Today, we’ll once again be examining how to make convincing-sounding pseudo-arguments in defense of murder. I want to focus in particular on how statistics can be manipulated, massaged, and misrepresented, and how talking points that seem powerful can actually be irrelevant or misleading.
Let’s start with a statistic from Charlie Kirk, the head of Turning Point USA, a conservative student group with over 350 chapters. (The Current Affairs Book & Periodical Discussion Society, by contrast, has only a single chapter.) Kirk is not a particularly reliable guy, and has shown himself willing to say blatantly false things to his 600,000 Twitter followers. I wouldn’t trust him as far as I can throw him, and trust me, I cannot throw him very far. (Alas.) Here he is explaining why discrimination by police against black people “doesn’t exist”:
A police officer is 18.5 times more likely to be shot by a black man, than an unarmed black man is to be shot by a cop. Black on black crime, not police brutality is the big issue facing black America. This all stems from a broken culture – failing schools & fatherless homes.
In the accompanying video, Kirk cites the “18.5″ statistic and angrily says “That’s a fact! That’s a fact!” He’s not the only one who has used that number: Here’s Tomi Lahren saying the same thing. And the statistic turns out to have originated with Manhattan Institute fellow Heather Mac Donald, who published it in the Wall Street Journal, the Washington Post, City Journal, and The Washington Examiner (headline: “Black Lies Matter”). Here she is explaining how she calculated the number:
The 36 unarmed black male victims of police shootings in 2015 measured against the total black male population [nearly 19 million in mid-2014, per the Census Bureau] amounts to a per capita rate of 0.0000018 unarmed fatalities by police. By comparison, 52 law enforcement officers were feloniously killed in 2015 while engaged in such duties as traffic stops and warrant service, according to the National Law Enforcement Officers Memorial Fund. The FBI counted nearly 628,000 full-time law enforcement officers in the United States in 2014. Assuming that the number of officers did not markedly increase in 2015, the per capita rate of officers being feloniously killed is 0.000082, or 45 times the rate at which unarmed black males are killed by cops. The Memorial Fund does not have data on the race of cop-killers in 2015, but applying the historical average over the last decade in which 40 percent of all cop-killers were black would yield 21 cops killed by blacks in 2015. An officer’s chance of getting killed by a black person is 0.000033, which is 18.5 times the chance of an unarmed black person getting killed by a cop. After this year’s 72 percent increase in felonious killings of police officers, these ratios will be even more lopsided.
You should be able to instantly notice that “18.5” can’t actually be correct for Mac Donald’s particular conclusion. She uses the “unarmed black male victims of police shootings” to calculate “the chance of an unarmed black person getting killed by a cop.” Hang on, though: Eric Garner didn’t die in a police shooting. Neither did Freddie Gray. If we consult the Guardian’s compilation of deaths, we discover that actually, if we include causes of death beyond shooting (like being choked, beaten, tased, or bitten by a police dog), 76 unarmed black men died at the hands of police in 2015, not 36. That’s more than twice as many, which means that Mac Donald is significantly undercounting the dead and is full of crap when she says she’s measured the chance of an “unarmed black person getting killed by a cop.”
But wait a second: There’s still something odd here. Unarmed black man. Philando Castile wasn’t unarmed. John Crawford had a BB gun, and Tamir Rice had an airsoft pistol. I thought we had a Second Amendment in this country! Merely to say that the victim “had a gun” is no different from saying they were “wearing a hat,” absent further facts. I mean, unless we say that the Second Amendment doesn’t apply to black men. So how many black men were killed total, armed and unarmed? 295, which is a lot more than 36! Now, I’m sure that plenty of these involve suspects who shot at police officers, and officers who returned fire in self defense. But that’s the number you want, because mere possession of a gun doesn’t suspend the rules of engagement or entitle police officers to kill you. And that number is going to be really difficult to obtain with precision, because it depends on us figuring out the truth of whether or not a black man with a gun was or was not a threat. Weirdly, this point sounds less compelling than it actually ought to be, but I happen to think it’s important to note that the fact a black man had a gun does not in and of itself justify his death. (Whenever there’s a “Police mistook a wallet for a gun” incident, I always think to myself “But what if it had been a gun?”)
We have good evidence, then, that Heather Mac Donald is intentionally excluding certain categories of deaths in order to minimize the number she has to include in her calculus. At this point, though, I want to show why the whole statistic is utterly silly. Let’s remember what it was:
A police officer is 18.5 times more likely to be killed by a black man than a black man is to be killed by a police officer.
18.5 is a lot, and that can certainly make it sound like police officers are really under threat from black men! The implication here is that police officers are showing admirable restraint under the circumstances. They are so much more likely to be killed than to kill, and given that they are besieged on all sides by killer black men, can we really be so surprised that 36 unarmed black men die per year? 18.5! What a ratio! How can Black Lives Matter (I’m sorry, Black Lies Matter) try to draw our attention to such a comparatively insignificant problem!
