how to lie with statistics about the white guy mass shooting problem

A talking point that’s been gaining traction lately is that, when you think about it, don’t we have more problems with white dudes being terrorists or rampage killers than anyone else? Culturally speaking, everyone knows about guns and creepy white dudes. It’s a thing, and you’re lying if you try to convince people otherwise.

That’s what Daniel Enger has done here, with statistics. The idea is to dazzle the reader with numbers, like this:

In a narrow sense, these stories are correct: The plurality of mass killers are white. But the notion that white men of privilege are disproportionately represented among mass shooters—indeed, that they make up “nearly all” of them—is a myth.

The myth began to spread in July 2012 after a 24-year-old white man, James Holmes, murdered 12 people in an Aurora, Colorado, movie theater. That summer, a team of researchers at Mother Jones put together a database of rampages in public settings. Among the 62 cases (and 64 total perpetrators) the team reviewed—instances dating back to 1982, in which at least four people had been killed in a public place—44 of the killers were white men and one was a white woman. In total, about 70 percent of the perpetrators in the database were white…

What those initial Mother Jones numbers showed, though, was that white people weren’t overrepresented among mass shooters. The media outlet had found that roughly 70 percent of the shooters in mass killings were white—certainly a majority. But according to Census Bureau estimates for 2012, whites accounted for 73.9 percent of all Americans…

Since 2012, Mother Jones has added 29 more mass-shooting events to its database (and tweaked its definition of the crime to fit with new federal guidelines that placed the threshold at three victims instead of four). In this bigger data set, the proportion of white mass shooters drops down to 56 percent, by my count. Judging by those newer numbers, and the most current census estimate that 76.9 percent of Americans are white, the whites-are-overrepresented-among-mass-shooters meme appears even less accurate. Perpetrators that Mother Jones classifies as Asian make up 7.4 percent of the data set, versus an estimated 5.7 percent of the population, while those MoJo identifies as black represent 17.0 percent of the mass shooters in the database versus an estimated 13.3 percent of the population. According to this data set, then, Asians and black Americans are overrepresented among mass shooters by about the same proportion (a bit more than one-fourth) that whites are underrepresented. This means the population rate of mass shootings by whites (at least according to the tiny sample measured in the MoJo database) is 0.021 per 100,000 people, while the corresponding rate of mass shootings by blacks is 1.7 times higher, at 0.037.

This disparity, which could be thought of as the statistical non-whiteness of mass shootings, is much smaller in magnitude than the one for killings nationwide. Overall murder rates among black Americans are 6.3 times higher than they are for whites, according to a report from the Bureau of Justice Statistics. The same report suggests white offenders made up just 45.3 percent of everyone who committed homicides between 1980 and 2008. In other words, white Americans may be somewhat underrepresented among mass shooters, but they’re even more underrepresented among all killers. In that limited sense, it would be fair to say that whites are responsible for more public massacres than you might expect. Does that mean their whiteness is a factor in these crimes?

Please click here and look at the actual spreadsheet we’re talking about. Spend a few minutes looking at the summaries to form your own independent impression.

This is the key: at the start of this post, when I was talking about the white guy mass shooter problem, it was entirely clear what I was talking about. Mother Jones is also explicit about what they’re talking about:

We honed our criteria accordingly:

  • The attack must have occurred essentially in a single incident, in a public place;
  • We excluded crimes of armed robbery, gang violence, or domestic violence in a home, focusing on cases in which the motive appeared to be indiscriminate mass murder;
  • The killer, in accordance with the FBI criterion, had to have taken the lives of at least four people.

This is very abstract! They’re clear about what they’re counting, but it’s unclear that this is a good way of counting “Aurora-like events”, which was their goal. Consider that all of these are grouped together:

Randy Stair, a 24-year-old worker at Weis grocery fatally shot three of his fellow employees. He reportedly fired 59 rounds with a pair of shotguns before turning the gun on himself as another co-worker fled the scene for help and law enforcement responded.

