Cluster: Crime and Criminal Justice
Martha Roditti, PhD, New Mexico State University , Madeline Gillette, MSW, MA, New Mexico State University , Michael DeMers, PhD, New Mexico State University and Zach Edwards, MA, New Mexico State University
Sunday, January 17, 2010: 10:45 AM-12:15 PM
Pacific Concourse I (Hyatt Regency)
Few studies have critically examined the spatio-temporal occurrences of family violence, even though there are many popularly held beliefs about when and where family violence is most likely to occur. For example some researchers suggest that special “drinking events” like the Super Bowl are associated with higher rates of intimate partner violence, yet others found no evidence of an increase in calls for service on such drinking holidays. Family violence usually occurs at home. Most crime analysis applications of GIS (Geographic Information Systems) have as their primary goal the apprehension of criminals by evaluating the overlap of criminal and victim activity spaces. Such “geographic profiling,” while increasingly robust, does nothing to assist in cases where the criminal and the victim activity space are essentially identical because they occur in the home. There are no published studies that demonstrate the potential of GIS to identify, characterize, and predict spatio-temporal patterns of domestic violence. This roundtable session will begin by discussing the following research questions; Research Question 1: Is family violence positively correlated with DWI, socioeconomic, or demographic variables? If so, to what extent are these variables correlated? Research Question 2: Are there temporal variations in family violence incidents by season, time of day, day of the week, or holiday? If so, what identifiable events or activities might prove correlative to increased domestic violence? Research Question 3: Using socio-economic and demographic information, is it possible to predict where family violence occurs in space and time? Research Question 4: Is there sufficient correlative data, at the necessary spatial and temporal resolutions, combined with robust computational algorithms to provide the best predictive models of domestic violence? One of our presenters will discuss the use of state-of-the-art (GIS) software to encode and analyze family violence rates over the entire 2007 time period for a Southwestern county. We will show how socioeconomic and demographic data from the US Census Bureau, law enforcement calls for family violence incidents from the city Police Department, and home addresses of DWI offenders gathered by County Health and Human Services as a part of their DWI Compliance Program informed the research. Maps will be available for discussion. The presenters will explain how the data were analyzed using both correlation and spatial regression. Spatial regression analysis, also referred to as Geographically Weighted Regression (GWR) returns regression results will be used for each geographic unit of analysis (i.e. census block, zip code, etc.). This analysis also compensates for spatial autocorrelation, which has historically been a fundamental impediment to spatial analysis when attempting to statistically predict spatial patterns. The roundtable goals are: (1) to stimulate input from the roundtable participants about a spatial-regression based model we are testing that can be used to predict where and under what circumstances family violence is most likely to occur; and (2) to promulgate the use of GIS as a powerful technique for planning and policy formation. Once we know where family violence is most likely to occur, we can take preventative measures.
See more of: Roundtables