Back to Annual Meeting

Back to Annual Meeting

APHA Scientific Session and Event Listing

Tinka Markham Piper, MPH, CSW¹, Sandro Galea, MD, DrPH², Steve Messner, PhD³, Angela M. Bucciarelli, MPH¹, Victoria Frye, DrPH⁴, Ken Tardiff, MD MPH⁵, and David Vlahov, PhD¹. (1) Center for Urban Epidemiologic Studies, New York Academy of Medicine, 1216 Fifth Avenue, New York, NY 10029, 212-822-7277, kmarkham@nyam.org, (2) Department of Epidemiology, University of Michigan School of Public Health, 1214 South University, Room 243, Ann Arbor, MI 48104, (3) Department of Sociology, University of Albany, 1400 Washington Ave., Arts & Sciences 351, Albany, NY 12222, (4) Center for Urban Epidemiological Studies, New York Academy of Medicine, 1216 Fifth Avenue, New York, NY 10029-5202, (5) Department of Psychiatry, Cornell University, 525 East 68th Street Box 140, New York, NY 10021

Homicide rates declined sharply in New York City (NYC) during the 1990's. Two theories may explain this trend: the decrease in crack cocaine and increased policing (typically called the "broken windows" hypothesis) during this period. We used data from three sources: the Office of the Chief Medical Examiner of NYC, the NYC Police Department, and the US Census Bureau. We examined overall homicide and gun- and non-gun-related homicides separately. We estimated the effects of measures of misdemeanor arrests, cocaine prevalence, and firearm availability on homicide to test these theories. The analyses were based on random effects multivariable models using pooled, cross-sectional, lagged time series data for 74 NYC precincts over the 1990-99 period. Final random intercept models showed that misdemeanor arrests had a negative effect (Beta=-0.0006, p °Ü 0.01) on subsequent homicide mortality rates; cocaine prevalence (Beta=0.1620, p °Ü 0.01) and firearm availability (Beta=0.0510, p °Ü 0.05) had a positive effect on subsequent homicide mortality rates. Further analyses of homicide disaggregated by gun or non-gun use showed that the effects of misdemeanor arrests and firearm availability were present for gun-related but not non-gun-related homicides (Beta=-0.0004, p °Ü 0.05; Beta=0.0001, p=NS, respectively); the measure of cocaine prevalence was significantly related to both forms of homicide (Beta=0.1228, p °Ü 0.01, and beta=0.0428, p °Ü 0.01 respectively). This study suggested that both the decrease in crack cocaine and the implementation of quality of life policing in NYC likely contributed to the drop in homicide in NYC during the 1990s, independently of one another.

Learning Objectives:

• 1. Describe homicide trends and correlates in NYC during the 1990s.
• 2. Discuss two theories that account for decline of homicide during the period (decrease in crack cocaine and increased policing, “broken windows” hypothesis)
• 3. Identify the effects and outcomes of misdemeanor arrests, cocaine prevalence, and firearm availability on homicide trends in NYC during the 1990s

Presenting author's disclosure statement:

Any relevant financial relationships? No