Does media violence lead to real violence, and do video games impair academic performance?

Cross-posted from the University of Michigan Press blog.

"Twilight of the Books," an essay of mine published in The New Yorker on 24 December 2007, has been honored by inclusion in The Best of Technology Writing 2008, edited by Clive Thompson. When The New Yorker published my essay, I posted on my blog a series of mini-bibliographies, for anyone who wanted to dig into the research behind my article and try to answer for themselves whether television impaired intellect or whether literary was declining (here's an index/overview to all these research posts). A month or so ago, when the University of Michigan Press, the publisher of The Best of Technology Writing 2008, invited me to write about my essay for their blog, I was afraid I didn't have any more to say. Also, alas, I was under deadline. But I have a breather now, and looking over my year-old notes, I realize that there were a couple of categories of research that I never posted about at the time, because the topics didn't happen to make it into my article's final draft.

This research tried to answer the questions, Does exposure to violence on television or in video games lead to aggressive behavior in the real world? and Do video games impair academic performance? I still think the questions are very interesting, though I must now offer my summaries with the caveat that they are somewhat dated. In fact, I know of some very interesting research recently published on the first question, some of which you can read about on the blog On Fiction. I'm afraid I haven't kept up with video games as closely, but I'm sure there's more research on them, too. I hope there is, at any rate, because when I looked, I found very little. (By research, in all cases, I meant peer-reviewed studies based on experimental or survey data, and not popular treatments.)

A few words of introduction. The historian Lynn Hunt has suggested in her book Inventing Human Rights that in the eighteenth century, the novel helped to change Europe's mind about torture by encouraging people to imagine suffering from the inside. As if in corroboration, some of the research summarized below suggests that the brain responds less sympathetically when it is perceives violence through electronic media. As you'll see, however, there is some ambiguity in the evidence, and the field is highly contested.

