Expert David DuBois Describes the Promise of Mentoring Programs for Delinquency Prevention

To realize the full potential of youth mentoring programs, it is critical to advance research on program effectiveness and population-level impact.

By David L. DuBois, Ph.D., Reprinted from the National Institute of Justice

David L. DuBois, “Mentoring Programs for Youth: A Promising Intervention for Delinquency Prevention,” May 26, 2021,

Mentoring programs are a prominent strategy in the United States for preventing negative outcomes and promoting resilience among at-risk youth.[1] Although diverse in their design and implementation, mentoring programs share a common aim of providing young people with structured support from older or more experienced people, such as adult volunteers or students at higher grade levels.

These programs date back to initiatives in the early 20th century that sought to engage men from local communities to be positive role models for boys from disadvantaged life circumstances and, in doing so, stem the tide of young males becoming involved in the justice system.[2] Today’s mentoring programs serve a wide range of age groups — from young children to older adolescents — and populations with diverse needs and risk factors — from poverty and neighborhood disadvantage to specific vulnerabilities such as disability, mental health challenges, or experiences of commercial sexual exploitation. Current program models and approaches differ according to the age of the mentor (e.g., older peers vs. adults), whether mentors are volunteers or paid staff, format (e.g., one-to-one vs. group), and location (e.g., school vs. community). Some programs focus on delinquency prevention while others promote mental health and academic achievement.

The Office of Juvenile Justice and Delinquency Prevention (OJJDP) is the largest federal funder of mentoring programs and awarded nearly $1 billion in grants to mentoring organizations from fiscal year (FY) 2008 to FY 2019. Between FY 2017 and the first half of FY 2019, OJJDP-funded programs recruited 95,000 new mentors and served more than 600,000 youth nationwide.[3]

For such a large and broad investment portfolio to yield the desired results, it must be informed by rigorous and actionable research. This includes identifying ways to enhance program effectiveness[4] and, in doing so, minimize the risk of unintended harm to any participating youth.[5] It is equally important, however, to use research to advance understanding of how to implement effective programs with sufficient scale and reach to make a measurable difference in delinquent behavior, juvenile arrest rates, victimization, and other outcomes at a community, state, regional, or national level.[6]

This article takes stock of the current state of mentoring research on program effectiveness and population-level impact. Each section reviews the research to date, notes key challenges and remaining questions, and highlights promising directions for addressing limitations in the current evidence base.

Mentoring Program Effectiveness

There is ample evidence that mentoring programs have the potential to contribute to positive outcomes for at-risk youth across a variety of demographic groups (e.g., age, gender, race/ethnicity) and program approaches, including cross-age peer, one-to-one, group, and both school- and community-based.[7] Studies find, in particular, that connecting youth to mentoring programs is a viable strategy for both preventing and reducing delinquent behavior.[8] In line with this research, CrimeSolutions — an initiative of the National Institute of Justice that reviews justice-related practices and programs for evidence of their effectiveness — has rated mentoring as “effective” for “reducing delinquency outcomes.”[9] CrimeSolutions has also rated several specific mentoring programs aimed at preventing delinquency or reducing recidivism for those with justice system involvement as “promising” or “effective.” These include, for example, Reading for Life, a group mentoring program that uses works of literature to facilitate moral development and character education as an alternative to court prosecution for first- and second-time juvenile offenders. A randomized controlled trial found statistically significant declines in rates of rearrest and number of arrests for a two-year period following program participation, with these impacts most evident for relatively serious felony offenses compared to misdemeanors.[10]

Research suggests that the effectiveness of mentoring programs tends to be enhanced by practices that are directed toward training and supporting mentors, as well as implementation of programs with fidelity.[11] Furthermore, a strong emotional bond with one’s mentor and related interpersonal experiences (e.g., when youth develop a sense that they matter) have emerged as important mechanisms through which mentoring relationships can promote positive outcomes,[12] including prevention of delinquent behavior.[13]

Conversely, findings also indicate a potential for program participation to be harmful under various conditions, such as when mentoring relationships end prematurely[14] or mentors fail to follow through on basic expectations for maintaining contact with youth.[15] Of particular relevance to delinquency prevention, one study found that participation in a mentoring program was associated with increased involvement in criminal behavior among youth who did not have significant prior arrest histories and who, due to the nature of the program, were exposed to youth who had been arrested.[16] Thus, even though mentoring programs like Big Brothers Big Sisters have been recommended as a way to minimize opportunities for peer contagion and deviancy training (e.g., peers modeling and rewarding deviant behavior),[17] they are not immune to this risk when they incorporate opportunities for peer interaction.

