Arches Transformative Mentoring Program

*Note: The National Mentoring Resource Center makes these “Insights for Mentoring Practitioners” available for each program or practice reviewed by the National Mentoring Resource Center Research Board. Their purpose is to give mentoring professionals additional information and understanding that can help them apply reviews to their own programs. You can read the full review on the Crime Solutions website.


In considering the key takeaways from the research on this program that other mentoring programs can apply to their work, it’s useful to reflect on the features and practices that might have influenced its rating as “No effects” (that is, a program that has strong evidence that it did not achieve justice-related goals).

1. Group mentoring combined with personal reflection and writing, seems like a promising approach to working with youth with justice system involvement.

Although the Arches program struggled to achieve some of its goals related to recidivism, its group mentoring approach is one with great appeal for working with populations that have experienced juvenile justice involvement and adverse life experiences. There is a hope that a group setting, especially a supportive and relaxed one, might make participants, especially boys, more open to sharing their experiences and feelings than might be the case in a more intimate one-on-one model. It is also assumed that the group setting will allow youth to not only learn from the shared life experiences of others, but also will facilitate peer support and new relationship formation, something that might be particularly valuable to youth looking to leave old habits and friends behind. The group environment might also allow reticent youth to ease their way into a mentoring relationship, to see how the adults in the space are going to act before committing, something that is hard to do when thrust into a two-person mentoring relationship. Lastly, it is hoped that a group environment will allow for more fun, a wider variety of activities, and a sense of comradery among the participants, resulting in a more enjoyable program experience overall.

In fact, we have seen other programs working with a similar population use a group approach. In particular, the Reading for Life program, which used small groups built around a book club format to reduce system involvement with 13-18 year olds. That program is rated “Promising” by Crime Solutions, suggesting that a group approach can be at least moderately successful in working with similar youth. In fact, Reading for Life also has an extensive journaling and peer-sharing component, not too dissimilar from the Interactive Journaling used at the Arches sites. These journaling exercises give youth something to contemplate between each session that reinforces key learnings and understandings from the program; they also give the mentoring sessions something to focus on that put youth voice at the center of the conversation and get youth opening up in ways that could easily be challenging without a prompt. Journaling between sessions, in addition to helping youth make meaning of what they are learning through the program, provides the backbone of their interactions and the connections they make with other youth through sharing and listening. 

The evaluation here suggests that the group mentoring model was implemented with varying levels of quality across the Arches sites. Some sites appear to have had very experienced mentors who knew how to facilitate a group, when to step in and out of conversations, and how to create a safe and open culture of sharing. Other sites are reported to have struggled with mentors who lacked facilitation skills or who tended to dominate the conversation with their own voices rather than teasing out the voices of the youth in the room. But in all, it seems like most participants valued the group approach, suggesting it is a good foundational structure for the work of the program. This group format is supplemented by ad hoc one-to-one mentoring for youth who want more support. But in general, Arches seems like yet another good example of a program using a group mentoring model, supported by supplemental journaling, to effectively support juvenile justice involved youth. The program’s outcomes may be enhanced by strengthening the facilitation of the groups, something noted in the conclusions of the evaluation itself.

2. Credible messengers, well supported, can also be critical to working with these youth.

One of the things noted as meaningful by youth participants was that the program placed great emphasis on recruiting mentors with backgrounds and experiences that, to the degree possible, mirrored the backgrounds of the youth being served. The evaluation notes that this similarity of experiences, culture, community, and ethnicity helped place youth at ease and let them feel like mentors could relate to their stories, opinions, and emotions.

However, using mentors with “lived experience” also came with some challenges. According to the report authors, some mentors lacked training in key skills, such as group facilitation, utilizing change talk or cognitive-behavioral principles, responding to trauma, and getting the most out of the journaling curriculum. Even though these were paid positions, not volunteers, many of the mentors seemed to struggle with basic aspects of group management and conversation facilitation. The mentors, it seems, often were exactly the right people to relate to the youth in the program, but may not have known how to build effectively on that “credible messenger” status. The program offered a number of supports in this area—in fact it contracted with an organization just to provide mentors with in-depth training and technical assistance on the skill gaps noted here. But that training was inconsistently utilized and not always adhered to when provided.

The design of Arches—the combination of group mentoring, journaling, interactions informed by cognitive-behavioral principles, and 24-hour-a-day support—seems tremendous on paper, but it faced challenges when using less skilled mentors as much of the delivery system. These credible messengers brought tremendous skills and relatability to the role, but found it challenging to implement some key aspects of the model. This might suggest the need for more staff involvement in program activities or, at least, more scaffolding provided to mentors in support of their work. It’s unclear why the training and technical assistance offered here was ineffective, but it’s clear that credible messengers may need additional support to be so deeply responsible for delivering what can be a fairly technical and nuanced intervention like Arches.

