Youth Tech. Enterprise, technology, community. A program of C4I Veteran Services.

The Mission

Reach a young person before the street does, and give them something better to be.

Youth Tech puts veterans and real enterprise technology in front of young people growing up around gunfire, and shows them the machines artificial intelligence actually runs on. Not a lecture about staying out of trouble. A rack, a network, a working model, and a person who has been somewhere and came back to say it is reachable.

Illustration of a Youth Tech session: six young people and two veteran instructors standing around a table with a retired enterprise server opened up, handling memory modules and hard drives. — IllustrationA session, as it is designed to run. A retired enterprise server comes off the shelf and onto the table, opened up, with the memory and the drives in their hands inside the first hour. The veteran teaching it is not demonstrating from the front of a room. He is standing in it with them.

The Idea

Technology Is the Most Interesting Thing in the World Right Now. We Intend to Use That.

Every teenager in America has used artificial intelligence this month. Almost none of them in the neighborhoods we are going into have seen the thing it runs on: a rack of servers, the power feeding it, the fiber leaving the building, an industry spending billions to build more. That gap is our opening and it is an unfair advantage. We do not have to convince a sixteen-year-old that this is interesting. We only have to put it in the room.

Then we bring the person. A veteran who left the service, moved into information technology, and is willing to spend an evening in a community center is a different messenger than a recruiter or a teacher. He is not selling anything. He has been through something the young person respects, he is doing the work right now, and he is proof that the distance between here and there can be walked.

The young people we are going after live somewhere that makes concentrating hard and planning harder. We are not going to pretend a technology program fixes that. What we are saying is narrower and, we believe, true: a young person who can picture a specific adult version of themselves, and who has one adult paying real attention, behaves differently. The research on that is below, and so is the research that argues against us.

What we are claiming

That every mechanism this program rests on has published evidence behind it, in populations like the one we serve.

Intensive, in-person technology programs raise STEM degree completion for underrepresented students.
Out-of-school programs work when they teach a defined skill, and do not work when they only occupy time.
A committed adult, not the activity, carries most of the effect.
Access on its own changes who takes computing seriously.

What we are not claiming is that this program reduces shootings. Nobody has shown that about a technology program, including us. The section further down says exactly where our evidence stops.

The Research

Why We Believe This Will Work

Graded honestly, the same way we do it on the veteran side. Every claim carries the study behind it and the caveat that goes with it, in the same block.

First, the ground we are standing on

Community violence is not only a safety problem. It is measurably a learning problem, it happens on a schedule, and it lands on the young people with the least access to computing.

0.5 to 0.66 SD: Drop in vocabulary and reading assessment scores among African American children tested within a week of a homicide in their own block group. A child does not have to witness a shooting for it to reach them. Sharkey, PNAS, 2010.
3 points: Lower English language arts passing rate for Black students exposed to violent crime on their own block in the week before the exam, across New York City public schools. Sharkey and others, 2014.
20%: Share of juvenile violent crime that happens in the four hours between 3pm and 7pm on school days, with the peak in the hour school lets out. OJJDP national report, analyzing 2001 incident data.
60%: Of United States public high schools offer a foundational computer science course. In North Carolina, 65% of rural schools do, against 78% of suburban schools. Code.org, 2025.
18.5%: Of children aged 3 to 18 in households earning under $10,000 have no internet access at home, against 7.9% in households over $100,000. Census Current Population Survey, November 2023.

One finding shaped this program more than any other. Across 16,146 New York City middle school students, the damage neighborhood violence does to test scores disappeared for students attending schools that felt safe and had a strong sense of community. The harm is real, and belonging to an institution that feels like yours appears to buffer it. That is a description of what we are trying to build in a community center on a Thursday night. Laurito and others, 2019.

Six Findings

What the Research Actually Supports

Each graded by the strongest study design available, with its limitation printed alongside rather than left for you to find.

