The Weed Violence Study Has a Prohibition Problem

Filed Under: Correlation Is Not a Conviction

---

A new cannabis study has a number built to travel: 2.05.

That is the pooled odds ratio researchers found for violence among cannabis users in the general population. Put plainly, people who used cannabis showed about twice the odds of committing violence compared with people who did not.

The review is not some junk survey scraped together for clicks. Published in Psychological Medicine, it combined 63 peer-reviewed studies involving 265,079 people.

That is serious evidence.

It is also a lot messier than the headline number makes it look.

The biggest warning sign sits right next to the 2.05. The review’s I² statistic was 98.6 percent, meaning the study estimates varied enormously from one another.

So the number everybody will remember is an average built from research that often looked very different under the hood.

The picture gets more interesting when the studies are separated by design.

Cross-sectional studies produced an odds ratio of 2.37. Those studies looked at cannabis use and violence within roughly the same period.

Longitudinal studies, which followed people over time, came in at 1.17.

The association remained, but its size dropped hard.

That matters more than another round of statistical throat clearing. A cross-sectional study can tell you that cannabis use and violence showed up together. It is far weaker at telling you which one came first, what caused what, or whether something else was driving both.

Following people over time gives researchers a better shot at sorting that out.

The study authors argue that the remaining longitudinal association means reverse causation cannot explain everything. Violent people simply being more likely to use cannabis is not enough to make the whole finding disappear.

Fair enough, but 1.17 is not 2.37.

That is a very different story from the one carried by “cannabis users have twice the odds of violence.”

Professor Simon Moore, director of Cardiff University’s Violence Research Group, went straight at the problem in an independent response to the review.

“Without this information, the review cannot say whether cannabis itself or the environments in which it is used explain the violence.”

That is the fight inside this paper.

Not whether an association exists.

It does.

The fight is over what is inside that association.

Start with the phrase “cannabis use.”

Some studies counted lifetime use. Others looked at recent use. Some relied on self-reports. Others used biological evidence.

Most did not have enough information to cleanly separate someone who used cannabis occasionally from someone consuming it every day. Detailed information about dose and potency was usually missing too.

That left the researchers unable to build a meaningful dose-response analysis.

A person who smoked a few times last year is not having the same exposure as someone using high-potency THC every day. Yet much of the research feeding this review could not separate those people well enough to show whether heavier use produced more violence.

One number is being asked to represent people who used cannabis once, people who used it every day, self-reported fights, sexual violence, and criminal records.

That number is doing a hell of a lot of work.

Violence itself was measured all over the place. The studies included physical violence, sexual violence, intimate violence, nonintimate violence, self-reported behavior and official records.

The review had to pour all of that into a pooled estimate.

That may be useful for spotting a broad signal, but it is a long way from proving that THC itself is turning people violent.

The researchers also tried to account for confounding, and cannabis advocates should not pretend otherwise.

Studies were excluded when cannabis could not be separated from combined drug use. The authors also compared adjusted and unadjusted results.

The association barely moved.

Adjusted studies came in at 2.01. Unadjusted studies came in at 2.19.

So no, this is not a case where one obvious variable blows the whole study apart.

But adjustment only works on what was actually measured.

---

M O R E F R O M P O T C U L T U R E M A G A Z I N E

---

The authors acknowledge that cannabis users often used other substances and that many of the underlying studies did not fully account for concurrent drug use.

Alcohol remains one of the unresolved pieces.

In a study about violence, alcohol is not background noise.

Then there is victimization.

Cannabis users were also more likely to be victims of violence, with an overall odds ratio of 1.49.

When researchers looked only at longitudinal studies, the number dropped to 1.05.

That finding does not erase the possibility that cannabis contributes to aggression. It makes the story harder to reduce to THC turning people violent.

The authors themselves point toward risky situations and violent social environments as possible pieces of the puzzle.

Now prohibition walks into the room.

