I've written previously about the work of Yale sociologist (and Chicago native) Andrew Papachristos, who studies how many urban gunshot victims in Chicago and elsewhere can be found within relatively small social networks. He looks for people who commit crimes together and get arrested for it, something he dubs "co-offending networks." In 2013, for instance, he found that in one neighborhood, 41 percent of all homicide victims were contained within a small portion of the community's social network; the risk of being killed increased massively the closer one got, in terms of social ties, to others who were killed.

He's out with a new paper, written with two Ph.D. candidates, Ben Green and Thibaut Horel of Harvard, that adds another dimension to these networks.

"This has a dimension of time. If you want to understand diffusion, you need to understand timing—who had the virus first, who has it second, how did they transmit it? This work is dynamic," Papachristos says.

The authors found that within the network they studied—138,000 people over eight years, representing almost 10,000 gunshot victims—63 percent of shootings "were attributable to social contagion." The victims were shot, on average, 125 days after their "infector," the "person most responsible for exposing the subject to gunshot violence," was shot.

"That's actually very important," Papachristos says. "We're talking three to four months. Whet gets shot, Andy gets shot, there's going to be three to four months where the people around us are at severe risk of being part of this contagion. That's the real difference, in terms of both the science and the applicability to violence prevention."

Using these co-offending networks, they were able to rebuild "cascades" of gun violence, in which multiple gunshot victims could be traced back through their social connections. Some of those cascades were quite large; one included 469 people, though the average was 2.7. The cascades are not necessarily chains of retribution—instead, they reflect an increased general risk. This knowledge could be used to prevent additional violence in the wake of a shooting, given the relatively short time window between shootings in a cascade.

"It's pretty quick, man. If you think about three months—speaking as a human being, having worked with gunshot victims and gang members in Chicago for years—the shooting happens, you spend the first month just reeling from the event itself. People are reliving the event, retelling the event, dealing with the aftermath, starting the mourning process. And then you get a little, tiny bit of a return of normalcy or whatever that might be, and then another shooting happens," Papachristos says. "I can't imagine the toll it takes psychologically, physically, emotionally, let alone the loss of life and health associated with this."

One way to prevent further shootings is direct intervention, like the Cure Violence method the authors reference in the paper. But there are a variety of approaches that can work in combination.

"By and large, if you insert the right people in the right place at the right time, we know how to do trauma reduction, we know how to do mediation, we know how to do outreach, we know how to do wraparound services," Papachristos says. "We use that as one example—including emergency room-based violence prevention—and other sorts of things that could benefit from a strategic, data-driven infusion. I don't think it should replace what they're doing, but if you had the ability to say, hey, let's send outreach workers here, or let's send trauma-care workers here, I think you'd experience less trauma."

Of course, not everyone in these networks becomes a victim of gunshot violence, and the cascades aren't infinite. They break on their own, whether by chance or something else, and Papachristos is trying to go beyond the raw data to figure out why.

"We're trying to understand who's at the highest risk and didn't get shot, and we're going to start an interview-based, qualitative study to see what these folks do that live in these neighborhoods, that experience these traumas, how do they get by?" Papachristos says. "We don't have scientific evidence for that."