March 29, 2006
Written Statement of
Professor Kevin W. Saunders, J.D., Ph.D.
Michigan State University College of Law
United States Senate Committee on the Judiciary
Subcommittee on the Constitution, Civil Rights, and Property Rights
March 29, 2006
Mr. Chairman, I want to thank you and the Committee for the opportunity to share my thoughts on the shielding of children from violent video games. My name is Kevin Saunders and I am a Professor of Law at Michigan State University. I have spent the last dozen years studying the constitutional issues surrounding attempts to limit the access of children to depictions of extreme violence and other negative media influences. I have been involved, in one way or another in most of the recent round of attempts by state and local governments to limit either arcade play by, or retail sales to, minors of violent video games. While the attempts have thus far been unsuccessful, when challenged in court, see Interactive Digital Software Ass'n v. St. Louis Co., 329 F.3d 954 (8th Cir. 2003); American Amusement Machines Ass'n v. Kendrick, 244 F.3d 572 (7th Cir.), cert. denied, 534 U.S. 994 (2001); Entertainment Software Ass'n v. Blagojevich, 404 F.Supp. 2d 1051 (N.D. Ill. 2005); Video Software Dealers Ass'n v. Schwarzenegger, 401 F.Supp. 2d 1034 (N.D. Cal. 2005); Video Software Dealers Ass'n v. Maleng, 325 F.Supp. 2d 1180, 1186 (W.D. Wash. 2004); see also Entertainment Software Ass'n v. Granholm, 2006 WL 148756 (E.D. Mich. 2006)(refusing to dismiss industry suit) , there are bases on which restrictions may overcome First Amendment limits and protect children from the dangers these products present.
The first two potential bases I will mention only briefly, because they have met with at best limited success in the courts, and later courts may well take these earlier decisions as authoritative. The bases are, however, not completely unavailable, since the Supreme Court has not ruled on the theories involved. One approach is to argue that sufficiently violent material, particularly when presented to children, may come within the obscenity exception to the First Amendment. This theory, argued for in my book Violence as Obscenity, was accepted by the federal district court in the Indianapolis litigation but was rejected by the Seventh Circuit, see American Amusement Machines Ass'n v. Kendrick, 115 F. Supp. 2d 943 (S.D. Ind. 2000), rev'd, 244 F.3d 572 (7th Cir.), cert. denied, 534 U.S. 994 (2001). It is important to note that the Supreme Court has never ruled that violent material cannot be restricted. In the only case directly raising the issue to reach the Court, Winters v. New York 333 U.S. 507 (1948), the statute was struck down as vague, but the Court specifically warned against taking its holding as a conclusion that a properly drawn statute could not stand up to constitutional scrutiny.
The second theory that has had some limited success is that video game play, like the play of pinball machines, is not an activity protected by the First Amendment. This theory was asserted by the district court in the St. Louis litigation but was rejected by the Eighth Circuit, see Interactive Digital Software Ass'n v. St. Louis County, 200 F. Supp.2d 1126 (E.D. Mo. 2002); rev'd, 329 F.3d 954, 959 (8th Cir. 2003). The important distinction here, one not well spelled out by either court, is between the creative activity of the programmer and the communication of the product of that activity to the player on the one hand and the player's playing of the game on the other. This sort of distinction was recently recognized by the Fourth Circuit in another context in Willis v. Town of Marshall, 426 F.3d 251 (4th Cir. 2005). That court distinguished between the creative performance of a band at a community dance and the activity of a dancer on the dance floor. The band was engaged in protected activity, but the dancer was said not to be engaged in communication and not to have the protection of the First Amendment for her sexually provocative dancing. The dancer actually has a better claim to First Amendment protection than the video game player; there are others who saw her performance, while there is no one there to whom the player communicates. This argument, which applies to arcade play rather than retail sales is spelled out in greater detail elsewhere, see Kevin W. Saunders, Regulating Youth Access to Violent Video Games: Three Responses to First Amendment Concerns," 2003 L. Rev. M.S.U.-D.C.L. 51, 93-105.
Lastly, before turning the issue analyzed at greatest length by the courts, I will mention one last, and as yet untested, theory. In a recent book, Saving our Children from the First Amendment, I argue against the assumption that the dissemination of material to children is as protected by the First Amendment as dissemination to adults is unwarranted. The harms to children are greater and the benefits to self-government and autonomy do not apply with the same strength.
