From BBBS.org

Research

Big Brothers Big Sisters of America

Todd Manwaring

on Oct 8, 2026

• 10 min read

Our Recommendation

Big Brothers Big Sisters of America has submitted the same program to two independent randomized controlled trials thirty years apart. The 1995 Public/Private Ventures study randomized 959 youth to a match or a waitlist and found, at 18 months, 46 percent lower drug initiation, 52 percent less school skipping, and 33 percent less hitting. The 2026 trial, run by the University of Illinois Chicago across 17 affiliates in 13 states and 1,353 youth over four years, pre-registered on the Open Science Framework before enrollment and published in the Annals of the New York Academy of Sciences, confirmed lower violence-related delinquency and recurring substance use and went further. It found significant effects on mental health, including positive affect and depressive symptoms, and on the parenting behavior of the youth's own caregiver.

Those are ultimate outcomes under our framework, measured against a randomized control group at four years, and the caregiver finding is remarkable: a program aimed at a child moved the adult.

Big Brothers Big Sisters of America'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)

Donate Now! 100% Goes to Big Brothers Big Sisters of America

Some young people have an adult outside the household who notices them, takes an interest, and stays. Others do not, and the difference shows up across their adolescence and into their adult lives.

The 1995 study framed the population as youth in single-parent families facing adversity. The researchers' own reading of their results points at the mechanism: what mattered to the children was not the activities but the fact that they had a caring adult in their lives, someone to confide in and to look up to, and because of that they were doing better at school and at home.

The consequences the two trials measured are the ones that follow from that absence. In the near term: skipping school and classes, falling academic performance, starting drugs and alcohol earlier, hitting people, drifting toward negative peers, and weak self-control when pressure comes. In the household: strained relationships with parents. Inside the young person: depressive symptoms and low positive affect, and the absence of confidence that anything they do will matter.

Further out sit the consequences the organization's 2026 report emphasizes: reduced college enrollment, lower lifetime earnings, greater dependency on social services, and an adult life narrowed by what happened between ages 10 and 18.

The Solution

The core model is community-based mentoring: a screened adult volunteer, called a Big, matched with a young person, called a Little, meeting regularly in ordinary settings. The 1995 study documented what the matches actually did, which was eating out, playing or attending sports, going to movies, sightseeing, and hanging out. The intervention is not a curriculum.

Professional infrastructure around the volunteer is what distinguishes this from informal mentoring. The trial registration describes the intervention precisely: one-to-one mentoring provided by an adult volunteer with training and ongoing monitoring and support from program staff. The 1995 researchers found that Big Brothers Big Sisters matches consistently spent more time together and lasted longer than those in other mentoring programs they had studied, and that the program focuses less on specific problems after they occur and more on meeting a young person's basic developmental needs.

A separate 2013 study of mentoring across seven programs tested that infrastructure directly and found that mentors who received early-match training and consistent program support met more frequently and sustained longer relationships, with their mentees reporting higher relationship quality. The same study found that youth with differing risk profiles had relationships of similar strength and duration and derived similar benefits, which is the available answer to the standard objection that mentoring programs succeed by selecting easy cases.

Dosage in the evaluated model: roughly three meetings a month over an average of a year in the 1995 study, with the current trial following youth for four years.

A federated structure. A national office in Tampa sets standards, youth safety policy, and measurement tools; local affiliate agencies recruit, screen, match, and support. Youth safety is treated as a named organizational function rather than an administrative one.

Mapped against a Theory of Change, the chain is short and has now been tested end to end: a consistent caring adult, then better self-control and better peer choices, then less substance use and delinquency and better school attendance, then better mental health and family relationships. The 2026 trial measured every link on that chain simultaneously, which is why its secondary outcomes matter as much as its primary ones.

Key Outputs

120 years of operation, one of the longest-running youth-serving organizations in the United States
A national federation of local affiliate agencies across the country; the 2026 trial alone drew on 17 affiliates in 13 states
$2,000 to $3,000 per youth per year in programming costs, per the organization's own estimate
Two randomized controlled trials: 959 youth at 8 agencies (1995) and 1,353 youth at 17 affiliates in 13 states over 4 years (2026), the latter registered as NCT03495635 with a pre-registered analysis plan
Game On, a sports-and-mentorship initiative launched in 2026 with a target of getting 500,000 young people "in the game" by 2035
Programs including an Alumni Network, a National Youth Council, Little Reporters, and Bigs and Littles of the Year

Key Intermediate Outcomes

From the 1995 Public/Private Ventures randomized waitlist-control trial (959 youth, 8 agencies, 18-month follow-up, surveys of both youth and parents at baseline and follow-up):

46 percent less likely to begin using illegal drugs
27 percent less likely to begin using alcohol
52 percent less likely to skip school, skipping half as many days
37 percent less likely to skip a class
33 percent less likely to hit someone
More confident about doing schoolwork

From the 2026 University of Illinois Chicago randomized trial (1,353 youth randomized three to one, 1,011 treatment and 342 control, across 17 affiliates, 4-year follow-up, pre-registered). Analysis was intent-to-treat across everyone assigned to treatment, 68 percent of whom had actually been matched with a mentor by year four, which means these figures understate what a completed match produces:

Significantly lower violence-related delinquent behavior, at 29.6 percent against 43.0 percent, and recurring substance use, at 18.2 percent against 31.4 percent
Significant effects on negative peer associations, aggressive behavior, self-control, social skills, and coping efficacy, academic performance, and occupational identity
Significantly lower property-related delinquency, at 26.4 percent against 34.1 percent in the control group
Arrest was nonsignificant, at 9.4 percent against 13.4 percent, with arrest records obtained for 73 percent of the sample
Four years out, mentored youth were significantly less likely to have dropped out of high school

Two things to the organization's credit, and one it should fix. The 2026 trial used administrative arrest records rather than self-report for its delinquency outcomes, which is the harder and better choice. The null results are published rather than omitted. What should change is the 15 percent lifetime earnings figure the organization quotes, which comes from a paper whose own randomized sample showed no significant effect.

