On the morning of August 12, 2013, nearly eight months after 20-year-old Adam Lanza shot his way into Sandy Hook Elementary School in Newtown, Conn., and killed 26 people, Michael Mudry, an investigator with the Connecticut State Police, drove to nearby Danbury to try to solve a little mystery. Police had found a Garmin GPS unit in Lanza's house, and its records showed that the gunman had driven to the same spot nine times in April, May and June 2012, arriving around midnight each time and staying for hours.

The GPS readout took Mudry to the vast parking lot of a suburban shopping center, about 14 miles west of Lanza's home. Workers at a movie theater there immediately recognized Lanza from a photograph. He was at the theater constantly, they told Mudry, but never to see movies. He came to the lobby to play an arcade game, the same one, over and over again, sometimes for eight to 10 hours a night. Witnesses said he would whip himself into a frenzy, and on occasion the theater manager had to unplug the game to get him to leave.

Police had been scouring Lanza's home since the shootings, and on his computer hard drive they found information on weapons magazine capacities, images of Columbine killers Eric Harris and Dylan Klebold, copies of the violent movies Bloody Wednesday and Rampage, and a list of ingredients for TNT. And like many teenaged boys, Lanza owned the typical first-person shooter, fighting and action games: Call of Duty, Dead or Alive, Grand Theft Auto.

But those weren't the games that possessed Lanza at the movie theater. The title that so consumed the Sandy Hook shooter? Dance Dance Revolution—an arcade staple that has players dance on colored squares to the rhythm of Asian techno-pop. That discovery not only surprised investigators, it also was at odds with overheated speculation in the media and around dinner tables that violent video games had helped turn Lanza into a killer.

Yet no one knows how any of these games—Dance Dance Revolution included—might have affected a kid who was clearly struggling. The truth is that decades of research have turned up no reliable causal link between playing violent video games and perpetrating actual violence. This is not to say that games have no effect. They're built to have an effect. It's just not necessarily the one that most people think.

A tradition of worry
The implicit connection between violent media and violent behavior is so old that, like a barnacle clinging to a hull, it's not easily dislodged. The notion dates at least to the Victorian era, when educators, tastemakers and clergymen began criticizing what was then a fairly raucous popular culture. Violent, sex-soaked dime novels and penny-dreadful magazines were immensely popular, and upstanding publications such as Harper's and the Atlantic Monthly took delight in denouncing them. Author and critic Harold Schechter, whose 2005 book Savage Pastimes lays out a social history of violent entertainment, notes that the trend divided the literati of the time. Ralph Waldo Emerson complained about his countrymen “reading all day murders & railroad accidents,” but Nathaniel Hawthorne loved the scandal sheets so much that he had a friend ship stacks of them to Liverpool, England, while he lived abroad as a U.S. consul. The belle of Amherst herself, Emily Dickinson, relished stories of “those funny accidents where railroads meet each other, and gentlemen in factories get their heads cut off quite informally.”

The 20th century saw criticism grow more robust. In 1936 Catholic scholar John K. Ryan laid out what he called the “mental food of American children,” as seen through the media they consumed. It was a long menu, one that included “sadism, cannibalism, bestiality. Crude eroticism. Torturing, killing, kidnapping.” He was talking about daily newspaper comic strips. In 1947 critic and actor John Houseman lodged similar complaints about cartoons on television. They “run red with horrible savagery,” he wrote.

Into this fray entered Stanford University psychologist Albert Bandura, now 89, whose experimental studies in the early 1960s established the theoretical basis for limiting kids' access to violent media. In a 1961 study, Bandura and his colleagues gathered 72 preschoolers. Laboratory assistants led the kids, one at a time, into a playroom, where they sat at a small table and received instruction on how to make potato-print pictures. Soon another adult entered the room and settled into the opposite corner with a Tinkertoy set, a mallet and a five-foot, inflated Bobo clown doll, the kind that rights itself if knocked over. The adult then either quietly assembled the Tinkertoys, ignoring Bobo, or turned to the doll and began “aggressing toward it”—punching it, sitting on it, kicking it around the room, all the while saying things such as “Sock him in the nose!” and “Pow!”

After 10 minutes, each child was led into another room and invited to play with some “relatively attractive toys,” such as a fire engine, a spinning top and a doll set. But after two minutes, a lab assistant announced that these were “her very best toys” and that she'd decided to reserve them for other children. The kids were swept into a third room that held more toys, both “aggressive and nonaggressive”: a tea set, crayons, dart guns, a mallet … and a three-foot Bobo doll. You see where this is going.

Faced with the frustration of having nice new toys suddenly snatched away, the preschoolers who had watched Bobo get mistreated were more likely than the others to take out their aggression on the mini Bobo. Bandura repeated the experiment in 1963, using film and cartoon depictions of Bobo's mistreatment, with similar results. The conclusions seemed clear: watching unchecked aggression in real life, on film or in cartoons makes us more aggressive because it provides us with “social scripts” to guide our behavior. Bandura's conclusions opened a floodgate of “media effects” research that continues today.

