Legalization & Mental Health — A State-Level Analysis

WHAT THE PHONE STARTED

If cannabis legalization were driving America's mental health crisis, the data would show it. It doesn't. Here's what the state-level evidence actually reveals — and what they're not telling you about the 2018 Farm Bill.
March 2026 · Cannabis Policy · mostlycbd.com
01

THE QUESTION EVERYONE'S ASKING WRONG

The national conversation frames it simply: cannabis legalization leads to more mental health problems. Headlines point to rising rates of cannabis use disorder, psychosis-related ER visits, and adolescent depression — and then draw a line to the map of states that have legalized.

That framing skips three questions that matter more than the headline.

First: did these mental health trends start before legalization expanded? Second: do they appear equally in states that never legalized? And third — the one almost nobody asks — are we even talking about the same product?

The 2018 Farm Bill split the cannabis conversation in two. Before it, cannabis meant cannabis — regulated or not, medical or recreational, but a known plant. After it, a legal loophole created an entirely separate market of synthetic, hemp-derived intoxicants with zero testing, zero age verification, and zero clinical oversight. And the data on "cannabis-related" harms after 2018 lumps both together.

This deep dive separates what the data actually shows from what the headlines claim. State by state. Timeline intact. Product distinction preserved.

02

THE MENTAL HEALTH CRISIS STARTED FIRST

America's youth mental health crisis is real. But it didn't start in a dispensary. It started on a phone.

33%
Increase in adolescent
depressive symptoms
2010–2015
65%
Increase in suicide rate
for teen girls
2010–2015
52%
Increase in adolescent
major depression
2005–2017

These increases began in the early 2010s — years before most recreational cannabis laws took effect. Colorado and Washington legalized in 2012, but retail sales didn't start until 2014. Most other states legalized between 2016 and 2022. The mental health trajectory was already set.

What did happen in 2010–2012? Smartphone adoption crossed 50%. Social media daily use among teens climbed from 50% to 85% between 2009 and 2017. In-person teen socializing — a consistent behavior for three decades — fell off sharply starting in 2010.

"Every indicator of mental health and psychological well-being has become more negative among teens and young adults since 2012." — Jean Twenge, San Diego State University, analyzing nationally representative data

The U.S. Surgeon General issued a 2023 advisory naming social media as a threat to youth mental health. No equivalent advisory exists naming cannabis legalization. The timeline makes clear why: the mental health crisis preceded legalization expansion. It accelerated during years when teen cannabis use was actually declining.

03

WHAT THE ALGORITHM OPTIMIZED FOR

The mental health crisis didn't just coincide with smartphone adoption. The mechanism is specific and worth naming — because once you understand it, blaming the dispensary becomes genuinely difficult to do with a straight face.

The comparison trap went infinite. Before 2012, a teenager's social reference group was roughly 30 classmates. After 2012, it became thousands of algorithmically curated profiles — selected not for accuracy, but for engagement. Filters, highlight reels, and influencer aesthetics set a standard nobody in real life clears. Social comparison is normal adolescent development. Social comparison against an infinite, curated, engagement-optimized feed is something structurally different.

Social acceptance got scored publicly. Adolescent brains treat belonging as a survival-level signal — rejection and exclusion activate the same neural circuits as physical pain in developing brains. Likes, follower counts, and comment sections turned that ancient, already high-stakes signal into a live public score, updating in real time, visible to everyone at school. The stakes didn't change. The exposure did.

31%
Higher depression risk
in smartphone owners
at age 12
62%
Higher insufficient
sleep risk in
smartphone owners at 12
57%
Greater odds of clinical
psychopathology after
acquiring a phone at 13

The algorithm has no wellbeing incentive. Engagement optimization is neutral — it doesn't care whether a user is happy or miserable, only whether they stay on platform. A sad, anxious teen scrolls longer than a satisfied one. The platform learns this. Content that generates envy, distress, and insecurity performs well by the only metric that matters. Nobody designed it to hurt teenagers. It just optimized for what it measured.

And sleep became a direct casualty. Phones in bedrooms, pull notifications, infinite scroll at 1am — the same study that tracked depression risk found that smartphone ownership at 12 carries a 62% higher risk of insufficient sleep. Among kids who had no phone at 12 but acquired one by 13, the odds of clinical-level psychopathology in the following year rose by 57%. Not kids with prior mental health histories — new cases, in the year after the phone arrived. Sleep deprivation alone is one of the strongest standalone predictors of depression, anxiety, and suicidal ideation. The mechanism is physiological: less sleep means less emotional regulation, full stop.

