Cannabis and violence: what a review of 265,000 people found, and what the link does and does not mean

Cannabis has a reputation as the drug that makes people placid. A large review from King’s College London, publicized at the end of September, cuts against that. It found that people who use cannabis are about twice as likely as those who do not to have committed a violent act, and more likely to have been on the receiving end of one.

The headline figure is real, and it needs unpacking. It comes mostly from studies that measured cannabis use and violence at the same moment. In studies that tracked people forward in time, the link was much weaker. And the review cannot separate cannabis from the other things that tend to come with heavy use.

This explainer goes through what was found and what it can and cannot support.

What the review of cannabis and violence found

The researchers gathered every study they could find and combined the results, a method called meta-analysis. They synthesized evidence from 63 peer-reviewed studies (N = 265,079), aiming to clarify the link between cannabis use and violence, either as a perpetrator or as a victim. N is the number of people across the studies. The review appeared in the journal Psychological Medicine.

Violence was defined widely. The researchers used the World Health Organization’s broad definition of violence, which includes physical, sexual and psychological violence toward another person.

Their reason for asking was partly clinical. Cannabis use is a known risk factor for adverse mental health outcomes, including psychosis, which is frequently associated with aggression. Psychosis is a state in which a person loses touch with reality, with delusions or hallucinations. Yet, as the authors put it, the specific impact of cannabis use on the risk of violence remains unclear.

Results are given as odds. Odds of 2 mean that violence was about twice as common among cannabis users as among non-users in the studies pooled.

Group or outcome	Odds among cannabis users, compared with non-users	Plausible range
Committing violence, general population	2.05 times	1.74 to 2.41
Committing violence, psychiatric patients	2.49 times	1.72 to 3.61
Violence with a criminal conviction, general population	3.75 times	2.54 to 5.53
Violence with a criminal conviction, psychiatric patients	4.21 times	2.86 to 6.22
Being a victim of violence	1.49 times	1.38 to 1.60

This association was observed in both psychiatric patients and the general population. There was an especially large effect for violence involving a criminal conviction.

Senior author Robin Murray summarized it. “In the general population, people who used cannabis were around twice as likely to have committed a violent act as those who didn’t; for those in psychiatric care, cannabis users were around two and a half times as likely to be violent as other psychiatric patients who didn’t use cannabis,” Murray said.

Why the cannabis link depends on the kind of study

One comparison inside the review matters more than any other, and it is easy to miss.

Most of the pooled studies were surveys that asked people, at a single point, about their cannabis use and their history of violence. Those found odds of 2.37. A smaller set of studies were longitudinal, meaning they recorded cannabis use first and then followed people to see who went on to be violent. Those found odds of 1.17.

So the odds were 137% higher in the snapshots and 17% higher in the follow-up studies. Murray acknowledged the gap. “The link held in studies that followed people over time, though it was smaller than in studies that looked at a single point in time,” Murray said.

The difference matters because a snapshot cannot show which came first. A person with a violent history may be more likely to take up cannabis. Something else, such as a chaotic upbringing, may lead to both. Following people over time addresses the first problem, though not necessarily the second, and when that was done, most of the association went away.

Does cannabis itself make people more aggressive?

One way to find out is to give people the drug and watch. A team at Maastricht University in the Netherlands did so in 2016.

They started from an open question. Alcohol consumption is known to facilitate aggression, they wrote, whereas a causal link between cannabis and aggression has not been clearly demonstrated. They recruited heavy alcohol (n = 20) and regular cannabis users (n = 21), and controls (n = 20). That is 20 heavy drinkers, 21 regular cannabis users and 20 people who were neither. Each user was tested once after their drug and once after a placebo, a dummy version, and all were put through tasks designed to provoke them.

The drugs pulled in opposite directions. Alcohol intoxication increased subjective aggression whereas cannabis decreased the subjective aggression following aggression exposure. Subjective aggression means how aggressive the volunteers said they felt.

That is a small experiment in experienced users, measuring feelings over a few hours in a laboratory. It says nothing about withdrawal, about years of heavy use, or about cannabis-related psychosis. But it does not support the simple idea that being stoned makes a person violent.