I can see how that would sound persuasive, actually, so let’s have a look at why it doesn’t actually show that police killings of black men aren’t a problem. Here’s the part where we get to have fun with numbers, and realize how easy it is to churn out meaningless factoids. Okay: A cop is 18.5 times more likely to be killed by a black man than a black man is to be killed by a cop. But wait, what are the ratios for white men? Well, let’s use Mac Donald’s sources and her calculations and find out.
Mac Donald cites official FBI statistics on police killings. The FBI actually says that 41 officers “died from injuries incurred in the line of duty during felonious incidents” in 2015, not 52. Mac Donald examines the “historical average” of offenders’ races, and I’ll do the same: From 2006-2015, 289 of 543 “cop-killers” were white, which is 53%. 53% of 41 deaths is 22 cops killed by white people. Out of 628,000 cops that’s .000035. Now: There were 45 unarmed white men killed by cops in 2015. Out of 100,000,000 million white men, that’s .00000045, which gives us: 77! A cop was 77 times more likely to be killed by a white man than an unarmed white man is to be killed by a cop. Holy hell, these white men are on the rampage. Now, I did that calculation incredibly sloppily and overlooked a bunch of stuff (e.g., how are they counting Hispanic men?) but no less sloppily than Mac Donald. Someone else looked at the numbers and concluded that a cop was 165 times more likely to be killed by a white man than an unarmed white man was to be killed by a cop. Computer scientist Joseph Atkins-Turkish concluded that for all non-black men, the ratio was 124-1. (Manufacturing dumb, inflammatory statistics is fun!)
“A police officer is 165 times more likely to be killed by a white man than an unarmed white man is to be killed by a cop.”
When you start doing these types of ratios, and are slippery enough with your terms, you can come up with all sorts of stuff. Because 0 police officers were killed by women in 2015, but 12 black women were killed by police officers, a black woman was ∞ times more likely to be killed by a police officer than an officer was to be killed by a black woman (or any woman). And because an officer can’t be shot by an unarmed person, an unarmed black man was ∞ times more likely to be shot by a cop than a cop was to be shot by an unarmed black man. How meaningful is any of this? Not very!
There’s something fishy about the whole enterprise of comparing the chances of cops being shot to the chances of people being shot by cops. After all, police officers are the portion of the population that deals with criminals. We’d expect them to have a far higher likelihood of being shot than any random person would have of being shot by a police officer. And it’s only by looking at these ratios that we’re able to obscure the fact that way, way more black people are killed by police officers (300 in 2015) than police officers killed by black people (17 in 2015). Here’s another way of framing the numbers: 18 times more black people are killed by police than police killed by black people!
Now, you could say, as I’m sure Mac Donald would, that since there are far more black people than police officers, this comparison is misleading, which is why it’s wise to adjust the statistics by the population sizes. But this whole enterprise is a very poor way of checking whether there’s something systematically unjust going on. After all, think of what it could conceal. Say we lived in a high-crime city, but occasionally the police just went out and shot a person at random to blow off a little steam. If there was a public outcry, the police could say “Well, we’re far more likely to get shot than any individual citizen is to be shot by us.” But of course that wouldn’t be a defense at all! Or say that for every 20 bombs the FBI received in the mail, it sent one to a random person’s mailbox. Would the “odds” of receiving such a bomb mean it was a defensible, proportionate response? They would not!
Conservative presentations of the facts on shootings are often misleading in this way, suggesting that if they are “comparatively rare” they are therefore not a problem. David French of the National Review, for example, has said that black people are “taught to hate and fear law enforcement, fed on a steady diet of lies about their own country” because factually speaking, “on the whole [the police] tend to use force appropriately.” A statement like that should sound ridiculous to us: Oh, good, cops tend not to murder people. Likewise all the stuff about “black on black crime.” Well, compared to the number of black people killed by other black people, police killings are minuscule. But that doesn’t actually justify extrajudicial executions by the state! (Also, by the way, conservatives who think black communities don’t care about everyday non-police murders are only showing how little they know about the concerns of black communities. The problem is not that they’re not speaking, but that you’re not listening.)
You can say it’s irrational to care about things that are “comparatively” rare statistically. But I don’t think so, because qualitative factors matter as well as quantitative ones, and if Donald Trump did actually shoot somebody on 5th Avenue, it wouldn’t really matter that any individual’s odds of actually being the person were one in 300,000,000 and that (as the NRA never tires of pointing out) far more people die by falling in swimming pools. The correct number of murders by law enforcement is zero murders by law enforcement! What can the statistics tell us about the death of Antwon Rose, for example? Was he not shot while running away unarmed? It doesn’t matter that statistically, it probably won’t be your son. What matters is that he is somebody’s son, and his life mattered, and it was taken from him for no reason by the government.
