Read our COVID-19 research and news.

A shooting range owner inspects a revolver in Guildford, Connecticut, in April 2013. Gun sales surged in the 4.5 months following the Sandy Hook Elementary School shooting in Newtown, Connecticut.


Accidental gun killings surged after Sandy Hook school shooting

A surge in gun buying in the months immediately following the 2012 mass shooting at Sandy Hook Elementary School in Newtown, Connecticut, corresponded with an increase in accidental gun deaths in the United States, one-third of them in children, according to an analysis published today in Science.

About 60 additional unintended shooting deaths, roughly 20 of them in children, occurred in the 5 months after the shooting, conclude the study’s authors, economists Phillip Levine and Robin McKnight of Wellesley College in Massachusetts. For all of the 2012 calendar year, there were 545 accidental shooting deaths, or about 45 per month, according to the U.S. Centers for Disease Control and Prevention. So a 60-death bump in a 5-month period is a considerable one.

Gun sales in the United States can increase after mass shootings. The authors of the new paper estimate that an additional 3 million guns were sold nationwide from December 2012 through April 2013, the 5-month window immediately before and for several months after a lone gunman killed 20 children and six adults at Sandy Hook on 14 December. But this is the first time researchers have found a coincident surge in accidental gun deaths: Levine and McKnight studied trends in firearms sales and accidental gun deaths from 2008 through 2015 and found no other instances of such a pattern.

“The fact that we see a spike in accidental firearm deaths precisely at this moment, it’s hard to imagine that’s a coincidence,” Levine says. He and McKnight suggest that increased exposure to guns after Sandy Hook, whether through new guns being purchased or through stored guns being taken out, handled, and cleaned, likely led to the increase. The authors found no corresponding increase in the number of gun homicides or suicides during the post–Sandy Hook shooting window, a finding that lines up with previous studies. They suspect that people who bought guns or pulled them out of storage—presumably over fears of a government clampdown—are “unlikely to be motivated by an intention to kill themselves or others.”

Levine said that he and McKnight decided to do the study after seeing data in an article in The New York Times that showed a dramatic surge in gun purchases after Sandy Hook and again after 14 people were gunned down on 2 December 2015 at a public health department Christmas party in San Bernardino, California. After both incidents, U.S. politicians led by then-President Barack Obama argued for tighter gun control measures. (Trends after the San Bernardino shooting weren’t captured in the current paper’s analysis.)

One expert not involved with the paper praised it for using “apparently rigorous” statistical analysis to wrest revealing information from the Sandy Hook tragedy. The take-home message, says Stephen Hargarten, an emergency physician who directs the Comprehensive Injury Research Center at the Medical College of Wisconsin in Milwaukee, is this: “Guns in the home present a risk.”

Critics see flaws

But the paper is drawing criticism from some researchers. “This study of a single mass shooting and a single type of gun violence amounts to little more than a statistical anecdote,” wrote Gary Kleck, a criminologist at Florida State University in Tallahassee, to Science. “Notwithstanding its prestigious outlet, this paper is junk science, and should never have been published.” In particular, Kleck finds the statistical associations implausible. He notes that the 3 million additional guns sold between December 2012 and April 2013 amounted to an increase of about 1% in a private U.S. arsenal that already totals roughly 300 million guns. That it should lead to a 27% increase in accidental firearm deaths and a 64% increase among children is “astounding,” he wrote. 

Decades of U.S. national data show a steady downward trend in accidental firearm deaths—from 1.55 per 100,000 people in 1948 to 0.18 per 100,000 in 2014. That has held true even as the number of guns in circulation has grown enormously, from 0.36 per person in 1948 year to 1.13 in 2014, notes David Kopel, an adjunct constitutional law professor at the University of Denver Sturm College of Law. “The talking point is going to be: If you have X more exposure to guns, you have Y more accidents,” he says. “And that is true, at most, in this unusual period that is the focus of the study. But it is certainly not true in general.”

Deborah Azrael, a health policy expert and statistician at the Harvard University Injury Control Research Center in Boston calls the study “imperfect, but clever.” Because the authors are limited by the lack of reliable data on U.S. gun ownership, she says, “they do something smart. They ask: ‘We know in this [post–Sandy Hook] period an exceptional number of guns were added to those already in circulation; did anything unusual happen?’”

But she is critical of at least one aspect of the study. Although the paper suggests that increased gun “exposure” caused the spike in accidental deaths, she says the authors don’t grapple with what they mean, precisely, by exposure. “How do they think it’s acting in this case? I think what they are suggesting is that there may have been something unique, not just in the spike of gun sales … after Sandy Hook, but also in the way people were accessing their guns.”

Authors respond

The authors have responses to many of the critiques. In response to Kleck’s concerns about the “astounding” increase in deaths, McKnight says that the new gun sales accounted for only part of the increased exposure after Sandy Hook; people removing their guns from storage to inspect or clean them could also have contributed. “The real issue is not the number of guns, but the number of guns that are stored improperly,” she wrote in an email. She also argued that buyers who feared new gun control laws and raced to buy firearms may have been new owners and less apt to handle guns safely. “For example, if they were less experienced gun owners, then they might be at greater risk” of accidents, she says.

The authors say there’s some evidence for the neophyte gun owner argument in data from California, a state that keeps permanent records of every gun acquisition. A study published this year in the Annals of Internal Medicine reports that so-called first-time transactions—the first time a buyer was documented to have obtained a weapon—accounted for 57% of the surge in gun acquisitions in that state in the first 6 weeks after Sandy Hook.

The authors also examined Google Trends data and found post–Sandy Hook surges in searches including the terms “buy gun” and “clean gun.” They call this “suggestive evidence” that gun exposure increased after the Sandy Hook killings. They hypothesized that people conducting such Google searches may be more likely to take out a stored firearm or buy a new one, thus increasing exposure, which Levine defined in an interview as “the idea that the gun is around.”

Levine and McKnight used the number of background checks reported to the government by federally licensed gun dealers as a proxy for gun sales. They chose the 5-month window because it corresponded with a surge in gun sales. The study period also bracketed a time during which Obama argued forcefully for, and issued executive orders aimed at, increasing federal gun regulation. Congress defeated related legislation on 17 April 2013, and the spike in sales ebbed. (Data from the entire months of December and April were included because the data are available only in monthly form.)

Levine and McKnight also gauged whether the increase in accidental gun deaths was more pronounced in geographical areas in which gun sales surged higher than elsewhere in the country after Sandy Hook. They divided U.S. states into two groups: those where more than 1000 additional guns were sold per 100,000 residents over the 5-month period, and those with fewer sales. Thirty-one states, including a northern band from Washington through Minnesota, as well as Oklahoma, Missouri, and Tennessee, fit into the first category. There, rates of additional accidental deaths of children under 15 years old were about 16 times higher than in the other states.

That stark finding, says Levine, “just strengthens support for our hypothesis. The fact that the increase in accidental deaths at precisely the time that gun sales spiked is concentrated in the states where that spike was the largest increases the likelihood that the impact was causal.”