Violence is exacerbated in small close-knit groups and through social identification processes. Furthermore, the homophily principle postulates that individuals tend to form relationships because of shared characteristics. We analyze the community structure of the Islamist co-offending network – a highly violent social network – and assess whether shared characteristics increase the probability to form co-offending ties. We rely on a new database on the individual attributes and the co-offending relationships of 503 Islamist offenders radicalized in the United States between 1948 and 2020. Using community detection algorithms, we show that the US Islamist co-offending network is highly clustered, modular, and includes only a few large communities. Furthermore, results from the exponential random graph modeling evidence the roles of spatial proximity, ideological affiliation, and shared social characteristics in driving co-offending among US Islamists. Overall, these findings demonstrate that small, local communities and homophily shape violent social networks.
Schwarzenbach, Anina, and Michael Jensen. 2023. “Extremists of a Feather Flock Together? Community Structure and Patterns of Homophily in the US Islamist Co-offending Network”. SocArXiv. May. https://doi.org/10.31235/osf.io/e3d2t.