Challenges and Unanswered Questions

Despite significant research on youth mentoring to date, a number of challenges and unanswered questions remain. One is how to account for the substantial variability in the effectiveness of programs that have received rigorous evaluation.[18] CrimeSolutions has reviewed and rated 55 programs that involve mentoring. Of these, nearly one-third (17) have a rating of “no effects”; the remainder are rated as either “promising” (30) or “effective” (8). Because a program must be implemented with fidelity to receive a rating, differences in the extent or quality of implementation are unlikely to fully account for this wide variation.

A second and related challenge is that efforts to incorporate new practices or activities into programs to help increase effectiveness have had limited success. A recent OJJDP-funded review of mentoring research looked at several studies that used randomized controlled designs to examine the effects of hypothesized enhancements to mentoring programs in areas such as mentor training, mentor-youth activities, staff support, and supervision of mentoring relationships. For the most part, the findings failed to reveal significant differences in youth outcomes based on whether they and their mentors had been selected to receive the new practices.[19] These results are concerning, in part, because most mentoring programs, even when demonstrating effectiveness, have been associated with only modest improvements in youth outcomes.[20]

Another challenge is the need for a deeper and more complete understanding of the specific mechanisms through which mentoring relationships influence youth outcomes in areas such as delinquent behavior. Both the lack of well-developed theories of change in the design and description of mentoring programs and the lack of measurement and analysis of potential mediators of outcomes have contributed to this limitation in the current knowledge base.[21] Research that illuminates the “black box problem” of what happens in mentoring relationships is likely to be key for better delineating sources of variation in youth outcomes within and across mentoring programs and then designing innovations that improve effectiveness.[22]

A final challenge worthy of note is that most research on mentoring programs to date focuses on their relatively immediate effects on the outcomes of participating youth.[23] Particularly striking is the limited investigation of the ability of programs to produce sustained, long-term effects on educational attainment, employment, arrests during adulthood, and other key outcomes.[24] Evidence that program effects can decay rapidly following program participation[25] underscores the need for greater understanding of this issue.

Conversely, many programs like Big Brothers Big Sisters have open-ended time frames for participation (e.g., until youth reach age 18) but have been evaluated largely with respect to only brief durations of involvement (e.g., one year). This limits our understanding of the effects that may accrue as youth receive “full doses” of mentoring over more extended periods of their development.

Promising Directions

Recent research provides promising directions for addressing these limitations in the knowledge base. One is to use evaluation to inform modifications to programs that have, in turn, resulted in greater evidence of their effectiveness.[26]

This potential avenue for strengthening mentoring programs is illustrated by recent research on the Quantum Opportunities program of the Eisenhower Foundation, an intensive, year-round, multicomponent intervention for high-risk minority high school students from inner-city neighborhoods. Youth receive both individual and group mentoring from paid staff. Following an initial evaluation that yielded mixed results, a randomized controlled evaluation of a subsequent iteration of the program found that participants had significantly higher grade point averages, high school graduation rates, and college acceptance rates.[27] For example, approximately 76% of program youth graduated from high school, compared with 40% of control youth. The stronger results were attributed, in part, to modifications made to the program, including a new “Deep Mentoring” training curriculum for staff that fosters more intensive and longer-lasting mentoring relationships with participating youth. The training includes an emphasis on mentors serving as advocates for youth — visiting their homes to discuss problems and find solutions, attending parent-teacher conferences, and standing in for parents when needed, for example.

Relative to the research referenced earlier, which tested potential improvements to mentoring programs with largely disappointing results, examples such as this point to a program-specific, data-driven, and iterative approach as more promising for increasing impacts on youth outcomes. This idea is well aligned with the tremendous diversity that exists across mentoring programs, both in their target populations and core models — a reality that makes one-size-fits-all enhancements seem unlikely.

