|
Abstract
|
VioLit summary:
OBJECTIVE:
The goal of this analysis by Rosenfeld and Messner was to examine the relationship of inequality and homicide cross-culturally by looking at a sample of small, nonindustrial societies.
METHODOLOGY:
A quasi-experimental, secondary analytical design was employed for this study. Data for the analysis were taken from the cross-cultural Probability Sample Files, a sample of distinct societies representative of major world cultural regions. These were developed by the Human Relations Area Files. 60 societies were represented in the data; 32 societies had homicide data. These 32 societies were used in the analysis. Homicide was defined as defiant homicide, that is, homicide that was disapproved of by a significant part of the society. Normative killing such as revenge killing, ceremonial sacrifice, execution, and killing of enemies during warfare were excluded. Homicide was coded in the data as a dichotomous variable (absent/infrequent or frequent). These values were based on ethnographers' and/or secondary analysts' rating of frequent. A variety of tests done by Naroll reported that the defiant homicide indicator did not show significant bias, though there were some nonsignificant indications that high levels of defiant homicide may have been associated with the ethnographer's length of stay, familiarity with the native language, and failure to take population size explicitly into account. It was theorized that social inequality and social disorganization would both be positively correlated with defiant homicide. These variables, in addition to societal complexity were included in the analysis. Inequality was measured with four measures. These were class stratification (degree of economic or occupational inequality), caste stratification (degree of status ascription and endogamy), political oppression (actual or perceived political or social domination of one group by another), and relative deprivation (composite measure of the difference between culturally prescribed and actually achieved wealth, power, and status). Disorganization was measured by seven variables: internal population density (ratio of mean number of occupants to mean number of rooms per dwelling), size (population) of largest settlement, typical settlement size (household, village, town, city), total population, change in moral codes (introduction of new moral regulations), change in traditional authority (traditional political leaders regarded as possessing less power than in the past), and change in subsistence occupation (change in the mode of subsistence of the majority of the group). Complexity of society was measured in five aspects: technological complexity, political authority (differentiation of political roles and centralization of authority), judicial authority (degree of dispute resolution formalization), degree of military organization, and organizational complexity (combined measure of judicial authority, political authority, and the number of jurisdictional levels in the society). All variables were either already coded or coded by the researchers into dichotomous variables into categories that were largely based on absent/infrequent vs. frequent. Cases were included, by variable, for those societies which had a homicide score and a score on that particular variable. Z-score analysis revealed that typical settlement size, change in traditional authority, and level of technological complexity were significantly (.10) related to loss of cases. The data were analyzed with gamma statistics and Kendall's tau-b.
FINDINGS/DISCUSSION:
Examination of the bivariate relationships revealed that the four measures of social inequality were not strongly related to defiant homicide. Class stratification and relative deprivation showed small and insignificant positive relationships while caste stratification showed an insignificant negative relationship and political oppression showed a significant (p=.10) negative relationship. Social disorganization measures were not found to be strongly and positively associated with defiant homicide. Internal population density and change in moral codes were the only two of seven variables showing positive relationships; these were insignificant. Two of the disorganizational variables showed significant (p=.10) negative associations with defiant homicide: typical settlement size and total population. It was said that these findings led to the conclusion that the correlates of homicide observed in complex societies cannot be readily generalized to simple nonindustrial societies. The finding that typical settlement size was negatively related to defiant homicide was said to suggest the possibility that the general process of societal development may reduce levels of nonnormative homicide. This hypothesis was explored further through analysis of the five indicators of societal complexity and defiant homicide. All five indicators were negatively associated with defiant homicide; two (political authority and military authority) were significant at p=.10. More complex forms of social organization were said to potentially contribute to the reduction of nonnormative homicide. Some assessment of the validity of the homicide data was done examining correlations between homicide and two other forms of violence: wife beating and drunken brawling. Homicide was found to be significantly and positively correlated with both, and both forms behaved much like homicide when placed into the previous analyses. Thus, it was concluded that defiant homicide did differentiate between more and less violent societies. Differential effects of social inequality across different types of societies were said to possibly result from differential meanings about inequality as opposed to inequality per se. It was said that rates of interpersonal violence are higher in some simple egalitarian societies because they lack the situationalized political arrangements and coercive forces which limit violence in more complex societies.
AUTHORS' RECOMMENDATIONS:
The authors argued that future research should identify the threshold of societal complexity after which social inequality becomes an important source of homicide. This, they said, would require researchers to look beyond nation-states units of analysis.
EVALUATION:
This analysis offers us a fuller perspective on homicide. We need more studies like this to show us what societal-level variables affect the frequency of lethal violence. The major weakness in the data was thoroughly covered by the researchers: lack of reliability in data collection. Unfortunately, cross-cultural study is inhibited by these problems with the data. The researchers, in dichotomizing the variables, did alleviate much of the problem. In doing so, they did lose valuable information, but the benefits outweigh the costs. The use of the .10 p-value is of some concern when examining the significance of the findings; there is a great deal of chance that is included in such a value. Overall, this study gives good direction to be taken and a solid study, given the data they had to work with.
(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)
Cross Cultural Analysis
Cross-Cultural Comparison
Cross Cultural Differences
Cross Cultural Studies
Cross National Analysis
Cross National Comparison
Cross National Differences
Cross National Studies
Violence Causes
Homicide Causes
Inequalitly
Social Class
Social Development
Social Organization
Adult Violence
Sociocultural Factors
Socioeconomic Factors
Countries Other Than USA
Political Factors
Demographic Factors
Class Factors
05-05 |