Jeffrey B. Bingenheimer, a doctoral student in health behavior and health education, analyzed five years of data from adolescents living in 78 neighborhoods in Chicago. Bingenheimer is lead author on a paper in this week's journal Science.
Using a statistical method called propensity stratification, Bingenheimer and coauthors Robert Brennan and Felton Earls aimed to establish a firm cause and effect relationship between exposure to gun violence and later perpetrating violence.
"We wanted to know whether these just tend to occur in the same populations or if one actually caused the other," Bingenheimer said. "These findings suggest there is a substantial cause and effect relationship between exposure and perpetration. Violence can be transmitted from person to person by means of exposure in the community."
Bingenheimer used data from the Project on Human Development in Chicago Neighborhoods, a longitudinal study that conducted extensive interviews with more than 6,000 young people and their primary care givers on everything from reading proficiency to family members with legal problems to delinquency in their peer groups. Bingenheimer used data from about 1,500 adolescents who were within six months of their 12th or 15th birthday at the time of their first interview.
The project is based at Harvard University, with involvement from Columbia University, Emory University, Johns Hopkins, and Pennsylvania State University, among others. Brennan and Earls are faculty members at Harvard.
Bingenheimer received a fellowship that involved working with Stephen Raudenbush, one of the project's scientific directors and a faculty member at the Survey Research Center at U-M's Institute for Social Research. That led Bingenheimer to apply his public health interests to firearm violence.
Bingenheimer said many researchers have looked at how violence begets violence. His use of the complex statistical approaches he learned from Raudenbush help make this paper unique.
While many investigations, such as clinical trials of new drugs, use randomized experiments to test effects, that is not possible in a real-life situation like examining the effects of gun violence. So instead they looked at information provided in questionnaires and used propensity stratification to simulate randomization as best they could.
Overall, Bingenheimer found that adolescents who were exposed to firearm violence were nearly four times as likely as unexposed adolescents to perpetrate violence over the next two years.
But exposed and unexposed adolescents differed on a wide range of demographic, socioeconomic, temperamental and other factors. These differences, rather than gun violence exposure itself, could account for the elevated rates of violence among exposed adolescents. Propensity stratification allowed Bingenheimer to control for the effects of more than 150 characteristics of the adolescents, their families and their neighborhoods.
Propensity stratification combines these variables into a single number, the predicted probability of being exposed to gun violence. If the association between actual exposure and perpetration persists among adolescents who had similar predicted probabilities of exposure, this is evidence of a cause and effect relationship, Bingenheimer said.
The study defined exposure to firearm violence as having been shot or shot at or seeing someone shot or shot at. It defined perpetrating violence as carrying a hidden weapon, attacking someone with a weapon, shooting someone, shooting at someone, or being in a gang fight.
Bingenheimer is scheduled to defend his dissertation Friday, the final step before receiving his doctorate. His dissertation topic is "Mortality from Competing Risks and the Spread of HIV-1 in Human Populations."
To receive a copy of Bingenheimer's paper, contact the AAAS Office of Public Programs at (202) 326-6440 or email@example.com
For more on the U-M School of Public Health: http://www.
For more about the Survey Research Center at U-M's Institute for Social Research: http://www.
Project on Human Development in Chicago Neighborhoods: http://www.
Producers: U-M has professional TV studios and uplink capabilities.