’re telling me, “Yes you are.” And I’m like, “No I’m not, I’m fine.” And on the way home that night the light bulb came on and I went, “I am mad at myself. I haven’t forgiven myself.
Group members may initially resist the feedback: “At first, when the group tried to give me feedback, I didn’t like it. I thought who are they to tell me what to do? Now I think they are probably right.” But over time, they are able to incorporate feedback into their view of self.
Mediating Effects
Group norms and behaviors collectively have the effect of uncovering and exposing offenders’ histories, beliefs, and motivations through a caring group experience that directly contrasts with what offenders heretofore have experienced. These two mechanisms mediate offenders’ intrapsychic processes and resultant behaviors related to changes in self-schema.
Uncovering
Repeatedly, offenders and victims make reference to unearthing hidden truths as a mechanism for articulating and acting on desired changes. A victim recapped an offender’s process:
He hasn’t told his story and this is an issue that he has been suppressing all of his life. He believes that he has to deal with it to make changes. It’s like, how can something that has happened many years ago be controlling him today?
Group members may describe the uncovering process as “digging up,” “unpeeling layers of an onion” (evoking tears), and even the opening of old wounds or scabbed over places. This includes the victim volunteers:
I’m working down into where all that center of the hurt and pain is for myself. It’s not just surface, it’s getting to the actual root because I think there’s so many layers of all that stuff . . . [M]any times [the men] have been hurt and just kind of . . . piled up and they have to start working through it. And so that’s what I hope . . . is to continue working through it until I’m all the way at the bottom where there’s nothing left that feels ugly, until I’m just (relieved sigh).
The new and often painful process of sharing is made possible by the supportive environment of the group and allows the meaningful personal reflection needed to reveal the true self. Ultimately, getting to the bottom of the pain means surrender and relief.
Contrast effect
Offender participants continually reflect on the noticeable contrast between their BTL and prior experiences which, though unexpected and perhaps disorienting, grabs their attention. One offender shared, “I’ve been to AA and NA but I’ve gotten bad vibes. It’s more of a heartache or hindrance to hear about people using. In BTL, people are real sincere.” The contrast makes differences more stark. For example, in comparing substance abuse counselors and BTL volunteers, some offenders noted the disingenuousness of substance abuse counselors’ who used alcohol in comparison with the integrity of BTL volunteers.
In other programs there are consequences for refusing to participate—serving more time, having to return to rehab, ending up in the same place all over again. The volunteers seem to have genuine concern. They are doing it from their hearts—not counselors trying to get into our brains. It is a human process. You are able to be yourself.
Many offenders have rarely experienced such supportive interactions. A volunteer observed, “Most of them have just never had anybody that was just a true friend that just really was just a friend. Most of their FRIENDS want drugs or there’s some connection to them being friends.” The contrast effect is also apparent in offenders’ experiences with expressing feelings. An offender who remarked that he had “become cold in prison because feelings were a sign of weakness” commented that BTL “finds that soft spot” in him and is starting to break down the wall to help himself.
Intrapsychic Processes
As group norms and behaviors produce experiences for the participants that include unexpected and divergent occurrences as well uncovering heretofore buried realities, offenders begin to feel shifts in their intrapsychic processes that ultimately result in self-system change. Intrapsychic processes can be defined as the internal reactions of participants to the behaviors of other group members in the BTL program. These internal processes include visioning the possible, feeling an obligation to respond, experiencing dissonance, and surrendering. An additional intrapsychic process, outside the influence of BTL, is the offender’s preexisting personal readiness.
Visioning the possible
Prompted by the suggestions of, and positive interactions with, other group members, offenders begin to envision new possibilities for themselves including new or hopeful thoughts. Offenders commonly talk about “conquering the old tyrant to become a new person” or moving from a vague goal to “fix myself” to more a more realistic picture of “making cabinets for [my] friend’s business and taking [my] grandkids fishing and camping.” Commenting on seeing himself differently, an offender said, “When I see myself in the mirror, I see something I like. Before I had so much hatred.”
