About . . . . . . Classes . . . . . . Books . . . . . . Vita . . . . . . . Links. . . . . . Blog

by Peter Moskos

April 12, 2015

Killed by Police (2 of 3): Race

[See parts 1 and 3 and NYC]

Using the data from killedbypolice.net, I looked at the race of those killed by police. Though before I give you these numbers, ask yourself this question: what percent of those whom cops kill do you think are white, black, and hispanic. Forgive the callousness, but we're talking numbers. And this figure will get to the core of the notion that cops are "gunning for" or "hunting" black men.

[Update: These are based on the best data from April, 2015. As of November, 2015, we have more data, but I don't believe the conclusions have changed substantively. See also the Guardian and also, slightly better for what they exclude, the Washington Post for more up-to-date numbers on those killed by police. All three data sources are more or less in sync.]

Given that blacks are 13 percent of the population and whites about 65 percent, what percentage would expect to see among those killed by police? Presumably police are more likely to shoot a murderer than an average Joe. So I wouldn't think it's reasonable to expect to find that blacks would be just 13 percent of those killed by police. What percentage would you expect to see. What percentage would mean there isn't a problem of police shooting African Americans in particular (as opposed to police just shooting too much in general)?

Presumably police are (and should be) more likely to kill those who are willing or trying to kill other people. [Update: others have used arrest rates as the denominator, which seems like a good way to analyze racial disparities vis-a-vis police interactions.] Nationwide, blacks are also about 50 percent of murder victims (and thus presumably murder perpetrators, since most homicides are intra-racial). Over the past 10 years, according to FBI data that looks at those who have feloniously killed a police officer, whites are 53 percent and blacks are 44 percent.



[For various reasons, none particularly good, FBI data seems to lumps all hispanics into the white category. The data at killedbypolice.net counts hispanics as separate. But it means that hispanics are invisible in the FBI data but counted in killedbypolice data. And I'm using both. So consider this a fair warning about comparing numbers and figures from different data sets.]

One would hope that the racial breakdown of police-involved homicides would be not out-of-wack with the racial breakdown of those who kill police. This indeed seems to be the case: 48 percent if those killed by police are white, 30 percent are black, and 18 percent hispanic (the doesn't-add-up-to-100-percent bit consists of Asian, Indian, Pacific Islander, and other.)



Though if you're so inclined, and I see so reason not to, you could spin the same data this way:



Now the data doesn't indicate which shootings are justified (the vast majority) and which are cold-blooded murder (not many, but some). And maybe that would vary by race. I don't know, but I doubt it.

Still, per capita, blacks are 3.5 times more likely than white men to die at the hands of police. This is now adjusted for population, and only includes men.



Keep in mind the homicide rate for the entire country of Canada is 1.6. So a homicide rate of 1.3 for black men just killed by police (!) is very high.

While it is a very damning figure for our country, it's not necessarily damning for police. There is a 6:1 (per capita) black-to-white homicide rate disparity and a 4:1 black-to-white disparity (per capita) among those who felonious kill police officers. Given disparate rates of violence, it would be naive to expect equal rates among those killed by police.

If one adjusts for the racial disparity in the homicide rate or the rate at which police are feloniously killed, whites are actually more likely to be killed by police than blacks.

Adjusted for the homicide rate, whites are 1.7 times more likely than blacks die at the hands of police. Adjusted for the racial disparity at which police are feloniously killed, whites are 1.3 times more likely than blacks to die at the hands of police.

Another statistic: A black man is 16 more likely to be killed by a cop than kill a cop. A white man is 20 times more likely.

Though it goes against the all-cops-are-racist narrative, it's not inconceivable, given an equal threat level, that a white person is actually more likely to be shot by police. I've gone into these reasons before, and it's just speculation, but two I want to highlight are 1) cops in more minority cities face more political fallout when they shoot, and thus receive better training and are less inclined to shoot, and 2) since cops in more dangerous neighborhoods are more used to danger; so other things being equal (though they rarely are), police in high-crime minority areas are less afraid and thus less likely to shoot. Based on experience, I suspect that police in high-crime areas deserve more credit than they get for not shooting. Some of the bad shootings I've seen recently... I can't imagine a cop in Baltimore being so damn scared for no good reason.

So I am saying that a guy with a gun in the ghetto might actually be less likely to be shot by a cop who is more used to danger and guns. But a guy with a gun who makes a sudden movement in a neighborhood with a cop who has never faced danger in the face? Boom.