The platform didn't know it was hurting teenagers. It just knew that hurt teenagers scroll longer.
04

TEEN CANNABIS USE WENT DOWN, NOT UP

If legalization drove youth mental health problems, you'd expect teen cannabis use to rise in legalization states. Across every major national survey — NSDUH, YRBS, Monitoring the Future — it didn't.

↓ 19%
Decline in teen current
cannabis use nationally
2014–2023 (NSDUH)
↓ 35%
Decline in teen use
in Colorado & Washington
since legalization
↓ 18%
Decline in teens
who ever tried cannabis
2014–2023 (NSDUH)

Colorado — the first state with retail cannabis sales — saw current use among high schoolers drop from 22% in 2011 to 12.8% in 2023. A 42% decline. Washington shows similar patterns across all surveyed grades.

The data from before and after the 2018 Farm Bill tells the same story. The Harvey dissertation (2020), analyzing NSDUH data from 2008 to 2017, found that adolescent marijuana use rates remained at 14% in both the first and last year of the study — flat — even as legalization expanded dramatically during that period.

A 2024 JAMA research letter confirmed: neither legalization nor the opening of retail stores led to increases in youth cannabis use. Multiple independent analyses from CDC, NSDUH, and Monitoring the Future converge on the same conclusion.

"We can state with some confidence that, even in states that have enacted marijuana liberalization policies, marijuana use among adolescents is not currently increasing. In fact, there is rather compelling evidence that adolescent marijuana use has steadily declined."

Regulated markets removed the "forbidden fruit" appeal. Licensed dispensaries verify age at rates the illegal market never could. In 2012, before any state had regulated sales, 40% of high schoolers reported knowing a student who sells marijuana at school. That pipeline operated without a single compliance check. Dispensaries face them routinely.

05

THE 2018 FARM BILL: THE LOOPHOLE NOBODY DESIGNED

On December 20, 2018, the Agriculture Improvement Act — the Farm Bill — redefined hemp as cannabis containing less than 0.3% delta-9 THC by dry weight. The intent was to support hemp agriculture. The effect was to open a legal doorway for an entirely new category of intoxicating products that nobody anticipated.

Delta-8 THC, delta-10 THC, THC-O, HHC, THC-JD, THC-B — all synthesized from hemp-derived CBD using chemical conversion processes. Psychoactive. Intoxicating. And, under the letter of the law, legal. No testing requirements. No age verification. No dosing standards. No packaging restrictions.

Pre-2018
Only 15 total delta-8 adverse event cases in the entire FDA FAERS database since tracking began. Cannabis products meant cannabis — plant material or regulated state dispensary products.
Dec 2018
Farm Bill passes. Hemp and its derivatives legalized federally. The 0.3% delta-9 threshold creates an unintended gateway for synthetic cannabinoids.
2020–2021
Delta-8 products flood gas stations, vape shops, and convenience stores. Products packaged as candy, marketed without warnings. FDA receives 104 adverse event reports by Feb 2022. 92% of all FAERS delta-8 cases occurred after the Farm Bill.
2021
Poison control centers log 660 delta-8 exposure cases in the first seven months alone — up from essentially zero. 41% involve children under 18. 18% require hospitalization, including pediatric ICU admissions.
2022
Delta-8 poison control calls spike 82% over 2021, reaching 3,358 exposure cases. A third involve children under six. Products sold as "Cannaburst" in Starburst-lookalike packaging. "Death by Gummies" brand sells 100mg doses to 14-year-olds.
2021–2023
Over 8,000 delta-8-related calls to poison control centers. New compounds — HHC, THC-JD, THC-B — emerge. By 2023, delta-8 accounts for 80% of hemp-product adverse events, with even newer synthetics making up the rest.
Late 2025
Congress amends the federal hemp definition to "total THC" — closing the delta-9-only loophole. But the new rule doesn't take effect until November 2026. In the meantime, the unregulated market continues.

Here's the critical distinction: none of this came from regulated cannabis dispensaries. This came from the unregulated hemp loophole. Gas stations. Vape shops. Convenience stores. Online retailers with no age gates. Products designed to look like candy, sold to anyone with cash.

But when researchers measure "cannabis-related" emergency department visits or "cannabis poisoning" trends, these hemp-derived products get folded into the same data. A 14-year-old hospitalized after eating synthetic delta-8 gummies from a gas station becomes a data point in a study about "cannabis legalization and mental health." The product wasn't cannabis. It wasn't from a dispensary. It was a direct consequence of the regulatory failure the Farm Bill loophole created.