Finding 1

Intensive technology programs move underrepresented students into STEM degrees

Strong

Study: A randomized controlled trial of residential STEM summer programs across three cohorts of applicants. 35% Black, 43% Hispanic, 4% Native American, a quarter of them first-generation college students.
Finding: Students randomly assigned to the intensive in-person programs completed STEM degrees at 64% against 52% for the control group, measured six years later. Enrollment at selective institutions rose 9 to 16 percentage points.
Source: Cohodes, Ho and Robles (2022, revised 2024), National Bureau of Economic Research working paper 30227.
The honest caveat: The applicants were already high achieving, averaging a 3.86 grade point average, and the authors say plainly the results may not apply to all applicants. It is still a working paper. And it cuts at us as much as for us: the online version of the same program produced no significant effect. If you take this study seriously you have to run the program in person, which is the expensive way.

Finding 2

Weekend and after-school programs work when they teach something

Strong

Study: Meta-analysis of 68 evaluations of after-school programs, sorted by whether the program used four practices: a sequenced set of activities, active hands-on learning, a component focused on a specific skill, and explicit targeting of that skill.
Finding: Programs using those practices produced an effect of 0.31. Programs that did not produced 0.07, statistically indistinguishable from zero. The pattern held for problem behaviors, positive social behavior, school grades and achievement test scores. In the authors' words, only the structured programs yielded significant effects on any outcome.
Also: A separate meta-analysis of 35 out-of-school-time studies found that whether a program ran after school or over the summer made no difference to its effectiveness. What the program did mattered. When it met did not. Lauer and others, 2006.
Source: Durlak, Weissberg and Pachan (2010), American Journal of Community Psychology 45(3 to 4), 294 to 309.
The honest caveat: The authors state that these data are correlational and they cannot conclude those practices caused the difference. The formal moderator test reached significance on four of eight outcomes. This is the best available answer to your question, not a settled one.

Finding 3

The adult is the mechanism, not the activity

Strong

Study: Randomized controlled trial of a summer jobs program. 1,634 students from 13 high-violence Chicago high schools, eight weeks of part-time work with adult job mentors at roughly one mentor to every ten young people.
Finding: Violent-crime arrests fell 43% over sixteen months, p = 0.022. The detail that matters: the decline appeared largely after the program ended, not during it, so keeping young people busy cannot explain it. Adding a social and emotional learning curriculum made no additional difference, which the author suggests points to what both groups shared: income, a caring adult, job skills and a connection to an employer.
Supporting: A randomized study of Big Brothers Big Sisters, 959 young people followed eighteen months, found those matched with a mentor were 46% less likely to start using illegal drugs, hit someone 32% less often, and skipped 52% fewer days of school.
Source: Heller (2014), Science 346(6214), 1219 to 1223. Tierney, Grossman and Resch (1995), Public/Private Ventures.
The honest caveat: The Big Brothers Big Sisters findings are all self-reported, and two widely quoted numbers from it, alcohol initiation and grade point average, were significant only at the ten percent level. The Chicago study improved no employment or schooling outcome, and a later pooled analysis found property-crime arrests went up.

Finding 4

Who the mentor is, and how long they stay, decides the result

Strong

Study: Systematic assessment of 73 independent evaluations of youth mentoring programs, plus a study of relationship duration inside the Big Brothers Big Sisters randomized sample.
Finding: Programs worked better under five specific conditions, and two of them describe this program exactly: a good fit between the mentor's occupational background and the goals of the program, and mentors structured to teach or advocate rather than simply befriend. Programs also worked better when the young people served had been exposed to significant environmental risk.
Source: DuBois, Portillo, Rhodes, Silverthorn and Valentine (2011), Psychological Science in the Public Interest 12(2). Grossman and Rhodes (2002), American Journal of Community Psychology 30(2).
The honest caveat: The average mentoring effect is modest, around 0.21, and has not improved across two decades. More seriously, young people whose matches ended quickly reported declines on several measures of functioning. A short or badly run match is worse than none, and that is the strongest argument we know of for running a smaller program that holds.