All this said, one should keep all this morbidity in perspective. The odds that any given black man will shoot and kill a police officer in any given year is slim to none, about one in a million. The odds for any given white man? One in four million. The odds that a black man will be shot and killed by a police officer is about 1 in 60,000. For a white man those odds are 1 in 200,000.

But the odds that any given police officer will have to shoot and kill somebody this year? 1 in 1,000. That is not negligible. Add when one adds in the times a cop was afraid for life and didn't shoot? Or a officer did shoot and missed? Or shot and wounded? And then you multiply that by 20 years? Those are odds most people would not accept in a job description.

[Go back to Part 1 or forward to Part 3.]

#

[My critique of ProPublica's misleading claims. More analysis from St. Louis NPR.]

[Corrections welcome. Please double check my work.]

Those last figures are based very roughly on 1 million cops, 1,000 killed by police, 20 million black men, 333 killed by police, 100 million white men, 500 killed by police. For everything else, feel free to check my math and excel formulas, if you can make sense of this:

For reference, the top-left cell is Row 13, column A. Annual rate =SUM((B14/2)*(24/23)). Rate: =SUM(C14/(E14/100000)). Rate LE killed by= =SUM(F14/(E14/100000)). Final rate: =SUM(D14/G14). Hom rate: =SUM(J14/(E14/100000)). Killed by adj by hom rate: =SUM(D14/K14).

31 comments:

CollegeCop said...

Well done

Chris Haugen said...

Very interesting post.

One thing I noticed:

One thing I noticed: In the pie chart showing the race of those who feloniously killed a police officer, there doesn't seem to be a separate category for Hispanics (or Asians). There is an "Other" but its tiny.

In the second pie chart, based on killedbypolice.net, Hispanics make up 19% of those killed by police.

Most likely, in the first chart, Hispanics are being counted under "white", as often seems to be the case. If that is the case, then the proportion of officers shot by non-Hispanic whites is probably significantly smaller than 53%.

Peter Moskos said...

Probably. But we don't know. Indeed, in the UCR data on police-involved justifiable homicides, all hispanics are white. But I don't know about this data set. My guess is that you are correct. But I just don't know.

Peter Moskos said...

The problem of hispanics in FBI is multi-fold. First of all they overlay it on white or black (like the census does). The problem is no police department in the nation does this (at least I don't think so).

Many police departments don't even offer hispanic as a possibility (Baltimore, back when I was there, had only three races: white, black, and other). Some do. Some don't. Some report crime stats to the FBI. Some don't.

So at each level you get compounding issues of problem. The end result is at the end of the day, I don't even want to touch UCR hispanic data.

Anonymous said...

Wikipedia has the same list:
http://en.wikipedia.org/wiki/List_of_killings_by_law_enforcement_officers_in_the_United_States,_March_2015

One good thing about lists of this sort is that they don't wear out -- the improve with use. Messy data can be cleaned up, and the more eyes, the better.

The homicides I checked from news reports seem to be quite regarding race. That is, the victims identified as Hispanic look very hispanic. Same with blacks, whites and asians.

Other thoughts ... If killed by law enforcement lists are 2x as large uniform crime statistics, isn't it reasonable that overall homicide rates are undercounted also?

And a general comment about fear of homicide. People are really afraid of being shot and killed by a stranger. But reading through some homicide lists, domestic and family disputes are a fairly large proportion of white homicides. I have the impression that homicide by stranger is much less common than the overall figures would suggest, especially for whites. And that in the better neighborhoods, it is as low as Europe or the other 'non violent' countries.

Peter Moskos said...

Nice. The wikipedia list has one great advantage: it lists the city (and not just the state). Because along with there being differences between states, I'm certain there are further big difference in cities within states. I'd bet my life that the Riverside PD shoots a lot more than people than San Fransisco, for instance.

As to the overall homicide rates being undercounted. It's not a big problem. I mean, there's *some* undercount, but it's minor. Departments that do not report police-involved homicides (like the NYPD) still report homicides in general. I'm not certain if the FBI has some stick they use to get voluntary compliance on the Part 1 crimes. I suspect they do. So the UCR homicide dataset is *much* more complete than the justifiable homicide issue. Remember, the latter is not a crime.

As to the stranger homicide, even in the worst neighborhoods there aren't a lot of strangers being killed. Of course there are some robberies gone bad and the occasional "stray" bullet. But as a rule of thumb, everybody knows the person who kills them. So, yeah, homicide by stranger is refreshing rare. Particularly for those who aren't messed up in drugs or doing anything illegal.

