With a homicide rate historically more than three times greater than the rest of the United States, Newark, N.J., isn’t a great vacation spot. But it’s a great place for a murder study.
Led by April Zeoli, an assistant professor of criminal justice, a group of researchers at Michigan State University tracked homicides around Newark from 1982 to 2008, using analytic software typically used by medical researchers to track the spread of diseases. They found that “homicide clusters” in Newark, as researchers called them, spread and move throughout a city much the same way diseases do. Murders, in other words, did not surface randomly—they began in the city center and moved in “diffusion-like processes” across the city.
The study also found that the there were areas of Newark that, despite being beset by violence on all sides, remained almost completely immune to the surrounding trends over the entire course of the 26 years studied. Despite the longitudinal nature of the study, Zeoli notes in a press release that the analytic software can be employed in real time, so that police might potentially identify problem areas as they are emerging—or perhaps, one imagines, before they emerge. The research is due to be published in a forthcoming issue of a journal called Justice Quarterly.
But why does violence spread this way? And why are some areas almost more immune than others? The answers are complicated and uncertain. But some clues emerge from the data.
Newark isn’t the most violent city in America. Flint and Detroit rank as more dangerous (numbers 1 and 2, respectively among America’s most violent cities, according to reports culled from recent FBI data). And they’re both within driving distance from the researchers' home base, the Michigan State campus at East Lansing. But the MSU researchers had data for Newark that they didn’t have from anywhere else thanks to the long-term efforts of the study’s co-author, Jesenia Pizarro, an MSU assistant professor of criminal justice, who was born and raised in Newark, and has been gathering Newark homicide data throughout her career.
From data both quantitative and empirical, Zeoli and her team were able to build some hypotheses based on previous research regarding why violence spreads in cities like Newark. Citing a 2000 study by British geographer, Peter Haggett, Zeoli’s study notes that crime, like disease, spreads when three conditions are met: First, “a source or infectious agent”; second a “mode of transmission”; and, third, “a population susceptible to transmission.”
Space-time clusters of overall homicides in Newark, NJ, USA, 1982-2008. (All clusters are significant at p < .01.) Between 1997 and 2000, there were no new clusters of overall homicides in Newark, which is consistent with the national and Newark trends of decrease in overall homicides.
In terms of violence, the “source” is the presence of violence itself, and of forms of social interaction—like drug dealing and maintaining street cred—that lend themselves readily to violence. Certain populations, like the inner-city areas plagued by violence in the study, are “susceptible to transmission” by virtue of things like poverty, high unemployment, poor relations with police, the long-term effects of racial and economic segregation, high levels of gang membership, and prevailing notions that aggression and stocking firearms are requisites for survival.
Prior research shows that the so-called “mode of transmission” is all about direct contact, just as it is with disease. Citing multiple previous studies of similar urban areas, the authors note that “for example, a single homicide of a gang member may precipitate multiple more homicides as the initial death is avenged, which in turn may lead to additional retaliatory homicides.” In other circumstances, “fear of homicide in nearby areas may precipitate arming by individuals who believe they must use lethal force to defend themselves because they perceive others are willing to resort to lethal measures, and they do not trust law enforcement’s ability to protect them.”
And, as importantly, the authors note, “homicide can spread without direct contact between parties involved in particular precipitating events, but also through the lasting impact of prior homicides on social interactions.”