Finding 5

Access on its own changes who takes computing seriously

Moderate to strong

Study: Synthetic difference-in-differences across 287 high schools over fifteen years, examining what happened when a school adopted the Advanced Placement Computer Science Principles course.
Finding: Computer science exams taken by Black or Hispanic students more than quadrupled, and exams taken by female students more than quadrupled, with a 29 percentage point increase in the chance a school had any participation at all. Nothing about the students changed. The opportunity did.
Supporting: An independent evaluation of a national coding program, comparing participants with waitlisted applicants, found participants 13.2 percentage points more likely to major in a computing field, with the largest gains among Black students, Hispanic students and students on free or reduced lunch.
Source: Ganelin and Dee (2025), Proceedings of the National Academy of Sciences 122(15). Lin, Joshi, Caton and Wilson (2024), American Institutes for Research.
The honest caveat: The course study covers one state and measures course taking, not careers. The coding program evaluation was not randomized, and its own evaluators urge caution about reading the findings as causal impacts. We are repeating their caution rather than dropping it.

Finding 6

A concrete picture of your own future changes behavior

Strong

Study: Randomized trial in three Detroit middle schools. 264 eighth graders, 72% African American and 17% Latino, two thirds on free or reduced lunch. A seven-week program helping students build a specific, believable picture of an adult self and the steps to reach it.
Finding: By the end of the year, standardized test scores up 0.36 standard deviations, grade point average up 0.25, absences down 0.73. Two years later the group still had higher grades, spent nearly a full hour more per week on homework, and showed markedly less disruptive behavior.
Source: Oyserman, Bybee and Terry (2006), Journal of Personality and Social Psychology.
The honest caveat: This was not a technology program. We cite it for the mechanism, not the outcome. Putting a teenager in front of a working engineer from a background like theirs is our attempt to produce that picture, and that is a reasoned bet rather than a replicated result. It is also one study, in one city, twenty years ago.

Why Veterans

The Messenger Is Part of the Message

The mentoring research is specific about what makes a match work, and two of the five conditions it names are the reason this program uses veterans rather than volunteers in general. Programs did better when there was a good fit between the mentor's occupational background and the goals of the program, and when mentors were structured to teach or advocate rather than simply befriend. A veteran moving into information technology, teaching information technology, is that fit exactly.

There is also a modest but real effect from seeing someone like you doing the work. Across 45 studies, role models in field settings produced an effect of about 0.20. Small, consistent, and pointing the same direction as everything else here.

We will be honest about the limit of that argument. A recent review of 42 studies found role model interventions can backfire, sometimes reinforcing the idea that anyone who did not make it simply did not try hard enough. That is a real risk in a room full of teenagers who have watched adults work hard and lose anyway. It is why our veterans are briefed to talk about the help they got, not only the effort they put in.

And why a community organization at all

A study of 264 American cities across more than twenty years found that every ten additional organizations focused on crime and community life, per hundred thousand residents, was associated with a 9% lower murder rate, a 6% lower violent crime rate and a 4% lower property crime rate.

That is measured across the density of many organizations at the city level. It cannot tell you what any single organization does, and the authors say plainly that a registered nonprofit is not the same as an effective one. We cite it as a reason to build community organizations, not as a claim about ours. Sharkey, Torrats-Espinosa and Takyar, 2017.

Illustration of a veteran instructor kneeling at an open network rack, pointing out the cabling to a young person standing beside him while others watch. — IllustrationThe teaching happens at the equipment, one young person at a time. This is the fit the mentoring research describes: an adult whose working background matches what the program is for, structured to teach rather than to keep company.

Straight Answers

What We Do Not Claim

If you are considering funding this, you deserve to know exactly where the evidence runs out. Here is where.

