The Supreme Court’s decision in Brown v. EMA held that California’s attempt to restrict minors’ access to violent video games violates the First Amendment. The law could not satisfy the strict scrutiny requirement of being “justified by a compelling government interest and  narrowly drawn to serve that interest.” (p. 11). In reaching this conclusion, the Court gave little credence to the scientific literature on video game violence, which claims “that exposure to violent video games causes an increase in the likelihood of aggressive behavior.” Appendix to Brief of Senator Leland Yee. Did the justices just not get the science? At least in part, they probably did not. It’s not the case that “nearly all of the research is based on correlation” rather than evidence of causation. (p. 13). Most of the research wasn’t in the record, however, either because lawyers think difficult research questions can be resolved with a small number of studies or because they think judges can’t or won’t read a large number of studies. (The latter concern may be accurate; judges are busy.) Nevertheless, the Court did recognize a serious problem with the scientific literature’s support for California’s law, and I agree that the Court reached the right outcome.
There are two basic views about when minors’ access to certain violent media may be regulated by the government. One is when the violence is obscene for minors. In Kendrick, which was probably the video game industry’s favorite decision until today, Judge Posner resisted treating violence as obscene, but still said, “If the [video] games used actors and simulated real death and mutilation convincingly, or if the games lacked any story line and were merely animated shooting galleries (as several of the games in the record appear to be), a more narrowly drawn ordinance might survive a constitutional challenge.” Am. Amusement Mach. Ass’n v. Kendrick, 244 F.3d 572, 579-80 (7th Cir. 2001). As games become more photorealistic, this suggestion in Kendrick could have come back to haunt the industry. I don’t see how this could happen now. Brown clearly says only sexual material can qualify as obscene. (pp. 5-10). The view that violence can be considered obscene is dead.
The other view about when access may be regulated is when the video game violence causes harm to minors. This view requires an assessment of the scientitic literature, but what to do in the face of scientific uncertainty? The lower courts consistently held that strict scrutiny applies to these video game regulations. The Supreme Court agreed. At the same time, the lower courts sometimes invoked the standard in Turner Broadcasting, which calls for courts to be deferential to legislative findings because legislatures are in a better position to analyze “vast amounts of data.” Turner Broad. Sys. v. FCC, 520 U.S. 180, 195 (1997). But the lower courts cited Turner’s standard without actually deferring to any state legislatures’ judgments about the video game violence literature. Brown clarifies that Turner’s standard is for intermediate scrutiny situations and does not apply to things like California’s video game law. (p. 12). So the lower courts were right not to be deferential (and probably right if they were unsure about Turner’s applicability).
But if the courts are not deferential to the legislature’s reading of the science, how certain must they be about the relationship between video game violence and harm to minors? The scientific evidence will always leave some uncertainty. Although the Supreme Court did not acknowledge it, there was a circuit split on this question. According to the Eighth Circuit, what is required is “incontrovertible proof of a causal relationship between the exposure to [video game] violence and subsequent psychological dysfunction” and “statistical certainty of causation.” Entm’t Software Ass'n v. Swanson, 519 F.3d 768, 772 (8th Cir. 2008). The Eighth Circuit recognized that this standard “may reflect a refined estrangement from reality,” but claimed binding circuit precedent required it. (It didn’t.) Understandably, the Ninth Circuit disagreed: “Although we do not require the State to demonstrate a ‘scientific certainty,’ the State must come forward with more than it has.” Video Software Dealers Ass’n v. Schwarzenegger, 556 F.3d 950, 964 (9th Cir. 2009).
Under the more realistic standard of the Ninth Circuit, the Supreme Court still got it right, even if you accept the consensus of the scientists who study media violence—i.e., that it causes aggression or other negative effects. For at least three related reasons, the law was not narrowly drawn to serve a compelling interest.
Reason No. 1
California’s statute targeted violence that a reasonable person would think “appeals to the deviant or morbid interest of minors,” that is “patently offensive” to community standards “as to what is suitable for minors,” and that “as a whole” lacks “serious literary, artistic, political, or scientific value for minors.” (p. 1). This is most likely to be extreme or graphic violence, not cartoonish or sanitized violence. As an additional restriction, the act covered only violence towards humans. Extreme violence towards, say, Vulcans was not included. (Oral Argument Transcript at 59). Apparently, the law could have been avoided by video game publishers simply declaring that every character in a game is a humanoid looking alien.