Obligation to respond
In response to group member support, offenders feel a sense of obligation to reciprocate the group’s care. Moreover, when they feel resistant to taking next steps, a desire to return what they have gotten from the group overtakes initial reluctance. The mechanism of obligation is frequently referred to by participants as a gift exchange or an offering. A victim said, “[S]ome of them want to make us proud. And that’s neat. Because that’s . . . a gift for somebody to want to seek your approval.” A research investigator made an observation about an offender who had had an altercation with a victim participant because of his reluctance to acknowledge the harm he had caused but addressed it in his final letter.
It [was]almost like a gift he gave to her (victim) . . . He had to know the kind of impact that would have, especially on her . . . so he had to come to terms with it for himself, but he also had to know . . . what he would be giving her to, to be that honest and to say those things. And to say them aloud in the group where they’d publicly had this altercation.
Offenders also speak often of plans to “give it away” or “give back” by sharing their story with others, mentoring youth, or taking care of their family in a new way.
Dissonance between selves
As group members begin to envision new possibilities for themselves, they experience conflict between who I have been and what is already formed and who I am just becoming, which is tenuous. Offenders feel the discord that accompanies incorporating new discoveries into their existing schema of “who I am.” An offender commented, “[I want to learn] what kind of person it took to do what I did. It is depressing and I feel ashamed when I contemplate life.” Another expressed similar confusion: “If I really loved my family, how could I stay locked up for 25 years? I’ve never been to my daughter’s grave. I have questions about lots of things. I don’t understand why I am the way I am.” Dissonance begins to resolve as offenders surrender to the change process disavowing allegiance to their prior self. One stated, “I still have the dark side but I’m not the dark side.”
Surrender to process
Indeed, this giving over, with active support from the group, is both a manifestation of the evolving self and an action that produces greater consonance in line with new beliefs about the self.
I had to deal with feelings I didn’t even know were there. I had to deal with them step by step. In the past I was a person I didn’t like. Now I love me. I learned I didn’t have to be that person that I didn’t like. I didn’t hold myself accountable. I’m ashamed that I looked for ways to cover up. The shame kept me from asking for help.
The process of reaching surrender may be messy or require significant effort to let go of resistance, shame, and past hurt. Offenders use metaphors for surrendering such as a weight or burden being lifted, walls coming down, or even undergoing a cleansing flood of tears: “I have held back for so long that everything came rushing to the front . . . it was like a dam burst. And I’m . . . just absolutely exhausted.” This language indicates a sense of breaking to allow something new and better to emerge.
Self-System Change
The intrapsychic processes triggered by the group result in behavioral self-system change. Our findings not only support but also add to Stein and Marcus’s theory of a desired self, which suggests that for meaningful changes to take place, individuals must alter their view of self and find ways to act in consistent ways with the new self. Volunteers witness the process of the emerging self as a gradual opening and unfolding:
These men, when they start changing and you see this change in them . . . [I]t is just so beautiful. It’s like a little seed you plant . . . the very first day and by the end of 12 weeks you’ve got this beautiful rose.
Self-system change involved three stages: (a) articulating counteractive selves, (b) articulating possible selves, and (c) acting on desired selves.
Articulating counteractive selves
In initial interviews, offenders frequently label themselves using negative self-concepts including loser, monster, and criminal. Identifying these tags as counteractive is the first step in self-system change. An offender shared a newfound recognition of how his past self held him back: “My family enabled me. They allowed me to get away with things. But I can’t blame them. I have to take responsibility. I was manipulative and put up a façade. I should have tried harder.” As offenders express an awareness of the counteractive self, they also begin to articulate that these characteristics no longer have to define them. An offender spoke about wanting to be different, but expressed the challenge of accessing a new way of being:
Just know I don’t want to be a hard hearted fool and that’s what I’ve become. In the pen feelings are a sign of weakness. I became cold. My feelings aren’t dead. They’re still there in protective custody. They’re just harder to get to.
Articulating possible selves
In recognition of their existing counteractive selves and through visioning what is now otherwise possible for them, offenders begin to articulate new possible selves both desired and feared. Possible selves often directly contradict their old counteractive labels generating dissonance.
