Background: Violence in acute psychiatric wards affects the safety of other patients and the effectiveness of treatment. However, there is a wide variation in reported rates of violence in acute psychiatric wards. Objectives: To use meta-analysis to estimate the pooled rate of violence in published studies, and examine the characteristics of the participants, and aspects of the studies themselves that might explain the variation in the reported rates of violence (moderators). Method: Systematic meta-analysis of studies published between January 1995 and December 2014, which reported rates of violence in acute psychiatric wards of general or psychiatric hospitals in high-income countries. Results: Of the 23,972 inpatients described in 35 studies, the pooled proportion of patients who committed at least one act of violence was 17% (95% confidence interval (CI) 14-20%). Studies with higher proportions of male patients, involuntary patients, patients with schizophrenia and patients with alcohol use disorder reported higher rates of inpatient violence. Conclusion: The findings of this study suggest that almost 1 in 5 patients admitted to acute psychiatric units may commit an act of violence. Factors associated with levels of violence in psychiatric units are similar to factors that are associated with violence among individual patients (male gender, diagnosis of schizophrenia, substance use and lifetime history of violence).