At the same time, notable progress has been made in delineating broader avenues for strengthening programs. Two meta-analyses have identified significant trends toward greater effectiveness for programs that feature support for mentors to provide youth with intentional teaching or guidance as well as advocacy.[28] These findings stand somewhat in contrast to earlier research that pointed to the potential for overly directive, prescriptive mentoring approaches to conflict with youths’ developmental needs for autonomy and constrain opportunities for emotional bonding between mentors and youth.[29] They also run counter to an emphasis in many programs (particularly those using volunteers) on the need for firm boundaries in mentor-youth relationships, presumably to minimize any risk of harm to participating youth.[30]

Further study is needed to understand the conditions under which supporting more encompassing and directive roles for mentors helps avoid pitfalls and improve outcomes. Tasking mentors with highly structured, curriculum-based approaches to guidance, for example, has not been associated with greater effectiveness,[31] suggesting the need for more nuanced and flexible ways of incorporating a teaching role. Recent advances in measuring the distinct processes involved in mentoring relationships offer a promising direction for helping to answer these questions.[32] Researchers, for example, recently reported initial validation research on measuring five mentoring intervention processes: identification with the mentor, social and emotional support, teaching and education, advocacy, and shared time and activity.[33] Examining these processes in relation to youth outcomes could be highly informative in the design and ongoing development of mentoring programs.

Finally, evaluations have emerged that examine the longer-term effects of mentoring on outcomes extending into adulthood. On the whole, the findings provide intriguing preliminary evidence that mentoring received through a program during childhood or adolescence can indeed foster improved functioning at least into early adulthood.[34] One study, for example, recently reported that elementary and high school students randomly assigned to receive school-based mentoring, combined with case management through Communities in Schools (CIS), had fewer arrests in adulthood and, among females, were more likely to attend post-secondary education compared to those receiving CIS case management alone.[35] Meanwhile, a follow-up study of participants in a randomized controlled trial of the Youth Nominated Support Team-Version II (YNST-II) intervention — which helps adults from family, school, and neighborhood or other community settings provide support to suicidal youth following psychiatric care — found that those in the program had significantly lower rates of overall mortality, as well as deaths due to suicide or drugs, at follow-up 11 to 14 years after receiving the program.[36] It is notable that the longer-term impacts of these two programs are evident despite limited evidence of their effectiveness when evaluating outcomes closer to the time of program participation. (The YNST-II program’s rating in CrimeSolutions has changed from “no effects” to “promising” based on the results of the follow-up study.) This pattern of results supports the idea that it can be important to examine the implications of mentoring program participation for later life outcomes, even when evidence of effects on more immediate outcomes is limited.

Scale and Population-Level Impact

It is critically important to consider the extent to which mentoring programs are reaching the youth who stand most to benefit from them, as well as the factors that may be inhibiting achievement of this goal.

Based on a 2013 survey of a nationally representative sample of youth between ages 18 and 21, researchers estimated that of the approximately 24 million at-risk young people, 15 million will have had an adult mentor at one or more points between ages 8 and 18. Structured mentoring relationships — that is, those established through programs — were substantially less common than informal mentoring ties with individuals such as neighbors or teachers. Nineteen percent of the surveyed youth reported having had a structured mentoring relationship, and 44% reported having had only an informal mentor.[37] The greater the number of risk factors reported, the more likely respondents were to recall a time when they did not have, but wished they had, an adult mentor (43% of those with two or more risk factors compared with 22% with no risk factors).

A recent national survey of mentoring programs[38] found that mentor recruitment was the most commonly reported challenge faced by programs (47%). More than 1 in 4 programs (28%) also reported program growth and sustainability as challenges. On average, programs reported that more than 50 youth were waiting to be matched with mentors, which is significant given that the average program served approximately 250 youth. Boys referred to programs were particularly likely to be on a waitlist and to have relatively long waits, with nearly half of programs reporting an average wait time of more than four months for boys.[39]

For youth involved in the juvenile justice system, the reach of mentoring programs has been limited.[40] A national study funded by OJJDP found that only about 6 in 10 juvenile justice settings provided mentoring to youth through their own embedded programs or services or referred youth to external mentoring programs.[41] Among the settings that did not use or refer youth to mentoring, the most common barrier cited (51%) was a lack of access to mentoring programs. Furthermore, more than one-third (39%) of juvenile justice settings reported that one-quarter or fewer of the youth they referred to outside programs were ultimately matched with a mentor. In line with the challenge of mentor recruitment, mentoring programs most commonly cited lack of mentor availability as a barrier to providing services to referred youth (50%). A substantial portion (27%) also reported that refusal or lack of acceptance of the referral by the youth or family was an issue.

