Presidential candidate Joe Biden writes:
Many police departments have reported an increase in criminals using assault weapons since 2004. And multiple analyses of the data around mass shootings provide evidence that, from 1994 to 2004, the years when assault weapons and high-capacity magazines were banned, there were fewer mass shootings — fewer deaths, fewer families needlessly destroyed.
Somewhat perplexingly, the links on “multiple analyses” and “fewer mass shootings” actually go to the same study, just hosted at two different websites. But it’s a new one I hadn’t seen before, and slightly different from the one I wrote about here, so let’s take a look.
First, a quick bit of history: The U.S. had a ban on “assault weapons” — semiautomatic guns with certain combinations of “tactical” cosmetic features, including pistol grips and folding stocks — from 1994 to 2004. The existing guns were not confiscated; it was just illegal to sell new ones. The same law banned “high-capacity” magazines, setting the threshold at ten rounds, which is actually below the standard capacity of many modern handguns. (Both of my own handguns come standard with 15-round magazines, for example.)
After the ban was allowed to expire, experts were more or less unanimous that it hadn’t had a strong effect on overall homicides. Certainly there were fewer of the banned items in circulation than otherwise would have been the case, but the law’s definition of “assault weapon” focused on features that don’t really affect a gun’s lethality; few murders are committed with long guns of any kind, to say nothing of “assault weapons” in particular; and while smaller magazines force more frequent reloading, it takes only a second or so to pop a new one in.
Recently, though, some have claimed that the law reduced mass shootings in particular, which account for a tiny fraction of overall homicides but command an incredibly disproportionate amount of public attention. What seems to be true is that the ban years were relatively peaceful on this front, despite covering the rash of school shootings that included Columbine. They were especially peaceful compared with the past ten years or so, which have seen an alarming rise in this form of terrorism.
However, the three most important facts about mass shootings are (A) they have historically been incredibly rare, with entire years passing without one sometimes; (B) they are contagious, with high-profile incidents inpiring copycats and competitors; and (C) they are incredibly variable in the number of fatalities, from a low bound of wherever the researcher chooses to set it (the study Biden cites uses four, not including the perpetrator) all the way up to 58 at Las Vegas. Trying to detect a pattern in data like this, and then attributing the pattern to a single law change that covered the entire country for a ten-year period, is madness.
The new study doesn’t add much to what we already know and is downright bizarre at times. For instance, the authors combine three different data sets of mass shootings, but instead of including all the incidents they found (to be as comprehensive as possible), they included only the incidents counted in all three databases. They also misuse the term “assault rifle,” which refers to a weapon capable of full-auto fire, indicating that not one of the nine authors and not one of the journal’s editors is all that familiar with firearms or the gun debate. (The weapons at issue in the assault-weapon ban are semiautomatic civilian versions of military-style guns.)
But at any rate, some charts from the study show the core problem here. Here is the total count of mass-shooting fatalities by year. As you can see, they go in spurts, and recent years in particular have been horrible, but this does not come close to showing the assault-weapon ban mattered. I don’t think anyone would look at this and say, unprompted, “obviously we had a really effective law on the books from 1994 to 2004.” The claim makes even less sense when you remember that used copies of the banned guns and magazines remained available.
Here’s another one, though it presents mass-shooting deaths as a percentage of overall gun homicides (which is odd because the two are rather different phenomena):
Same problem. Is the 1994–2004 period really the one that stands out? Or is it the frightening increase in fatalities more recently?
Also worth reading: the RAND Corporation’s summary last year of the high-quality evidence on assault-weapon bans and mass shootings. There’s one study suggesting they work and another finding no effect, both with some methodological issues, leading to a verdict of “inconclusive.”
Update: I should also point out that Louis Klarevas, who wrote the pro-assault-weapon-ban study I discussed in my earlier post, strongly criticized the new study, especially for overstating how many mass shootings involve the “assault weapons” targeted by the ban. Hat tip to the Twitter account GunStats.