The theory to which the courts have paid the most attention is based on the claim that, even if violent video game play by children comes within the protection of the First Amendment, restrictions may stand, based on the danger the games pose. As in other areas of constitutional protection, the infringement of a constitutional right does not mean that the limitations are automatically struck down. Instead, the existence of a constitutional right requires that the restriction meet strict scrutiny. The government must show that the limits are necessary to, or narrowly tailored to, a compelling governmental objective. All the courts to examine the issue of violent video games, or indeed of violent depictions in other contexts, have accepted that the state has a compelling interest in the physical and psychological well-being of youth. The courts have not, however, been willing to find the restrictions at issue in those cases to be necessary to that interest.
Courts have questioned the scientific evidence and have been concerned that, even accepting evidence of correlation between violent video game play and real world violence, correlation does not demonstrate causation of that real world violence. Actually, the evidence is not only correlational. There are experimental and longitudinal studies that more directly demonstrate causation. The courts' unwillingness to find causation is particularly interesting, in the face of this evidence, given the overwhelming consensus of the health and science community that media violence causes real world violence. See, e.g., Amer. Acad. of Pediatrics, et al., Joint Statement on the Impact of Entertainment Violence on Children, Statement to the Cong. Pub. Health Summit (July 26, 2000). Further, the courts' difficulty connecting the correlation demonstrated by empirical studies with causation is despite the fact that the correlation found is stronger than for other areas in which there is no real dispute over causation. As the work of Professors Anderson and Bushman demonstrates, the correlation of media violence with real world violence is as strong as that for second hand smoke and lung cancer, lead exposure in children and lower IQs, use of the nicotine patch and smoking cessation, and asbestos exposure and cancer of the larynx. See Brad J. Bushman & Craig A. Anderson, Media Violence and the American Public: Scientific Facts and Media Misinformation, 56 Am. Psychol. 477, 481 (2001). While this reluctance to accept causation here, despite the acceptance in these other areas, may stem from the seemingly more mysterious processes of brains compared, for example, to lungs, the causal route is becoming more understandable through increased understanding of the development process taking place in the judgment and inhibition region of the adolescent brain and the effect of the environment on that development. See Barbara Strauch, The Primal Teen: What the New Discoveries About the Teenage Brain Tell Us About Our Kids (2003)(presentation of the neuroscience developments in a way accessible to a lay audience).
Despite this agreement in the health and scientific community and the continuing development of the relevant science, the courts have been unwilling to accept these conclusions. It is interesting to note that, among the first three video game cases, the district courts have been more receptive than the appellate courts. In both the Indianapolis and St. Louis litigation, the district court found danger in allowing children to play these games. The district judge in the Washington case also seemed to credit the evidence that video games generally present a danger. See Video Software Dealers Ass'n v. Maleng, 325 F. Supp. 2d at 1188 (The court held that the State had "presented research and expert opinions from which one could reasonably infer that the depictions of violence with which we are constantly bombarded . . . have some immediate and measurable effect on the level of aggression experienced by some viewers and that the unique characteristics of video games . . . makes video games potentially more harmful to the psychological well-being of minors than other forms of media. In addition, virtually all of the experts agree that prolonged exposure to violent entertainment media is one of the constellation of risk factors for aggressive or anti-social behavior . . .") What the judge rejected was the focus of that state's statute on games in which the players shoot law enforcement officers. There was no evidence of special psychological impact for those games. Where the protective effort met its real resistance in that first round of cases was at the appellate level. Judge Posner in holding the Indianapolis ordinance unconstitutional rejected the empirical evidence, concluding that "[c]ommon sense says that the City's claim of harm to its citizens from these games is implausible, at best wildly speculative." American Amusement Machines Ass'n v. Kendrick, 244 F.3d 579. This is in sharp contrast to the St. Louis district court judge's statement that "For plaintiffs to . . . argue that violent video games are not harmful to minors is simply incredible." Interactive Digital Software Ass'n v. St. Louis Co., 200 F. Supp. 2d at 1138.
Of course, this is not simply a matter of common sense. The issue of danger is a matter of science and the courts will need to understand that science. As indicated, some courts have accepted that science. The most recent examinations of the issue, perhaps affected by the appellate court reversals of the Indianapolis and St. Louis district court opinions, see Entertainment Software Ass'n v. Blagojevich, 404 F. Supp. 2d at 1059 n.3 (quoting the Seventh Circuit's rejection of the evidence in the Indianapolis litigation), have been less receptive. In the Illinois litigation, Judge Kennelly did, to his credit, examine the evidence with some thoroughness. The Blagojevich opinion expressed concern over effect size, that is the size of the increase shown in testing the effect of exposure to violent media. While that effect size might well indicate that a normal child will not become homicidal after playing a violent video game, no one makes that claim. Even a small effect size acting on the whole range of susceptibilities may have serious consequences in the more susceptible population.
