The campaign worked

A national counter-marketing campaign accounted for roughly 22% of the decline in US youth smoking between 1999 and 2002 — measured against ~50,000 students a year.

Farrelly MC, Davis KC, Haviland ML, Messeri P, Healton CG (2005). American Journal of Public Health 95(3):425–431. PMC1449196
Followed longitudinally rather than across a trend line: a cohort of 8,904 adolescents, tracked annually for seven years. Exposure lowered the risk of ever starting.

Farrelly MC, Nonnemaker J, Davis KC, Hussin A (2009). American Journal of Preventive Medicine 36(5):379–384. DOI 10.1016/j.amepre.2009.01.019 Read
$324M to build, run, evaluate and litigate the campaign. Just under $1.9B in medical costs averted for society — possibly as much as $5.4B.

Holtgrave DR, Wunderink KA, Vallone DM, Healton CG (2009). American Journal of Preventive Medicine 36(5):385–388. PMID 19211214 PubMed
Head to head against the industry’s own youth campaign. Counter-marketing moved attitudes. “Think. Don’t Smoke.” did not.

Farrelly MC, Healton CG, Davis KC, Messeri P, Hersey JC, Haviland ML (2002). American Journal of Public Health 92(6):901–907. PMC1447480
The state pilot that came first. Florida earmarked its own settlement, ran it for two years, and measured the effect on initiation — which is why the national fund had a working programme to copy.

Sly DF, Hopkins RS, Trapido E, Ray S (2001). American Journal of Public Health 91(2):233–238.
It replicated. A federal campaign averted an estimated 175,941 young people from becoming established smokers, at $1,337 per quality-adjusted life year and $128 returned per $1 spent.

Cost-Effectiveness Analysis of The Real Cost Campaign’s Effect on Smoking Prevention (2018). American Journal of Preventive Medicine. Read
The same maths run against a fragmented, digital-first media environment rather than broadcast television.

Hair EC, Holtgrave DR, Romberg AR, et al. (2019). IJERPH 16(22):4312. DOI 10.3390/ijerph16224312 DOI
The canonical review, across tobacco, alcohol, drugs and road safety: campaigns can change behaviour at population scale while competing against pervasive product marketing and behaviour driven by addiction or habit.

Wakefield MA, Loken B, Hornik RC (2010). The Lancet 376(9748):1261–1271. PMID 20933263 PubMed

What the campaign should be about

Why it worked when lecturing didn’t. Campaigns that turned adolescent autonomy against the manipulation outperformed campaigns about health consequences.

Farrelly MC, Niederdeppe J, Yarsevich J (2003). Tobacco Control 12(suppl 1):i35–i47.
Nearly 4,300 children tracked from age eight. Addictive use patterns predicted worse mental health and suicidal behaviour. Total screen time did not. Around 40% showed high or rising addictive use.

Xiao Y, Mann JJ, et al. (2025). JAMA, 18 June 2025. Weill Cornell · Columbia · UC Berkeley. Summary

Two things worth saying out loud. Every efficacy paper here is about tobacco. Nobody has yet run a controlled counter-marketing campaign against engineered compulsion and measured it — which is the gap, and the reason a proof market and a matched control market sit inside the ask rather than beside it. And the last paper shows association, not causation. Its authors say so. The narrower finding is the one that counts.