Friends of the Children

Our Recommendation

Friends of the Children is a strong organization, particularly for funders who recognize that breaking the cycle of generational poverty requires long-term human relationships rather than short interventions. Its model is explicitly built toward ultimate, whole-life outcomes such as: belonging, self-esteem, emotional health, hope, and agency. It measures both the intermediate milestones (graduation, avoiding the justice system, delaying parenthood) and several ultimate outcomes along the way.

This organization did something that takes real institutional courage: it submitted a 12-year relational intervention to a multi-site randomized controlled trial with a referral-only control group, starting when the children were in kindergarten, and it has kept that trial running for more than a decade as participants move into adulthood. It also commissioned quasi-experimental studies with ICF using school and child welfare administrative records across seven chapters. What those studies found is honest and modest. By the end of elementary school, caregivers rated program children as showing significantly less growth in school-related troubles and significantly more growth in positive school behavior, with every comparison favoring the intervention group. Emotional and behavioral strengths showed no difference in the early phase and clinically meaningful gains later, with effect sizes the researchers describe as typical of the youth mentoring field. The School Study found most outcomes not statistically significant in the early years.

Friends of the Children's Fierce Certification score is 100/100 points based on our criteria:
✔ Has Ultimate Outcome Goals (20 pts)
✔ Measures Intermediate Outcomes (15 pts)
✔ Measures Ultimate Outcomes (20 pts)
✔ Measures Intermediate and/or Ultimate Counterfactual (20 pts)
✔ Shows Continual Learning & Adaptation (25 pts)

A note on what this score means. Our framework rewards the quality of an organization's measurement, not the size of its results. Friends of the Children scores 100 because it submitted a twelve-year relational intervention to a randomized trial with a referral-only control and has run it for over a decade, not because the findings so far are large. They are not. The five-year results are a trend on emotional strengths and a significant effect on school behavior, with effect sizes typical of the mentoring field. The study that will answer the real question is still running.

The Social Problem

Friends of the Children selects children who face multiple systemic obstacles, identified in kindergarten or first grade through partnerships with public school districts and child welfare systems. The organization's founding premise, held since 1993, is that the binding constraint for these children is not a service they lack but a relationship: one stable, long-term adult who does not leave.

The consequences of that absence compound in a documented sequence. In the immediate term: trouble in school and disciplinary involvement, falling behind academically, entry into foster care and repeated placements once there. In adolescence: dropping out before a diploma, juvenile justice involvement, becoming a parent before adulthood. In young adulthood: no route into post-secondary education or work.

Underneath sit the harms that matter in themselves and that the sequence above is a proxy for. Untreated mental health difficulty and emotional distress. Exposure to violence, neglect, and abuse. Housing instability. Unmet medical need. The loss of self-worth that comes from concluding one is not worth investing in. Isolation. And the erosion of any expectation that things will change, which the organization names directly when it describes amplifying children's voices as they write their own stories of hope and resilience.

Around the child sits a caregiver carrying too much with too little, which is why the organization added a two-generation approach.

The Solution

The commitment. One child, one Friend, 12 or more years, no matter what. Children are selected in pre-kindergarten or kindergarten and matched with a paid professional mentor, called a Friend, who stays through high school graduation. The mentors are full-time salaried staff with small caseloads, trained and supervised, not volunteers. That distinction is the organization's central design choice and the reason its evidence is not interchangeable with the volunteer mentoring literature.

The dosage is real and measured. In the randomized trial, children averaged more than 300 mentor contacts over five years, 92 percent of children in the intervention condition met with their assigned mentor, and 81 percent still had a mentor at the five-year mark.

Four program strands built around the core relationship (How It Works): the Core Assets framework that structures what mentors work on; foster care work targeting entry and re-entry prevention; a mental health program; and a two-generation approach serving caregivers alongside children.

A national chapter model. Roughly 30 chapters operate as independent nonprofits delivering the model locally, supported by a national office and the Duncan and Cindy Campbell National Center of Excellence for Professional Youth Mentoring.