Cannabis users as victims of violence

The review also turned the question around. Cannabis users were more likely to have been attacked or abused, with odds about 1.5 times those of non-users. The risk was more pronounced in females and mixed-gender cohorts compared to males, cohorts being the groups of people studied.

This half of the result fits awkwardly with the idea that cannabis drives violence chemically. It fits more easily with the idea that people who use cannabis heavily are more often in settings where violence happens, as perpetrators and as victims alike. Heavy alcohol use, other drug use and difficult social circumstances would all do that.

Where this fits: earlier studies of cannabis and violence

The new review confirms an earlier one from an independent group. In 2020 researchers in Montreal pooled studies of people under 30. After screening the literature, 30 study arms were included, yielding a total of 296,815 adolescents and young adults. Their overall figure was nearly identical, odds of 2.11.

That analysis looked at dose. The risk of violence was higher for persistent heavy users than for people who had used in the past year or at some time in their lives. It also showed the effect of accounting for other influences: the odds were 2.62 in studies that made no adjustments and 2.01 in those that adjusted for factors such as income and other drug use. The authors’ summary was that these results demonstrate a moderate association between cannabis use and physical violence.

A Norwegian study took a stricter approach. Its authors began by noting that it is not clear to what extent the association is causal. They followed 2,681 young people and compared each one with themselves over time, asking whether a person’s violence rose in the years when their own cannabis use rose. Comparing people with themselves means that fixed traits such as temperament and family background cannot distort the result. In the paper’s words, the effects of time-invariant confounders were hence eliminated.

A link remained, and it was small. A 10% increase in cannabis use frequency is associated with a 0.4% increase in frequency of violence, they reported.

The longest follow-up is British. Young males from the Cambridge Study of Delinquent Development (n = 411) were followed up between the ages of 8 and 56 years. Those who kept using cannabis through adulthood stood out. Continued exposure to cannabis (use at age 18, 32 and 48 years) was associated with a higher risk of subsequent violent offending, the researchers found. They also found the arrow running both ways: these analyses uncovered a bi-directional relationship between cannabis use and violence. Violent young men were more likely to become cannabis users, as well as the reverse.

Taken together, the studies agree that a link exists and that it survives some adjustment. They disagree about its size. Comparisons of users with non-users put the odds at about double. The Norwegian study, which compared people with themselves over time, found that a 10% rise in use went with a 0.4% rise in violence.

Limits of the cannabis and violence research

It cannot show cause. The university’s own summary says so: the review identifies an association and does not prove that cannabis directly causes violence. Cannabis use often occurs alongside other factors that can also increase the risk of violent behavior or victimization, including alcohol use, other drug use and difficult social circumstances.

Dose is mostly missing. Only a small number of the studies measured how much cannabis people used or how often they used it. A person who smoked twice at college and a person who smokes all day are, in many of these studies, both simply users.

The conviction figures are the largest and the least secure. According to the university’s summary, they were based on a relatively limited number of studies. Convictions also depend on who gets arrested and charged, which is not the same as who is violent.

The cannabis has changed. Most of the research was conducted before high potency cannabis became common. The authors suspect that daily use of stronger products could carry greater risk. That is a hypothesis the pooled studies cannot test.

Much of the violence was self-reported. The review separated what people said about their own behavior from criminal convictions, and self-report depends on what people recall and are willing to say.

The authors were conscious of how results like these can be misused. “We’ve been cautious for a long time about how far we could go in describing this link, but a review of this size lets us speak with more confidence about the risks, while being careful not to stigmatize people who use cannabis more broadly,” said co-author Marta Di Forti, a psychiatrist at King’s College London.

Their practical conclusion is aimed at clinics. The review highlights the potential of cannabis use as a target for interventions to decrease violence, and calls for psychiatric services to ask routinely about it.

Across 63 studies, people who use cannabis were about twice as likely to have committed violence and half again as likely to have suffered it, though the odds were only 17% higher when people were followed over time, and the studies cannot separate the link from alcohol, other drugs and the circumstances of heavy users.