Still, even if one just looks at the white US, we're still killing each other much more than people in Western Europe (many of whom are not white, of course). Country trumps race, I think. There isn't much homicide even in the "bad" European neighborhoods. A poor black person in England, I would guess, is less likely to be murdered than a middle-class white American.

Peter Moskos said...

Alas, that list on wikipedia is not nearly as complete. Too bad.

Dan Hogan said...

Instead of using homicide rates as a comparison point, did you think about using something else? How about general police contacts? Or maybe number of total arrests?

You could start with a model that any contentious interaction between police in public has the potential to become a police-involved killing.

Peter Moskos said...

Total number of arrests would be interesting. And arrest numbers are pretty well recorded (unlike police contacts).

Adam said...

For someone else's (absurd) take on the new numbers, see:
http://www.cnn.com/videos/tv/2015/04/15/ctn-sam-singyangwe-police-violence-reports.cnn

Apparently police are "targeting" blacks because, by this guy's count, 26% of those killed by police are black. He also claims police killings are dramatically on the rise based on his comparison of numbers from Feb. 2015 and March 2015. Ahhh-mazing.

Anonymous said...

(Part I) Hmm...First, please stop justifying any statistical representation of the deaths of black men as "untrue". The statistics may not be or present a clear picture, there may clarifications or there there may need to be parsing of details, but some of your language is offensive and could blind people, the key audiences you want to reach beyond the law enforcement community, from your message. This is a matter that concerns real human beings who live in mortal fear of law enforcement - this is a feeling of betrayal especially by members of a community who are constantly attesting to their purchase of respect and obedience of the law, constitution and the American way of life. Please stop doing that.

Second, we know law enforcement serves a critical role in our society that's double edged - heroes & guardians of the sanctity of life.

But, I will posit that a lot of the reactions of many black people and men specifically stems from the failure of the law enforcement community to present themselves as their heroes rather than as their lethal disciplinarians reinforcing inequality [or oppressors].

This is especially stark, when law enforcement union's aren't vocal and public about pushing politicians to adopt policies that improve the quality of life in areas that double as law enforcement's most crappy workplaces. For example, I didn't see any major law enforcement union actively and publicly agitate for the gun control legislation following the Sandy Hook Elementary school shooting. T

This is not to say that they don't send letters to weigh in on the matter, it's only to say politics of the law enforcement community aren't consistent.

Now, onto the meat of my comment:

I am not sure about your conclusions simply because when counting the number a civilian deaths resulting from police shootings - there isn't a reliable, official date set of the total numbers because these are not required to be collected by the Department of Justice. That would suggest to me that any conclusions you draw **could be** inherently flawed. That does no one any benefit.

Given the sensitivity of your subject, you **could** inadvertently be contributing to the sustenance of crippling institutional biases, re-enforcing flawed methodologies that clearly neither add value nor preserve life and healthy relationships. In short, you could really be doing debilitating harm to the subject matter and providing aid and comfort to the worst impulses that contribute to social disorder.

Because of your status as an authority, this could also result in take-aways that influence actions of grave consequence.

Further, even the timeline of your data set [http://www.killedbypolice.net/] is problematic because the matter under consideration is a **historical issue** that would be best captured within a decennary time scale and not an annual one.

For example, police killings of black men did not begin to occur in 2013. For example, the Watts riots in 1965 stand as an example for why the time frame within a historic context is key** and the most correct for the delivery of intellectual outcomes of consideration.

What I do also find questionable about your conclusions is that there are three other sources of data on police shootings of black men that receive no mention or inclusion in your blog post.

It should be noted that Nate Silver is extra ordinarily reliable given his previous statistical analysis for the NYT. Please see the link to 538's numbers of black men being killed by the police [a again there are 3 sources of numbers here]: [http://fivethirtyeight.com/datalab/another-much-higher-count-of-police-homicides/]

Another factor that I find problematic in your data is: what are the absolute numbers of assigned police officers by geography. In black neighborhoods these numbers are likely to be disproportionately higher than in neighborhoods of other racial demographics, except **some Hispanic** ones.

Anonymous said...

(PartII) Yes, as you have done, you can make some inferences from simple demographic data. But, those numbers could be reinforcing only the assumptions of the police but not factors that should also inform them.

Neighborhoods that are "perceived" to be more dangerous are likely going to be more heavily policed, which then also provides greater opportunities to detect and counter crime. Little mentioned is the possible conflict of interest for careerist law enforcement officers willing to weigh the risks of working in challenging environments for the sake of pay, pension and promotion.