The media violence research is aimed at violence that is defined more broadly than just graphic violence, and it includes non-humans. While the definitions of media violence vary in the literature, one prominent definition is that it “refers to media depictions of aggressive and violent behavior directed at characters in the media story. Those characters can be human or nonhuman, cartoonish or visually realistic. Fictional, unrealistic, or animated violence is still considered violence if it meets the above definitions.” Gentile, Saleem & Anderson, Public Policy and the Effects of Media Violence on Children, 1 Social Issues and Policy Review 15, 17 (2007).
The public’s definition of violence is likely different. According to Gentile, Saleem, and Anderson, “Research reveals that the public tends to focus on the graphic aspects and intensity of a scene in determining its level of violence, contrary to scientific and legal definitions. Ironically, research has shown that children’s cartoons have the highest level of violence, but the public does not view cartoons as violent.” Id (citations omitted). See also Anderson, Gentile & Buckley, Violent Video Game Effects on Children and Adolescents 99 (Oxford 2007).
With such a broad approach to violence, media researchers find violence where most members of the public probably would not. Anderson and Ford describe Zaxxon (1982) as “somewhat violent” and “highly aggressive.” Anderson & Ford, “Affect of the Game Player,” 12 Personality and Social Psychology Bulletin 390, 395 (1987). In another study, the authors calculate the “percentage of game play depicting violence” in Galaxian (1979) and Galaga (1982) to be 100%, with an average of 34.5 and 62.1 deaths per minute, respectively. See Thompson & Haninger, “Violence in E-Rated Video Games,” 286 JAMA 591, 596 (2001). In still another study, Anderson uses Super Mario Brothers (1985) as an example of violence in children’s games. Anderson, “Violent Video Games and Aggressive Thoughts, Feelings, and Behaviors” in Children in the Digital Age 101 (Praeger 2002). The Court mentioned Sonic the Hedgehog as an example of a game defined as violent. (p. 13). Violence in games is pervasive, they say.
Media violence researchers are not entirely consistent on this point and sometimes they seem to accept that violence is really about what is extreme or graphic. Surprisingly, researchers sometimes rely on their research subjects’ self-reports of exposure to video game violence even though members of the public are likely to define violence differently than the researchers. See, e.g., Anderson, Gentile & Buckley at 81-82, 99. Anderson says, “Truly violent video games came of age in the 1990s[.]”). Anderson (2002) at 102. More recently, Anderson, Gentile, and Buckley similarly claim that “video games didn’t become very violent until the early 1990s.” Anderson, Gentile & Buckley at 16. Yet older and more repetitious games like Space Invaders and Galaxian involve a greater percentage of game time devoted to violence than more modern games like Grand Theft Auto: San Andreas.
(Anderson, Gentile, and Buckley also put the ultra-violent game Manhunt on the cover of their book, rather than something like the more recognizable but supposedly violent Super Mario Brothers.)
Despite some apparent inconsistencies, the bottom line is that “[p]rofound differences exist between how social scientists conceptualize violence and how the public sees it.” Potter, On Media Violence 75 (Sage 1999). Plenty of modern games and even plenty of pre-1990 games that the public would consider safe for children are in fact violent, if we use the definitions of media violence researchers. According to the research, these games may therefore have negative effects on children, but no politician would ever try to restrict access to Galaxian or Super Mario Brothers.
The California statute is tied to the public’s standards, not the standards of media violence researchers. Suppose you accept the consensus view of media researchers that media violence is a cause of aggression, not the single cause of aggression in the world, but one of many causes (and maybe even the very difficult to accept view that the violence in a game like Galaxian contributes to aggression in society). Statutes like the California one reinforce the view that graphic violence is the problem when media violence researchers actually think a broader category of violent media should be of concern. The Supreme Court noted this problem by describing the statute as “wildly underinclusive when judged against its asserted justifications.” (pp. 13-14). Given this, I’m not sure why media violence researchers are sympathetic to these statutes, as some seem to be, and I don’t know where media violence researchers have addressed what looks like a serious disconnect between the findings of their research and the types of games targeted by these statutes.
The problem of underinclusiveness, however, is more complicated than the Court’s discussion reveals. While the science suggests that more than just graphic violence is an issue, not all violence is equal.
