Note
2021-08-12: It has come to my attention that when this paper appeared in print--some months after I wrote this post--it appeared with a note from the editor:
"After the original version of this manuscript was published online, it was found to contain substantial personal commentary that went beyond what was justified by the results of the study and the state of the literature. Because our primary focus is on results of original research and their direct implications, the additional commentary was deemed to be out of place and was removed from the print version of the manuscript (Ferguson, C. J. [2015]. Does Movie or Video Game Violence Predict Societal Violence? It Depends on What You Look at and When. Journal of Communication, 65(1), 193–212. doi:10.1111/jcom.12142)."
At least some of my concern, then, seems to have been justified. A quick glance indicates that the alteration amounted to reducing the "Policy implications" section, which occupied somewhat more than a page in the original, to a brief paragraph in the "Theoretical implications" section (now renamed simply "Implications").
My attention was directed to a paper by Christopher J. Ferguson, to be published in the Journal of Communication, which argues that the common belief that there is a positive correlation between media violence and societal violence is false. The abstract:
This article presents 2 studies of the association of media violence rates with societal violence rates. In the first study, movie violence and homicide rates are examined across the 20th century and into the 21st (1920–2005). Throughout the mid-20th century small-to-moderate correlational relationships can be observed between movie violence and homicide rates in the United States. This trend reversed in the early and latter 20th century, with movie violence rates inversely related to homicide rates. In the second study, videogame violence consumption is examined against youth violence rates in the previous 2 decades. Videogame consumption is associated with a decline in youth violence rates. Results suggest that societal consumption of media violence is not predictive of increased societal violence rates.
Since the full article is about twenty pages, I'll summarize it here, with a bit more detail than the abstract.
Ferguson begins by establishing that there is not general agreement in the literature as to correlation (much less causation) between media violence and societal violence. He particularly argues that that laboratory studies which show an increase in aggression (e.g. an increased tendency to fill in missing letters in words to create violent words) following consumption of violent media may not correspond to increased societal violence (e.g. homicide).
The first study analyzes the relationship between violence in movies and homicide, between 1920 and 2005. To assess violence in movies, the top five grossing movies per year were taken at five-year intervals, and graded for violent content. This data was compared to statistics for murder and (non-negligent) homicide during the same years. Ferguson concludes that:
Taken together these data suggest that perceived correlations between movie and societal violence were associated with a chance concordance during the mid-20th century. Given that these phenomena were not in concordance in either the early or latter 20th century, it appears that efforts to establish causal connections between movie and societal violence based on a select set of decades were an ecological fallacy.
The second study analyzes the relationship between violence in video games and youth violence for the years 1996 to 2011. Ferguson uses sales data from IMDb and the ESA and takes the ratings assigned by the ESRB as a proxy for violent content. He concludes that there is a "remarkably strong" inverse relationship between violent video game consumption and youth violence and that, although this is no indication of causality, "Evidence from societal data does not support claims of dramatic videogame violence effects on violence among youth."
Finally, Ferguson argues that alternate methods should be used in the future when assessing the effect of media consumption on behavior, and that professional organizations such as the APA should "retire their policy statements on media violence as such statements tend to be misleading and may cause more harm than good." Furthermore, "Arguably, given that the results from the effects paradigm have been weak and inconsistent, it may be time for scholars to make less rather than more conclusive statements to news media regarding media effects on society."
Ferguson's conclusion in the first study (namely that positive correlation between media violence and societal violence is limited to a mid-century concordance) seems undeniable. I am not similarly satisfied with the second study.
The most obvious problem with the second study is that it addresses a period of only sixteen years, and in particular that it does not even consider whether the change in rate of youth violence differed before and after the introduction of video games. Given that Ferguson strongly criticized reliance on data from a limited time span in the first study, this is fairly damning.
Second, the second study does not argue that generally increased video game consumption corresponds to an increase in the consumption of violent video games by youths. I freely admit that it is likely that this is true, but the study makes no mention of it. In general, I am not convinced by Ferguson's 'just look at the top five games according to IMDb' approach to analyzing the state of gaming.
Third, the study blithely dismisses any correlation between increased incarceration and decreased youth violence, on strength of a paper by Stahlkopf, Males, and Macallair. Based on its abstract, that paper only concludes that increased juvenile incarceration does not lead to reduced crime (specifically in California). It is possible that the full paper has more useful (for Ferguson) conclusions, but I do not have access to it.
Fourth, the graphs in the paper are unclear as to what exactly is depicted. Figure 3, "Societal videogame violence consumption and societal youth violence, 1996-2011.", is a double chart with yearly figures for the two measures in question, with scales ranging from 0-9000 for one and 0-40 for the other. 9000 what? 40 what? I have no idea.
Fifth, without commenting on the sensibility of the suggestions Ferguson makes, I would say that his statements in the "Theoretical implications" section are not supported by the studies. Ferguson argues that different theoretical approaches should be used in studying the relationship between media violence and societal violence, essentially by mentioning a variety of alternatives without arguing for their particular effectiveness. Such arguments would be better suited to a paper which actually considers the relative merit of the various approaches.
Sixth, the "Policy implications" section reads like a conspiracy theory. Selections:
As a matter of policy, consistent with the statement by the Consortium of Scholars (2013)...
...media-based policy statements released by professional organizations have so often been revealed to be flawed (e.g., Ferguson, 2013;Magid, 2011).
...it has been revealed that past policy statements were typically developed by specially selected researchers heavily invested in antimedia views, with no dissenting voices (Ferguson, 2013).
That is, problems have been 'revealed' by the author of this paper. The 'statement by the Consortium of Scholars' mentioned is an open letter of which Ferguson is a signatory, and the source of which is Ferguson's personal website.1
The reference to Magid links to a paper published in Pediatrics (by O'Keeffe and Clarke-Pearson), but the actual reference listed is to an article in Huffington Post which responds to (and disputes) the journal article. Why Ferguson would obfuscatingly link to the article in Pediatrics rather than the one in HuffPo to which he actually referred is a mystery.
Additionally, take note of "...some scholars have argued...(Males, 2013)." That reference is not to a scholarly paper, but what seems to be simply an opinion piece published by YouthFacts.org.2
The results of the actual studies in the paper seem fairly inoffensive, but do not support the discussion. It seems very like Ferguson wanted to pontificate about perceived problems in his field, and conducted a couple of very minor studies in order to have an excuse for publication.
First, I have not studied psychology. I studied math at uni, so except insofar as the paper uses statistical analyses, I have no formal education relevant to this paper.
Second, as I do not have access to a university library, I have not been able to review all of the paper's references. It is possible that some or all of my objections were answered by the cited papers.
Third, I am (obviously) a gamer, and inclined to agree with the results of the studies in this paper, despite my objections to the academic rigour of the paper. It is possible that I have been too lenient with the studies for this reason.
Which, incidentally, looks more like the geocities homepage of a teenager than the website of a professional academic. ↩
And a broken link, besides--here's a working one. ↩