When gauging the potential of mentoring programs for population-level impact, it is important to consider whether programs can be effective when implemented widely throughout a community (e.g., in a school system) or nationally. Several multisite randomized controlled trial evaluations of mentoring programs have reported evidence of their ability to positively influence youth outcomes. These include the Big Brothers Big Sisters community- and school-based mentoring programs,[42] Friends of the Children,[43] and the National Guard Youth Challenge program.[44] It should be noted, however, that the results of these evaluations have been somewhat mixed. For example, the National Guard Youth Challenge evaluation reported impacts on outcomes such as receiving a high school degree, but not on justice outcomes such as arrest.

These types of studies often place restrictions on site eligibility in ways that may limit the generalizability of findings to the full population of youth served by the program across all sites nationally. For example, affiliate agencies for the Big Brothers Big Sisters school-based mentoring trial were required to have at least four years of experience delivering the program, strong agency leadership, and strong established relationships with participating schools.[45]

Challenges and Unanswered Questions

It is clear from existing research that mentor recruitment is a pervasive challenge that substantially limits the reach and scale of many mentoring programs. However, investigation of this problem — particularly the effectiveness of different recruitment strategies — is strikingly limited. The OJJDP-funded National Mentoring Resource Center reviewed research on the effectiveness of male mentor recruitment practices, for example, and identified only one study that met methodological criteria for rigor.[46]

A second key challenge is the unknown effectiveness of most of the mentoring programs that have successfully scaled up to a regional or national level and are thus serving the largest numbers of youth. Furthermore, as noted above, the demonstrated effectiveness of widely disseminated or scaled programs that have undergone rigorous evaluation is mixed. These studies have noted challenges with maintaining fidelity of implementation within and across sites, as is common with scaled-up programs.[47] The burgeoning field of implementation science[48] offers frameworks and methods for cultivating a deeper understanding of such issues and developing and testing approaches to address them. However, for the most part, implementation science has not been integrated into research on youth mentoring. The companion area of dissemination science[49] likewise provides an opportunity to explore conditions and strategies that can encourage broader uptake of mentoring programs that show robust evidence of efficacy when implemented on a smaller scale.

Promising Directions

Broadening the range of people who are engaged as mentors is one promising direction for increasing the reach of mentoring programs. Some programs use mentors whose backgrounds may not necessarily align with conventional views or criteria for mentor eligibility or appropriateness, but whose life experiences align with those of participating youth in ways that are thought to make them “credible messengers.” For example, the Arches Transformative Mentoring Program, a group mentoring program that seeks to reduce recidivism among youth on probation in New York City, often uses mentors who have been formerly involved in the justice system, are from the same neighborhood as participants, and have been recipients of similar types of programs or services. A quasi-experimental evaluation of the program found statistically significant reductions in felony reconvictions for program participants compared with comparison group youth at 24 months.[50] However, there were no statistically significant differences in arrests, felony arrests, or reconvictions.

A conceptually related approach involves engaging existing members of the youth’s social network — people who are already involved in his or her day-to-day life. The previously referenced Youth Nominated Support Team-Version II and National Guard Youth Challenge programs employ this strategy. Youth recruit mentors from their own social networks; specific socialization agents, such as teachers in the youth’s school[51] or coaches,[52] can also serve as mentors. Such programs provide a promising approach for expanding the pool of adults involved in mentoring youth by actively engaging those who might not otherwise be considered appropriate for the role or seek it on their own.

Some programs provide mentoring to all youth within a given setting (e.g., a school). Sources of Strength, for example, is a school-based suicide prevention program that uses youth opinion leaders from diverse social cliques to develop and deliver, with adult mentoring, messaging aimed at changing the norms and behaviors of their peers within the entire school population. A cluster-randomized trial[53] of the program involving 18 high schools found significant improvements in perceptions and behaviors pertaining to suicide and in social connectedness among students in program schools.[54] CrimeSolutions rates the program as “promising.” This type of “whole setting” approach has received only limited evaluation to date, and not all results have been clearly supportive.[55] Yet, in view of its potential to greatly increase the number of young people whom structured mentoring programs can reach, it is a strong candidate for further investigation.