So, for example, if you have 25 cops per block of 20 feet in a neighborhood perceived to dangerous vs. 2 cops per block of 20 feet in a neighborhood perceived to be less dangerous then you'll have opportunities for catch more people breaking the law. Not only that, there's likely to be more opportunities for overtime and recognized professional merit leading to promotions as well.

Mind you, the above doesn't even get into the demographics of class within specific neighborhoods [especially black neighborhoods], which are likely to be highly dependent upon bias within the real estate industry as well as other data-sets of disparity that have been excluded from police statistics as being irrelevant. The problem with only viewing the issue through specific demographic factors is that they are in fact highly relevant if we're relying on them rather than absolute numbers of offense and arrests, which is the current status quo for police and law enforcement data collection [I could be wrong, but that's what I see presented].

Regarding the economic diversity within a black neighborhood that can affect crime rates, please see: [http://articles.chicagotribune.com/1997-02-02/news/9702020273_1_middle-class-middle-class-neighborhoods]; [http://www.epi.org/publication/making-ferguson/]; [http://www.theatlantic.com/business/archive/2014/05/the-racist-housing-policy-that-made-your-neighborhood/371439/]

In other words, as an example there could be 3 blocks of economically middle class black households whose rates of crime incidence aligns more so with similar economically middle class neighborhoods as defined by the demographics, except race white, which can skew the numbers. Thereby, the data as collected and unrefined presents outcomes that appear more equivocal but are in reality extraordinarily disparate. And, vice versa if you have a poor neighborhood surrounded by middle class households.

The problem with this is that it inflates absolute numbers and statistical outcomes that are 1. not only correct but, 2. so numerically unrepresentative of that the information therein derived is meaningless because it reinforces not only institutionally racist outcomes but misleads the allocation of law enforcement resources in actual operation.

This in turn can be abused by law enforcement which rather than being an actual contributor to social stability finds itself participating in the destabilization of urban areas. By extension, this makes law enforcement as dangerous as the malcontents which contributes to the unsettling of national security in the larger scheme of things.

As an example please see [http://apps.washingtonpost.com/g/documents/national/department-of-justice-report-on-the-ferguson-mo-police-department/1435/]; [http://gothamist.com/2014/12/22/photos_mourning_cops.php].

Another absent factor in your analysis - at lease from parts 2 of 3, is that there's no mention of likely offenders by demographic. Please see: [http://ccrjustice.org/floyd]; [http://www.newyorker.com/news/john-cassidy/the-statistical-debate-behind-the-stop-and-frisk-verdict]. See the numbers on whites and blacks and offenses.

Anonymous said...

(Part III)

Your analysis also fails to account for the presence of law enforcement in other areas of people's lives, specifically black people's lives, which also contributes to the increased likely hood of unfavorable interactions, which may lead to the destabilization or entrenched negative perception by law enforcement which may result in the deaths of black men.

An example is the area of education. Please see: [http://www.npr.org/blogs/codeswitch/2014/03/21/292456211/black-preschoolers-far-more-likely-to-be-suspended]; [http://www.pbs.org/wnet/tavissmiley/tsr/education-under-arrest/school-to-prison-pipeline-fact-sheet/]; [https://www.aclu.org/school-prison-pipeline]; [http://www.theatlantic.com/education/archive/2015/02/race-and-overreaction-on-the-streets-and-in-schools/385076/]; [http://journalistsresource.org/studies/government/criminal-justice/police-reasonable-force-brutality-race-research-review-statistics]

Maybe you don't see any of the above as relevant, but I would encourage to you do so because you have a platform that can influence how people look at these issues and can also positively increase police officer safety as well as their greater harmony with black people in general and black men specifically.

Moreover, there needs to be a paradigm shift and without considering additional data resources and information, we're not move forward as a country. That's not ok.

Anonymous said...

(PartI) Hmm...First, please stop justifying any statistical representation of the deaths of black men as "untrue". The statistics may not be or present a clear picture, there may clarifications or there there may need to be parsing of details, but some of your language is offensive and could blind people, the key audiences you want to reach beyond the law enforcement community, from your message. This is a matter that concerns real human beings who live in mortal fear of law enforcement - this is a feeling of betrayal especially by members of a community who are constantly attesting to their purchase of respect and obedience of the law, constitution and the American way of life. Please stop doing that.

Second, we know law enforcement serves a critical role in our society that's double edged - heroes & guardians of the sanctity of life.

But, I will posit that a lot of the reactions of many black people and men specifically stems from the failure of the law enforcement community to present themselves as their heroes rather than as their lethal disciplinarians reinforcing inequality [or oppressors].