Seven things we will not tell you

We do not claim this program reduces shootings. No technology program has been shown to do that. The violence research on this page tested paid summer jobs and mentoring, not technology education. We have borrowed their mechanisms, and borrowing a mechanism is not the same as replicating a result.
After-school programs, on average, do not reduce delinquency. Two independent meta-analyses found small and statistically insignificant effects, and the largest federal randomized evaluation found no achievement gains and modestly worse behavior in the treatment group. The evidence favors structured skill-building specifically, not after-school programming in general, and we will not blur that distinction to make ourselves look better.
A bad program is worse than no program. Young people whose mentoring matches ended quickly reported declines on several measures of functioning. That is the single most important sentence on this page, and it is why we would rather run something small that holds than something large that collapses in month four.
Mentoring effects are modest. The average across seventy studies and more than twenty-five thousand young people is around 0.21, and it has not improved across two decades of investment. Anyone promising transformation from mentoring alone is selling something.
The strongest study we cite studied high achievers. The randomized STEM summer program trial drew applicants averaging a 3.86 grade point average. Whether the same effect appears in young people who are much further from school is an open question, and it happens to be the exact question we intend to answer.
Nobody has studied veterans teaching technology to young people. We looked. There is no published research supporting or contradicting it in either direction. We think it will work, and we are building measurement in from the first class so that in a few years there is an answer.
We have not run a class yet. We have the design, the evidence base, the measurement plan, the safeguarding requirements and the equipment pipeline. We do not have results, because we do not have participants yet. Any organization our age telling you otherwise is guessing.

The narrower claim we will defend: every mechanism this program rests on has published evidence behind it in a population like the one we serve, and the three design features the research says matter most, in-person intensity, a defined skill, and one committed adult who stays, are the three features we built the program around.

How We Will Know

We Will Publish the Answer Either Way

From the first class forward, measured at intake and at fixed intervals: attendance and who stops coming, credentials earned, employment and wage at six and twelve months, school enrollment where a family consents, and a validated wellbeing measure collected as a program outcome and never interpreted clinically, because we are not qualified to.

If it does not work, that goes on this page. A program that only reports its wins is not measuring anything. And if a researcher at a university or a VA facility wants our class as a study population, the door is open and the data goes with them.

Sources

References

Every figure on this page traces to one of these. Where a study is a working paper, was not randomized, or has been disputed, we say so beside the claim rather than burying it here.