Mapped against a Theory of Change, the chain runs: a consistent professional adult relationship, sustained over 12 years, produces better behavior and stronger emotional resources in childhood, which produce educational persistence and avoidance of justice and early parenthood in adolescence, which produce a better adult life.

Key Outputs

• Roughly 30 chapters across the United States, including Portland, Seattle, Tacoma, Spokane, Southwest Washington, Chicago, Detroit, Austin, Tampa Bay, Los Angeles, Coachella Valley, and Flathead Nation, which launched following a 2023 resolution by the Confederated Salish and Kootenai Tribal Council
• Founded 1993; the Duncan and Cindy Campbell National Center of Excellence for Professional Youth Mentoring now anchors the model
• 12-plus year commitment per child, beginning in pre-kindergarten or kindergarten
• Trial-verified delivery: 92 percent of intervention children met their mentor, over 300 contacts per child across five years, 81 percent still matched at five years
• Foster care goal: a 10 percent reduction in five-year-olds entering foster care by 2030, with 12 percent fewer age-5 entries reported as of the most recent data

Key Intermediate Outcomes

The intermediary layer divides sharply into results with a comparison group and results without one, and a donor should read them differently.

With a randomized control group, from the multi-site trial of 278 children. Random assignment was at the individual level, blocked on school and child sex, with the control group receiving services as usual, with annual assessments across five years and intent-to-treat growth curve analysis:

• Significantly less growth in school-related troubles among program children
• Significantly more growth in positive school behavior
• Favorable Phase II effects on externalizing problem behaviors
• Every comparison favored the intervention group, and the pattern held despite real-world disruption: during the Great Recession one of the four chapters temporarily replaced professional mentors with supervised social work students

With a quasi-experimental comparison group, from the ICF studies across seven chapters using administrative records:

• The Child Welfare Study found a promising pattern related to reduced time in care for program youth
• The School Study found that most outcomes did not reach statistical significance in the early years of program involvement, and the researchers reported significant challenges obtaining and working with administrative data, which they note has implications for their conclusions

Without any comparison group, Friends of the Children reports four headline statistics. The organization publishes them in two places with slightly different values, so we give the range spanning both: 92 percent entering post-secondary education, service, or the workforce with a living-wage job; 83 to 90 percent earning a diploma or General Educational Development credential; 93 to 96 percent free from juvenile justice involvement; and 95 to 98 percent waiting to parent past the teen years.

A note on the foster care figures. The 12 percent reduction in five-year-olds entering foster care is a national population trend measured against the organization's own target, not an attributable effect. The claim that children at highest risk in the Los Angeles and Austin pilots avoided foster care 100 percent of the time is described as preliminary, has no stated denominator, and has no comparison group. A perfect rate in a small pilot is an early signal, not a finding.

Key Ultimate Outcomes

Emotional and behavioral strengths were measured with validated instruments against a randomized control, and the picture is genuinely mixed.

The trial administered the Child Behavior Checklist and the Behavioral and Emotional Rating Scale annually for five years to both conditions. The trial was ongoing and reported in phases. Phase I covered the program's initial years and found no differences between mentored children and controls. Phase II, the same 278 children assessed later, yielded clinically meaningful findings favoring the Friends of the Children participants. Because these are the same children measured further along rather than two different samples, the pattern is a statement about duration. The effect emerged as time in the program accumulated, which is what a twelve-year model would predict and what a shorter evaluation would have missed.

The randomized trial continues as the Young Adult Study, which is the phase designed to test whether childhood gains translate into adult wellbeing and quality of life. Under this framework, unfinished counterfactuals are excluded from scoring but count as evidence of ongoing learning, and this one is substantial: a randomized trial of a 12-year intervention followed into adulthood is a rare and valuable thing for the field regardless of what it finds.

Continual Learning & Adaptation

It submitted a relational intervention to randomization, which almost nobody does. Long-term mentoring is exactly the kind of program that is usually defended as too relational, too individualized, and too long to test experimentally. Friends of the Children randomized 278 children at the individual level, blocked on school and sex, with a referral-only control, and has run the study for well over a decade.

