Homicide as Infectious Disease: Using Public Health Methods to Investigate the Diffusion of Homicide

Abstract

This study examined the spatial and temporal movement of homicide in Newark, New Jersey from January 1982 through September 2008. We hypothesized that homicide would diffuse in a similar process to an infectious disease with firearms and gangs operating as the infectious agents. A total of 2,366 homicide incidents were analyzed using SaTScan v.9.0, a cluster detection software. The results revealed spatio-temporal patterns of expansion diffusion: overall, firearm and gang homicide clusters in Newark evolved from a common area in the center of the city and spread southward and westward over the course of two decades. This pattern of movement has implications in regards to the susceptibility of populations to homicide, particularly because northern and eastern Newark remained largely immune to homicide clusters. The theoretical and practical implications of the findings, as well as recommendations for future research, are discussed.

Keywords:

• homicide
• public health
• gangs
• firearms

Notes

1. While we do not intend a literal application, drawing parallels between the elements of disease diffusion to known risk factors for homicide is instructive in understanding how homicide may move through communities and what factors may be appropriate for prevention and intervention efforts.

2. Given our temporal unit of months, using census tracts as the spatial unit resulted in 28,890 spatio-temporal observations. Due to the complexities of the analytic strategy employed, a smaller spatial unit would have been more intensive to model.

3. The temporal unit used to investigate diffusion was one month. We did not put constraints on the models as to how long a clustering of homicides was allowed to last.

4. Future research should conduct a linear interpolation to reduce temporal measurement error.

5. It is possible that the absence of clusters in previously covered parts of the Central and West Wards after 1997 is the result of significant negative spatial autocorrelation (false-negative finding). However, the abrupt and persistent absence of clusters in these locations through 2008 is more suggestive of changing homicide-population processes internal to these areas or the introduction of an external intervention resulting in reduced homicide compared to the West and South Wards. More specifically, we suspect that this might be due to changes in public housing in the city, which resulted from the tearing down of high-rises in the areas during the late 1990s. Nonetheless, to fully reconcile the presence of negative spatial autocorrelation in trajectories of homicide diffusion requires testing for significant heterogeneity in homicides in these areas across the time period. In our study, we implemented the local G-statistic (Lee & Wong, 2001) on the homicide data from 1982 to 2008 and did not find significant negative spatial autocorrelation in these parts of Central and West Wards, supporting our hypothesis of relocation diffusion.

6. The airport appears as part of homicide clusters due to the modeling parameter of constraining the ellipse window to 1.5 miles. The airport is contained in a large census tract in the East Ward and if the centroid of that tract falls within a 1.5 buffer positioned on a contiguous tract centroid, then it is included in a cluster if that cluster is significant.

7. While our data suggest that the diffusion of homicide from around the Central Ward started before the advent of crack cocaine in Newark neighborhoods, conclusions related to such processes must be made with caution, as the use of 1982 as the start date of our study is arbitrary (i.e. it was the extent of our available data). If homicide indeed acts as an infectious disease, and is thus part of an ongoing process, processes set in motion prior to 1982 likely influenced the patterns witnessed in our earliest years of available data. However, given that crack markets did not emerge in the city until 1986, our data is well suited for detecting any appreciable change in the pattern of homicide that took place as a result of the introduction of this drug.

8. The American Public Health Association defines the word epidemic to mean “the occurrence in a community or region of cases of an illness (or an outbreak) with a frequency clearly in excess of normal expectancy” (Chin, 2000). Given that our clusters specifically identified areas and time periods in which the number of cases of homicide exceeded the expected level, and given our parallel between homicide and disease, the term epidemic is an appropriate term to use in describing the problem of homicide in Newark.

Additional information

Notes on contributors

April M. Zeoli

April M. Zeoli is an assistant professor in the School of Criminal Justice at Michigan State University. She obtained a PhD in health and public policy from the Johns Hopkins Bloomberg School of Public Health. Her main field of investigation is the prevention of intimate partner violence and homicide through public policy. Her work has appeared in the Journal of Interpersonal Violence, Evaluation Review, and Injury Prevention.

Jesenia M. Pizarro

Jesenia M. Pizarro is an associate professor in the School of Criminal Justice at Michigan State University. She obtained a PhD in Criminal Justice from the School of Criminal Justice at Rutgers University, Newark, NJ in 2005. Her research focuses on the social ecology of violence, homicide, and corrections policy. Her work has appeared in Justice Quarterly, Criminal Justice and Behavior, and Homicide Studies.

Sue C. Grady

Sue C. Grady is an associate professor in the Department of Geography and adjunct assistant professor in the Department of Epidemiology at Michigan State University. She obtained a PhD in Earth and Environmental Science at the City University of New York. Her research on maternal and infant health has been published in the journals Social Science and Medicine, International Journal of Health and Geographics, and Journal of Epidemiology and Community Health.

Christopher Melde

Christopher Melde is an assistant professor and coordinator of undergraduate studies in the School of Criminal Justice at Michigan State University. He received his PhD in criminology and criminal justice from the University of Missouri, St Louis in 2007. His primary research interests include program evaluation, juvenile delinquency and victimization, gangs, perceptions of crime and victimization risk, and criminological theory. His recent work has appeared in such outlets as Criminology, Journal of Research in Crime and Delinquency, Justice Quarterly, and Journal of Quantitative Criminology.