This is especially stark, when law enforcement union's aren't vocal and public about pushing politicians to adopt policies that improve the quality of life in areas that double as law enforcement's most crappy workplaces. For example, I didn't see any major law enforcement union actively and publicly agitate for the gun control legislation following the Sandy Hook Elementary school shooting. T

This is not to say that they don't send letters to weigh in on the matter, it's only to say politics of the law enforcement community aren't consistent.

Now, onto the meat of my comment:

I am not sure about your conclusions simply because when counting the number a civilian deaths resulting from police shootings - there isn't a reliable, official date set of the total numbers because these are not required to be collected by the Department of Justice. That would suggest to me that any conclusions you draw **could be** inherently flawed. That does no one any benefit.

Given the sensitivity of your subject, you **could** inadvertently be contributing to the sustenance of crippling institutional biases, re-enforcing flawed methodologies that clearly neither add value nor preserve life and healthy relationships. In short, you could really be doing debilitating harm to the subject matter and providing aid and comfort to the worst impulses that contribute to social disorder.

Because of your status as an authority, this could also result in take-aways that influence actions of grave consequence.

Further, even the timeline of your data set [http://www.killedbypolice.net/] is problematic because the matter under consideration is a **historical issue** that would be best captured within a decennary time scale and not an annual one.

For example, police killings of black men did not begin to occur in 2013. For example, the Watts riots in 1965 stand as an example for why the time frame within a historic context is key** and the most correct for the delivery of intellectual outcomes of consideration.

What I do also find questionable about your conclusions is that there are three other sources of data on police shootings of black men that receive no mention or inclusion in your blog post.

It should be noted that Nate Silver is extra ordinarily reliable given his previous statistical analysis for the NYT. Please see the link to 538's numbers of black men being killed by the police [a again there are 3 sources of numbers here]: [http://fivethirtyeight.com/datalab/another-much-higher-count-of-police-homicides/]

Another factor that I find problematic in your data is: what are the absolute numbers of assigned police officers by geography. In black neighborhoods these numbers are likely to be disproportionately higher than in neighborhoods of other racial demographics, except **some Hispanic** ones.

Peter Moskos said...

Certainly increased police presence matters in arrests and traffic tickets and interactions in general. I wrote a book about that (and other things). I'm not going to restate it every time I make a blog post.

You've also linked to the entire DOJ report on Ferguson, which isn't very helpful. But I've read the whole thing and commented on it.

And I do talk about "likely offenders by demographic," at least in terms of homicide. I liked to compare apples to apples, and I don't trust a lot of other crime data.

And yes, I look through things with a police prism. That's what I do. It seems like you haven't read much else I've what I've written outside of this post, which makes your points a bit superficial.

Robert VerBruggen said...

I've done some similar analyses myself, including some that remove those counted as Hispanic in the UCR data. (A lot of the time the ethnicity data is just blank, so this is hardly ideal, but I did this to mirror what ProPublica did in its "21 times" article.) I also did a department-by-department breakdown focusing on departments that (according to a Wall Street Journal report) consistently report their police homicides, which helps to address a lot of the data-quality problems.

Here's a wrap-up I wrote, with links to earlier pieces: http://www.realclearpolicy.com/blog/2015/01/09/race_and_police_killings_additional_thoughts_1166.html

Basically, no matter what data you use, the killed-by-police disparity is easily explained by differences in violent crime rates. Good to see the same thing confirmed with the killedbypolice.net data.

Anonymous said...

None of my comments were aimed to project aspersion but to recommend that you consider expanding your data sets - at lead detail why the other resources are problematic and can't be substituted with alternative resources to bring matters to sharper focus. It doesn't matter.

Not "trusting" legitimate datasets is concerning not only because it may imply you selectively choose those resources that suit your experiences, which also may inform what could be argued as fixing the underlying data for prejudicial outcomes. Maybe that;s not what you meant.

Another thing, if there are other areas of the site that contain relevant reference information, link to that so reader's have that cross cutting information; it's appreciated.

The DOJ report aside, there are numerous resources provided you - generalized unqualified refutations leaves the wonder if something's being omitted in your observations.

As you know better than I, being a thought leader is a tasking responsibility - tiresome even. But, if someone expresses an interest, suggests avenues of research with cross references - in their post - then they want to know more and I humbly brought something to the table - not waste your time. No one likes to be rebuffed and that the effort they made to be interested is wasted time.