Cohodes, S., Ho, H., and Robles, S. (2022, revised 2024). STEM Summer Programs for Underrepresented Youth Increase STEM Degrees. National Bureau of Economic Research working paper 30227.
Durlak, J.A., Weissberg, R.P., and Pachan, M. (2010). A meta-analysis of after-school programs that seek to promote personal and social skills in children and adolescents. American Journal of Community Psychology, 45(3 to 4), 294 to 309.
Lauer, P.A., Akiba, M., Wilkerson, S.B., Apthorp, H.S., Snow, D., and Martin-Glenn, M.L. (2006). Out-of-school-time programs: a meta-analysis of effects for at-risk students. Review of Educational Research, 76(2), 275 to 313.
Heller, S.B. (2014). Summer jobs reduce violence among disadvantaged youth. Science, 346(6214), 1219 to 1223.
Davis, J.M.V., and Heller, S.B. (2020). Rethinking the benefits of youth employment programs: the heterogeneous effects of summer jobs. Review of Economics and Statistics, 102(4), 664 to 677.
Tierney, J.P., Grossman, J.B., and Resch, N.L. (1995). Making a Difference: An Impact Study of Big Brothers/Big Sisters. Public/Private Ventures. 959 young people, randomized.
DuBois, D.L., Portillo, N., Rhodes, J.E., Silverthorn, N., and Valentine, J.C. (2011). How effective are mentoring programs for youth? A systematic assessment of the evidence. Psychological Science in the Public Interest, 12(2), 57 to 91.
Grossman, J.B., and Rhodes, J.E. (2002). The test of time: predictors and effects of duration in youth mentoring relationships. American Journal of Community Psychology, 30(2), 199 to 219.
Raposa, E.B., and others (2019). The effects of youth mentoring programs: a meta-analysis of outcome studies. Journal of Youth and Adolescence, 48. 70 studies, 25,286 young people.
Ganelin, D., and Dee, T. (2025). New Advanced Placement course designed to broaden access promotes participation and demographic diversity in computer science education. Proceedings of the National Academy of Sciences, 122(15).
Lin, S., Joshi, M., Caton, K., and Wilson, J.P. (2024). Girls Who Code Program Evaluation: Final Report. American Institutes for Research.
Oyserman, D., Bybee, D., and Terry, K. (2006). Possible selves and academic outcomes: how and when possible selves impel action. Journal of Personality and Social Psychology.
Lawner, E.K., Quinn, D.M., Camacho, G., Johnson, B.T., and Pan-Weisz, B. (2019). Ingroup role models and underrepresented students' performance and interest in STEM: a meta-analysis of lab and field studies. Social Psychology of Education, 22(5), 1169 to 1195.
Verniers, C., and others (2024). The double-edged sword of role models. Review of Research in Education.
Vandell, D.L., Reisner, E.R., and Pierce, K.M. (2007). Outcomes Linked to High-Quality Afterschool Programs. Study of Promising After-School Programs, 2,914 students across 35 programs. Not randomized.
Taheri, S.A., and Welsh, B.C. (2016). After-school programs for delinquency prevention: a systematic review and meta-analysis. Youth Violence and Juvenile Justice, 14(3), 272 to 290.
James-Burdumy, S., Dynarski, M., Moore, M., Deke, J., Mansfield, W., and Pistorino, C. (2005). When Schools Stay Open Late: The National Evaluation of the 21st Century Community Learning Centers Program. US Department of Education.
Sharkey, P. (2010). The acute effect of local homicides on children's cognitive performance. Proceedings of the National Academy of Sciences, 107(26), 11733 to 11738.
Sharkey, P., Schwartz, A.E., Ellen, I.G., and Lacoe, J. (2014). High stakes in the classroom, high stakes on the street. Sociological Science, 1, 199 to 220.
Laurito, A., Lacoe, J., Schwartz, A.E., Sharkey, P., and Ellen, I.G. (2019). School climate and the impact of neighborhood crime on test scores. RSF: The Russell Sage Foundation Journal of the Social Sciences, 5(2), 141 to 166.
Sharkey, P., Torrats-Espinosa, G., and Takyar, D. (2017). Community and the crime decline: the causal effect of local nonprofits on violent crime. American Sociological Review, 82(6), 1214 to 1240.
Snyder, H.N., and Sickmund, M. (2006). Juvenile Offenders and Victims: 2006 National Report. Office of Juvenile Justice and Delinquency Prevention, analyzing 2001 National Incident-Based Reporting System data.
Code.org Advocacy Coalition, Computer Science Teachers Association, and ECEP Alliance (2025). 2025 State of AI and Computer Science Education.
National Center for Education Statistics (2024). Digest of Education Statistics, table 702.40, from the Census Bureau Current Population Survey, November 2023.

C4I Veteran Services Corp is a workforce development and youth education organization. We are not a medical or mental health provider and we do not treat any condition. Nothing on this page is medical advice. If you are a veteran in crisis, dial 988 and press 1.

Program Administration

Youth Safety Documents

Youth safety comes first. Every adult who works with participants under 18 is screened, background checked by Checkr, Inc., and trained before any contact. No adult is ever alone with a minor. Every participant under 18 needs a signed parent or guardian consent form before the first session.

Youth Protection Policy

Screening, supervision rules, communication limits and reporting requirements.

Read the Policy

Adult Code of Conduct

Signed by every staff member, mentor, volunteer and partner before working with youth.

Download PDF

Parent/Guardian Consent Form

Required for every participant under 18. Print, sign and return to C4I staff.

Download PDF

Questions about youth safety, or to report a concern: Donnell Nichols, (202) 855-0703, dn22@c4iveteranservices.org. If a child is in immediate danger, call 911.

Put this in front of the young people who need it

A room, a power outlet and twenty teenagers is all it takes to run the first one. If you have that, or you can fund it, we will bring the rest.

Host a Session Fund a Class (919) 443-5998