Mentoring to Reduce Community Violence

Table of contents

Impact Assessment

The findings below synthesize the results of the studies on mentoring programs across three domains of measurement:

‍

Healthcare Cost, Utilization & Value: More evidence is needed on the impact of mentoring programs on healthcare utilization-related outcomes for adolescents and young adults. Only one descriptive study was identified examining the time spent in a behavioral health facility among youth who received mentoring. Future evidence could examine the impacts of mentoring on healthcare utilization related to injury, violence, as well as behavioral health.
Health: More evidence is needed on the impact of mentoring programs on health-related outcomes. A randomized controlled trial indicated a positive relationship between a high-quality mentoring relationship and self-rated health. However, evidence from a rapid evidence assessment found that the evidence on the impact of mentoring and health was inconclusive. An observational study found no relationship between mentoring and behavioral health outcomes such as suicidal ideation.‍
Social: There is strong evidence on the impact of mentoring programs on social outcomes. Across randomized controlled trials, observational studies, and reviews, participation in mentoring programs was associated with reduced delinquency, measured through a wide range of outcomes such as arrests and re-arrests, gang involvement, and engagement with the juvenile justice system. Positive impacts were also found on academic related outcomes such as truancy and post-secondary attendance, as well as social and psychological well-being. Aspects of mentoring, such as the length of involvement, the focus of mentoring (e.g., on academic skills development) and the quality of mentoring, were all important in shaping mentorship outcomes.

Background of the Need / Need Impact on Health

Community violence is “violence [that] happens between unrelated individuals, who may or may not know each other, generally outside the home[1].” Examples include assaults or fights among groups and gun-related incidents in public places, such as schools and on the streets. Young people are disproportionately impacted by violence in their communities, with homicide being the second leading cause of death for young people ages 10-24[2]. In addition, Black youth are more likely to experience the most severe forms of violence, such as homicides, fights with injuries, and aggravated assaults. LGBTQ+ youth are more likely to experience multiple forms of violence, including physical fighting, being threatened with a weapon, physical dating violence, sexual violence, and bullying, compared to their peers who do not identify as part of the LGBTQ+ community[3].

‍

From 2019 to 2020, the United States (U.S.) national homicide rate increased by 30%. Males from historically marginalized communities, such as Black, American Indian and Alaskan Native, and Hispanic or Latino communities, tend to have higher homicide rates than other racial and ethnic demographic groups[4].

‍

Beyond direct injury, exposure to community violence is associated with a wide range of negative health impacts. They include increased risk of substance use, smoking, and high-risk sexual behaviors, increased mental health conditions, including depression and suicidality, as well as poorer academic outcomes, increased academic difficulty, and a higher risk of dropping out of school[5].

‍

Exposure to community violence also has a direct negative impact on the social drivers of health[6]. It has been found that “differences in a neighborhood’s violent crime predicted up to a 3% difference in Math and English Language Arts test scores. Students living on high-crime streets scored an additional 1% lower than neighbors on safer streets[7],[8].”

Background on the Intervention

According to the Centers for Disease Control and Prevention (CDC), youth homicides and nonfatal physical assault-related injuries cost an estimated $122 billion annually in medical care, lost economic productivity, and quality of life, excluding the costs to the criminal justice system[9]. In 2018, the American Public Health Association (APHA) issued a policy statement framing violence as a public health crisis[10]. APHA’s policy statement draws attention to the need to comprehensively address community violence, using a public health response that also includes fostering the development of healthy relationships for young people, promoting cross-sector collaboration, and providing more positive support and comprehensive services for at-risk communities.

‍

Mentoring has been embraced as an approach to addressing and reducing community violence. Healthy People 2030 has a stated objective of increasing the proportion of adolescents who have an adult they can talk to about serious problems[11].

‍

There are many kinds of mentorship programs, and relationships may vary considerably in aims, content, and the nature of the mentor-mentee relationship. Mentoring is a term generally used to describe a relationship between a less experienced individual (mentee) and a more experienced individual (mentor). Traditionally, mentoring is viewed as a dyadic, face-to-face, long-term relationship between a supervisory mentor (generally an adult) and a mentee (generally a younger person) that fosters the mentee’s professional, academic, or personal development[12].

Additional Research and Tools

Table of contents

Cheng et al. (2008)

Youth aged 10 to 15 years, presenting with peer assault injury.