The review did not separate people buying cannabis legally from people obtaining it through illicit markets. It did not sort its violence estimates according to whether cannabis was legal, decriminalized, medically regulated, or prohibited where people lived.

That is not a small omission when much of the research comes from prohibition-era settings.

Moore specifically notes that illegal markets can bring cannabis users into contact with people and places where violence is more likely.

A person walking into a licensed dispensary and a person buying cannabis through an illicit market may be using the same drug while moving through very different environments.

That does not establish prohibition as the culprit, but it does establish prohibition as something the meta-analysis cannot isolate.

The conviction numbers make the problem even harder to ignore.

Among the general population, cannabis use was associated with 3.75 times the odds of violence when the outcome was measured through a criminal conviction.

That is the biggest general-population number in the paper.

It came from three studies.

Conviction records have value. People may lie about violent behavior, forget it, or refuse to report it. A conviction gives researchers an official record.

But a criminal conviction is not a blood test for aggression.

It is also a record of contact with police, prosecutors, and courts.

That makes conviction a different kind of outcome from someone privately admitting to violent behavior.

The review was not built to tell us whether cannabis criminalization changes a person’s chances of moving through that system.

So the 3.75 figure deserves attention.

It does not deserve to be treated like a laboratory reading of what THC does to the human brain.

The quality of the evidence is mixed too.

Only 13 of the 63 studies received a good quality rating. Forty-five were rated fair, and five were rated poor.

Removing the five poor studies barely changed the overall estimate, from 2.05 to 2.02.

That actually strengthens the review. The association is not being propped up by a handful of terrible studies.

The researchers also found evidence of publication bias or small-study effects and ran additional analyses to account for it.

The association stayed.

So there is no honest way to write this off as bad science.

There is also no honest way to pretend 2.05 settled the case.

The psychiatric findings deserve their own lane.

Among psychiatric patients, cannabis users had 2.49 times the odds of perpetrating violence compared with psychiatric patients who did not use cannabis.

Professor Marta Di Forti of King’s College London, one of the study authors, said in the university’s release:

“I’ve seen how heavy cannabis use can be a factor in violence.”

That is a legitimate clinical concern, especially when doctors are already assessing violence risk.

It is not evidence that the average cannabis user becomes dangerous.

King’s College London itself describes the finding as an association and says the review cannot prove cannabis causes violence.

Legalization gives us another real-world check.

A 2024 study in the Journal of Public Economics found that recreational legalization increased adult cannabis use and reduced marijuana arrests, yet researchers found little evidence of an increase in violent or property-crime arrests.

Another 2024 study found roughly a 20 percent increase in intimate-partner violence incidents following recreational legalization.

---

---

Anybody looking for a clean legalization victory lap or a clean cannabis panic story is going to leave disappointed.

The studies measured different things, and the results went in different directions.

That is exactly why one clean number cannot carry this whole argument.

The new meta-analysis found a real association between cannabis use and violence. Cannabis publications should not duck that because they dislike the implication.

But the causation case is nowhere near as clean as the headline number suggests.

The longitudinal effect is much smaller. The research often cannot distinguish occasional use from heavy use. Potency is poorly measured. Alcohol and other drugs remain incompletely controlled. Social environment is tangled inside the exposure.

And decades of prohibition are tangled in there too.

Prohibition is not an alibi for cannabis. It is an unmeasured suspect.

Until researchers can separate the drug from the world built around its use, “weed causes violence” is not a scientific conclusion. It is a slogan looking for evidence.

The study found a relationship worth investigating.

It did not convict the plant.

---

^{©2026 Pot Culture Magazine. All rights reserved. This content is the exclusive property of Pot Culture Magazine and may not be reproduced, distributed, or transmitted in any form or by any means without prior written permission from the publisher, except for brief quotations in critical reviews.}

---

F O R T H E C U L T U R E B Y T H E C U L T U R E