Friends to the End: How the Wrong Social Network Increases Risk of Homicide Victimization

Every year thousands of people are killed with firearms in the United States, but these killings are distributed disproportionately across certain places and populations. Research has shown that young people, men, African Americans, and gang members are more likely to be murdered than their counterparts in the general population. We also know that individuals in socially and economically disadvantaged neighborhoods are more likely to be murdered than people in less disadvantaged neighborhoods. But when it comes to individuals with similar profiles and shared risk factors, is there anything that makes one person more likely to be murdered than another?

To address this question, researchers Andrew V. Papachristos and Christopher Wildeman co-authored “Network Exposure and Homicide Victimization in an African American Community” in the November 2013 American Journal of Public Health. They examine whether a person’s social network can make them more likely to be killed than other people in the same community who have similar risk factors but different social networks. Specifically, they look into the effect of closeness to homicide victims and whether having social ties to a homicide victim makes one more likely to also become a homicide victim.

The authors first study a specific area in Chicago which had a much higher annual homicide rate than the rest of the city during the time period from 2006 to 2011 (55.2 homicides per 100,000 people, compared to 14.7 in the city as a whole) and then restrict their sample further to focus on people who lived in the community between 2006 and 2011 and who were arrested during that time period. To establish social ties, they find groups of people who had a documented history of engaging in risky behavior together, by looking for arrests where at least two people were arrested together.

This procedure narrowed the number of studied individuals to 8,222, from a total community of approximately 82,000 residents; these studied individuals were arrested in one of the 1,732 arrests (“components”) in which two or more offenders were captured. During the studied time period, 41 percent of the gun homicides in the community victimized a member of this co-offending network. Of these 103 homicide victims, each of them was also a member of one of a subset of 75 components. This network of 75 components only included 3,718 people—less than five percent of the total community population—but accounted for 41 percent of the homicide victims in the community.

The authors measure several other network properties for each individual in the study, including geographic distance in feet from the nearest homicide victim by street address, number of co-offending ties, and proportion of co-offending associates who were also tied to each other. They study these factors in three different regression models: one that only includes general risk factors, a second that includes variables based on a person’s social network, and a third using neighborhood fixed effects to compare different neighborhoods in the city.

Using the second model, they find that increasing social proximity and connections to homicide victims through shared, risk-increasing behavior increases a person’s likelihood of being murdered. The first model confirms the previous research on risk factors, while the third model confirms that these findings should apply to other neighborhood contexts.

The authors note that they worked with a small and homogenous sample, only studied one specific type of social connection, and that they did not identify a cause for selection into high-risk networks. They would like to see future research on networks that encompass more types of shared social behaviors and relationships, examinations of lower-risk communities, and research that determines how these high-risk networks are formed. This work and further research should prove very useful in lowering the risk of homicide victimization among people who face the greatest risk.

Feature Photo: cc/(J.Knecht)