The youth in the intervention group were matched with a mentor who implemented a six-session problem-solving curriculum, while parents received three home visits with a health educator to discuss family needs and facilitate service use and parental monitoring. The comparison group received a list of community resources with two follow-up phone calls to facilitate service use.

Pre-post analysis. N=118 participants who completed both youth and parent follow-up interviews. N=113 in the final analysis. 89 randomized to the intervention group and 79 to the comparison group.

Social: After adjustment for baseline differences, there was a trend toward a significant program effect, including reducing misdemeanor activity (rate ratio RR=0.29, confidence interval CI 0.08–0.98) and youth-reported aggression scores (RR=0.63, CI 0.4–1.00), while increasing youth self-efficacy (beta=2.28, p<0.05).

‍

The program’s impact was associated with the number of intervention sessions received.

DuBois et al. (2018)

Youth between 10 and 16 years old, enrolled from Philadelphia, PA; Rochester, NY; Minneapolis, MN; Columbus, OH; Wichita, KS; Houston, TX; San Antonio, TX; and Phoenix, AZ. 93% were between 10 and 14 while 7% were 15 to 16. Just over 60% were boys, and about 70% were African American. Almost all lived with one parent (the mother, in most cases), the rest with a guardian or relatives. Many were from low-income households, and a significant number came from households with a prior history of either family violence or substance abuse.

One-to-one mentoring provided by the Big Brothers Big Sisters of America (BBBSA) program.

Randomized controlled trial. N=1,138.

Health: Reports of a high-quality mentoring relationship predicted greater reported self-rated health.

‍

Social: Having been matched with a BB/BS for one year or longer was a statistically significant predictor of fewer total arrests later in life.

‍

Among members from racial/ethnic minority groups, one-year match history with a mentor was a statistically significant predictor of a greater likelihood of post-secondary attendance, a lower likelihood of a property offense, and fewer total offenses.

‍

Among participants whose parents had a low level of formal education (e.g., who did not complete the equivalent of high school), one-year match history predicted a greater likelihood of post-secondary attendance. Among younger participants, it predicted a greater likelihood of post-secondary degree completion.

‍

One-year match history also predicted a lower likelihood of a property offense among females. However, among White participants, it was associated with an increased likelihood of one or more property and person offenses, and among White females in particular, it predicted a lower likelihood of receiving a post-secondary degree.

‍

Reports of a high-quality mentoring relationship were a predictor of a lower likelihood of a juvenile arrest, and less reported stealing, greater reported grit (self-reported perseverance in achievement of goals), emotional, psychological, and social well-being, and less alcohol use during adulthood.

‍

A high-quality mentoring relationship was a relatively strong predictor of grit among females.

Karcher (2020)

Students from 20 different elementary, middle, and high schools in San Antonio, Texas.

The Study of Mentoring in the Learning Environment, which took place between 2003 and 2007.

Randomized controlled trial. N=516 enrolled, with follow-up data collected for 466 participants.

Social: In an analysis of outcomes 10 years after, those who were assigned to the standard of care services and a mentor group were half as likely to have been arrested for a misdemeanor and were 10% more likely to have pursued some form of post-secondary education by age 21, as compared to those who received only standard services.

King et al. (2021)

Youth (66.5% girls), aged 12 to 15, who reported peer victimization, bullying perpetration, and/or low social connectedness.

Let’s Connect (LC), a community mentorship program based on the positive youth development model. The program linked youth to community mentors who connected with the youth and facilitated their involvement in social growth activities over a 16-month period.

Randomized controlled trial. N=218.

Social: In intent-to-treat analyses, LC was associated with modest positive effects for social connectedness, self-esteem, and depression. It had no effects on suicidal ideation or behavior.

‍

Results suggest that LC has the potential to positively impact the developmental trajectories of youth dealing with the interpersonal challenges of victimization, bullying perpetration, or limited social connectedness.

Kuperminc et al. (2018)

Eighth-grade youth at risk for juvenile justice system involvement in San Francisco, California. 32% were African American and 35% Latino.

Project Arrive, a group mentoring program that matched participants in mentoring groups led by school-based mentors.

Pre-post analysis. N=77 youth, of which 68 completed the program.

