Media Violence and Its Effect on Aggression: Assessing the Scientific Evidence
Freedman, Jonathon L., Crooks, Claire V., Canadian Psychology
JONATHON L. FREEDMAN
Toronto, ON: University of Toronto Press, 2002, 227 pages
(ISBN 0-8020-3553-1 C$50 Hardcover)
Reviewed by CLAIRE CROOKS
Among the tenets that most psychologists embrace as self-evident, the negative effects of media violence on children holds a special place. Virtually every major professional organization concerned with the development of children has issued an unequivocal policy statement about the harmful effects of violent media. In this well-written and informative book, Dr. Jonathon Freedman portrays this stance as one that has gained great momentum and consensus despite inconsistent research findings and widespread methodological flaws. He describes a situation where everyone is citing everyone else's position paper, but the actual scientific literature has received little scrutiny. Some of his findings are shocking - while one well-respected professional organization referred to the "thousands" of studies that indicate the deleterious effects of media violence, Freedman estimates that the actual number of studies is closer to 200, and that the results of these are far from convincing. In this daring and original book, he endeavours to hold the research up under a microscope and, to challenge the conclusions that have been drawn from the literature as a whole. Although he makes many excellent points, in the end he undermines his contribution with faulty logic; if the quality of research has been poor, then the take-home message should be that we simply do not understand the effects of media on children. By arguing that there is justification to reject the hypothesis that media violence is harmful for children on the basis of the literature he criticizes so adeptly, Freedman detracts from the otherwise noteworthy contributions of his book.
In his book, previous research efforts are categorized as one of five types (i.e., survey studies, laboratory experiments, field studies, longitudinal studies, and "communities with versus communities without" studies). Each of these types is evaluated in terms of the evidence it has produced in support of the media violence hypothesis. He outlines his methodology in a clear and transparent fashion, and attempts to give readers enough information to critically evaluate his conclusions. Overall, his analysis of each type is systematic, detailed, and easy to digest.
His evaluation of survey research includes a thorough review of potential research pitfalls, none of which is unique to this field of study. He concludes that there is evidence of a positive correlation between media violence and aggression, but that this relationship is much more modest than typically portrayed.
The section on laboratory experiments suggests that while these studies are typically cited as supporting the media violence hypothesis, a careful review indicates that they do not unanimously support the view of media violence as harmful. Freedman is critical of whether or not "confounds" such as arousal were controlled for in particular studies, but I disagree with the contention that this is necessary, as the arousal response is part and parcel of the violence. While reading this section, I was struck by how dated the studies seemed - the Three Stooges was used as the "aggressive film with real people" condition in one study. Given the graphic and explicit nature of today's media, the antics of Curly, Larry, and Moe seem quaint by comparison. In a time when one of the most popular video games features a storyline where players work for the mafia, shoot police and bystanders, and club prostitutes to death (Grand Theft Auto - 3), it is not clear that these earlier studies are still relevant.
The discussion about field studies, longitudinal studies, and "have" versus "have not" studies follows similar lines of argument to the first two sections. Freedman highlights that although researchers often focus on the one or two significant findings, many more are not significant, or even show an effect in the opposite direction. …