Skip to main content

Community Policing & Criminal Justice

What Networks Reveal About Police Who Shoot

Andrew Papachristos
IPR sociologist Andrew Papachristos applies network science to the study of gun violence, police misconduct, illegal gun markets, street gangs, and urban neighborhoods.

Is police violence associated with characteristics of an officer’s social networks and his or her placement within those networks? In the Annals of the Academy of the Political and Social Science, IPR sociologist Andrew Papachristos and Linda Zhao of Harvard University look at the relationship between police networks and police who shoot. They investigate Chicago police officers who are brokers—or people within a network who occupy an important position in a network and connect its disconnected parts. They examine these “broker” officers in networks of complainants between 2000 and 2003 to see if they are linked to shootings of civilians between 2004 and 2016. A total of 388 officers fired their weapons at some point between 2004 and 2016, including 46 officers who fired on multiple occasions. Their findings show that although only a small percentage of all CPD officers shoot, those officers who do shoot appear to occupy the unique structural position of a network broker. This finding appears to be true even when considering factors such as age, race, gender, and activity. Brokerage roles put officers at a higher risk for shooting civilians, and when they are transferred to reduce misconduct, they may actually “spread” aggressive behavior. The researchers’ findings suggest that policies and interventions aimed at curbing police shootings should include not only individual assessments of risk but also an understanding of officers’ positions within larger social networks and the department as a whole.

Identifying the “Rules” of Gang Violence

Those immersed in gang culture frequently refer to it as a “game,” with rules that amount to matters of life and death. In Social Forces, Papachristos and the University of California, San Diego’s Kevin Lewis use network science to explore what behaviors comprise those rules. Papachristos and Lewis created exponential random graph models of data on more than 600 homicides committed in the late 1990s, which enabled them to better understand the sophisticated reciprocal nature of interactions among almost 70 gangs. They identify 25 discrete, empirically verified “rules,” including that gang homicide is highly reciprocal, especially within Black gangs; that two given gangs were unlikely to collaborate to prey on a third gang; that larger gangs were more likely to both commit and be victims of homicide; and that there were not many gangs who were mostly victims of homicide (or “sinks”). They also find rules are largely contingent on the racial makeup of each gang, and that Black and Latino gangs were more likely to commit violence among their own racial group than to others. They discover, however, that such rules are inconsistent over time and may reflect opportunism more than strategy. Still, identifying them may be useful in preventing gun violence.