Social: Results collected after early program implementation (during year 2) found that participants had a significant improvement in unexcused absences from school (28% reduction) and a 48% improvement in their grades.

‍

Eight years post-intervention implementation, participation had significant effects on aspects of external and internal resilience assets including school support, school belonging, meaningful participation in school and at home, caring peer relationships, having prosocial peers, and problem solving.

‍

Project Arrive participants earned significantly more credits relative to comparisons by the end of 9^th and 10^th grade. Additionally, by the end of 10^th grade, 60% of Project Arrive participants were on track toward graduation, in contrast to only 49% of comparisons.

‍

However, differences in grade point average (GPA) between Project Arrive students and comparison students did not reach significance by the end of 9^th or 10^th grade (average GPA for both groups was below 2.0 throughout the study period).

‍

Overall, 13% of Project Arrive students and 10% of comparisons had a juvenile justice record at any time from the 8^th to 10^th grade. Although risk ratios in the analysis of juvenile justice records were not statistically significant, the analysis found that Project Arrive may have contributed to a short-term reduction in risk for juvenile justice involvement during the 9^th grade.

Erdem et al. (2018)

Families living in Toronto, Vancouver, and Montreal, Canada. Enrollees were 9.8 years old on average; 64% girls, 56% White and had received one-to-one community-based Big Brother Big Sister (BBBS) mentoring for at least three months.

BBBS mentoring program.

Pre-post analysis. N=427.

Social: Mentor support predicted positive changes in youth academic adjustment (e.g., school attitude, academic self-efficacy, assistance seeking, and problem solving) mainly when mentees already reported high support from their mothers.

Lennon et al. (2020)

10‐ to 15‐year‐old adolescents who were evaluated in two urban, pediatric emergency departments (Baltimore, MD; Philadelphia, PA) for a peer fight‐related injury between June 2014 and June 2016.

Take Charge!, a mentor-implemented and research-informed violence prevention program that partners with one-on-one community-based mentoring agencies. Mentors were recruited by the community‐based mentoring organization, Big Brothers Big Sisters. The intervention included six violence prevention mentor-mentee sessions and three parent sessions.

Randomized controlled trial. N=49 matched youth and 49 unmatched youth.

Social: Youth who were successfully matched with a mentor were more likely to perceive their injury as very serious or somewhat serious, compared with unmatched youth (95.9% vs. 79.6%, p=0.28).

‍

All other factors (belief that future injury can be avoided, household chaos) were not significantly associated with successful mentor matching.

Lindstrom Johnson et al. (2022)

Youth ages 10-15 treated in the emergency department for an assault injury as a result of peer-to-peer violence.

Take Charge!, an evidence-based violence prevention mentoring program. The program used a one-on-one, community-based mentoring model to assign adults to meet with youth over two to six months and deliver a six-session violence prevention curriculum as well as engage in relationship-building activities. In addition, parents were offered three home visits focused on parental involvement and monitoring.

Randomized controlled trial. N=188, of which 98 were randomized into the intervention condition and 90 into the control.

Social: Intent-to-treat analyses showed statistically significant improvements in conflict avoidance and self-efficacy for the intervention group at nine months and reductions in fighting at 21 months, but an increase in parental report of aggression at nine months.

‍

Complier average causal effect models revealed evidence of an additional effect for reduced problem behavior at 21 months for intervention adolescents who received a mentor.

‍

No effects were found for youth-reported aggression, retaliatory attitudes, deviance acceptance, or commitment to learning. Sensitivity analyses suggested increased aggressive behavior for adolescents in the intervention group who did not receive a mentor.

Lynch et al. (2018)

Young adult probation clients aged 16 to 24 years in New York City (NYC).

Arches, a group mentoring program, started serving young adult probation clients. Arches uses an evidence-based interactive journaling curriculum centered on cognitive behavioral principles, delivered by mentors with backgrounds similar to those of their mentees, known as “credible messengers,” as well as direct service professionals with backgrounds similar to the populations they serve, often including prior criminal justice system involvement. Launched in 2012 as part of the NYC Young Men’s Initiative, with private funding from Bloomberg Philanthropies and oversight from the Mayor’s Office for Economic Opportunity (NYC Opportunity), Arches is managed by the NYC Department of Probation and currently operates with City funding at 13 sites across the five boroughs.

