|
Abstract
|
Research on the causes of homicide has focused on two competing theoretical models, socioeconomic and subcultural. After a review of these two theoretical perspectives and a number of recent empirical studies of homicide derived from these perspectives, several theoretically derived hypotheses relating types of homicides as classified by the victim/offender relationship and these perspectives are presented and discussed. The analysis is designed to evaluate these hypotheses, and to address some of the discrepant findings in previous studies. The results indicate the importance of specification issues such as measurement and aggregation in understanding previous research, the primacy of the socioeconomic approach for understanding the causes of homicide, and the importance of further classification of homicide for disentangling potential subcultural effects from socioeconomic effects. Finally, a recommendation for further research on the subcultural approach is discussed.
VioLit summary:
OBJECTIVE:
The purpose of this study by Parker was to examine two important issues in the study of homicide: methodological problems such as measurement and aggregation and the competing theoretical models of subculture and socioeconomic explanations of homicide rate. Using Messner's (1983) study, the author tested subcultural and socioeconomic explanations for total and disaggregated homicide rates by victim/offender relationship.
METHODOLOGY:
A quasi-experimental, cross-sectional design was employed using secondary data from official sources for 299 cities. Messner's results for 204 SMSAs were also reported for purposes of comparison. Data for homicide rate were obtained from the Supplemental Homicide Report for 1973-1975 which were averaged for the comparison equation and totaled for the others. The total homicide rate was disaggregated by victim/offender relationship. These categories were as follows: 1) homicide committed during commission of robbery, 2) during commission of other felony, 3) between persons who were acquainted but not sexually intimate, and 4) between spouses. Independent variables examined were structural poverty, racial composition, and the city's "southness" in the year 1973. Structural poverty was a composite measure including the following: l) percentage of families with $3,000 or less in income, 2) percentage of children living with only one parent, 3) percentage in the state where a city is located who failed the armed forces mental test for recruits, 4) percentage of those over 25 years of age with less than five years of education, and 5) infant mortality rate. Racial composition was measured by percent black. Southernness was measured using Confederacy as a guideline. Control variables were l) percentage aged 20-34, 2) the natural log of population size, 3) the natural log of density per square mile, and 4) Gini index of income inequality. All data for independent and control variables were taken from U.S. Census Bureau sources. The data were analyzed using l) OLS regression to replicate Messner's analysis with 1973-1975 data, 2) LISREL analysis of the total homicide rate, and 3) LISREL analysis of the four disaggregated homicide rates.
FINDINGS/DISCUSSION:
Significance was determined by whether the coefficient was at least l.6 times its standard error for a one-tailed test. Structural poverty, percent black (indicator for racial subculture of violence), and population (log) were significantly and positively related in the OLS regression for average homicide rate showing that aggregation to city rather than SMSA changed the significance of Southernness (regional subculture of violence) and Gini inequality. The LISREL model (using appropriate controls for measurement error and correlated measurement error in homicide and structural poverty) for total homicide revealed only structural poverty and population (log) and density (log) as significant predictors (in expected positive direction). When homicide rate was disaggregated into victim/offender relationship types, the overall pattern changed. Poverty was significantly and positively related to three of the four types of homicide (other felony, primary nonintimate, and family intimate) and in two of the three, family intimate and other felony, it was the dominant predictor. Percent black was significantly related to robbery and primary nonintimate homicide. Southernness was significantly related to robbery and other felony but in a negative direction. Population size was significantly and positively related to every equation except family intimate. Population density was negatively related to family intimate homicide.
AUTHOR'S RECOMMENDATIONS:
The author concluded by saying that homicide should be broadened in both theoretical and logical terms, particularly in the area of victim/offender typology. He also argued that current examination of subculture of violence theory in homicide research is misguided in both theoretical terms and operationalization in empirical measures. Measurement issues of aggregation and indications of homicide and other independent variables also need further research attention.
EVALUATION:
The central theme in this work is as important as it is difficult. The study of homicide has traditionally met with measurement and theoretical difficulties. The error inherent in the data we use is compounded by the effects of different levels of aggregation of these data. The author reveals, to the benefit of this kind of research, that reliance upon a singular indicator of homicide masks several differences that exist between different types of homicide. Also, the difference between a city and a SMSA is great and reveals theoretical differences between them. The exploration of more precise tools for statistical analysis is also a great step forward. There is continued neglect, however, of gender in this work. Homicide rates are assumed to be affected by the same factors whether the rates are for total homicide or for gender-specific homicide. The author ultimately questions the usefulness of subcultural theory and the use of demographic indicators such as percent black to proxy for subculture of violence in statistical models, but he uses percent black in his equations and does not make the argument forcefully enough. Theoretically these practices are not sound, and ethically, use of percent black as an indicator of violent subculture is questionable at best and racist at worst. (CSPV Abstract - Copyright © 1992-2007 by the Center for the Study and Prevention of Violence, Institute of Behavioral Science, Regents of the University of Colorado)
N1 - Call Number: F-228, AB-228
KW - Violence Causes
KW - Adult Violence
KW - Adult Offender
KW - Homicide Causes
KW - Homicide Offender
KW - Subculture Theory
KW - Subculture of Violence
KW - Socioeconomic Factors
KW - Socioeconomic Status
KW - Poverty
KW - Homicide Rates
KW - Comparative Analysis
KW - Racial Differences
KW - Sociocultural Factors
KW - Domestic Homicide
KW - Victim Offender Relations |