Longitudinal Data Show Only 17% Higher Odds of Violence Among Cannabis Users After Confounder Adjustment

The review pooled over 265,000 participants yet prioritized minimally adjusted estimates, leaving poverty, alcohol co-use, and illegal-market exposure unaddressed. Cross-sectional studies (46 of 63) captured simultaneous snapshots and reported more than doubled odds, whereas the fewer longitudinal cohorts tracked temporal order and yielded the attenuated 17% figure. This pattern indicates that unmeasured common causes, not direct pharmacological effects, likely explain most of the crude association. Policy discussions that treat the Telegraph headline as settled risk distorting resource allocation away from proven violence drivers such as alcohol and socioeconomic deprivation.