Citation

Grimes T, Bergen L. Am. Behav. Sci. 2008; 51(8): 1137-1154.

Copyright

(Copyright © 2008, SAGE Publishing)

DOI

10.1177/0002764207312008

PMID

unavailable

Abstract

Much of the media violence research that has occurred, principally in the United States, has been based on a fundamental epistemological error: A correlation between variables-- that is, between the consumption of media violence and acted-out aggression--has been inferred by many researchers to be a causal relationship. This article suggests that scholars who infer causation from correlation miss an important point. They propose to rescue a deeply flawed literary corpus by applying statistical principles that do not account for the theoretical and methodological disarray of much of the media violence literature, a disarray that invalidates their probabilistic calculations. This article suggests that rather than argue a causal hypothesis, it is more productive to untangle the conceptual and methodological confusion that hobbles this study area. A by-product could be the dissolution of the causal argument: The putative causal relationship between society's consumption of media violence and social aggression may disappear.