Judge Kennelly also expressed concern over the size of the community of those studying the issue and the relationships among the scientists. He noted that, of the seventeen research articles relied on by the Illinois General Assembly, fourteen were authored or co-authored by Professor Craig Anderson, one by a colleague of Professor Anderson, and two by a scientist who relied on Professor Anderson's research in designing his own studies. This concern might be eased by recognizing that the articles all survived peer review, but the concern might simply transfer to the peer review process and the small community from which referees might be drawn. It should, however, be noted that Professors Anderson's and Bushman's meta-analysis of the research in the field included studies by a significant number of scientists unaffiliated with Professor Anderson. See Craig A. Anderson & Brad J. Bushman, Effects of Violent Video Games on Aggressive Behavior, Aggressive Cognition, Aggressive Effect, Physiological Arousal and Prosocial Behavior: A Meta-Analytic View of the Scientific Literature, 12 Psychol. Sci. 353 (2001): Craig A. Anderson, An Update on the Effects of Playing Violent Video Games, 27 J. Of Adolescence 113 (2004). While these concerns of the court do not even currently seem valid, the continuing development of this area of scholarship and the attention paid by an increased number of scientists should eventually overcome the perceived shortcoming.
A second complaint may reflect less understanding of statistical inferences. The court noted that not all studies show a correlation between media violence and real world aggressiveness and that some even demonstrate a negative correlation. Correlation does not require that every sample demonstrate the same relationship. Differences in results from relatively small samples in particular are likely to occur, as indeed they do in studies of smoking and lung cancer. Such studies may be combined through meta-analysis to reach a conclusion across the body of research in a way conceptually similar to a comparison of season batting averages being a better basis for comparing players than their performances in any particular game.
Judge Kennelly also rejected the legislature's reliance on a functional magnetic resonance imaging study that purports to show a decrease among children exposed to violent media in the activity of the region of the brain that controls behavior. The video game industry offered an expert who suggested that decreased activity in the part of the brain studied could indicate an alternative method for accomplishing the work of that region. Even if true, the study did seem to show a difference between the control group and those who had more exposure to violence, and it was a difference that matched that for those with disruptive behavioral disorders. There are also several other studies that have found similar results in brain activity, using different experimental methods.
I am not a scientist and do not have the understanding of the issues that others testifying today have. It also seems likely that none of the judges involved in the cases have been scientists. I, legislators, and the judiciary, must be open to recognizing the continuing development of this area of research. From a legal point of view, it is important to note that the courts' decisions on the scientific issues can have no long term precedential effect. Unlike any conclusions of law, which may have precedential value or at least be taken as authority in future cases, the conclusions on the science are contingent. A conclusion by an earlier court that the science fails to establish the danger perceived by the public and the legislature is only a conclusion that the science at that time was lacking. It does not establish the conclusion that the science at the time of any future legislation or litigation is also lacking. Each time the issue arises, the courts must consider the science anew.
Lastly, returning to the sort of argument offered in Saving Our Children from the First Amendment, I think it important to consider the costs of the two possible errors here. If, in fact, violent video games are causing an increase in aggressiveness and real world violence and the courts refuse to allow limitations, the cost is psychologically, and perhaps neurologically, damaged children and, in the extreme, the deaths of the victims of that real world violence. The other possible error would be to allow restrictions, when media violence does not in fact have the effects suggested by the science. The costs there would seem to be to the values behind the First Amendment's expression clauses. But, the most important values behind protecting expression are those tied to self-governance. To be self-governing, we must have access to information, but children do not vote. True, as Judge Posner pointed out in the Indianapolis litigation, they need to be competent voters when they turn eighteen. That is why I have counseled legislators to limit only those under seventeen. That allows seventeen year olds a year to play as many violent video games as it takes to become a competent voter. The other strongest value thought by some to underlie the Expression Clauses is the autonomy found in making one's own decisions as to what to read, see or perhaps play. But, we do not really believe in autonomy for children, or we would allow them to smoke, drink and engages in any number of vices legal for adults.
I hope that legislatures will continue in their efforts to protect children from this serious danger. Absent a Supreme Court decision on the issues, at least some lower courts may consider the constitutional theories suggested. Even with a negative Supreme Court opinion on all the positions asserted, a failure to find adequate science at any one point does not bar relitigation at a later point. Despite past losses, as the science continues to develop, the effort can continue and the danger theory is never permanently dismissed.