I used to live by the code of the street. You don’t snitch. You live, eat and die by the rules of the street. Now . . . I’m living on the right path. I’m not a criminal. That’s what I’m learning now.
Repeatedly, offenders describe this new self as an external being that they are changing into or moving toward:
I look at this as an opportunity to be strong and become who God intended me to be. I see this as a rebirth, an awakening. I feel like a new person. It is enlightening and I am real happy.
Offenders also express concern about the possibility of returning to the past. Speaking about his feared self, an offender expressed that his life used to be boring without getting high. “[I] wonder what has changed and will it not still be the case.”
Testing and acting on desired selves
The final step in self-system change is practice. Although offenders are limited by what they can do while in prison, the act of claiming a desired self first takes the form of “testing the waters.” Offenders may do that verbally, as one said, “You are going to play how you practice. If I want to be productive, responsible and contributing I have to change my way of life.” Another offender declared, “I took a look and evaluated my file cabinet. I took inventory and now I’m ready to build.” Inside the program, they test out desired selves by behaviorally acting on the possibilities. For example, through storytelling, the writing of unsent accountability letters to victims and family members, and responding accountably to feedback and suggestions from the group, offenders test out their desired selves. These behaviors have weight because, as one volunteer stated, “It’s not just words in group, it’s really actions. And they get out and they change and there’s nothing better than that.” Offenders experience feelings of revelation, excitement, and pride when they take action. An offender got up in front of the large group to speak at graduation: “Oh, man—it was great! I was so nervous. I told myself I was coming from my heart. Me crying—I was so proud. It made me feel good.” The positive feedback and good feelings while testing desired selves progressively move offenders closer to others. With this reinforcement, the desired self grows and takes over.
Synchronicity
The processes within BTL seem to unfold spontaneously, often in a way that is perceived as serendipitous. Team members observe a number of “coincidences” and begin to profess that “there are no accidents.” Frequently, events take place that push forward a group member who is ready or hold back a member who is not yet. “It’s more than luck is all I can say,” said a volunteer. Indeed, it can seem as though an unseen force is directing the current within the group, allowing the right things to emerge at the right time. A research investigator remarked, “It’s just sort of this magic, synchronous unfolding. Unpredictable unfolding.” Some of the group members attribute the mechanism of synchronicity as divine intervention by God. An experienced volunteer shared, “I’ve never had a big religious background. But since I’ve been going through BTL I’ve been seeing these things that can only be described as miracles.” This belief contributes to the ability of members to have faith in the process regardless of an offender’s resistance and to stay through conflict and disruption. The unpredictable element of synchronicity also contributes to the release of members’ efforts to control outcomes, and the ability of group members to surrender to the processes taking place.
Discussion
The results of this study suggest a unique series of mechanisms, which support offender behavioral change in the BTL program, resulting in recidivism reduction. These mechanisms include key elements of restorative justice processes, such as deep caring and listening, leveling of traditional barriers, staying through difficulties, and demand for authenticity and accountability, as well as elements of group cohesion including trust, cooperation, caring, and sustained interest. Certain components in the results are consistent with prior research on offender’s identity shifts and their propensity to desist from crime.
Paternoster and Bushway (2009) contend that the combination of a perceived sense of a future or possible self as a non-offender, coupled with fear that no change guarantees an undesirable future, propels the movement associated with intentional self-change. Indeed, the high scores on offender’s readiness and expectations for change suggest some commitment to self-change before the program begins. They further argue that the strength of the possible self and commitment to change rests on having a road map about how to achieve that change. Similarly, the growing field of desistance research makes a distinction between primary and secondary desistance, with the latter representing an identity shift in which the offender incorporates desistance from crime into his or her self-perception (Maruna & Farrall, 2004; McNeill, 2006). Maruna’s (2001) work in particular highlights the role of narrative and storytelling in the process of secondary desistance. In addition, Maruna suggests that opportunities to engage in generative activities, or to “make good” help inspire and shape offenders’ life scripts.