1. Does exposure to violence on television or in video games lead to aggressive behavior in the real world?

  • In a summary of pre-2006 research, John P. Murray pointed to experiments in the 1960s by Albert Bandura, showing that children tend to mimic violent behavior they have just seen on screen and to a number of studies in the early 1970s that found correlations between watching violence and participating in aggressive behavior or showing an increased willingness to harm others. In 1982, a panel commissioned by the Surgeon General to survey existing research asserted that "violence on television does lead to aggressive behavior," and in 1992, a similar panel commissioned by the American Psychological Association reported "clear evidence that television violence can cause aggressive behavior." One mechanism may be through television's ability to convince people that the world is dangerous and cruel, in what is known as the "mean world syndrome." Murray claims that a twenty-two-year longitudinal study in Columbia County, New York, run by Huesmann and Eron, which was begun under the auspices of the Surgeon General's office, has linked boys' exposure to television violence at age eight to aggressive and antisocial behavior at age eighteen and to involvement in violent crime by age thirty; in fact, a 1972 study by Huesmann et al. did link boys' exposure at eight to aggressive behavior at eighteen, but the 1984 study cited by Murray linked violent crime at age thirty to aggressive behavior at age eight and said nothing about exposure to televised violence. In an unrelated study, when television was introduced in Canada, children's levels of aggression increased. [John P. Murray, "TV Violence: Research and Controversy," Children and Television: Fifty Years of Research, Lawrence Erlbaum Associates, 2007. L. Rowell Huesmann, Leonard D. Eron, Monroe M. Lefkowitz, and Leopold O. Walder, "Stability of Aggression Over Time and Generations," Developmental Psychology 1984. For a synopsis of Huesmann's 1972 study, see Steven J. Kirsh, Children, Adolescents, and Media Violence: A Critical Look at the Research, Sage Publications, 2006, p. 208.]
  • A longitudinal study of 450 Chicago-area children was begun in 1977 when the children were between six and eight years old, and continued in 1992-1995, when they were between twenty-one and twenty-three years old. As children, the subjects were asked about their favorite television programs, whether they identified with the characters, and how true-to-life they thought the shows were. Fifteen years later, it emerged that watching violent shows, identifying with aggressive characters of the same sex, and believing that the shows were realistic correlated with adult aggression, including physical aggression. The effect was present even after controlling for such factors as initial childhood aggression, intellectual capacity, socioeconomic status, and parents' level of emotional support. (Note that in the opinion of the researchers, the Six Million Dollar Man was considered a "very violent" show, and that the heroine of the Bionic Woman was considered an aggressive character.) [L. Rowell Huesmann, Jessica Moise-Titus, Cheryl-Lynn Podolski, and Leonard D. Eron, "Longitudinal Relations between Children's Exposure to TV Violence and Their Aggressive and Violent Behavior in Young Adulthood, 1977-1992," Developmental Psychology, 2003. Cf. Kirsh , p. 209.]
  • In a 2006 textbook about the relation between media violence and aggressive behavior, author Steven J. Kirsh notes that a 1994 meta-analysis of the link between television violence and aggression estimated the size of the effect to be r = .31. "The effect sizes for media violence and aggression are stronger than the effect sizes for condom use and sexually transmitted HIV, passive smoking and lung cancer at work, exposure to lead and IQ scores in children, nicotine patch and smoking cessation, and calcium intake and bone mass," Kirsh wrote. A 2004 meta-analysis found that the correlation between video game violence and aggressive behavior was r = .26. To put the effect sizes in perspective, Kirsh notes that they are greater than the link between testosterone levels and aggression, but weaker than the link between having antisocial peers and delinquency. In surveying the research on video games, Kirsh makes the point that there is little research as yet, and that most of it was done in what he calls the "Atari age," when the games were fairly innocuous; almost no one has experimentally tested the effects on children and teens of the new-generation, highly realistic and gory first-person shooter games. [Steven J. Kirsh, Children, Adolescents, and Media Violence: A Critical Look at the Research, Sage Publications, 2006.]
  • In a 2007 summary of research, three scientists asserted that there was "unequivocal evidence that media violence increases the likelihood of aggressive and violent behavior in both immediate and long-term contexts," and noted that the link between television violence and aggression had been proved by studies in both the laboratory and the field, and by both cross-sectional and longitudinal studies. Video games were not as well documented, but in the opinion of the scientists, the preliminary evidence suggested that their effect would be similar. Playing violent video games has been shown to increase physiological arousal. Measurements of skin conductance and heart rate show that people have less of an aversion to images of real violence, if they have previously been exposed to violent television or violent video games. Measurements of event-related brain potentials (ERPs) and functional magnetic resonance imaging (fRMI) allow researchers to look with new precision at the magnitude of brain processes that occur at particular times and at the activation of specific regions of the brain. A 2006 study by Bartholow et al., for example, showed that exposure to violent video games reduces aversion to scenes of real violence, as measured by a blip of voltage that typically occurs 300 milliseconds after sight of a gory image. A 2006 study by Murray et al. (see below) showed that violent scenes of television activated parts of the brain associated with emotion, memory, and motor activity. Yet another 2006 study, by Weber et al., showed that while players were engaged in violence during a video game, a brain region associated with emotional processing was suppressed, and one associated with cognitive processing was aroused, perhaps in order to reduce empathy and thereby improve game performance. In a 2005 study by Matthews et al., chronic adolescent players of violent video games scored the same as adolescents with disruptive behavior disorders on a test designed to assess a brain region responsible for inhibition and error correction. Attempting to explain the results of the various studies under review, the authors write: "Initial results suggest that, although video-game players are aware that they are engaging in fictitious actions, preconscious neural mechanisms might not differentiate fantasy from reality." [Nicholas L. Carnagey, Craig A. Anderson, and Bruce D. Bartholow, "Media Violence and Social Neuroscience," Currents Directions in Psychological Science, 2007.]
  • While a functional magnetic resonance imaging (fMRI) device monitored their brain activity, eight children watched a video montage that included boxing scenes from Rocky IV and part of a National Geographic animal program for children, among other clips. The violent scenes activated many brain regions that the nonviolent scenes did not, mostly in the right hemisphere. These regions have been associated by other researchers with emotion, attention and arousal, detection of threat, episodic memory, and fight or flight response. The authors of the study speculate that "though the child may not be aware of the threat posed by TV violence at a conscious level . . . a more primitive system within his or her brain (amygdala, pulvinar) may not discriminate between real violence and entertainment fictional violence." In the activation of regions associated with long-term memory, the researchers saw a suggestion that the television violence might have long-term effects on the viewer. [John P. Murray, etal. "Children's Brain Activations While Viewing Televised Violence Revealed by fMRI," Media Psychology, 2006.]
  • In a 2005 study, 213 video-game novices with an average age of twenty-eight were divided into two groups, and one group spent a month playing an average of 56 hours of a violent multi-player fantasy role-playing video game. Participants completed questionnaires to assess their aggression-related beliefs before and after the test month, and were asked before and after whether they had argued with a friend and whether they had argued with a romantic partner. The data showed no significant correlation between hours of game play and the measures of aggression, once the results were controlled for age, gender, and pre-test aggression scores. The authors note that there might be an effect too small for their study to detect, and that adults might be less sensitive to the exposure than children or adolescents. [Dmitri Williams and Marko Skoric, "Internet Fantasy Violence: A Test of Aggression in an Online Game," Communication Monographs, June 2005. Andrea Lynn, "No Strong Link Seen Between Violent Video Games and Aggression," News Bureau, University of Illinois at Urbana-Champaign, 9 August 2005.]
  • A 2007 book presented three studies of video-game violence's effect on school-age children. In the first study, 161 nine- to twelve-year-olds and 354 college students were asked to play one of several video games—either a nonviolent game, a violent game with a happy and cartoonish presentation, or a violent game with a gory presentation—and then to play a second game, during which they were told they could punish other player with blasts of noise (the blasts were not, in fact, delivered). Those who played violent games, whether cartoonish or gory, were more likely to administer punishments during the second game; playing violent games at home also raised the likelihood of punishing others. Children and college students behaved similarly. In the second study, 189 high school students were given questionnaires designed to assess their media usage and personality. The more often the students reported playing violent video games, the more likely they were to have hostile personalities, to believe that violence was normal, and to behave aggressively, and the less likely they were to feel forgiving toward others. The correlation between game playing and violent behavior held even when the researchers controlled for gender and aggressive beliefs and attitudes. The more time that students spent in front of screens (whether televisions or video games), the lower their grades. In the third study, 430 elementary school children were surveyed twice, at a five-month interval, about their exposure to violent media, beliefs about the world, and whether they had been in fights. Students were asked to rate one another's sociability and aggressiveness, and teachers were asked to comment on these traits and on academic performance. In just five months, children who played more video games darkened in their outlook on the world, and peers and teachers noticed that they became more aggressive and less amiable. The effect was independent of gender and of the children's level of aggression at the first measurement. Screen time impaired the academic performance of these students, too; they only became more aggressive, however, when the content they saw during the screen time was violent. [Craig A. Anderson, Douglas A. Gentile, and Katherine E. Buckley, Violent Video Game Effects on Children and Adolescents: Theory, Research, and Public Policy, Oxford University Press, 2007.]

