Quantifying criminal collaborations: A systematic review and Bayesian meta-analysis of co-offending characteristics

Co-offending involves two or more individuals committing a crime together. It is influenced by social processes such as peer influence, group cohesion and strategic cooperation that shape the social dynamics of criminal behaviour. Studying co-offending helps to reveal how social interactions and relationships influence individuals’ decisions to engage in crime. Here we conducted a systematic review and Bayesian meta-analysis on two measures of co-offending, prevalence and group size, using data from 110 included studies from 17 countries. The pooled effect of the prevalence of co-offences was 0.29 (95% credible interval (CrI) 0.22–0.35) based on a total sample of 45,973,445 offences, and the pooled effect of the prevalence of co-offenders was 0.50 (95% CrI 0.43–0.56) based on a total sample of 3,459,133 offenders. On average, most of the offences were committed by two people with a pooled average group size of 2.73 (95% CrI 2.63–2.84). Property crimes had a higher prevalence of co-offences and co-offenders compared with violent crimes, while the average group size was larger for violent offences. Young offenders were more likely to engage in co-offending compared with adult offenders, and offences committed by youth usually involved larger groups. Furthermore, male offenders tended to form larger groups than females. No notable differences were observed between male and female co-offenders in terms of the prevalence of co-offenders. Variations in reported measures across studies pose challenges for cross-study comparison, highlighting the importance of standardized reporting. The research plan was preregistered on the Open Science Framework platform (RK7Y3).