3. A “family” atmosphere may help participant engagement and reduce barriers to deeper involvement.

One of the clear strengths of Arches appears to be the notion of creating a sense of “family” at many of the sites. This is noted throughout the evaluation and the concept of family may be especially salient for the youth who Arches tends to serve. Again and again, the evaluation noted that participants recognized and valued the family atmosphere at each site. This family-type environment was likely very appealing to youth who were challenged by their home environment or lack of parental involvement. It also may have offered an easy way to develop a new “crew” of friends for youth whose delinquent behavior was facilitated by peers in an attempt to find a sense of belonging or connection to a group. The family environment also may well have boosted attendance and let youth know Arches was a safe space for them.

There were several things that Arches did to facilitate this family atmosphere. The program offered food each mentoring session as not only an incentive for attending but also to build a sense of family. Mentors and other staff were available 24 hours a day, seven days a week, something very rare in mentoring programs and much more akin to the role of a natural mentor. Alumni were encouraged to come back to the program and become mentors themselves or at least participate in the sessions and offer their guidance. The program also offered transportation assistance, referrals to other services, and a number of other features all designed to help mentees feel embraced and cared for. And the group mentoring formally built a sense of togetherness and connection that one can imagine must have felt hard earned by these youth.

Unfortunately, this emphasis on a family atmosphere didn’t have quite the influence one would have hoped as judged by results of the evaluation. The program had retention issues at many sites, with at least some of that reported to be the result of inappropriate youth being referred to the program. But some of it was also assessed to be the result of some youth not feeling like a good fit for the program. Some reportedly found the program to be too “male centric” and not as accepting of girls as they hoped. Others thought the journaling curriculum, which was originally intended for incarcerated youth, was off-putting. Still others just struggled to get to the bi-weekly sessions and wished the program had more flexible scheduling. It’s hard to make a program that appeals to, and works for, everyone. But the family atmosphere that Arches was aiming for seems like a good fit overall for these youth.

4. Think about whether your program has sneaky impacts.

There are several possible reasons that Arches wound up getting a rating of “No Effects” in its formal review. Perhaps none was more important than the distinction between the different, but related, outcomes of arrests and convictions. Arches seemed to have a rather significant impact on conviction rates a year and two years after starting the program. In fact, two years out from the beginning of the program. Arches youth were half as likely to have been convicted as the comparison group. But that’s not the whole story…

It turns out that they were more likely to have been arrested in that same time period. Now, some of this was due to the Arches youth having higher levels of risk to begin with—in theory, they should have been more likely to be arrested if the program never happened. But even when controlling for that, the Arches youth were slightly more likely to be arrested at the 12 or 24 month mark. But not convicted. Actually being convicted was far less likely than the comparison group.

One wonders if being in the Arches program itself didn’t influence the conviction rates for some of these youth. Perhaps being in Arches allowed parents, lawyers, or other advocates on behalf of the youth to argue more effectively against prosecution or conviction. Maybe they could argue that these youth were trying to turn their lives around by participating in this mentoring program and deserved another chance at continuing to improve themselves. This type of “hidden” impact might be at work here, where simple participation in the program acts as a bit of shield, even if the program isn’t necessarily making a huge difference for you individually.

The discrepancy between arrest rates and conviction rates is not explained in the evaluation report. It could be that Arches youth were arrested for less serious infractions that tend to be dropped without criminal prosecution. Or it could be that they had a lot of cases of mistaken identity or false arrests. Maybe they simply engaged in crime that was harder to prove and more easily dropped by overworked district attorneys. But those explanations seem unlikely, as they would require these youth to be still getting in trouble, but have that trouble be radically different than their comparison group of pretty similar youth (who were starting from a point of less criminality in the first place).

It’s unclear as to what’s behind that mystery in conviction rates here, but programs may want to ask themselves if there are hidden benefits for their participants—in the perceptions of others, in how they are treated by systems—that might influence their evaluation results. Because the good news on convictions for Arches is certainly confounded by the fact that their mentees were more likely to get arrested. That doesn’t seem much like the reduction in juvenile justice involvement they were hoping for.


For more information on research-informed program practices and tools for implementation, be sure to consult the Elements of Effective Practice for Mentoring™ and the "Resources" section of the National Mentoring Resource Center site.

Request no-cost help for your program

Advanced Search