Reason No. 2
Although the media violence literature is aimed at violence, broadly defined, media violence researchers do accept that the context in which violence is portrayed is relevant. Cartoonish violence and graphic violence might have different effects, but the research is less advanced when it comes to many claims beyond the general one that media violence (broadly defined) causes aggression (broadly defined). For example, showing the consequences of violence may desensitize viewers, but it may also (or instead) demonstrate that violence is an undesirable way to resolve problems. Potter & Smith, “The Context of Graphic Portrayals of Television,” Violence, J. of Broadcasting & Electronic Media 301, 319 (2000).
Although there is research on the role of context or “moderator” variables, as recently as 2008 one study noted that “researchers have not yet answered a simple question: What kinds of violent video games are problematic for children?” Shibuya, Sakamoto, Ihori & Yukawa, “The Effects of the Presence and Contexts of Video Game Violence on Children: A Longitudinal Study in Japan,” 39 Simulation & Gaming 528, 528 (2008). See also Barlett, Anderson & Swing, Video Game Effects--Confirmed, Suspected, and Speculative, 40 Simulation & Gaming 377, 383-386 (2009). This acknowledgment is important. A requirement for surviving strict scrutiny is that the government’s solution is supposed to be geared towards solving a problem. (pp. 11-12).
Despite the confidence of media violence researchers that media violence causes aggression, they should be far less confident about the effects of violence in different contexts and about how different variables interact, e.g., cartoonish versus graphic violence, humorous versus non-humorous violence, violence in fanciful versus realistic settings, violence that is rewarded versus punished, etc., etc. Kevin Saunders, who argued in favor of reversing the Ninth Circuit, acknowledges that figuring out the role of many different variables on viewers is likely to be very difficult. He thinks it might be impossible. Kevin Saunders, Violence as Obscenity 44 (Duke 1996). Media violence researchers are actively interested in these contextual questions, but a very large number of studies will be needed to sort out the role of assorted contextual variables, and the existing number of studies on particular variables remains relatively small.
California’s statute was not tied to a substantial body of research about the effects of specific types of violence. The statute was instead aimed at what the public typically finds offensive. What the public finds offensive and what is actually problematic might not be the same thing. In some ways, the statute was wildly underinclusive, but how much so is not clear. It may have been overinclusive in some ways too. More research is needed, but it’s unlikely any statute will ever be broad enough to sweep up games like Galaxian, Super Mario Brothers, or Sonic the Hedgehog.
Reason No. 3
Insofar as the government’s solution is supposed to solve a problem, it’s not clear that restrictions on the sale of violent video games will actually reduce minors’ exposure to video game violence. During the past decade, retailers have become more vigilant over time about selling M-rated games to minors. FTC, Marketing Violent Entertainment to Children 28 (6th ed. 2009). I’m not aware of any research evaluating whether minors’ exposure to M-rated games has actually decreased over this period of time. The courts have not discussed any such research. Minors can still get games from friends or parents. It may be that these retail restrictions have accomplished little to nothing other than to insulate the video game industry from criticism and to make some parents feel better because “something” should be done and these privately enforced restrictions are something.
So is this the end of legislative efforts to regulate minors’ access to violent video games? A substantial increase in the amount, specificity, and persuasiveness of the scientific research is one way the issue may arise again, but the Supreme Court left open another possibility in footnote 3. The Court said, “[P]erhaps . . . the state has the power to enforce parental prohibitions -- to require, for example, that the promoters of a rock concert exclude those minors whose parents have advised the promoters that their children are forbidden to attend.” (p. 7). Politicians do like registries, and I suppose this footnote leaves legislatures with a Round 2 option in the form of some sort of retail opt-out registry. Parents could register their children as being disallowed from buying certain games. Politicians might create one big registry or try to make each retail chain establish its own. On the other hand, perhaps politicians will just move on, given the lack of success so far and the lack of success with the V-chip (a very individualized opt-out program).
Justice Breyer claimed in dissent that this case was more about education than censorship. He said we need to raise future generations in a way that facilitates constitutional self-government, which requires a committment to public cooperation. But does he really see the value in things like restrictions on video game purchases in terms of promoting public cooperation? By teaching children that choices are sometimes made for them “by the people acting democratically through their governments”? (Dissent p. 19).
Children are already likely to grow up well aware of how their governments will make numerous choices for them about numerous things. Occasionally, we should also remind future generations that there are some limits.
Note: A previous entry on video game violence in court is here.