Intersections of Race, Gender, Disadvantage, and Violence: Applying Intersectionality to the Macro-Level Study of Female Homicide

Abstract

Intersectionality implies that multiple socially constructed categories (race, gender, and class) interact, but also operate at many levels when contributing to inequality. This paper provides an illustration of the intersection of race, gender, and disadvantage in the study of race-specific female homicide rates. In this research, we are interested in how race- and gender-specific forms of disadvantage differentially influence rates of homicide offending by white and African-American women in cross-sectional and change models. We account for the availability of domestic violence resources and crime control policies, in addition to the structural forces that have been found to influence female offending specifically. Methodological techniques are used to amend caveats in the data, such as the rare nature of female homicide offending. Our findings reveal similarities in the way economic marginalization and divorce influence female homicides across racial groups but a number of differences in the role of crime control policies and availability of domestic violence resources on race-specific female homicides cross-sectionally and over time. Overall, our study produced results which advocate the intersectional framework.

Keywords:

• intersectionality
• race
• gender
• homicide offending
• macro-level

Notes

1. Chattanooga, TN, Citrus Heights, CA, East Los Angeles, CA, Kansas City, KS, Lowell, MA, Metairie, LA, Omaha, NB, Overland Park, KS, Paradise, NV, Springfield, IL, Sterling Heights, MI, Worchester, NY, and all cities in the state of Florida were dropped from the analysis due to not reporting at all during either the period 1989-1991 or 1999-2001. Chicago, IL was the final city dropped due to being an extreme outlier.

2. SHR are a primary source of data for homicide researchers (Fox, 2004; Fox & Swatt, 2009b; Pampel & William, 2000). Although missing data are common, particularly as it relates to race of the offender and victim-offender relationships, the most recent version of the Fox and Swatt SHRs (2009a) employ a multiple imputation technique to address missing data on the race of the offender. The merits of multiple imputation method, over other strategies like dropping cases and use of mean substitutions, have been discussed in the criminological literature (Brame & Paternoster, 2003; Fox & Swatt, 2009b). Thus, the latest Fox and Swatt data with multiple imputation on the race of the homicide offender has been utilized in this study.

3. The authors thank Professor Laura Dugan, University of Maryland, for providing earlier Domestic Violence Service Directories for data collection purposes.

4. It is important to note that these two dependent variables can be related, as argued by Steffensmeier and Haynie (2000a). When that happens, the error terms of the regressions are not independently and identically distributed as you would assume if you run them in separate regressions. Failures to account for this dependency could result in both biased standard errors (biased downwards) and larger coefficients. The method commonly used to account for this methodological concern is seemingly unrelated regression (SUR). SUR modeling technique has been used by others examining gender-specific homicide rates (see Steffensmeier & Haynie, 2000a). We estimated SUR models and the results produced are identical to the results presented here. That is, given the non-normal distribution in the dependent variable and the fact that the SUR results are the same, we present the only the results for our negative binomial regression estimates.

Additional information

Notes on contributors

Karen F. Parker

Karen F. Parker is professor in the Department of Sociology and Criminal Justice at the University of Delaware. Her current research interests include exploring the influence of macro-level constructs on urban violence, particularly labor markets, racial segregation, immigration and concentrated disadvantage. Much of her recent work has incorporated longitudinal models to examine how changes in the local urban economy differentially influence race-specific homicide rates over time. Research on these topics has been published in Criminology, Social Science Research, Feminist Criminology and Journal of Quantitative Criminology.

M. Kristen Hefner

M. Kristen Hefner is a doctoral student in the Department of Sociology & Criminal Justice at the University of Delaware. She received a BA in psychology from North Carolina State University and an MA in sociology/criminology from the University of North Carolina at Greensboro. Her primary research interests include gender, law, criminal justice policy, feminist theory, and prison institutions.