Playing Games with Numbers

Stanford Law Professor John Donohue and Theodora Boulouta, a Stanford undergrad, wrote an opinion piece in the New York Times that claims their new research shows the 1994 “assault weapons” ban really did work. 

Their finding is that, “Compared with the decade before its adoption, the federal assault weapon ban in effect from September 1994 through 2004 was associated with a 25 percent drop in gun massacres (from eight to six) and a 40 percent drop in fatalities (from 81 to 49).” 

For some reason, Donohue and Boulouta chose to use six fatalities as the minimum, instead of the common definition of four or even the Obama-era attempt to reduce the threshold to three. Jacob Sullum at Reason has a pretty good guess as to why Donohue and Boulouta chose six: “Yet they chose to focus on cases with six or more fatalities, for no obvious reason except that it exaggerates the changes they attribute to the "assault weapon" ban that expired in 2004.” Sullum’s article is titled, “A Suspiciously Selective, Logically Shaky Analysis of Mass Shooting Data Claims the Federal ‘Assault Weapon’ Ban ‘Really Did Work’.”  

Sullum uses the full Mother Jones dataset, including incidents in which three people were murdered (instead of the long-accepted threshold of four fatalities) and incidents that Mother Jones categorizes as a spree killing. So, while Sullum’s point about Donohue and Boulouta’s anything-but-arbitrary selection process is valid, his totals could be further refined to better meet the commonly accepted definition of a public mass shooting.

So that’s what we did. We limited the dataset to those identified by Mother Jonesas mass shootings and those with four or more fatalities. There were 15% fewer mass shootings during the so-called “assault weapons” ban than in the previous decade…but we’re talking about rare events so the 15% decrease is a shift from 13 to 11. There was a 32% decrease in fatalities. The prior decade included the Luby’s Cafeteria attack in Killeen, Texas. Excluding that case alters the decreases to 8% in incidents and 12% in fatalities. 

See how a single incident among rare events can alter the calculated statistics? 

Donohue and Boulouta appear to have used some additional case selection process in their analysis. Using the same database and looking only at mass shootings with six or more fatalities, we have seven incidents in the decade before the ban and six during the ban. Excluding weapons not legally acquired shifts the totals for each period down by one event (to six and five, respectively). There were still 48% more fatalities before the ban than during. Let’s look at the six cases from the decade before the shooting. Two involved semi-automatic variants of AK-style rifles and another involved a pair of Intratec DC-9 pistols. Here are what the other incidents involved:

•      Three semi-automatic handguns – a .22 and two Colt 1911s
•      Two semi-automatic handguns, a bolt-action rifle, two revolvers, and two shotguns
•      An M1 carbine and a revolver
•      Two semi-automatic handguns

Which of those firearms would have been prohibited under the so-called “assault weapons” ban?

Donohue and Boulouta deliberately chose a definition that allowed them to deliver the most dramatic change possible. They present this rudimentary analysis as evidence that the “assault weapons” ban is responsible for any change at all. They assert that the change must be because the law and the subsequent sunsetting allowed “the gun industry to flood the market with increasingly powerful weapons that allow for faster killing.” 

Except none of that is true. Semi-automatic rifles in the years before the ban, those that were explicitly banned, those that were available during the ban, and those now available are functionally the same. Rates of fire and muzzle velocity have not changed. Donohue argues that, “assault weapons are semiautomatic firearms designed for rapid fire and combat use, and large capacity magazines increase the number of rounds that can be fired without reloading.” Semi-automatic firearms, as anyone familiar with guns or anyone not looking to manipulate an argument knows, fire a single round per trigger pull. They’re also not designed for combat use. Military forces generally use select-fire weapons that are capable of firing more than one round per trigger pull. 

What is noticeably absent from this analysis is any consideration for the known contagion effect of mass shooters. Mass shooters seek fame, and media reportshelp deliver this to their twisted minds. One of the key differences between now and the 1980s, 1990s, and even the early 2000s is the proliferation of internet access and content available digitally. Information on mass shooters is readily available, allowing disturbed and deranged monsters the ability to obsess over mass shooters and helping their own twisted fantasies fester. The contagion effect is real, and should be considered in any sort of analysis – as well as a warning sign for potential mass shooters. 

As mentioned above, Donohue is a Stanford Law professor. This doesn’t stop him from repeating the absurd anti-gun talking points about the Protection of Lawful Commerce in Arms Act. He claims that, “no other industry is allowed to act so recklessly without facing legal challenge. But a 2005 law immunized gun manufacturers for harm caused by the criminal use of firearms.” What industry is held responsible for the criminal use of its legally sold products, Professor?  

Warning signs are often strung together after the fact – or ignored in real-time by authorities more interested in funding than safety, as in Parkland. No system – not even an “assault weapons” ban – will work if efforts are not made to utilize it.

Frankly, we expected more from Donohue and his newest co-author. Donohue has, in the past, attempted to position himself as a researcher. He appears willing now to out himself as an activist.  While we’re not sure if this opinion article is a precursor to an actual study, the choices made throughout the article suggest that – yet again – gun controllers let their desired outcome dictate the research design.