A lot of conservative defenses of police shootings are designed to distract attention away from examining the facts of the cases and the lives of the people who were killed. For example, they will point to crime rates, to show that black people are supposedly only killed in proportion to the amount of crimes they commit. But that is no defense at all, even on the assumption it’s statistically accurate: You’re not supposed to be summarily executed for committing a crime! In fact, a number of those who have been killed were “criminals”: Eric Garner was selling loose cigarettes, Antwon Rose was resisting arrest, Laquan McDonald was holding a knife and refused to drop it. Unless we live in a country where the penalty for petty crime is instantaneous death, “crime rates” cannot tell us anything about police shootings.
Of course, you won’t be shocked to learn that people like Mac Donald only pretend to care about statistics, and invoke them whenever they can prove that black victims deserve their fate. (Mac Donald is worth paying particular attention to because she’s probably the foremost conservative intellectual specializing in policing, she is highly-credentialed, and her work is cited all over the place.) In a City Journal article on her favorite topic (“the anti-cop” left), Mac Donald uses statistics to raise fears of black “cop killers,” but then immediately turns anecdotal in order to argue that actually, “inner city” people love the police:
There is huge unacknowledged support for the police in the inner city: “They’re due respect because they put their lives every day on the line to protect and serve. I hope they don’t back off from policing,” a woman told me on Thursday night, two nights before the assassination, on the street in Staten Island where Eric Garner was killed.
Why does Mac Donald cite some random woman from Staten Island to prove “huge unacknowledged support”? Because here, if she used the statistics, she’d have to deal with the fact that more than ⅔ of black people think police do a “poor” job of treating races equally, and only ⅓ of black people think police can be relied on to use force appropriately and will be held accountable if they don’t. (About ¾ of white people are confident in police, on the other hand.) The polling makes conservatives uncomfortable, because they have to argue that black people are actually delusional about their own personal experiences, that even though most of them claim direct knowledge of police racism, they are wrong about their own lives. That’s what you have to believe in order to accept Charlie Kirk’s contention that discrimination “doesn’t exist” or is the product of a few bad apples.
There’s one more argument I hear a lot, and it’s kind of an amusing one. It’s that actually police shoot more white people than black people, therefore they’re not racist. Now, the most obvious response is that since there are more white people than black people, that’s not surprising, but I don’t think that’s the right response. The right response is that it wouldn’t really make a difference. Say we did prove that police weren’t actually “racist,” they were just “trigger-happy,” and everyone had an equal chance of being shot, but due to the increased police presence in black communities, black people got shot disproportionately. Should we be satisfied? No, we should not. In fact, there have been outrageous police killings of white people, from the beating of homeless man Kelly Thomas to the choking of Robert Saylor (who had Down syndrome) to that man in Las Vegas who was shot while crying and crawling across the floor. If police “equalized” the number of killings, or if black police officers were equally complicit (most of the officers who killed Freddie Gray were black), it would show only that American police brutalize homeless, disabled, and mentally challenged people equally across races. (And wasn’t that Martin Luther King’s dream?)
There are plenty of robust empirical debates among criminologists, economists, and sociologists about whether and to what extent police use of force is racially biased. I won’t get into those here, partially because people like Kirk and Mac Donald don’t actually want to have an honest conversation about them. They’re not trying to get to the truth, to figure out the real data, they just want to discredit the left by any means necessary, making up meaningless numbers and shifting between statistics and anecdotes as it suits them. But the second reason is that while the social science debate is important in figuring out what the roots of the injustice actually are (Is it implicit bias by police? Is it militarization and draconian laws? Is poverty more relevant than race?), none of its findings are going to make it any more legitimate to have shot Antwon Rose in the back, or to have taken Philando Castile away from the children who adored him. Oscar Grant won’t magically come back to life if someone can successfully challenge the methodology of the Implicit Association Test. Sandra Bland was my age when she died alone in a jail cell, and I don’t care what you call her death, it shouldn’t have happened and it only did happen because we are a fucked-up country that doesn’t care about its people. Erica Garner buried her dad, then worked herself to death trying to get him justice. There is no statistical evidence you can offer that either changes this or justifies it.
Nonsense talking points are easy to manufacture, and numbers are easy to manipulate. Lies, damn lies, et cetera. But even though facts should be vigorously checked, and dishonest arguments disproven, let’s never take our eyes off what actually matters: 1,000 people, each year, who never get to see another day, whose families will never again hear their voices. If that doesn’t make you enraged, then you don’t understand facts, and certainly don’t have feelings.
Photo: NYC 2015 / infootage.com
If you appreciate our work, please consider making a donation, purchasing a subscription, or supporting our podcast on Patreon. Current Affairs is not for profit and carries no outside advertising. We are an independent media institution funded entirely by subscribers and small donors, and we depend on you in order to continue to produce high-quality work.