2. Do video games impair academic performance?

  • In a 2004 survey of 2,032 school-age children, there were statistically significant differences in print and video-game use between students earning As and Bs and those earning Cs and below. On average, A-B students had read for pleasure 46 minutes and played video games for 48 minutes the previous day; C-and-below students had read for pleasure 29 minutes and played video games for 1 hour 9 minutes. Television watching seemed constant between the groups. [Donald F. Roberts, Ulla G. Foehr, and Victoria Rideout, Generation M: Media in the Lives of 8-18 Year-Olds, The Henry J. Kaiser Family Foundation, March 2005, page 47.]
  • A 2007 book presented results of a study in which 189 high school students were given questionnaires designed to assess their media usage and personality. The more time that students spent in front of screens (whether televisions or video games), the lower their grades. In a related and similar study, 430 elementary school children were surveyed twice, at a five-month interval, and screen time impaired the academic performance of these students, too. [Craig A. Anderson, Douglas A. Gentile, and Katherine E. Buckley, Violent Video Game Effects on Children and Adolescents: Theory, Research, and Public Policy, Oxford University Press, 2007.]

UPDATE (27 Feb. 2009): For ease in navigating, here's a list of all the blog posts I wrote to supplement my New Yorker article "Twilight of the Books":

Notebook: "Twilight of the Books" (overview)
Are Americans Reading Less?
Are Americans Spending Less on Reading?
Is Literacy Declining?
Does Television Impair Intellect?
Does Internet Use Compromise Reading Time?
Is Reading Online Worse Than Reading Print?
I also later talked about the article on WNYC's Brian Lehrer Show and on KUER's Radio West.
And, as a bonus round: Does media violence lead to real violence, and do video games impair academic performance?


