Cannabis use appears twice as likely to predict violence, a Daily Telegraph headline claims, but the evidence behind that number is far weaker than it suggests.
Simon C Moore, Professor of Public Health Research, Co-Director of Crime and Security Research Institute and Director of the Violence Research Group, Cardiff University
The claim stems from a systematic review by King’s College London that pooled data from 63 studies covering more than 265,000 individuals, according to the researchers.
Researchers found a statistical association between cannabis consumption and violent behavior, yet the headline overstates the certainty of that link.
Confounders such as poverty, alcohol use, other drug exposure, and a troubled upbringing independently increase violence risk, and cannabis purchase often places users in high‑risk environments, according to the review.
The review attempted to control for these factors, but it preferentially selected studies that adjusted for the fewest confounders, limiting its ability to rule out alternative explanations.
Cross‑sectional studies, which measured cannabis use and violence at a single point in time, accounted for 46 of the 63 included investigations and produced odds ratios more than twice the baseline, the review reported.
Longitudinal studies, which follow participants over time, showed a 17% higher odds of violence among cannabis users, a weaker association that suggests temporal direction is uncertain.
Twice the odds does not equal twice the likelihood; odds compare the ratio of those with violence to those without, while likelihood expresses the proportion of the total population experiencing violence, the authors clarified.
When violence is rare, odds and likelihood diverge only slightly; when it is common, the difference widens dramatically, as illustrated by the example calculations in the review.
Systematic bias can also distort the findings: studies differ on how they define cannabis use and violence, grouping domestic incidents, sexual assault, fights and robbery together, each with distinct drivers, the review noted.
Results varied widely across the 63 studies, with odds ratios ranging from about 1.3 to nearly 10 in separate analyses of the same national survey, reflecting divergent definitions and populations, the authors said.
Publication bias may inflate the apparent link, since studies showing little or no association are less likely to be published, the review warned, and its statistical adjustments could not eliminate that concern.
The review assigned a very low certainty rating to the evidence, indicating limited confidence in the size of the estimated effect, a nuance absent from the Telegraph coverage, the authors added.
Overall, while cannabis and violence may co‑occur, the data do not support the claim that users are twice as likely to commit violent acts, the researchers concluded.
Support our Nation today
For the price of a cup of coffee a month you can help us create an independent, not‑for‑profit, national news service for the people of Wales, by the people of Wales.
Get more trusted Welsh news
Choose Nation.Cymru as a preferred source in Google News to see more of our journalism.