In this study, however, the results suggest that offender participants’ motivation for change is fueled first by the dissonance in perception of self, which is created not only by personal readiness but also by how “free world” participants view them. Areas of dissonance include the unlikeliness of a caring connection with victims who are similar to persons the offender has harmed, the realization that others see strengths and possibilities that the offender has never envisioned for himself, the group demand to remaining in discomfort to work the conflict through to a positive end, the unexpected safety that ensues when the power hierarchy over others is made irrelevant, and the dissolution of weakness otherwise associated with being authentic, being held accountable, and accepting feedback. The discomfort caused by these dissonance-generating experiences or contradictory conditions is alleviated as offender participants gradually move closer to and act on their desired selves in the BTL sessions, thereby aligning themselves with new cognitions about themselves and the world while pulling themselves further away from their feared self. In essence, the group process—and the unique, non-adversarial group norms consistent with the restorative approach—provides essential support for the activities of self creation in BTL. Motivation for commitment to a possible self is activated as well by the satisfying relationships forged as a result of interaction with others on the basis of that identity.
These findings support Fox’s (2015) recent work on the role of community volunteers in promoting desistance. In her study of a restorative justice-based program known as Circles of Support and Accountability (CoSA), Fox found that voluntary community partners can offer a shared moral space with recently released offenders, and through actions of voluntary acceptance, decrease the perceived distance between offenders and community members. Although evidence suggests that caring and respectful relationships with helping professionals provide motivation for offender change (Burnett & McNeill, 2005; Rex, 1999), the current study supports Fox’s (2015) proposal that voluntary commitments in reintegration work offer added value to the therapeutic relationship and increase intervention effectiveness. As Fox notes, the role of communities in offender reintegration is “undertheorized and understudied.” Through a theory-building approach, the current study clarifies the specific norms and behaviors that assist community members in building a shared moral space and promoting crime desistance with offender populations, thereby creating a therapeutic road map for doing so.
Fox suggests that the formation of positive self-identity can be facilitated by community engagement. Likewise, Farrall (2002) highlights the role of social relationships and social circumstances as “both the object of the intervention and the medium through which change can be achieved” (p. 212). Farrall’s work looks beyond the therapeutic relationship to the value of building social capital in communities. This is a logical next step for the continuation of reintegration efforts started in in-prison programs such as BTL. McNeill (2006) advises that “it is not enough to build capacities for change where change depends on opportunities to exercise capacities” (p. 3).
Albeit with the support of the group, the decision to test and act on a new self is self-determined. This act of choice reinforces offenders’ abilities to control their behavior, a condition that in itself has been found to amplify the manifestations of cognitive dissonance (Galinsky, Magee, Gruenfeld, Whitson, & Liljenquist, 2008). The cognitive dissonance generated by BTL has also been recognized in motivational interviewing as a discrepancy-generating model for understanding the theory’s motivational processes (Miller & Rollnick, 1991). Although not examined in this study, attitude and behavior changes resulting from cognitive dissonance have been found to be mediated by impression management, an activity by which an individual claims that his attitudes have changed to resolve seeming dissonance in the eyes of others (Tedeschi, 1981). Likewise, Gaes, Kalle, and Tedeschi (1978) find greater attitude change when individuals were publically committed to their statements, as occurs in the BTL group context, instead of making them anonymously. Moreover, research shows that dissonance is increased when there is a conflict between a person’s attitudes and something, such as crime, that they freely choose to do (Wicklund & Brehm, 1976). In other words, the person is responsible for the inconsistency. This suggests that public acknowledgment of responsibility, which is an expected part of an offender’s crime story, may be a pre-requisite for the creation or maximizing of cognitive dissonance in the BTL mechanisms of action model. Studies also indicate that either the dissonant state must be maintained over time or the inconsistency must be reintroduced to obtain a long-term effect (Draycott & Dabbs, 1998). The infusion of dissonance into BTL processes over the course of 3 months likely assists in the response effects and reinforcing and sustaining possible selves.
