|
Abstract
|
VioLit summary:
OBJECTIVE:
The aim of this study by Centerwall was to evaluate television as a risk factor for violence by examining the homicide rates in South Africa, Canada and the United States.
METHODOLOGY:
The author followed a quasi-experimental design in which the homicide rates in Canada and the United States were compared retrospectively with those of South Africa. From an epidemiological orientation, the author sought to determine the effect of an intervention, the introduction of television, on homicide rates by using as a control group a population which was not subjected to the intervention. In Canada and the United States television was introduced in 1945 while in South Africa, television was not introduced until 1974. South Africa was selected as the control population since it did not receive the intervention, television, until almost thirty years after Canada and the United States. The author compared only the homicide rates of the white populations of each country.
FINDINGS/DISCUSSION:
The author found that between 1945 and 1974 the annual incidence of white homicide deaths in the United States and Canada increased 93 per cent and 92 per cent, respectively. In the United States, there were 3.0 homicides per 100,000 whites in 1945 compared with 5.8 homicides per 100,000 in 1974. In Canada, there were 1.3 homicides per 100,000 in 1945 compared with 2.5 homicides per 100,000 whites in 1974. The incidence of white homicide deaths in South Africa during this same time period decreased by 7 per cent; there were 2.7 homicides per 100,000 whites in 1943-1948 compared with 2.5 homicides per 100,000 in 1974.
Since violence is a multifactor phenomenon involving many variables, the author suggested that the attained rate of violence reflects exposure to television, the particular country's baseline rate of violence, as well as the effect of changes in other risk factors, such as socioeconomic conditions. The author further argued that this type of population intervention study is free of the ecological fallacy. The ecological fallacy occurs when the investigator infers that the unobserved effect of one variable on another at the individual level is the same as the observed effect at the group level. In the present study, the author suggested that the issue of ecological fallacy is rendered moot insofar as "the link between risk factor and outcome variable has already been established at the level of the individual" (p. 646). Each white South African was equally without television in the same way as each white Canadian and American was with television between 1945 and 1974. The author also acknowledged that in conducting this type of comparative study there are "rough edges" insofar as prior to 1975 there could have been South African whites who were exposed to television outside of South Africa. Specification bias was reduced by choosing a control population, South Africa, with a similar social and historical background to Canada and the United States.
The author examined potentially confounding third variables, but found none adequately explained the observed homicide trends. Potential confounding third variables were controlled through the use of multiple control groups. However, the author was unable to control for the effect of local inaccessible confounders since there were no other television-free countries similar to South Africa. The author found that the two major etiologic hypotheses were substantiated by the data. One hypothesis stated that the exposure to television is followed by an increase in rates of violence. Another hypothesis stated that the timing of the introduction of television predicts the timing of the subsequent increase in the rates of violence. In conclusion, the author calculated that exposure to television is etiologically related to about one half of the homicides committed in the United States.
EVALUATION:
The author presents an interesting comparison of homicide rates in the United States, Canada and South Africa as a function of each countries' relative exposure to television. However, the task of explaining an entire country's rate of homicide by exposure to television alone seems problematic for several reasons. The author does not convincingly justify the exclusion of the effects from other confounding variables from the relationship. For example, the relative effects of exposure to other mediums in each country is not discussed. Finally, the author assumes that the backgrounds of all three countries are comparable and that there was similar television programming both in Canada and the United States during the period studied. (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-433, AB-433
KW - South Africa
KW - Canada
KW - US Foreign Comparison
KW - Countries Other Than USA
KW - Homicide Rates
KW - Television Viewing
KW - Television Violence
KW - Comparative Analysis
KW - Violence Risk Factors
KW - Violence Causes
KW - Media Violence Effects
KW - Caucasian Adult
KW - Caucasian Victim
KW - Adult Victim
KW - Cross National Comparison
KW - Cross National Differences
KW - 1940s
KW - 1950s
KW - 1960s
KW - 1970s |