Merely stating "I've covered this" or making a statement highlighting the commenters' ignorance takes away from your thought leadership and suggests you're about PR; not academic rigor. Not about policy advocacy that improves neighborhoods and thereby the working conditions for officers. None of this may be true, but it could be a take away.

Thanks for plugging your book (will check it out), generally dismissing the rest of my comment as either irrelevant or insubstantial value. Also, noting having read something doesn't mean "here's what I think in light of" the killedbypolicenet.data, for example.

It leaves me with the thought that maybe the "siege" mentality within "cop culture" and anti-empirical study runs so deep that perhaps they have abdicated any other role for themselves that goes beyond executing actions of organized violence for pay, pension & promotion, advancing blue collar labor union politics and right wing paramilitary cultural politics topped off with PR of legitimate good deeds to deflect legitimate concerns about operational procedure that disproportionately results in the deaths of a specifically defined, historically victimized community: black men and boys.

Violent crime rates and the proportionate provision of law enforcement resources cannot be debated.

What is on the table is if whether the thinking within the law enforcement community it doing more self-harm than not; if they're observant of the paradigm shifts within communities and if they can play a far greater and constructive socio-economic-political role beyond budget requests and federal funding for more tactical equipment.

It's not a one way street and leveraging communal malaise for professional benefit, no matter how justified by self, not independent, study means being and asserting non-solution solutions.

We all have to get up and justify our professional roles - how the law enforcement community is conducting itself in these high-profile and egregious incidents overwhelming points to their active investment in the social-political-economic inequalities -no different from, say, slave catchers in the antebellum South - to disqualify them a credible stakeholders for harmony.

Cops killed in the line of duty whose names are etched in marble can be easily countered by the narrative of the enforcement of oppressive policies that in outcome kills black people and reenforces the worst conditions of our society.

You may mean none of it and sincerely so. It fails my confidence to see that result that upholds the otherwise.

Thank you for your time and consideration.

Peter Moskos said...

You make many good points. I just don't have the time to click on every link put here anonymously and analyze or critique it. You might be a brilliant statistician. You might be a kook. I have no idea. I don't know who you are. I don't know your ability to judge data.

If you're going to ask for my time, give me one or two good links. Not ten. Tell me what why it's important and what I can get from it. Help me out.

It would be good if I could link to everything I've covered before. But it's just a logistic nightmare. I don't mean this in a snarky tone, but when I say "I've written about it," I mean you can search for it and find it. I'd prefer you to search for it than me to search for what I've already written.

[Also, your longer comments got labeled as spam, which is why they've only just appeared.]

I'm not going to respond point by point. But what I'm trying to do here is very initial analysis of new -- seemingly good -- data.

This is the first time we have ever had even half-decent data on people who die in police custody! That's important.

So I've started to look at it. No it's not historical. Of course this picture is bigger. So I'm taking what I can from it. Nothing more.

And thank you for your time in writing an intelligent and well thought comment. As I said, you raise many good points.

Anonymous said...

So, 44% of cop killers are black? Yet, blacks are 13% of the US population. Do you wonder why cops are more nervous and trigger-happy around blacks than anyone else? It may not be right, but it's the way to bet -- especially if it's your life at stake.

Peter Moskos said...

But roughly 999,999 out of 1,000,000 black men are not cop killers. So you need to have a bit of perspective.

CBlack said...

Peter, Can you tell me where you got this from: "6:1 (per capita) black-to-white homicide rate disparity"?

NegroAchievement said...

I've been involved in occasional activism around NYC for some time, I've also worked with Michael Meyers of the New York Civil Rights Coalition and learned a lot about police relations from him.

I've been inspired by your analysis of these events to fire up my own blog. I wrote a very opinionated piece about black lives matter but I was wondering if you could have a quick read and see if my stats are wrong on the surface. Iæm not much of a numbers guy but I gave it a go.

Jean said...

I am a bit concerned that you take some very damning ratios in the racial makeup of those killed by law enforcement officers and turn them upside down using the black-to-white disparities in homicide rates and felonious killings of police all based on your statement that "Presumably police are (and should be) more likely to kill those who are willing or trying to kill other people.") I really think this is quite a jump.

I'm not saying that we shouldn't consider these two disparities but there are many factors that play into who is killed by police. Using just these disparities to turn these numbers around and state that whites are more likely to be killed by police than blacks is misleading. By doing so, it reduces the problem to JUST these disparities and limits further investigation.

