Deaths in police custody, research results

The limited data available do not suggest a recent overall increase in the number of homicides by police or the racial composition of those killed, despite the high-profile cases and controversies of 2014-2015, according to a New York Times analysis. But a January 2015 report published in the Harvard Public Health Review, “Trends in U.S. Deaths due to Legal Intervention among Black and White men, Age 15-34 Years, by County Income Level: 1960-2010,” suggests persistent differences in risks for violent encounters with police: “The rate ratio for black vs. white men for death due to legal intervention always exceeded 2.5 (median: 4.5) and ranged from 2.6 (95% confidence interval [CI] 2.1, 3.1) in 2001 to 10.1 (95% CI 8.7, 11.7) in 1969, with the relative and absolute excess evident in all county income quintiles.”

And this:

For the most recent period where statistics are available (2003-2009), the BJS found that 4,813 persons “died during or shortly after law enforcement personnel attempted to arrest or restrain them… About 60% of arrest-related deaths (2,931) were classified as homicides by law enforcement personnel.” However, among these 2,931 homicides by law enforcement personnel, 75.3% were reported to have taken place in response to a violent offense — constituting a force-on-force situation, such as an intervention with an ongoing assault, robbery or murder: “Arrests for alleged violent crimes were involved in three of every four reported homicides by law enforcement personnel.” Still, 7.9% took place in the context of a public-order offense, 2.7% involved a drug offense, and among 9.2% of all homicides by police no specific context was reported.

There is much more of interest at the Harvard Kennedy School link.


This is not surprising. What has increased is the awareness of these issues due to social media.

I think cellphone video is probably the big driver in the visibility of police brutality cases. Without video, you have a criminal making one claim, and a cop making a contradictory claim, and no evidence either way. (The "criminal" may not actually be a criminal in the sense of having done anything wrong, but he probably is in the sense that he's been charged with some crime, if only resisting arrest.) With video, you have actual evidence of the cops lying about what happened, often without the cops actually facing any consequences at all for lying on official reports and beating someone up.

Also, with video, you have something that you can show on TV, which does pretty well at grabbing eyeballs.

With video, you have actual evidence of the cops lying about what happened, often without the cops actually facing any consequences at all for lying on official reports and beating someone up.

You had video in the Rodney King case. The news media helpfully deleted the first few seconds of the video in their broadcasts so you would not see King charging the officers. The media was also fairly studious about burying an interesting tidbit: that King's companions were sitting in the car unmolested by the police.

The sociopathic character of the media is a problem here.

FWIW, I've followed the Freddie Gray case via the BBC. It wasn't until I read an account on another website that I learned Gray was a heroin dealer with a long rap sheet. (Obviously, I'm not saying he deserved to die in this manner.)

I can't recall anyone suggesting that things have gotten worse recently. The fact that things have stayed the same, i.e., bad, is precisely the problem.

Actually, the probability of being killed by cops has come down a lot since the 1960s according to this study. The black-white disparity has decreased a lot as well.

Decreasing lethality does not necessarily mean that the police have become less trigger-happy. Violent crime is down, and medical advances have saved lots of people.

Violent crime is down making use of 1980 as a benchmark. Less so making use of 1969.

Yeah, but way way down if you use 1925.

The rule of thumb here in the Third World (Philippines) is never to call the police, not that they would come promptly anyway. If you see a problem, avoid it. Unfortunately sometimes you get rare cases such as serial killers who prey on their neighbors (happened last year in Manila, it was some brothers who were thrill killers and preyed on a poor neighborhood who was terrified of them) but mostly crime here is for a reason (contract killing, honor killing, killing of journalists for some expose, killing for money) and the cops do a decent job of tracking criminals "in hot pursuit" (their detective work is shoddy however). The USA would do well to emulate this tactic, instead of having police on every corner looking to arrest somebody. Cab drivers are often a bigger deterrent to crime than a policeman on every corner. Ditto CCTV.

The increased efficiency of police is a potential problem. Especially for things like bench warrants, probation violations, and the like.

Honor killings? You talking about Mindanao or something?

Well, the first two comments make the same point - there is nothing recent about any of this. Only the documenting is becoming harder to dismiss, since it no longer requires a broadcaster to make information available to a large audience that can then decide whether to believe their eyes or the police lies.

There are no 'police lies'. The two subpopulations differ according to the commonality of violent behavior therein. No one who's looked at descriptive statistics on the comparative prevalence of homicide, rape, or robbery would be surprised that there are comparatively more violent encounters between police and populace. You'd only be surprised if you fancied the whole phenomenon was driven by police aggression. That's not a police lie. That's your lie.