The problem is that many of the findings, especially when applied to children's media and play, are misleading at best. Critic Gerard Jones, whose 2003 book Killing Monsters makes a case for giving kids access to “make-believe violence,” writes: “There is no evidence to suggest that punching an inflatable clown has any connection to real-life violence.” In many cases, he and others say, researchers mistake natural competitiveness or the effects of discomfort for aggression or mislabel the subjects' temporary aggression as behavior that holds the potential for violence. In an often quoted 1976 study led by Brian Coates at Washington State University, researchers found that preschoolers who watched the famously mild Mister Rogers' Neighborhood were three times more aggressive afterward. Jones suggests that the experiment itself may have made kids anxious or even angry by compelling them to “sit in a hard plastic chair in a strange room” and watch TV on cue.

It was the 1999 Columbine High School shootings that got many Americans thinking about violent video games. After the attacks, victims' families sued more than two dozen game makers, saying titles such as Doom, a first-person shooter that the two teen gunmen played, desensitized them to violence. A judge dismissed the lawsuits, but the post-Columbine uproar led more researchers to begin dissecting games, much as Bandura did for TV, in search of the roots of aggression.

Deciphering the data
A few studies tried to draw distinctions between good and bad games. In a 2010 experiment, Tobias Greitemeyer, then at the University of Sussex in England, and Silvia Osswald of Ludwig Maximilian University in Germany asked subjects to play one of three video games—either a “prosocial” game, an “aggressive” game or the “neutral” game Tetris. After eight minutes, an experimenter reached for a stack of questionnaires but “accidentally” knocked a cup of pencils off the table and onto the floor. Participants who had played the prosocial game were twice as likely to help pick up the pencils as those who played the neutral or aggressive game.

Others have tried to tease out the aftereffects of playing violent games. In a 2012 study, André Melzer of the University of Luxembourg, along with Mario Gollwitzer of Philipps University Marburg in Germany, found that inexperienced players felt a need to “cleanse” themselves after playing a violent video game (the so-called Macbeth effect: “Out, damned spot!”). Researchers asked subjects to play either a driving game or the mayhem-heavy Grand Theft Auto for 15 minutes, then pick gifts from an assortment, half “hygienic” (shower gel, deodorant, toothpaste) and half nonhygienic (gummy bears, Post-it notes, a box of tea). Inexperienced players who played Grand Theft Auto were more likely to pick out hygienic products than were experienced players or inexperienced players who had played the driving game.

But neither of those studies make the case that these games lead to real-word violence. Although drawing conclusions about small population subgroups—such as kids at risk of violence—from broad population trends can be dicey, it is still worth noting that as violent video games proliferated in recent years, the number of violent youthful offenders fell—by more than half between 1994 and 2010, according to the U.S. Department of Justice. This trend is not what you would expect if these games had the power to make good boys go bad. Indeed, in a 2011 analysis of game sales from 2004 to 2008, A. Scott Cunningham of Baylor University, Benjamin Engelstätter of the Center for European Economic Research in Mannheim, Germany, and Michael R. Ward of the University of Texas at Arlington found that higher rates of violent game sales actually coincided with a drop in crimes, especially violent crimes. They concluded that any negative behavioral effects playing violent games might have are more than offset because violent people are drawn to such games, and the more they play, the less time they have for crime.

Even if violent video games are not turning people into killers, we might still wonder if they are harming our kids in subtler ways. As psychologist Douglas A. Gentile of Iowa State University puts it, whatever we practice repeatedly affects the brain. If we practice aggressive ways of thinking, feeling and reacting, he writes, “then we will get better at those.” In a 2008 survey on the gaming habits of about 2,500 young people, Gentile and his father, psychologist J. Ronald Gentile, found that children and adolescents who played more violent games were likelier to report “aggressive cognitions and behaviors.” They concluded that violent video games “appear to be exemplary teachers of aggression.” They also found that eighth and ninth graders who played violent games more frequently displayed greater “hostile attribution bias” (being vigilant for enemies) and got into more arguments with teachers.

The greatest worry is the impact on children who are already at risk. “Media is most powerful in our lives when it reinforces our existing values,” media scholar Henry Jenkins, now at the University of Southern California, said in a 2003 episode of Religion & Ethics Newsweekly. Indeed, Jenkins argued in an essay for PBS, a child who responds to a video game the same way he or she does to a real-world trauma could be showing symptoms of an emotional disturbance. So used in the right setting, a violent game could actually serve as a diagnostic tool.

But beyond such special circumstances, media effects research, with its Bobo dolls as markers of real-world aggression, is problematic. The fighting kids do in physical games and video games alike is just a simulation. In other words, it is play. It looks like fighting, wrote Brian Sutton-Smith, the late renowned play theorist, in his book The Ambiguity of Play, “but it is also the opposite of fighting … carried on by those who are not enemies and who do not intend to harm each other.”

In a way, we are pointing fingers at the wrong people. When we worry that a violent game is going to turn our kids into killers, aren't we the ones who can't tell fantasy from reality? Kids already know the difference.

Adapted from The Game Believes in You: How Digital Play Can Make Our Kids Smarter, with permission from Palgrave Macmillan, a division of St. Martin's Press. Copyright © 2015.