A review of 265,000 people ties cannabis use to twice the odds of violence

Pool 63 studies and more than 265,000 participants, and cannabis users come out about twice as likely as non-users to commit violent acts in the general population. King’s College London researchers reached that figure in a systematic review and meta-analysis in Psychological Medicine, and they put the word association in front of every version of it.

“As cannabis becomes more commercialized and legalized around the world, the conversation about its risks has focused mostly on the person using it,” said Professor Sir Robin Murray of the Institute of Psychiatry, Psychology & Neuroscience, who co-led the work with Dr. Marta Di Forti. Violence, committing it and suffering it, belongs in that conversation too.

The general-population doubling and its steeper subgroups

The doubling is the broadest ratio. In the university’s release, general-population cannabis users were roughly twice as likely to commit violent acts, while cannabis-using psychiatric patients were 2.5 times as likely to be violent as patients who did not use the drug. The comparison group matters: each ratio sets users against non-users within the same kind of population, not against the public at large.

Criminal convictions produced the largest multiples. Cannabis users were three to four times as likely to be convicted of a violence-related offense, though the Sense About Science write-up of the paper flags that this estimate rests on a limited pool of studies. Convictions measure what the justice system recorded, which can differ from what people did, so that figure carries more uncertainty than the two ratios above it.

The review’s reach is what separates it from the single-country studies that came before. Pooling 63 papers lets the team compare cross-sectional surveys, cohort studies and registry-based work in one framework, and it is published as volume 56 of Psychological Medicine, the Cambridge University Press journal. Because individual studies measured violence in different ways, with conviction records at one end of the range, the pooled ratios blend several yardsticks, and no single figure describes a typical user’s risk.

Violence done to users

Users were also about 1.5 times as likely to be victims of violence, stronger in women, per the EurekAlert version of the release. The authors treat victimization as part of the same finding.

Di Forti said the size of the evidence base changed how far the team was prepared to go. After a long period of caution about how strongly the link could be described, she said, a review of this size allows more confident statements about the risks. On that footing, the team recommends that mental health services routinely ask about cannabis use when assessing violence risk, and that public health information on cannabis mention the association.

Confounding, study design and potency

The authors stress that the data cannot show cannabis causes violence. Use tends to travel with alcohol, other drugs and difficult social circumstances, each of which raises violence risk independently, and the review has no way to peel those influences apart. A person who uses cannabis and is violent may be responding to the same underlying stress, or violent behavior may precede the drug use rather than follow it.

Two features of the evidence pull in opposing directions, and both are spelled out by the authors. According to the Sense About Science summary, the link held in studies that followed people over time but was smaller than in studies that looked at a single moment, which suggests the snapshot designs overstate the relationship. In the other direction, few of the included studies recorded how often or how much people used, and most predate today’s high-potency products, so the numbers may understate what modern daily use looks like. The team lists this as a limit of the evidence base, not as a result: few included studies tested today’s concentrated products, and the review therefore cannot say whether the real-world association has grown, shrunk or stayed put since the older data were gathered.

Outside the London group, federal guidance is similarly cautious. America’s National Institute on Drug Abuse notes that many factors besides the drug, such as genes, trauma and stress, influence both mental health and drug use, and says prospective, longitudinal research is needed to learn whether, to what extent and for whom cannabis contributes to poor outcomes. That is the same gap the London review leaves open.

The ScienceDaily summary credits the work to the National Institute for Health and Care Research Biomedical Research Centre and the Medical Research Council (grant MR/T007818/1). For readers weighing the number, the practical reading is narrow: a pooled association of about twofold in general-population users, rising to 2.5-fold in psychiatric settings, drawn largely from studies that cannot say which came first. Murray and Di Forti, working from the Institute of Psychiatry, Psychology & Neuroscience, frame the result as grounds for asking clinical questions, not for drawing causal conclusions.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview