In the movie Moneyball, based on the 2003 book by Michael Lewis, a young Yale-educated economist played by Jonah Hill explains to Oakland Athletics general manager Billy Bean, played by Brad Pitt, how baseball doesn’t know itself. “There is an epidemic failure within the game to understand what is really happening,” he says. This failure is found in a consistently flawed evaluation of individual performance where star players fetch salaries wildly out of proportion to their actual added value. He goes on:
People who run ball clubs, they think in terms of buying players. Your goal shouldn’t be to buy players, your goal should be to buy wins. And in order to buy wins, you need to buy runs.
He cites as an example Johnny Damon, the A’s star outfielder whom the Boston Red Sox had just pinched for $7.5 million. “When I see Johnny Damon,” he says, “what I see is an imperfect understanding of where runs come from.”
Hill’s character is loosely based on a real person, Paul DePodesta, a Harvard-trained economist who also played baseball for the Crimson. DePodesta and Bean built a championship team in a down market with a payroll that was literally a fraction of the richest teams in Major League Baseball. They did this by applying a more critical methodology to evaluating baseball players and their contribution to a winning record. Their approach was based on the eccentric scholarship of Bill James, a sausage plant security guard who self-published books about baseball statistics in his spare time in the 1970s and 1980s.
James dissected baseball’s ingrained assumptions about the relationship between measured player performance and game outcomes. He pitched virtually all fielding statistics (an “error,” he noted caustically, “is a record of what an observer thinks should have been accomplished”) as too noisy to draw any meaningful conclusions about its relationship to winning. He determined that some of the most-cited metrics of offensive performance such as RBIs, stolen bases, even batting average, had little to no direct relationship to runs and thus to wins. How can a low batting average not predict runs? Because it ignores walks and hits-by-pitching. The most reliable predictor of runs, James wrote, is on-base and slugging percentage.
By now the story of Billy Bean’s A’s and the revolution in baseball statistics is well-known. But flawed methodology for determining cause and effect lives on, notoriously in areas of social policy. Nowhere is this more the case than in the debate about firearms in the United States. Applying baseball statistics to gun violence may appear frivolous given the gravity of the subject. But the same reframing and close analysis applied in Moneyball can and should be applied to our approach to gun violence.
The focus on guns, like the focus of baseball on players, is misplaced in the debate over violence and death. Because, to borrow from Moneyball, the focus of gun control should not be the firearms themselves. The focus should be on the deaths caused by gun violence. The two are related, of course, just as buying baseball players is related to winning games. But the focus on gun control ignores what we’re really trying to do, which is to reduce violence that ends in fatalities. If we widen or shift our focus slightly, the relationships between violence, injury, death, and weapons become much more clear.
For example, it is a common assumption, especially in gun control advocacy, that more guns means more deaths. That appears to be borne out in readily available statistics, especially when we compare gun ownership in United States to similarly developed countries like Australia and the United Kingdom. But this relationship breaks down on closer examination. Here is where my Moneyball analogy applies. If we focus on deaths per population instead of firearms per population, we find the world upside down: The continental United States doesn’t make the top 10 among the most violent countries in the world.
Another examination of basic statistics demonstrates a weak relationship at best between the number of firearms and the death rate in the United States. It is estimated that the number of firearms has grown to about 400 million in the U.S. This proxy chart demonstrates the explosion in background checks, which are pegged to firearm sales, in just the last 20 years:
If we assume a direct relationship between firearm ownership and death, then the homicide and suicide rate should track gun sales. But they do not. In fact, U.S. deaths per population have remained in a fairly narrow band for more than 50 years:
More insight into this paradox can be found in ownership records. The percentage of households owning firearms has also remained relatively steady for the last 50 years (although it is on a slow downward trend since 1960). So where are all the new guns going? To households that already own guns. This is relevant because it further decouples the number of weapons from death rates. Unless households are getting bigger, there is little practical difference between an owner with one gun and an owner with ten. There is also little practical difference between someone killed with a gun and somebody killed with 10: the outcome is the same. Put simply, more guns in the same households does not result in more deaths.
This may appear to be a pro-gun argument but it is not. By shifting the attention from guns to deaths, we can get by the impasse between gun control advocacy and the pro-gun lobby. The lobby does not see gun control as the solution to the problem of gun violence. They believe more guns are the solution to gun violence: “the only answer to a bad guy with a gun is a good guy with a gun.” The lobby has legislated this mantra through the expansion of the castle doctrine, open carry, and stand your ground. In short, the answer to every mass shooting, family annihilation, homicide, or domestic abuse death is more guns. But this is also plainly not true. If it were, the trendline in deaths would be inverted to rising gun ownership.
By focusing on deaths, we do three critical things to the policy debate: 1) we decouple guns from the political discussion about how to prevent death, 2) we see more clearly the problem set as one that involves firearms but is not limited to them, and 3) we have much more innovative and creative ways available to us to reduce deaths than legal restrictions and regulation.
It is regularly noted in this debate that while mass shootings, especially at schools, capture the public’s attention, these events are comparatively rare. That is true. But again we may be missing the point hidden in plain sight: each shooting incident is different. A school shooting is different from a gangland homicide. A suicide is different from an accident. Intimate partner violence is different from a police shooting. These types of incidents have factors unique to themselves beyond the weapon of choice.
By breaking down the types of death, the linkage between firearm ownership and specific deaths becomes much, much more clear than the aggregate statistics outlined above:
- Suicides are the leading cause of death involving firearms
- Half of all suicides involve a firearm
- Suicide attempts with a firearm are 90 percent fatal
This may seem paradoxical given my previous argument. How can guns be unrelated to overall death rates but be closely tied to individual deaths? The answer is we have much more meaningful statistics based on the type of incident. Put more practically, 400 million guns is not as relevant as the handgun in your depressed uncle’s dresser.
A suicide prevention social marketing campaign can urge a disturbed person and their family to remove guns from their possession. But it is not limited to that: it can provide mental health resources, early warning signs, risk factors, and prevention measures. (For example, a dramatic cut in suicides could be made by focusing on men aged 25-64, which not incidentally are the primary demographic for gun ownership.)
Similarly, the relationship between firearm deaths and intimate partner violence is very clear:
- Half of all women killed in the United States are killed by an intimate partner
- Half of all intimate partner deaths are committed with a firearm
- Women are five times more likely to die due to intimate partner violence if that partner has a gun
- The vast majority of women shot or killed with a firearm experienced stalking or other abuse prior to their death
A campaign focused on domestic abuse victims and their families could urge the separation of the gun from the abuser. But it is not limited to that: getting out of the home, finding shelter and mental health resources, accessing law enforcement, and restricting the abuser’s movements are all appropriate measures to put distance between the victim and the man who wields a weapon.
While school attacks and active shooters grip the imagination, we spend much more time preparing to hide, run, or fight than we do on early warning and prevention. There is substantial evidence that active shooters leave an extensive forensic trail well before their death spree. A social marketing campaign (see example below) could raise awareness of these warning signs and how to intervene.
While I am advocating for a change in policy – to include the popular and effective means of violence prevention including age restrictions, weapons registration, safety training, magazine caps, and limiting assault weapons – this task should not be left to legislation alone. I am encouraging others, especially foundations, advocacy organizations, and think tanks, to develop creative initiatives that demonstrably lower violent death. It is hard to imagine the pro-gun lobby like the National Rifle Association cooperating with such a campaign – although this is possible, especially when it comes to suicide prevention and weapons safety – but a social campaign can be successful regardless. Under the current dynamics, the lobby has a monopoly on the legislative process. While the lobby pursues its more-guns agenda, a mass social marketing campaign can operate more creatively, independently, and without interference.
There is ample precedent for the efficacy of extensive, well-resourced, and sustained social marketing campaigns to change individual behavior and social norms. In the case of the anti-smoking movement in the 1960s and 1970s and Mothers Against Drunk Driving (MADD) in the 1980s, the campaigns paced social policy; the campaigns were so successful, the laws that followed were virtually moot.
MADD is an exemplar for the kind of campaign I am advocating. MADD did not promote restrictions on alcohol or cars but focused instead on social behavior that decoupled those two proximate causes, dramatically lowering fatalities and injuries. Similarly, while the anti-smoking campaign included regulating and restricting tobacco and its advertising, the campaign’s ultimate aim was not to eliminate cigarettes but to reduce smoking-related disease. And that was accomplished by changing social mores and individual behaviors so dramatically that lung cancer rates have fallen to 1950s levels.
I wrote this essay because in addition to the legislative stalemate, the unyielding positions of debate around gun violence, and the endless iteration of the same tepid proposals to restrict firearms, there is a sense of social despair that we can’t solve this problem. Too many are demanding action but too few are offering workable solutions. I feel responsible as a citizen and a parent to contribute to this debate.
An open-ended and creative approach to violence prevention could be more successful than decades of thwarted legislation. As in Moneyball, the positions in this debate are so dug in and invested that they can’t actually see the problem set for what it is. We need a change in perspective and a fresh approach. At stake are the lives of tens of thousands of Americans every year.