John Jay College Of Criminal Justice | The City University of New York

 

By Jennifer Nislow

With tremendous ability to analyze the interstices among members of a group, but a shortage of rich data sets to plumb, the past few years have been somewhat frustrating for social scientists eager to apply the technique of Social Network Analysis (SNA) to their research.

SNA is used to analyze existing networks of social contacts, according to Professor Kirk Dombrowski, a member of John Jay’s Department of Anthropology. It is a fairly old methodology that began with the creation of sociograms that diagrammed the connections between people. Anthropologists used SNA as part of kinship studies, he added.          

The software used by social network researchers like Dombrowski is not to be confused with the social network services programs that run Facebook and other online communities.

“Some social networking sites, like Facebook, use personality profiles to match people,” he said. “They do not analyze their network profile except in the most rudimentary way — usually to show people the web of their social contacts.”

Dombrowski offered this analogy: social network services software is like a typewriter that can be used to write numbers; SNA software is a calculator that performs calculations.       

 One of the major barriers to the greater expansion of SNA in the social sciences has been the “leapfrogging” of data and analytic ability, he explained. Data sets were created before the software made it possible in the last few years to analyze the “micro structures” of social networks in a millisecond.

“In the last four or five years, we have this huge analytical ability and so few data sets because data sets that had been generated before have all been worked over and it costs so much money to do network research,” said Dombrowski.

Indeed, that issue was explored at a symposium hosted in August 2007 at John Jay by Dombrowski, Professor Ric Curtis, who chairs the College’s Department of Anthropology, and Professor Bilal Khan of its Department of Mathematics and Computer Science.  Its purpose was to link SNA with a social sampling technique for use on hard-to-reach populations called Respondent Driven Sampling (RDS). 

“Because RDS makes practical use of the sorts of social networks that SNA theorists are most interested in to recruit population samples, and because RDS produces data on network populations both quickly and reliably, the link between the two would seem ideal,” said Dombrowski.

Finding ways to produce network data from new sources — quickly, easily, cheaply — remains a major hurdle for social network researchers, he added.

There is one field, however, where the data sets have not been picked clean — criminal justice.

Characterizing criminal-justice data as “low-hanging fruit,” Dombrowski said, “there are loads of data in criminal justice where these kinds of questions about local structures are really relevant and there’s all this capacity on the social networking side, if we can just connect them.”           

The 1990s were a peak time for social network theory, he explained. Vastly enhanced computer capability was met by millions of dollars in funding for AIDS research. One of the projects that benefited from this confluence, he noted, was a study that examined the social network of 767 injecting drug users and the risk factors for HIV among them.

“It’s one of those moments you look back on with such nostalgia because you’ll never again get $4 million to interview 700 people,” said Dombrowski.

Recently, a National Science Foundation (NSF) grant was awarded to Dombrowski and Khan, an expert in graph theory, to do social network modeling on HIV stabilization rates among infected drug users. They will be testing structural factors that influence the spread of the virus within that particular community and trying to discover why the rate of infection has stabilized within the group well below the point of saturation.

“What we know about epidemiology is that these things should spread through the whole network fairly rapidly,” said Dombrowski. “But for some reason, they don’t — and it’s not because people change their behavior, necessarily.”

Dombrowski and Curtis, have just completed a social network project on the sexual exploitation of underage sex workers in New York City.

Another study in the works is one with Professor David Kennedy, director of John Jay’s Center for Crime Prevention and Control, which will examine immigrant victims of violence.

“We’re taking a theory that’s already fairly well established and research protocols that are established, and applying them to data in criminal justice that is very amenable to that kind of analysis, but has not been done,” said Dombrowski. “We think it has huge potential. That’s what makes it so exciting.”

Jennifer Nislow is assistant publications director at John Jay College.