Observational study with a comparison group, pre-post analysis, and qualitative analysis involving focus groups, interviews with program stakeholders, observations of group mentoring sessions, and use of the curriculum. N=279 participants. They were compared with a group of 682 young people who began probation at approximately the same time but did not participate in Arches.

Social: Arches participants were significantly less likely to be reconvicted of a crime. Relative to their peers, felony reconviction rates among Arches participants were 69% lower 12 months after beginning probation and 57% lower 24 months after beginning probation. This impact was found to be largely driven by reductions among participants under age 18.

Sheehan et al. (1999)

Children between seven and 13 in Chicago, Illinois.

An inner-city peer-mentoring program.

Observational study with a comparison group. N=50 for cases and N=75 for controls.

Social: After the intervention period, the participants who received mentoring indicated less support for violence than those who were not mentored. Additionally, there was a statistically significant difference in the behavioral scores of the two groups, with the behavior of youth who were not mentored worsening.

Stewart & DuBois (2020)

A national sample of adolescents from the National Longitudinal Study of Adolescent to Adult Health who were in grades seven through 12 when enrolled at the start of data collection during the 1994-1995 school year.

Youth mentoring.

Observational study with a comparison group. N=4,177.

Healthcare Cost, Utilization & Value: A small proportion of respondents (2%) reported mental health conditions, including having spent time in a mental health facility at Wave III (when study participants were between 17 and 26 years old).

‍

Health: In an analysis of the relationship between mentoring and mental health conditions, a small proportion of respondents reported mental health problems, including past-year suicidal ideation (6.3%) when participants were between the ages of 25 and 34. Having a mentor did not exhibit an association with reports of suicidal ideation.

‍

‍Social: Around the five-year follow-up, analysis found that about the same proportion of mentored and non-mentored youth reported a juvenile arrest (4.5%). However, at around the eight-year follow-up, mentored participants differed from non-mentored participants in that they were 1.9% less likely to have been arrested as a juvenile, 3.2% less likely to have one or more arrests, and 4.7% less likely to have two or more arrests.

Hawkins et al. (2020)

Young people, including those at risk of delinquency.

Mentoring relationships.

Literature review.

Social: Mentoring relationships, both those provided through programs and naturally occurring, have a well-established potential to contribute to reduced delinquent behavior and juvenile justice system involvement among youth. However, the estimated benefits are typically small in magnitude and are not evident consistently across studies.

‍

Evidence is consistent with potential benefits of mentoring program participation both (a) among those who are at-risk for, but not yet demonstrating, delinquent behavior or having juvenile justice system involvement (primary prevention); and (b) for youth who have already exhibited initial delinquent behavior or having justice system involvement (e.g., reduced likelihood of rearrest; secondary prevention).

‍

The available evidence is not sufficient in scope or consistency to support even a preliminary conclusion regarding typical benefits of mentoring for longer-term outcomes, such as arrest during adulthood (tertiary prevention). It does appear, however, that this may be possible.

Lyonsa & Chan (2021)

Young people.

Youth mentoring programs.

Literature review.

Social: The empirical studies indicate that mentoring programs have a “small-to-moderate” impact on mentees’ academic outcomes. However, even the small effects on academic performance and outcomes of educational attainment can have a cumulative impact due to the relatively large number of mentoring programs impacting youth.

‍

Mentoring practices that target school-related skills linked to improved grades and high school graduation are more likely to have a larger-than-average impact on mentees’ academic performance.

Plourde et al. (2020)

Adolescent boys and young men (ABYM).

Mentoring interventions.

Literature review. The review included literature published between 2008 and 2019; a total of 29 articles were included for analysis.

Social: Results found that mentoring approaches demonstrate promise for improving soft skills (positive self-concept, self-control, higher order thinking, social skills, communication, goal orientation, empathy, negotiation, self-efficacy, and decision making) and social assets (such as social networks, peer relationships, and social support) among ABYM—which are two factors that are thought to contribute to positive youth development outcomes for reducing violence perpetration.

Raposa (2013)

Youth.

Mentoring programs.

Literature review.

Social: Mentoring was found to be effective at reducing crime and delinquency for youth and promising at reducing youth use of alcohol and drugs. It was also found to be promising with a small positive effect on multiple educational outcomes, including school attendance, grades, and academic achievement test scores. Additionally, it was found to be promising in improving social skills and peer relationships.

