Big Brothers Big Sisters Mentoring Reduced Loneliness in a 4-Year Randomized Trial
DuBois, D. L., & Herrera, C. (2026). Investigation of the effects of Big Brothers Big Sisters community-based mentoring on loneliness during adolescence: A randomized controlled trial. Youth, 6(4), 152. https://doi.org/10.3390/youth6040152
Summarized by Emily Buss
About the Study
Loneliness is now widely treated as a public health concern, and adolescence appears to be a period of heightened vulnerability. Mentoring programs would seem well positioned to help, since they are built around a sustained, caring relationship with an adult. Yet, as David DuBois and Carla Herrera point out, almost no rigorous research has tested whether mentoring actually reduces young people’s loneliness. This study addresses that gap using data from a large randomized controlled trial of the Big Brothers Big Sisters of America (BBBSA) community-based mentoring program.
The trial enrolled 1,353 adolescents with an average age of 12.3 and followed them for 4 years. Youth were randomly assigned either to the program or to a control group. About 68% of youth assigned to the program were actually matched with a mentor. The researchers estimated 2 effects on youth-reported loneliness at the end of the study. The first was the effect of being assigned to the program, regardless of whether a match occurred (the intention-to-treat effect). The second was the effect of actually being matched with a mentor (the complier average causal effect). They also ran exploratory analyses to see whether effects differed by gender, age, whether youth already had a very important nonfamily adult at baseline, and how lonely youth were at the start.
Both effects favored the mentoring group. Youth assigned to the program reported lower loneliness at the end of the study than youth in the control group (standardized mean difference of 0.17), and the estimated effect for youth who were actually matched was somewhat larger (0.25). These are modest effects, but they come from a randomized design over a 4-year period. The exploratory analyses generally found no differences by gender, baseline loneliness, or whether youth already had an important adult in their lives. The one exception was age. There was preliminary evidence that the program had stronger effects for older youth.
Key Findings
The authors describe their findings as initial support for mentoring as a strategy for reducing loneliness in adolescence. Because the analyses of subgroups were exploratory and intended to generate hypotheses, the age finding should be treated as a lead for future research rather than a settled conclusion. Readers should also note that the full article is currently available as an early access version, and this summary is based on the published abstract.
They indicate that BBBS mentoring significantly decreased feelings of loneliness for adolescent participants. (DuBois & Herrera, 2026, Discussion)
Implications for Mentoring
For programs, the most important practical point may be the gap between assignment and matching. Nearly a third of youth assigned to the program never received a mentor, and the estimated effect of receiving a match was larger. This estimate depends on additional statistical assumptions and is different from simply comparing matched and unmatched youth. Reducing waitlists and failed matches is therefore part of the loneliness story. The findings also give programs a credible, evidence-based way to talk about their work at a time when funders and policymakers are paying close attention to youth social disconnection. Finally, the hint of stronger effects for older adolescents suggests that programs should not assume that middle and high school youth are too old to benefit from a traditional mentoring relationship.
Several limitations are important. Loneliness was not a prespecified outcome in the trial registration, and the study used 2 items rather than a standard loneliness scale. Follow-up loneliness data were missing for 28% of participants. The researchers used multiple imputation and complete-case analyses, but neither approach eliminates uncertainty about attrition. The estimated effect of receiving a match also relies on assumptions that may not be fully met.
The follow-up overlapped with the COVID-19 pandemic, and results may differ under other conditions. The study examined one community-based model and did not establish that school-based, group, online, or professional mentoring would have the same effects.