Any study analyzing cannabis-related harms after 2018 that doesn't control for the Farm Bill loophole market is measuring two completely different things and calling them one.

06

WHAT THE STATE-LEVEL RESEARCH ACTUALLY SHOWS

The largest study to examine cannabis legalization and psychosis across all 50 states — using commercial and Medicare claims data from 2003 to 2017 (entirely pre-Farm Bill) — found no association between state cannabis legalization policies and psychosis-related healthcare utilization.

That study, published in JAMA Network Open in 2023, used state-level panel fixed effects to model within-state changes. Medical cannabis laws, recreational cannabis laws, and commercial sales — none showed a statistically significant association with increased psychosis-related diagnoses or antipsychotic prescriptions.

Other findings from the state-level evidence base:

What the data supports

A 2024 JAMA Psychiatry study (Jayawardhana et al.) found medical cannabis laws associated with a 42.7% increase in cannabis use disorder diagnoses and 88.6% increase in cannabis poisoning diagnoses from 2011 to 2021. Recreational laws were associated with a 31.6% increase in cannabis poisoning. These are real findings that deserve honest engagement. But the study period extends well past the 2018 Farm Bill, and the data does not distinguish between dispensary cannabis and gas-station synthetics. It also cannot determine whether increased diagnoses reflect increased use, increased harm, or — critically — increased detection and reporting as cannabis became legal and healthcare providers began screening for it.

What the data doesn't support

The claim that legalization causes psychosis at the population level. The 2003–2017 nationwide analysis found no such association. The claim that legalization increases teen use — every major national survey contradicts it. The claim that legalization worsens population-level mental health — the mental health crisis accelerated in all states, regardless of cannabis policy status.

What the data complicates

A 2025 study examining policy bundles found that both "pharmaceutical" (medically-oriented) and "permissive" cannabis policy designs were associated with mental health improvements among youth. Only "fiscalization" — tax-heavy, revenue-focused policy design — had a negative impact on mental health. The policy details matter more than the legalization status itself.

Regulated Market (dispensary)

Age verification at point of sale. Product testing for potency, contaminants, pesticides. Labeled dosing information. Trained staff. State compliance checks. Tax revenue funding education and treatment.

Colorado and Washington: teen use down 35%+ since legalization.

Unregulated Market (Farm Bill loophole)

No age verification. No product testing. No dosing standards. Candy-like packaging targeting youth. Synthetic conversion from CBD with unknown byproducts. Gas station and convenience store distribution.

8,000+ poison control calls (2021–2023). 41% involving children.

07

THE INVERTED RELATIONSHIP

Mendelian randomization studies — which use genetic markers to trace direction of cause and effect — have found that a predisposition for schizophrenia predicts cannabis use more strongly than cannabis use predicts schizophrenia. The vulnerability came first. The plant came second.

This doesn't mean cannabis is harmless. High-potency, unbalanced products carry real risks — especially for people already genetically vulnerable. But it reframes the question. Instead of asking "does cannabis cause mental illness?", the evidence asks: "why are people with mental health vulnerabilities turning to cannabis instead of getting clinical help?"

67
Median days to see
a psychiatrist in the U.S.
60%
Of Americans locked out
of mental health care
by cost
9%
Of U.S. medical schools
that include cannabis
education

The dispensary door was open when the clinic door was closed. That's not a cannabis problem. That's a healthcare crisis wearing a cannabis mask.

Meanwhile, the data on what happens when people do gain regulated cannabis access tells a different story. The Chen, Maclean & French (2026) study — covering all 50 states from 2010 to 2022 — found that after recreational cannabis legalization, parents reported 8.9% fewer days of bad mental health, 6.8% fewer days of bad physical health, and 11.3% less likelihood of reporting fair or poor health. Substantiated child physical abuse reports declined 26.7%. Child injury-related deaths declined 4.1%.

Anderson, Rees & Sabin (2014) found decreased suicides among men aged 20–39 in states with medical marijuana laws. Denson & Earleywine (2006) found that weekly-or-less cannabis users had less depressed mood, more positive affect, and fewer somatic complaints than non-users.

The picture isn't one-dimensional. It never was. But the honest picture is: regulated access to balanced cannabis products, within a functioning healthcare infrastructure, produces better outcomes than prohibition, which produces underground markets, uneducated providers, and patients with nowhere else to turn.