Kori Ali Muhammad, 39, opened fire along a street in downtown Fresno, killing three people randomly in an alleged hate crime prior to being apprehended by police. Muhammad, who is black, killed three white victims and later described his attack as being racially motivated; he also reportedly yelled ‘Allahu Akbar’ at the time he was arrested, but authorities indicated they found no links to Islamist terrorism.

Micah Xavier Johnson, a 25-year-old Army veteran, targeted police at a peaceful Black Lives Matter protest, killing five officers and injuring nine others as well as two civilians.

Robert Lewis Dear, 57, shot and killed a police officer and two citizens when he opened fire at a Planned Parenthood health clinic in Colorado Springs, Colorado.

Dennis Clark III, 27, shot and killed his girlfriend in their shared apartment, and then shot two witnesses in the building’s parking lot and a third victim in another apartment, before being killed by police.

They’re grouping together targeted killings with political motives, killing coworkers, shooting random people, etc. The count is based on body count.

I don’t think it’s intuitively obvious that these events all belong in the same category. When Engber is talking about percentages, he’s talking about percentages of this list, which is an estimate of mass shootings with questionable validity. He gets a ton of propaganda mileage out of this spreadsheet, though.

The way he tells it, these numbers mean that white people are disproportionately peaceful. Blacks and Asians are suspicious and dangerous. 1.7 times. Or 6 times. Or something. That was the entire point of the article, which is obvious from what was the top comment until a couple minutes ago:

I like it that he debunked the silly ‘white mass murders’ meme pretty well, and correctly points out that the black murder rate is ~6x the white murder rate.

…And then he says that white people are still responsible for the black murder rate being so high, because privilege.

Yes, trolls love this kind of shit:

It’s possible—but given all the numbers above, I think it makes more sense to ask why those classified as non-white might be disproportionately represented among killers, from mass shooters down the line. The answer there would seem to have everything to do with privilege. Structural inequalities related to education, employment, housing, and health care, along with de facto segregation and a history of discrimination and bias, create conditions under which black Americans in particular are more likely to be both the perpetrators and the victims of this violence. More than half of those committing homicides in the BJS data set are black, and close to half of the victims of those homicides are black. These statistics show us that in a global sense, a lack of privilege contributes to killing and that white privilege kills, at least in part, through the reciprocal cost it imposes on to other groups.

Be afraid of poor niggers.

You can feel the self-satisfaction coming off that paragraph. Of course it’s a true observation, also made by Primo Levi:

It is naive, absurd, and historically false to believe that an infernal system such as National Socialism sanctifies its victims; on the contrary, it degrades them, it makes them resemble itself.

What’s important is that the white people never, ever concede this point in other contexts. Engber only concedes it here because it helps make black people look like dangerous savages. If the topic was reparations, or structural inequality, reasons “black people are lazy” doesn’t explain things, the white people would never take responsibility for the ghetto, at least not anymore.

Buried deep in the article is the part where he has to admit the truth:

But Lankford found some nuance in the data, particularly when it came to the most horrific massacres. According to the study, white and Asian mass murderers perpetrated crimes with more victims, on average, and they were more likely to carry out those crimes in public places. Nearly one-fourth of the white mass murderers and one-fifth of the Asians in the group engaged in public killings. Among the black mass murderers, this proportion was just 6 percent. Lankford suggests the relative whiteness of public killings, in particular, could indeed result from structural advantage and “aggrieved entitlement.” At the same time, he says, those public crimes get far more media coverage than any others—a factor that might reinforce the false belief that mass murder is a mostly white phenomenon.

The straw man is the last sentence. This conversation is happening because the Las Vegas shootings force attention to glaring problems with mainstream culture. That gets subtly turned into “the false belief that mass murder is a mostly white phenomenon.”

When we limit the data to what we’re talking about, there’s a clear white guy problem. The way to obfuscate that is to add a bunch of noise to the data by including killings other than the ones we’re talking about on a common sense level. None of what Engber wrote changes the fact that, yeah, why the fuck are white guys so into Rambo fantasies?