Are you really saying the police never lie in these cases (police brutality, police shootings, etc.)? I assume you're trying to make some other point, because it would be pretty amazing if police training and hiring procedures were so perfect, nobody capable of telling a lie ever got hold of a badge.

Are you really saying the police never lie in these cases (police brutality, police shootings, etc.)?



He is saying this subject is suffused with lies by police. These data do not show that police are 'lying' or that the phenomenon in question has aught but contours you'd expect if the phenomenon of violent crime in this country were familiar to you.

But even the baseline is screwed up. Police lie just as much as anyone else, and they are automatically the more "reliable" source in situations of alleged misconduct, so their story typically prevails.

Police lie just as much as anyone else,

And we have it on your authority.

“The rate ratio for black vs. white men for death due to legal intervention always exceeded 2.5 (median: 4.5)

The homicide rates in and among the two subpopulations also differ (by a larger ratio).

Actually, 4.5 is probably a reasonable cut at the difference in rates of criminal behavior between blacks and whites, so we might be looking at a difference in (say) how often blacks vs whites are arrested driving the difference in how often they're killed by the police. Imagine if each time you're arrested, there's some small probability you end up dead. That model might plausibly explain the difference in black vs white rates of being killed by the police.

Now, if that model is right (I'm sure it's way oversimplified--think of differences in kinds of crimes, neighborhoods, etc.), then we may have a problem with the police being broadly too brutal, or too willing to shoot fleeing suspects, or something. But it's probably not primarily about race.

That's important, because the kinds of solutions you'd want for an epidemic of racist policemen looking to shoot black kids are completely different from the kinds of solutions you'd want for an epidemic of unnecessary police shootings.

You're presuming good faith on the part of various parties. If the party is Mark Kleiman, that might be reasonable. If the party is ericholder/alsharpton/bencrump/ryanjulison, nope.

Wagers on the table: we're not having these controversies because there is an unusual problem. We have them because a concatenation of characters despise law enforcement and need a wedge issue.

Conspiracy theories abound. They just hate law enforcement! Why? Just because!

Conspiracy theories

That term does not mean what you think it means.


While we're at it, Ben Crump and his sidekick Ryan Julison are conspirators.

Mark Kleiman is interesting. For a reasonable data-driven centrist on drug and police matters, over at Reality-Based Community he spends alot of time taking potshots at Republicans. Other contributors there do not. I wonder if he's compensating for something.

The mortality rates at the hands of the police of 15-34-year-old black men compared to white men changed over time in this way:

1965: 7.65 x the white rate

1975: 7.56 x the white rate

1985: 4.48 x the white rate

1995: 3.37 x the white rate

2005: 2.57 x the white rate

My hypothesis would be that in the 1960s and 1970s, given an objectively similar situation, the police were more likely to kill a black suspect than a white suspect. But this changed by the 1980s, and nowadays it may be true that cops hesitate more with lethal force when dealing with blacks (there's some recent experimental evidence that supports this).

I think the changes over time in the racial composition of those who kill cops might be a useful proxy for studying this. If the police are more likely to shoot first and ask questions later when dealing with one race versus the other, then members of the latter race will probably more often have an opportunity to shoot back. I noted below that the black share of cop killers is nowadays higher than the black share of those killed by cops, while the opposite is true of whites. I don't know if this was the case a few decades ago.

My experience has lead me to believe that some police believe that night sticks are for the backs of trouble makers. They seem to believe that they are doing the citizens a service. I think that they are wrong but I am open minded enough to know that there is a small possibility that they are correct.

I knew a fireman who told me about some teens were throwing rocks at the firemen while they were putting out a fire. They called the cops and the the senior cop a huge guy said "I will show you how to handle these punks". The firemen said "He was tossing those teens around rag dolls".

I once saw a cop purposely hit a bad kid that I new with car sending him sprawling on pavement. I was once harassed and held up by a cop for an hour and then let go. The cop said I matched the description of a guy breaking and entering. Driving by the scene of an accident one night a cop pulled me over and tapped me on the chin with a night stick.

I think it is problem and they should try to stay within the laws but that as long as there are cops you will have stuff like this and even they will occasionally kill someone and maybe even an innocent person. We should try to reduce this. I think ending the war on drugs might help and black cops for majority black areas might help also.

I have little sympathy for "teens [who] were throwing rocks at the firemen while they were putting out a fire".