“The Terror Last Time,” my article about the 1886 trial of Chicago’s Haymarket anarchists, which is in part a review of James Green’s new book Death in the Haymarket, is published in the 13 March 2006 New Yorker. As it happens, there are many Haymarket resources on the web, so I thought I’d link to a few of them. What follows will seem a little scattered unless you read my article first (ahem), but if you’ve done that, then . . .

If you want to read the witnesses’ testimony yourself, the Chicago Historical Society has published the trial transcript in the Haymarket Affair Digital Collection. The collection has all sorts of neat tidbits. If you thought my description of Louis Lingg’s beauty was a bit too breathless, for example, you can judge for yourself here. If you want to see exactly how nut and bolt screwed together to make a bomb, look here, for a bomb allegedly Lingg’s. The historical society also collaborated with Northwestern University to create Dramas of the Haymarket, a sort of online guided tour of the archival holdings.

The 2003 re-analysis of the Haymarket bomb fragments and evidence was described in this article by Timothy Messer-Kruse, James O. Eckert Jr., Pannee Burckel, and Jeffrey Dunn in a 2005 issue of the journal Labor: Studies in Working-Class History of the Americas.

The night before Parsons, Spies, Engel, and Fischer were hanged, Parsons sang the Scotch ballad “Annie Laurie.” There’s no recording of Parsons himself singing it, but there’s a period recording of the song by the Edison Male Quartette in the UC Santa Barbara Cylinder Preservation and Digitization Project. As I mention in the article, the next morning, just a few hours before they were hanged, the men sang the “Workers’ Marseillaise” together. The three German speakers may well have sung in German, and I strongly suspect that that’s what’s being sung in this period recording. I’m not sure, though, because my German comprehension is extremely poor; it’s the right tune, certainly, and someone has catalogued it under the title Arbeiter, i.e., “workers.”

Neither new nor rare

Historians have debated at length the question of when certain ideas about homosexuality came into currency. Here’s a piece of evidence, not previously reported to my knowledge:

In the Supreme Judicial Court of Massachusetts on 11 December 1868, a young man named Samuel M. Andrews was tried for murder. He pleaded not guilty by reason of “transitory insanity.” He was driven mad, he said, by Cornelius Holmes’s attempt to have sex with him.

It was a strange, sad case. According to the prosecuting attorney, Cornelius Holmes was not like other men. He was fifty-three years old at the time of his death and weighed 225 pounds. A bachelor, he lived alone in a boarding house. Though he had no occupation, he was said to be worth $20,000. The prosecution claimed that “The only person with whom he was at all intimate was the prisoner,” Andrews, who had killed him. The defense more or less agreed with this characterization of their relationship. Andrews “was almost the first young man Cornelius Holmes had ever met, who had not slurred him, & hooted at him,” Andrews’s attorney said. He added that Holmes “was not an idiot; . . . He was rather slow; played in childhood with children younger than himself.”

How intimate were they? The lawyers tried to bring this out. There was a telling exchange during the examination of a witness named Lysander Bartlett, a ship’s carpenter:

Question. Were Cornelius & Andrews intimate friends?

Witness. Mr. Andrews can tell you better than I can.

Chief Justice. You should not say to counsel that Andrews can tell you that better than I can. It is improper.

Witness. ‘Tis, eh? I knew they were intimate. . .

After Andrews was arrested for the murder, witnesses saw him kiss Holmes’s body.

In court, Andrews had a difficult task: he had to establish the nature of Holmes’s interest in him, and he had to make it sound plausible that he had resisted Holmes’s advances despite their persistence. He didn’t quite manage. He testified that “about nine years ago one stormy evening,” he and Holmes had shared a bed. “After talking awhile he turned towards me, & tried to put a part of his person between my legs, behind. I left the bed.” Though Holmes made other advances, the friendship continued. In fact there were hints that Holmes had thought of leaving his money to Andrews.

The murder came about one day when Holmes beckoned Andrews to follow him off the road and into the forest. Once they were alone, Holmes threw him down, tore open his pantaloons, put his hand in a relevant place, and said, “Now I’m going to have some, this time.” In a panic, Andrews grabbed a stone . . . That, at any rate, is what Andrews claimed in court, though he also claimed, confusedly, that he was “entirely unconscious of what took place.”