08

WHERE WE GO FROM HERE

The mental health crisis is real. It deserves real solutions — better access to psychiatric care, honest research into every substance people use to cope, and policies built on what the evidence actually says.

What it doesn't deserve is a scapegoat.

The data doesn't show that legalization broke mental health. It shows that mental health was already breaking — and that unregulated markets made it worse while regulated ones didn't.

That distinction matters for everything that comes next. The problem isn't that cannabis exists in a legal market. The problem is the absence of good regulation, the persistence of the research blockade, and a healthcare system stretched too thin to meet the demand.

The Farm Bill loophole is what no regulation looks like in practice — 1,000mg THC gummies on gas station shelves, no age verification, no dosing standards, no accountability. Prohibition showed us the other failure mode: black markets, mass incarceration, and zero clinical infrastructure. Neither extreme protected anyone.

The middle path is regulation built on evidence. Here's what that looks like in practice:

RFER

Reschedule · Fund Trials · Educate · Regulate

RReschedule
Reschedule cannabis federally. End the Schedule I classification that blocks the very research needed to answer these questions. You cannot demand clinical evidence while simultaneously making it illegal to produce.
FFund Trials
Fund real clinical trials. Not observational studies that can't determine directionality — randomized controlled trials that can. If the goal is to protect youth, build the science that actually tells us how.
EEducate
Train the people doing the advising. Cannabis education is absent from 91% of medical school curricula. Patients are making consequential decisions without clinical guidance — not because they don't want it, but because their providers were never taught.
RRegulate
Regulate by evidence, not fear. Protect functional patient access for the millions using cannabis therapeutically — 63% report substituting it for prescription drugs. Close the Farm Bill loophole completely: age verification, product testing, dosing standards, packaging restrictions. The regulated market already does this. The unregulated one does none of it.

Sources & References
  1. Twenge et al. (2018). "Increases in Depressive Symptoms, Suicide-Related Outcomes, and Suicide Rates Among U.S. Adolescents After 2010." Clinical Psychological Science 6(1).
  2. U.S. Surgeon General's Advisory (2023). "Social Media and Youth Mental Health."
  3. Pew Research Center (2022–2025). Teen social media surveys.
  4. SAMHSA/NSDUH (2024). 2023 National Survey on Drug Use and Health. Federal data on teen cannabis use 2014–2023.
  5. Colorado Healthy Kids Survey (2011–2023). State teen substance use data.
  6. Marijuana Policy Project (2025). Analysis of CDC/YRBS/MTF data across legalization states.
  7. Harvey (2020). Dissertation analyzing NSDUH 2008–2017 cohorts.
  8. FDA Consumer Update (2022). "5 Things to Know about Delta-8 Tetrahydrocannabinol." 104 adverse event reports, 2,362 poison control cases by Feb 2022.
  9. Lo et al. (2023). "Delta-8, a Cannabis-Derived THC Isomer: Evaluating FAERS Database." PMC. 92% of cases post-Farm Bill.
  10. CNN (2024). Delta-8 poison control calls up 82% from 2021 to 2022.
  11. Fuentes (2026). Tampa Bay Times. Florida enforcement operations; 155,000+ illegal hemp packages removed.
  12. Elser et al. (2023). "State Cannabis Legalization and Psychosis-Related Health Care Utilization." JAMA Network Open. 2003–2017 data, all 50 states. No association found.
  13. Jayawardhana et al. (2024). "Association of State Cannabis Legalization With Cannabis Use Disorder and Cannabis Poisoning." JAMA Psychiatry. 2011–2021 data.
  14. Mallinson & Hannah (2025). "Adding Nuance: Policy Bundles and Youth Mental Health." Substance Use & Misuse.
  15. Mendelian randomization data: Multiple studies on genetic liability for schizophrenia and cannabis use (referenced in Harvey 2020).
  16. Chen, Maclean & French (2026). "Legalizing Cannabis: Implications for Child Maltreatment." NBER Working Paper No. 34673.
  17. Anderson, Rees & Sabin (2014). Decreased suicides in medical marijuana law states.
  18. Denson & Earleywine (2006). Cannabis users and depressed mood.
  19. "Smartphone Ownership, Age of Smartphone Acquisition, and Adolescent Health Outcomes." Pediatrics 157(1), 2025. Adolescent Brain Cognitive Development Study, n=10,588. Smartphone ownership at age 12 associated with OR 1.31 for depression, OR 1.62 for insufficient sleep; acquisition by age 13 associated with OR 1.57 for clinical-level psychopathology in the following year. → publications.aap.org