Key Ultimate Outcomes

Four ultimate outcomes are measured against randomized control groups.

Mental health. The 2026 trial found significant effects favoring the treatment group on positive affect and depressive symptoms at four-year follow-up. The four-year crime and delinquency report adds a finding the published paper does not emphasize: suicidal ideation was significantly lower among mentored youth. Under this framework, genuine mental health and wellbeing are ultimate outcomes, and these are randomized estimates of them.

Hope for the future. The four-year report found significantly higher hope for the future among mentored youth. Hope and belief in one's own future is named as an ultimate outcome in our framework, and this is the rarest kind of evidence for it: a randomized estimate, four years out, of whether a young person expects their life to go well.

Family relationships. The 1995 trial found Littles getting along better with their families and reported improved parental relationships. The 2026 trial found significant effects on the parenting behavior of the youth's caregiver. That second finding deserves emphasis: a volunteer matched with a child produced a measurable change in how the child's parent parented. It is a spillover nobody designed for, it points at a mechanism the field has largely ignored, and we could not find it anywhere on the organization's website.

Peer relationships. Improved in the 1995 trial. We count this separately from the 2026 finding on negative peer associations listed above: avoiding harmful peers is an intermediate result, while belonging and genuine social connection are valued in themselves.

Three further claims rest on 2009 opinion surveys with no comparison group. Self-confidence, at 94 percent of Littles expressing confidence in achieving their goals and 90 percent of alumni agreeing their Big made them feel better about themselves. Friendship satisfaction at 72 percent. And wellbeing, at 86 percent of alumni agreeing they lead a fulfilling life.

Arnold Ventures' Social Programs That Work rates the program highly promising rather than top tier, citing attrition in both trials that could undermine the randomization. Arnold Ventures also funded the 2026 trial, which it discloses, so the organization assessing the evidence is the one that paid for it. We score the quality of the measurement rather than the strength of the verdict, and by that standard two pre-registered randomized trials thirty years apart is strong.

Continual Learning & Adaptation

It submitted the same program to randomized evaluation twice, thirty years apart. In 1995 and again in 2026. Most organizations with a famous favorable result from the 1990s would have lived on it. This one commissioned a replication, at larger scale, with a longer follow-up, on harder outcomes. The longer follow-up earned its cost. Several outcomes showed no effect at 18 months and emerged as significant by year four, which means a study on the 1995 timeline would have missed them.

The second trial was designed to be harder to pass. Pre-registered on the Open Science Framework before enrollment began. Registered on ClinicalTrials.gov. Administrative arrest records rather than self-report as a primary outcome. Intent-to-treat analysis. Multi-source, multi-informant measurement with parents reporting separately on outcomes they would observe. And a mid-study analysis plan revision, dated July 2022, that added a more conservative adjustment for siblings randomized within the same family, explicitly noted as appropriate to add despite outcome analyses already being under way. Tightening your own statistics mid-study, in the direction that makes significance harder to reach, is a small act of integrity.

The findings are published in full, including the nulls. The peer-reviewed abstract states plainly that arrest was nonsignificant and that there were numerous outcomes for which effects were nonsignificant. The researchers offer their own explanation for the arrest null, pointing to systemic bias in arrest patterns and incomplete records, with arrest data available for 73 percent of the sample. The independent Arnold Ventures review, which names the limitations of both studies including attrition that could undermine randomization, is public.

The findings are split across two documents. Several of the strongest results, including lower suicidal ideation and higher hope for the future, appear in the four-year crime and delinquency report rather than in the peer-reviewed paper, which focuses on social-emotional, behavioral, and academic outcomes. A reader relying on either document alone gets an incomplete picture of what this trial found. Both are public, which is the thing that matters most.

It publishes the paper that undercuts its own headline numbers. The Long-Term Impacts of Mentors working paper, with its statement that no significant effect was detected in the randomized sample, is hosted at bbbs.org. The organization made the disconfirming detail available while quoting the favorable number, which is a strange combination but is at least not concealment.

The model evolved into new delivery formats over 120 years, including school-based and site-based variants alongside the community-based program, with Game On launched in 2026 as a sports-based route targeting 500,000 young people by 2035.

Youth safety is a standing organizational function, with a dedicated section and policy infrastructure, in a program category where the risk is obvious and where other organizations have failed badly.

Big Brothers Big Sisters of America

Todd Manwaring

Our Recommendation

The Solution

Key Outputs

Key Intermediate Outcomes

Key Ultimate Outcomes

Continual Learning & Adaptation

Further Reading

Written by

Todd Manwaring

Our Recommendation

The Social Problem

The Solution

Key Outputs

Key Intermediate Outcomes

Key Ultimate Outcomes

Continual Learning & Adaptation

Further Reading

Be the first to know when a new organization is certified—and why it meets the standard.

Written by