For instance, integrating poverty statistics into your analysis might show that living in poverty is more significant than your disparities. In a quick analysis I determined that if you limit the population of both races to just those living in poverty, one in 18,960 whites are killed by cops while 1 in 17,274 blacks are killed by cops. Of course, this assumes that all of the 2,177 people killed by cops are living in poverty (no different than the assumption that all of the 2,177 people killed by cops were trying to kill another person or trying to kill a cop). From this it could be claimed that poverty, not race or criminal behavior, is what drives killings by cops.

So, while both poverty and the disparities you cite presumably have an impact on who is killed by a cop, unless this data is being tracked and can be tied directly into the statistical analysis, broad generalizations should not be made. As four months have gone by since you released your analysis and conclusions, I'm sure you're amazed at the "claims" that are being made in online media outlets and social media based on your conclusions. I know I am.

Peter Moskos said...

That poverty twist is quite clever. I hadn't thought of that. I would assume that the vast majority of those killed by cops *are* living in poverty.

I certainly do believe that class is more important than race with regards to police/public interactions (and shootings). I don't think this limits further investigation at all. Why would it?

But I don't think I'm making "broad generalizations" except that violence rates are related to imminent threats are related to police-involved shootings. That's not much of a stretch. I don't think for a second everybody shot by police was trying to kill somebody. I do think the vast majority are threats. That's a *much* lower threshold.

And I'm not certain what claims you are referring to that are being based on these data. But I do think we need to move away from the falst narrative that police-involved shootings are exclusively a white-cop shooting innocent-black-man problem.

j m said...

I have no problem moving from the narrative to which you refer and I believe that narrative can hinder finding a solution. However, your statistical analysis is flawed. If you are going to apply the disparity ratios then the information that the ratios is based on has to have been tracked in your data set (i.e., whether the killings by police happened during an attempted homicide or an attempt to kill police). Nothing about these disparities was tracked in your data set. You make an assumption and treat it as fact.

You say that you are not making broad generalizations except the one that I claim you are making. Again, I have no problem with you linking violence, imminent threat and police shootings. My problem is that you apply it statistically as fact.

With regard to claims, here are a couple:

"Study: More Whites Killed By Police Than Blacks" --
http://www.thefederalistpapers.org/us/study-more-whites-killed-by-police-than-blacks

"Police kill more whites than blacks, but minority deaths generate more outrage: Analysis contradicts widespread views about racial targets" --
http://www.washingtontimes.com/news/2015/apr/21/police-kill-more-whites-than-blacks-but-minority-d/?page=all

These, in turn, are propagandised into social media posts that claim, as fact, that more whites than blacks are killed by police... a fact that neither you nor anyone else has proven because it is not true. In fact, the only valid statistical analysis that you make in this blog post results in a black being 3.5 times more likely to be killed by police than an white.

The way this limits further investigation is that when you pose it as a valid statistical analysis then the majority of the general population believes it and doesn't think there is a further need for investigation since it really ISN'T a problem. That is, if the actual truth is that more whites are killed by police than blacks why would we investigate why more blacks are killed by police than whites? We wouldn't since it's not true. Unfortunately, it is true.

If you are a responsible researcher/reporter, you would retract your findings.

Peter Moskos said...

I don't follow.

It would be nice to have the data you refer to in your first paragraph, but no, it's not necessary. Can you explain why you would expect that missing data to be non-random?

I'm not even certain what assumption you think I'm making. Please explain. What do I say that isn't true? I'm honestly baffled. I'm not going to defend either the federalist papers or the Washington Times. I don't read them. I don't really care what they say. What do they have to do with me and my data analysis? You seem to be making a straw man out of me, and assuming far to much.

As to blacks being 3.5 times more likely to be killed by police, that's a pretty big point! I'm trying to explain what that might mean.

Peter Moskos said...

Also, to be clear, the point of this and related posts was to point out that the ProPublica finding -- widely reported as "young black men are 21 times more likely than whites to be killed by police" -- is complete and utter bullshit. I strongly stand by that.

First we need facts based on the best data we have. And the facts show that blacks are 3.5 times more likely than whites to be killed by police. That's a big difference. And once we agree on that, we can discuss its significance and what it means.

I surmise that this racial disparity is in fact smaller than one might expect, based on overall racial disparities in violence. Yes, this is based on the assumption, my belief based on my knowledge and experience that police do, in fact, most of the time, shoot people in response to imminent threats. You give me different facts and I might change my opinion. Sure. You might disagree with the conclusions I draw from these data, but I think they're reasonable. Reasonable people can disagree. "Retractions" are based on errors in fact. What do I saw that is demonstrably wrong? My fear is that you (and you're not the only one) falsely assume things about my beliefs based more on ideology than on anything I ever said.