@Bostonian Yes but do you think that the police should be judge, jury and executioner of the punishment? Anyway if we are to be truthful about Baltimore we need to discus this honestly. How about a couple of cops who ran into a teen with a known criminal record in their car, he was OK but had his head hit the cement he would have been dead?
Are you also comfortable with that? I think some cops think it is their job to rough up and intimidate that element.

I was once harassed and held up by a cop for an hour and then let go.

Quit toking on the street corner, and that won't happen to you, Spicoli.

I was not much of user and did not that day.

My father in law was a cop years ago and felt a bit sorry for the kids now. Back then they could whack them (both black and white) with a backhand or scare the crap out of them.

Now they get processed and have a record over some minor things.

That is a good point, the criminal justice system is like Russian roulette. You could get nothing but you get serious time for selling a little marijuana/Crack/Acid etc.. I few whacks with the night stick would be preferable. But it is my opinion that if the cops played by the law the laws might be changed.

It reminds my Philip Crosby in his book about quality, saying that machinists should make the part according to drawing even if he know the drawing is wrong so that it is fixed at the right level or else quality will suffer. The laws need to be fixed. If cops were filling jails with the children of the middle class arrested for using selling recreational drugs the war on drugs might end quickly.

Good point. I do still see a place for the police to make minor corrections on the fly though. Like warning tickets.

Imagine what could have happened if you were black.

A few years ago on an African game preserve, some juvenile male elephants were rampaging, goring rhinos to death and generally being out of control. Solution: the wardens imported some fully-grown bull elephants who proceeded to stomp the juveniles until they stopped their antisocial behavior.

To channel Bloomberg, if you look at the race of people who fire guns at police versus the race of people killed by police, whites and Asians/other are the most victimized group.

Is that data for New York City? Here's some national data from the Fatal Encounters project. Blacks comprise a larger share of those who kill cops than of those who are killed by cops (although the absolutely number of the latter is much greater).

What fraction of Asians arrested by police are killed by them? What fraction of women? I bet Asians are less likely than whites and women less likely than men to be killed after being arrested, because those groups are less likely to resist arrest.

I can't explain it. Cops don't seem to kill Chinese, Japanese, Navahos, Pakistanis, East Asian Indians, Egyptians, Inuits, . . .

What about Icelanders?

lol. Because Icelanders have "stones." They told the ECU that Iceland's taxpayers weren't going to (likley could not) cover europeans' credit losses in Icelandic banks. The ECU creditors could get in line with the rest of the general creitors of Icelandic banks. Have you been watching "Vikings" on History Channel? I wouldn't want to be the cop if one of them decided to resist arrest.

Latinos? What about the 17% of the US?

I think 8, J, Bostonian are bringing up issues of subtle controls. I have a more basic question: Is this study controlling for the arrest rate?
That difference alone may be high enough to explain the fatality differential.

Yeah it seems like the reference population includes non-criminals as well. It's a bad sign when the "best and brightest" (Harvard) are so blatantly biased.

When have the “best and brightest” ever not been blatantly biased? In every human society, there's always been some body of dogma/doctrine the holding of which is a necessary precondition to membership in the elite, and which speaking against is generally punished. (When such bodies of dogma include supernatural claims, as they did in pre-modern times, they are called official religions, and the speaking against them blasphemy or heresy.) The "bias" here is simply a result of the positions mandated by the official doctrine of our elite.

They aren't biased. They aren't attempting to explain racial differences. They just document them.

Bias is 90% presentation - that is, what you choose to or to not document. In fact, I would say selective documentation it is the worst form of bias because most will never even realize they're only being informed about a small part of the story. If national media were to only run stories that covered the murders of Black men by Chinese men, it would no doubt create an amazingly distorted view of reality. Would you call that unbiased? Because strictly speaking, that would only consist of "documentation".

The quoting from the BJS site is unfortunate and perpetuates a common misconception when discussing this topic. The full quote is:

"Of reported arrest-related deaths, 61% (2,931) were classified as homicides by law enforcement personnel, 11% (541) were suicides, 11% (525) were due to intoxication, 6% (272) were accidental injuries, and 5% (244) were attributed to natural causes."

The immediate response of most when they hear a death is classified as a homicide is to assume that means criminal. But the "homicide" classification simply means that it was death that happened as the result of the direct action of human. Technically, suicides are homicides, but are separated out for statistical purposes.

What the quote obfuscates is how many of the deaths classified as homicide were due to legally justifiable use of deadly force (self defense/defense of others), a result of injuries sustained under permitted use of force by law enforcement, negligence, willful disregard or wrongful acts. We may also want to know how many deaths due to permitted use of law enforcement force get classified under accidental or natural causes (aggravated medical conditions).

Comments for this post are closed