Although examination of faith-based treatments and the role of faith in offender change has been largely neglected (Mears, Roman, Wolff, & Buck, 2006), “synchronistic experiences” in group therapy have been found to strengthen the therapeutic relationship as well as generate insights about self and others (Roxburgh, Ridgway, & Roe, 2015). Focus on synchronicity may provide a novel avenue by which to explore the effects of meaningful coincidences and faith in the unknown as experienced in spiritually integrated interventions. Jung (1972) first proposed the term synchronicity in reference to the making of an acausal connection through meaning. He formulated the notion of archetypes as structuring principles that underlay synchronistic events connecting the inner world of the psyche to the outer world of matter (Mansfield, 1995). This suggested unity between psyche and matter is embodied today in the field of quantum theory (Peat, 1987). In this study, synchronicity refers to participants’ explanations for coincidences that were too common and meaningful to be accounted for by chance. The term also captures how participants used the power of such coincidences as a reminder to let go of the illusion of control and remain engaged in the change process. Whether participants assigned credit to God, the power of the BTL program, or some other source, the emphasis was on the psychological principle of surrender or letting go as a pathway away from resistances and toward change. The principle of living in the present, which is central to Gestalt therapy (Perls, 1969), or surrender as noted in Eastern meditative and mindfulness practices, is not unique but arguably reflects a dynamic necessary for psychological change.
Limitations
This study offers a unique exploration of change processes within the closed system of a prison. Although the prison site restricted data collection methods, the triangulation of possible data sources—including self-reports, group member reports, and prolonged participant observation—supports the validity of the results. In the absence of member checks with offender participants who had been released prior to the completion of the study, BTL staff and community partners provided feedback about the results. Notably, the small sample does not allow for the power to draw conclusions about statistical significance or to generalize to the whole population of BTL offenders, restorative justice participants, or therapeutic groups. Instead, we provide an initial exploratory model of change, which deepens the understanding of desistance mechanisms and advances the literature by providing direction for future research with these populations.
Implications for Practice and Research
The study results have possible application to restorative justice practices, namely, victim–offender mediation, victim–offender dialogue, family group conferencing, and restorative circles (Umbreit & Armour, 2010), as well as other in-prison and group therapeutic programs. Specifically, the group norms and related group behaviors are prominent in restorative practices but have not been previously identified as mechanisms of action or contributory, as a whole, to offender change. Moreover, the caring, respect, and visioning from victims or other persons who are accorded influence in an offender’s life are common elements in restorative justice processes, which serve to challenge the offender’s self-schema and negative behavioral trajectory. So too are the leveling of differences between victim and offender that allows for the humanity that joins them in new ways.
The mechanisms of action identified in this study highlight the dynamics in restorative justice that move change forward. As such, these mechanisms provide a framework for evaluating and troubleshooting restorative justice practices that are limited in their effectiveness, as well as a conceptual tool for designing restorative justice programs and monitoring fidelity. This study offers similar utility to other prison-based programming and group therapy processes utilizing cognitive dissonance and self-schemas as a path to behavioral change.
This exploratory model could be tested across sites using a time series design. Moreover, there should be further assessments of mediators and mechanisms, with attention to operationalization of hard-to-capture constructs such as “visioning” and “dissonance,” as well as manipulation of the proposed mechanisms to measure their effects. In addition, research should test application of the framework to other restorative justice practices as well as expand the work to address the impact of these mechanisms of action on victim participants.
Conclusion
BTL is a 15-year-old restorative justice program that has a demonstrably low reincarceration rate, significant change in offender’s empathy, vengefulness, forgiveness, and relationships with others, and cost savings based on annual costs in Texas facilities. This study examined change in offenders’ self-concept, social identity, and motivation to change their criminal behavior, and identified key components in this program that influence this change. This work has broad applications for the understanding of criminal justice interventions and opens the door to investigation of the underlying mechanisms that create behavioral change across restorative justice practices.
Footnotes
Appendix
Authors’ Note
The contents of this document reflect the views of the authors and do not necessarily reflect the views or policies of the Texas Department of Criminal Justice.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