It kept the trial running through conditions that would have ended most studies. The Great Recession forced one of the four participating chapters to temporarily swap professional mentors for supervised social work students partway through, a fidelity break that the researchers disclosed rather than concealed, and which the published analysis accounts for.

It measured its own delivery, not just its outcomes. The proportion of children who met their mentor, the proportion with the same mentor across five years, the proportion still matched at the end, and the average number of contacts are all published. Those figures are how a reader knows whether the thing being tested was actually delivered, and most organizations do not report them.

It commissioned quasi-experimental studies that came back mostly null, and the results are in the literature. The ICF School Study found most outcomes not statistically significant in the early years, and the researchers documented the administrative data problems that limited their conclusions. That work appears alongside the favorable findings in the same 2026 handbook.

It used qualitative research to change the program. Five qualitative studies examined the youth experience, the mentor experience, and the caregiver and parent role. The published summary states plainly that findings from these studies have been used by Friends of the Children to adapt and improve the program across time, and frames this as a model for how child- and family-serving programs can listen and learn, specifically to be more inclusive of parents and caregivers and to attend better to the needs of front-line mentors.

The program changed in response. The two-generation approach serving caregivers, and the dedicated foster care and mental health strands, are downstream of that qualitative work and of the recognition that the child cannot be helped in isolation.

It has set a falsifiable external goal. A 10 percent reduction in five-year-olds entering foster care by 2030 is a public target against which the organization can be measured and can fail.

Further Reading

Organization Sources

• Friends of the Children national homepage, source of the 92 percent, 90 percent, 96 percent, and 95 percent outcome figures and the model description. friendsofthechildren.org
• Friends of the Children, Impact Report 2024-2025. impact.friendsofthechildren.org and full PDF
• Friends of the Children, Long-Term Outcomes section of the Impact Report, source of the 92 percent, 83 percent, 93 percent, and 98 percent figures, the foster care reduction goal and the 12 percent result, and the Los Angeles and Austin pilot claim. impact.friendsofthechildren.org
• Friends of the Children, How It Works, Core Assets, Power of One, Foster Care, Mental Health, and Two-Generation Approach. friendsofthechildren.org/how-it-works
• Friends of the Children, Resources and Locations. Resources and Locations
• Friends of the Children, Financial Snapshot, Impact Report 2024-2025. impact.friendsofthechildren.org

Randomized and Quasi-Experiment Evidence

• Eddy, J. M., Martinez, C. R. Jr., Grossman, J. B., et al. (2017). "A Randomized Controlled Trial of a Long-Term Professional Mentoring Program for Children at Risk: Outcomes Across the First 5 Years." Prevention Science. Source of the school behavior findings, the strengths trend, and the effect size characterization. Springer
• Eddy, J. M., Martinez, C. R. Jr., Herrera, D., Wheeler, A., Cearley, J. J., Gau, J. M., and Grossman, J. B. (2026). "The Child Study: The Launch of a Randomized Controlled Trial of the Friends of the Children Program." In Handbook of Professional Youth Mentoring. Source of the sample size, random assignment design, mentor match and retention figures, and the continuation as the Young Adult Study. Springer
• "Measuring Mentoring Outcomes Using Administrative Data: Results from a Child Welfare Study and a School Study." In Handbook of Professional Youth Mentoring (2026). Source of the ICF quasi-experimental findings, the reduced time in care pattern, and the null school outcomes. Springer
• "Summary of the Findings of Qualitative Studies of the Friends of the Children Professional Youth Mentoring Program." In Handbook of Professional Youth Mentoring (2026). Source of the statement that qualitative findings have been used to adapt and improve the program. Springer
• Office of Juvenile Justice and Delinquency Prevention, "Twelve-Year Professional Youth Mentoring Program for High Risk Youth: Continuation of a Longitudinal Randomized Controlled Trial." Source of the Phase I and Phase II characterization and the tentative framing. ojjdp.ojp.gov
• National Institute of Justice abstract of the five-year randomized controlled trial results. ojp.gov

Independent Commentary

• Evidence Based Mentoring, "Decade-Long Mentoring Program Shows Small Wins in Childhood, Bigger Test Still to Come," August 2026. evidencebasedmentoring.org

• Share
• Tweet