Also notable is a promising strategy from the broader prevention field that involves using technical assistance to help communities select, implement, and sustain evidence-supported prevention and promotion programs that are matched to their local needs and resources.[56] Cluster-randomized trials of these approaches have indicated sustained positive effects on youth outcomes, such as violence-related behavior and substance misuse.[57] Applying these approaches to youth mentoring programs could encourage greater uptake of evidence-supported mentoring programs within communities and other settings (e.g., schools, juvenile justice systems), especially given the wide range of program parameters that must be considered and the reliance of programs on local resources (e.g., types of available mentors) in making these decisions.

Research and Practice Going Forward

To realize the potential of youth mentoring programs, we must advance the knowledge bases required for optimizing both program effectiveness and the capacity for achieving broad, population-level impacts.

Several topics stand out as worthy priorities in the area of effectiveness research. First, there needs to be more intensive investigation of the change mechanisms that are most important in driving youth outcomes. The National Institutes of Health recently established a funding priority for investigations of mechanisms of change based, in part, on the prospects that such studies could help unify research on behavior change strategies and better delineate key targets for intervention.[58] Extending this approach to youth mentoring research, including its applications to juvenile justice, holds similar promise and could be supported through more consistent measurement of common relationship processes in evaluation studies.[59]

Second, greater attention should be given to the ongoing development of mentoring programs to optimize their effectiveness. Iterative cycles of development, rigorous evaluation, and program refinement appear particularly promising in this regard. This type of research can help better delineate the outcomes and youth who are most likely to benefit from a given mentoring program. Clearly, no mentoring program will serve all purposes or benefit all youth. Greater understanding of which types of mentoring (e.g., one-to-one, group, or peer) are best suited for different purposes and youth would provide a valuable foundation of knowledge for research-informed matching of individual youth with specific programs.

Greater investigation of the longer-term effects of mentoring program participation also merits priority status. This is especially true given the research findings that suggest that some effects occur or continue several years after program participation on important justice-related outcomes. Data already collected in evaluations of shorter-term outcomes could be leveraged to extend the scope and examine program effects at later points in time in a relatively cost- and time-efficient manner. It is clear, furthermore, that this type of follow-up may be useful even when programs have demonstrated limited signs of initial effectiveness.

Advancing the knowledge base for population-level impact should include rigorous impact evaluations of mentoring programs currently being implemented at relatively large scale (e.g., on a regional or nationwide basis). To optimize generalizability of findings, these evaluations need to be designed with representativeness of program sites and participants in mind. Such studies also should carefully examine factors that facilitate or constrain implementation of key program components (e.g., mentor training) both across and within sites; this information can be leveraged to design approaches to improving the quality and consistency of delivery that then can be tested rigorously in the contexts of dissemination and scale-up.

Research that can help expand the reach of local programs also is needed. The relative efficacy of different strategies for recruiting mentors, especially those who are most often in short supply (e.g., males), is one area that is clearly ripe for investigation. Another is the development and evaluation of approaches for reaching larger numbers of youth, including using nontraditional mentors and infusing widely available opportunities for mentoring into sites such as schools and correctional settings.

The future directions of research are meaningful only if they can be applied to future practice decisions and programming structures. Research must not just explain what has been observed, but also provide a systematic and structured path to applying that information in the dynamic reality of everyday practice. In keeping with these considerations, it is important to bear in mind that advancing the foundations of knowledge required for program effectiveness and population-level impact — although discussed separately here — stand to be mutually informative and synergistic in ways that support effective translation of research into practice. Consider, for example, strategies for encouraging uptake within communities and other settings of evidence-supported mentoring programs that are tailored to their specific needs and resources. Such approaches offer the promise of increasing the dissemination and reach of programs, thus furthering their potential for broader impact. At the same time, the viability of this type of strategy clearly depends on continued investment in research on program effectiveness. This research will be vital for ensuring that a robust menu of options for evidence-supported mentoring programs exists for those working on the ground in communities to leverage mentoring as a strategy for addressing the needs of young people.

A final and related point to underscore is the field’s need for overarching initiatives and infrastructure to support mutually informing connections between research and practice. These connections are essential for translating research findings into practice — thus ensuring that new knowledge makes a meaningful difference. They are equally important for keeping research appropriately aligned with the most pressing needs of programs and the communities they serve, thereby avoiding gaps in areas of knowledge that are critical for supporting practice. To that end, the National Mentoring Resource Center, funded by OJJDP, has the goal of connecting research and practice through a variety of mechanisms, including reviews of the research evidence to support different program practices and a curated repository of resources that facilitates sharing of practitioner innovations.

To access the resource, please click here.