VIOLENT AND VINDICTIVE WOMEN: A RE-ANALYSIS OF THE SOUTHERN SUBCULTURE OF VIOLENCE

Abstract

In order to further develop an understanding of the Southern subculture of violence and its association with female-perpetrated homicide, new techniques must be developed that sufficiently operationalize Southern cultural influence. For this study, we created the Southern Subculture Index, a factor combining a county's population born in the South, Evangelical Christian, and of Scots-Irish ancestry. Specifically, we hypothesize that by using this index, the relationship between female-perpetrated homicide and Southern culture will be visible. Results from a negative binomial regression model confirm the veracity of a positive and significant effect of the Southern Subculture Index on female-perpetrated homicide.

The authors would like to thank Dr. Matthew R. Lee, the editors, and reviewers for helpful comments on prior drafts of this manuscript. A version of this article was presented at the annual meeting of the American Society of Criminology, St. Louis, Missouri, in November 2008.

Notes

¹Of the 3,141 U.S. counties, only 3,062 had comprehensive data available for all variables from all sources used in the study.

²Infants and young children have not been included, as they are generally killed for reasons not related to the present study, that is, Southern culture may not be predictive of instances where a distressed mother kills her child in a fit of exhaustion and rage.

³The “South,” as defined by the U.S. Census Bureau, includes Alabama, Arkansas, Delaware, the District of Columbia, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Virginia, and West Virginia.

N = 3,062.

⁴Because some counties had exceptionally high homicide rates, Cook's distance test for influential cases was computed. Of those cases that were found to be influential, all counties contained extremely large cities with historically high homicide rates (i.e., New York County; Los Angeles County; Harris County, TX (Houston); Philadelphia County; Dallas County; Bronx County; and Baltimore City).

⁵To be sure of the accuracy of our index, we used both maximum likelihood and generalized least squares extraction methods in prior regression analyses. All came to the same results and significance levels as the principal components analysis. Therefore, because principal components analysis is most often used in this type of data reduction and because its “components reflect both common and unique variance of the variables” and are thus more robust, it was chosen as the preferred technique for this analysis (Garson 2009).

⁶The bivariate correlations between the dependent variables, explanatory variables, and indices are as follows: Male Homicide and Born in South, Evangelical, and Scots-Irish are −.022, −.076, and −.036, respectively; Male Homicide and the Southern Subculture Index is −.057; Female Homicide and Born in South, Evangelical, and Scots-Irish are .002, −.056, and −.025, respectively; and Female Homicide and the Southern Subculture Index is −.032.

*p ≤ .05; **p ≤ .01; ***p ≤ .001.