It feels like everyone your age is betting. Before we look at any data — go with your gut.
You guessed first, so this is about your number — not a stranger's.
Here's the honest part: no one has a clean national count of how many under-18s bet on sports. It's underage and hidden, so it's hard to measure — and that gap is part of how "everyone's doing it" survives. But every careful signal points the same way: well below the loud number. Adults looking back report 23% ever placed a sports bet before age 21 — and parents almost never see it (only 2% think their own teen bets online).
When researchers actually surveyed high-schoolers, about 1 in 4 (25%) had gambled in any form in the past year. Not "everyone." About a quarter — and that's counting cards and scratch-offs, not just sports betting.
This isn't a one-off. Across study after study, the number people picture runs far higher than the number people actually report — and here's the part that matters: the bigger someone's overestimate, the more likely they are to bet themselves. People quietly calibrate to the crowd they think they see.
Your number wasn't random. Four things inflated it — and you've met some of them before.
You see betting on every broadcast, stream, and feed (that was 3.1). Constant exposure makes it feel universal.
People screenshot the bet that hit. Nobody posts the ten that didn't. The visible slice is rigged toward "everyone's winning."
A few heavy bettors are very visible. The majority who don't bet are invisible — silence doesn't trend.
Sportsbooks win when "everyone's in" feels true. The crowd in your head is partly something that was sold to you.
Normative misperception is believing a behavior is far more common than it actually is. With gambling, almost everyone overestimates how many of their peers bet. And it's not harmless: the bigger your overestimate, the more likely you are to bet yourself — because people quietly adjust to the crowd they think they're in. Correct the picture, and the pull weakens.
Here's the line to hold onto. "Everyone's doing it" is not a fact you observed — it's a number that was sold to you. The real crowd is smaller and quieter than the one in your head.
When did you last get the feeling that "everyone" is doing something — betting, or anything else? Where did that feeling actually come from?
National Council on Problem Gambling / The Harris Poll (March 2026). National Survey Finds Widespread Gambling Participation Before Age 21. ncpgambling.org
Neighbors, C., et al. (2015). Efficacy of personalized normative feedback as a brief intervention for college student gambling: A randomized controlled trial. Journal of Consulting and Clinical Psychology, 83(3), 500–511. pubmed.ncbi.nlm.nih.gov
On the "~25% any past-year gambling" anchor: Connecticut adolescent Youth Risk Behavior Survey — 25.4% of high-school students reported past-year gambling; ~2.1% met a problem-gambling screen (SOGS-RA). State sample; the most defensible recent U.S. high-school measurement. ncbi.nlm.nih.gov
On "23% ever bet on sports before 21 / 2% of parents": NCPG–Harris Poll (Feb 2026, n=2,072 adults) reports retrospective adult recall, not a direct youth count; C.S. Mott National Poll on Children's Health (Jan 2024, parents of teens 14–18) found only 2% of parents think their teen has used an online betting platform. mottpoll.org
On the perception gap & PNF mechanism: Neighbors et al. (2015) RCT — perceived peer gambling exceeded actual; personalized normative feedback lowered perceived norms and reduced gambling frequency/amount. Larimer et al. (2012); Berkowitz (2004), social-norms approach.
Evidence caveat (important): Every gambling PNF randomized trial used samples aged 18+. The corrective-norm effect is unvalidated under 18, and clean U.S. youth sports-betting prevalence data is genuinely thin. The robust, replicated finding is the direction of the gap (perceived > actual). Exact youth figures are presented as estimates, not settled fact — which is itself the point of this unit.
Note for the build team: keep the headline claim at "people overestimate peer gambling, and the overestimate predicts their own betting" — well supported. Do not assert a precise national under-18 sports-betting rate; none exists cleanly. Date-stamp the NCPG and prevalence figures and re-verify on update. In the integrated build, swap the national "~25%" anchor for the live/prior-cohort aggregate where available.