Stewart & DuBois (2021)

Young people.

Mentoring of youth by police and other law-enforcement personnel. This includes community-based mentoring programs that provide opportunities for law enforcement and youth to interact with one another in informal community settings and school-based mentoring that brings youth into contact with law enforcement during school hours.

Scoping review. A total of 4,286 studies were screened, and 28 were selected for full review.

Social: A randomized trial following 102 youth between the ages of 10 and 16 in disadvantaged metropolitan areas in Queensland, Australia found that youth who participated in a school-based police mentoring program had a significant reduction in truancy, compared to youth at schools where truancy policies were implemented as usual. Evidence is lacking to draw conclusions about the effect (if any) of police mentoring on other targeted outcomes, such as youth attendance at school, justice system involvement, and youth social and emotional skills.

Taylor et al. (2013)

Young people.

Mentoring interventions. These were defined to include peer mentoring schemes, programs in prison, the community, and ‘through the gate’; programs in which mentoring is the only or main service provided and/or those where it is delivered to support other kinds of interventions.

Rapid evidence assessment. The evidence assessment was limited to studies published between 1990 and 2012, related to mentoring for adults (aged 18 years and above). 23 studies, conducted in the U.K., U.S., Canada, and Australia were included in the review.

Health: There is no strong evidence of the effects of mentoring on health outcomes, as the evidence was inconclusive. One study showed no effect, but two qualitative studies with small sample sizes reported mentees’ accounts of drawing great support from mentors at times when they were contemplating suicide.

‍

Social: Of eight studies, six detected statistically significant positive impacts of programs involving mentoring on reoffending, rearrest, and time-to-rearrest. However, overall evidence of the effect of mentoring on reoffending was inconclusive: in some studies, mentoring was only one element of the program, and the available evidence does not allow the impact of mentoring alone to be isolated.

‍

Of seven studies that examined employment outcomes, all but one found that mentoring was associated with an improvement in employment outcomes.

‍

Evidence from one study found that mentoring may be associated with increased participation in other programs designed to reduce reoffending.

‍

Available empirical evidence of the effect of mentoring on housing was inconclusive.

‍

All studies reviewed as part of the rapid assessment suggested a positive direction of change for substance misuse-related outcomes. One study showed an association between mentoring and reduced substance misuse, but the findings were not statistically significant. Another study found that mentoring increased contact with specialist drug treatment agencies. Furthermore, there is a strong, wider evidence base demonstrating a link between drug use and reoffending.

‍

All five studies which looked at the effects of mentoring on coping with stress and mentees’ perception of their life problems found positive outcomes. Though the evidence is not conclusive, it does offer some support to a hypothesis that mentoring might improve these psychological skills.

‍

Two studies reported that mentoring led to reductions in pro-criminal attitudes. Positive improvements in self-efficacy, motivation, and feelings of self-worth were reported in two studies.

‍

Two studies showed no effect of mentoring on exposure to negative peers or frequency of community activities. One study found improvements in mentees’ perceived social support, and three qualitative studies provided accounts of mentoring improving mentees’ understanding of relationships and social capital.

Strong Evidence

There is strong evidence that the intervention will produce the intended outcomes.

At least one well-conducted systematic review or meta-analysis (including two or more large, randomized trials) showing a significant and clinically meaningful health effect; and
Consistent findings of health effects from other studies (cohort, case-control, and other designs).

Sufficient Evidence

There is sufficient evidence that the intervention will produce the intended outcomes.

At least one well-conducted systematic review or meta-analysis (including two or more large, randomized trials) showing a significant and clinically meaningful health effect, but inconsistent findings in other studies; or
Consistent findings from at least three non-randomized controlled trial studies (cohorts, practical trials, analysis of secondary data); or
A single, sufficiently large well-conducted randomized controlled trial demonstrating clinically meaningful health effect and consistent evidence from other studies; or
Multiple expert opinions/government agencies supporting the intervention.

More Evidence Needed or Mixed Evidence

There is insufficient evidence that the intervention will produce the intended outcomes, however the results may indicate potential impact.

Lack of demonstration of improved health outcomes based on any of the following: (a) a systematic review or meta-analysis; (b) a large randomized controlled trial; (c) consistent positive results from multiple studies in high-quality journals; or (d) multiple expert opinions or government agencies supporting the intervention.
An insufficient evidence rating does not mean there is no evidence, or that the intervention is unsafe or ineffective.
In many cases, there is a need for more research or longer-term follow-up.