Noumenon said...

social media posts that claim, as fact, that more whites than blacks are killed by police... a fact that neither you nor anyone else has proven because it is not true.

The first article you linked mentions Politifact, which says this is a fact:

The Centers for Disease Control and Prevention keeps data on fatal injuries from 1999 to 2011 and one category is homicides by legal intervention. The term "legal intervention" covers any situation when a person dies at the hands of anyone authorized to use deadly force in the line of duty.

Over the span of more than a decade, 2,151 whites died by being shot by police compared to 1,130 blacks.


Also, there's not much you can do about people who are too innumerate to understand that a minority can be targeted at a high rate and still have fewer overall deaths because there aren't as many of them. People like that are looking for slanted facts to support their beliefs, not picking their beliefs because of slanted facts.

Peter Moskos said...

Yeah, more whites *are* killed by police than blacks are killed by police. Not by rate, of course, but by number. But we're both numerate, right?

As soon as I can correct "social media," I'll do so.

mlittlemeyer said...

Peter you have done a great job presenting the data which I, too, have been presenting. Data that counterbalances the false narratives and consensus-realities manufactured by the media and other institutions, with purpose. Funny how people on here keep telling you that you don’t have enough data to support your claims or inferences, which you do when they aren’t misunderstanding or misrepresenting them, however they continue to draw on even narrower data that focuses only on racial disparities that support their mythological, current, racist White Power Structure thesis thinking that their conclusions are automatically in full accordance with reality. Since that is the reality they have been taught.

The links being left you are just endless wastes of time that all have the same fundamental problem: they present data elucidating racial disparities from which racial discrimination and bias is inferred by default without any actual evidence of it. There is an endless pile of these garbage “studies” and written pieces because the media narratives and the “academic” institutions have centered their entire philosophies around this cultural Marxist version of reality. It is secular religious dogma. You can't really fight religious dogmas amongst its adherents. Its futile.

Just like an above poster left you several links such as links showing disparities in the rates at which Black children are disciplined compared to Whites. It’s the same nonsense by which it is assumed that everything else is equal so therefore instead of there being cultural and behavioral differences leading to these disparate outcomes that it MUST be racial bias instead. No other explanation is even acceptable under the current cultural orthodoxy. Particularly where the bubble world of academia is concerned.

Does anyone really think that the same types of disparities can’t be found between different White-American ethnicities, cultures and groups when compared? Does anyone believe that Eastern Europeans such as Serbian-Americans will have the same rates of everything when measured against German, Irish or Italian American etc.? Between different Black-American ethnicities, cultures and groups when compared, such as between black Martinican-Americans like my wife to other Black Americans? Between different Asian-American ethnicities, cultures and groups such as the Vietnamese and Cambodians when compared to Eastern Asians? Of course not.

We don’t really know much about these disparities because we only focus on Black vs White. For all we know disparities between different ethnicities within our racial groups are far larger than between group disparities at the broader racial level. Wherever and however one wants to divide up individuals into groups disparities will exist because that is natural.

There is no equality in nature at the individual or group level. And who pronouncing the greatness of diversity would want there to be? The scariest thought would be a day in which disparities didn’t exist because something quite sinister would surely be underfoot. Such uniform results could only come through unimaginable top down control and oppression or a world in which the population have become completely controlled automatons or something. It would take the most evil forms of anti-humanity and social-engineering to accomplish equality of outcome. Something which should not be a desirable goal by any sane human being.

mlittlemeyer said...

And speaking of per capita disparities, the disparities between genders are far larger than between Whites and Blacks, when it comes to every element of the criminal justice system; including the rates at which men are shot by cops, arrested, prosecuted, incarcerated, and the amount of time they receive for criminal offenses. Literally, almost all police shootings, justified and unjustified, are of males. So if all of these things were evidence of structural biases and discrimination, the anti-male biases and discrimination in our system would be much, much, much stronger. The argument that males are being targeted disproportionately, is much stronger statistically, than is any argument for racial bias. Yet it is funny how those screaming about these racial disparities being evidence of racial biases and discrimination accept that because males commit more violent crime that all of these same disparities as it relates to gender, just makes sense. Because males are not a favored group under the current cultural orthodoxy, anti-male bias doesn't fit their narrative. So they dismiss it out of hand as quickly as they accept racial bias without actual evidence. The assumption always that it is there even if and when it can’t be demonstrated.

At any rate, thanks for all of the great posts.