In summing up, the defense stressed how common Holmes’s tendency was:

Nor, gentlemen, is this any new crime. Go to our soldiers & sailors, inquire of our naval officers & see whether it is a new crime. The government would have you believe that this is an improbable story, because the crime is rare; gentlemen, this story is more than probable, for it is a crime which has always existed.

The prosecution, on the other hand, stressed how unlikely it was that Andrews would have had to kill Holmes in order to defend himself from rape. “Rape,” the prosecution insisted, wasn’t even the right word; in fact, the proper word didn’t exist, because the thing was impossible. The whole question of sex between men, the prosecution suggested, was probably a red herring: “There was in the present case no adequate evidence of any voluntary acts of indecency between the parties,” the prosecution argued, “but even if such existed, these had no tendency to prove an attempt to commit this act by force.”

In his instructions to the jury, the judge observed that Andrews had taken Holmes’s earlier advances quite calmly and had remained his close friend despite them. Taking the hint, the jury found Andrews guilty of manslaughter, and he was sentenced to twenty years.

(Source: Report of the Trial of Samuel M. Andrews, Indicted for the Murder of Cornelius Holmes, before the Supreme Judicial Court of Massachusetts, December 11, 1868, including the Rulings of the Court upon Many Questions of the Law, and a Full Statement of Authorities upon the Subject of Transitory Insanity. By Charles G. Davis, of Counsel for the Prisoner. New York: Hurd & Houghton. Cambridge: Riverside Press, 1869.)

‘I’m glad we got that straight’

. . . says Tom Cruise to Kelly McGillis, just before he doesn’t make love to her. Last night Peter and I rented Top Gun, not for its topicality, which was mildly unnerving, but because we had put it on our Netflix queue ages ago. Neither of us saw it when it came out in 1986—in my case, because I was closeted and wary of the homo subtext that the movie was rumored to have. The homo subtext is indeed formidable (rather often a male hand trails rather loosely across another man’s shirtless back), but what struck me most was the crass instruction that the movie offered to insecure straight men. (In this it’s much like Jerry Maguire.)

The most salient feature of Maverick (Cruise) in Top Gun is how short he is. There’s scarcely a frame in which he isn’t the tiniest person, Charlie (McGillis) included. His fellow pilots tower over him, and his enormous smile looks not so much like generosity of spirit as the defensive gesture of someone cornered. He flashes it when attacked. The dilemma of being Maverick is this: if you are short, have no father, lack impulse control, and cannot sing, how will you ever get laid? The answer is that you will turn two of your vulnerabilities into strengths: low impulse control will become a willingness to take risks that others won’t. This transformation is easy for Maverick, and he’s a little too fond of it. But the second is trickier: His wish for a father is to become an ability to bond tightly to his comrades. Herein lies the lesson, and it is unlovely. Maverick’s bonding is too personal; it’s really only to Goose, his “rear man,” whom the filmmakers have made ugly, to deflect suspicion. Maverick must eliminate the component of his love that is attached to a particular beloved. He must learn not to care when Goose dies.

Thus the emotional nonsense of the movie’s opening scene, when a pilot named Cougar freaks out, during a dogfight, and can’t pull himself together to land his plane but merely stares fixedly at a photo of his wife and children. I say nonsense, because presumably the thought of his wife and children would be an incentive to Cougar to rejoin the living. But instead the thought of them seems to tempt him to die at sea. That’s because the scene has no coherence of its own but is merely a setup to Maverick’s decision, in a later dogfight, not to die at sea as Goose did—not to “follow his leader,” as Melville might have put it.

If Tom can do this—and “this” is not mourning, it’s something quicker and more brutal, like cauterization—then he can have Charlie (McGillis). Note that for as long as it seems that he can’t do this, she rejects him, with a heartlessness that would be a flashing red alarm if a real woman were to display it. But she isn’t a real woman, she is the goddess of reproductive success, available to all who can accurately simulate alpha-male status. Maverick isn’t really an alpha male (note his prompt decision to become a professor once his warrior bona fides are proven), which is why the audience roots for him in his struggle to appear to be one.