There is strong evidence that the intervention will produce the intended outcomes.

There is sufficient evidence that the intervention will produce the intended outcomes.

There is insufficient evidence that the intervention will produce the intended outcomes, however the results may indicate potential impact.

At least one well-conducted systematic review or meta-analysis (including two or more large, randomized trials) showing a significant and clinically meaningful health effect; and
Consistent findings of health effects from other studies (cohort, case-control, and other designs).

At least one well-conducted systematic review or meta-analysis (including two or more large, randomized trials) showing a significant and clinically meaningful health effect, but inconsistent findings in other studies; or
Consistent findings from at least three non-randomized controlled trial studies (cohorts, practical trials, analysis of secondary data); or
A single, sufficiently large well-conducted randomized controlled trial demonstrating clinically meaningful health effect and consistent evidence from other studies; or
Multiple expert opinions/government agencies supporting the intervention.

Lack of demonstration of improved health outcomes based on any of the following: (a) a systematic review or meta-analysis; (b) a large randomized controlled trial; (c) consistent positive results from multiple studies in high-quality journals; or (d) multiple expert opinions or government agencies supporting the intervention.
An insufficient evidence rating does not mean there is no evidence, or that the intervention is unsafe or ineffective.
In many cases, there is a need for more research or longer-term follow-up.

[1] CDC. (2024, July 28). About Community Violence. Community Violence Prevention. https://www.cdc.gov/community-violence/about/index.html

[2] CDC. (2024, July 27). About Community Violence. Community Violence Prevention. https://www.cdc.gov/community-violence/about/index.html

[3] CDC. (2024, February 15). About Youth Violence. Youth Violence Prevention. https://www.cdc.gov/youth-violence/about/?CDC_AAref_Val=https://www.cdc.gov/violenceprevention/youthviolence/fastfact.html

[4] GBD US Health Disparities Collaborators. (2025). Homicide Rates Across County, Race, Ethnicity, Age, and Sex in the US: A Global Burden of Disease Study. JAMA Netw Open. 2025;8(2):e2462069. doi:10.1001/jamanetworkopen.2024.62069

[5] David-Ferdon, C., Clayton, H. B., Dahlberg, L. L., Simon, T. R., Holland, K. M., Brener, N., Matjasko, J. L., D’Inverno, A. S., Robin, L., & Gervin, D. (2021). Vital Signs: Prevalence of Multiple Forms of Violence and Increased Health Risk Behaviors and Conditions Among Youths — United States, 2019. MMWR. Morbidity and Mortality Weekly Report, 70(5), 167–173. https://doi.org/10.15585/mmwr.mm7005a4

[6] Sharkey, P., Schwartz, A. E., Ellen, I. G., & Lacoe, J. (2014). High stakes in the classroom, high stakes on the street: The effects of community violence on student's standardized test performance. Sociological Science, 1, 199-220.

[7] Richards, M.H., McCrea, K.T., DiClemente, C., et al. Cross-age peer mentoring to enhance resilience among low-income urban youth living in high violence Chicago communities. September 2019. Retrieved from: https://www.icpsr.umich.edu/web/NACJD/studies/37494/versions/. Last accessed on: October 25, 2022.

[8] O'Brien, D.T., Hill, N.E., Contreras, M. (2021). Community violence and academic achievement: high-crime neighborhoods, hotspot streets, and the geographic scale of "community." PLoS One.16(11): e0258577.

[9] CDC. (2024, July 27). About Community Violence. Community Violence Prevention. https://www.cdc.gov/community-violence/about/index.html

[10] Violence is a Public Health Issue. (2018, November 12). American Public Health Association. https://www.apha.org/policy-and-advocacy/public-health-policy-briefs/policy-database/2019/01/28/violence-is-a-public-health-issue

[11] AHealthy People 2030. U.S. Department of Health and Human Service. https://odphp.health.gov/healthypeople/objectives-and-data/browse-objectives/adolescents

[12] USDA Mentoring Program. (2025, November 13). Usda.gov. https://www.usda.gov/directives/dr-4740-001

Study Characteristics and Contextual Tags

Share Your Needs or Insights