Thank you for this data. I'm going to be using it on my Race Haven Podcast. Race Haven is solution focused dialogue about race relations in America.

Thank you so much for this. Really good example for how to teach statistics using real data.

Using your logic men are killed at an alarming rate compared to their proportion of the population(95% of victims vs about 49% of population). Where is the outcry? You'll then point out I am using the wrong denominator because men commit more crime, are more violent, and therefore are more likely to be confronted by police. Possible you got the other denominator incorrect? When is the appropriate denominator. It sure isn't percent of the general population.

I would be interested to see someone drill a little deeper into this dataset. In particular I think it would be worthwhile to look racial makeup of victims vs total population within a defined geographic location ie. City or neighborhood where the killing took place. The stats above would be more relevant if racial makeup was a normally distributed population across the nation. It seems to me that narrowing our area of interest (within reason) would remove some potential bias from the analysis.

What is the breakdown of number of crimes REPORTED in black, white, hispanic etc neighborhoods.

THANK YOU! The next time someone says that African Americans are the only innocent ones getting gunned down by police I'll be sure to show them this!

I found this because in the wake of the killing of Philando Castile, somebody shared it on Facebook with the caption "Data doesn't lie." Data doesn't "lie" but it can easily be misinterpreted. I'm not a sociologist, statistician, or PhD, but I have some comments to offer.

The author's conclusion, based on the data, is "This data further solidifies the concerns and demands of the Black Lives Matters movement; there is a pattern of police using deadly force disproportionally on blacks that needs to be addressed by policy makers and police departments across the nation."

But that is an interpretation superimposed on the data; the data doesn't explicity say that, and is actually insufficient to support that claim. (Note that I am not saying the claim is false, only that the data as presented isn't enough to prove it.)

An equally valid interpretation, according to the author's methodology, is that "there is a pattern of police using deadly force disproportionally on MALES that needs to be addressed by policy makers and police departments across the nation". After all, the data show that "Most, 95.2%, of those killed by police in 2015 were men. The pie chart below shows that only 4.8% of the victims identified as anything other than male." That is a far greater disproportion than the racial breakdown.

So what gives? Obviously there must be some additional factor(s) not used in the correlation of the data. The author gives a hat tip to it. But he doesn't factor it in before reaching a conclusion. Among those killed by police - of any race, of any age, by any means - what were the circumstances? Was it during the commission or investigation of a crime? Did the victim resist? Did the officer happen upon the victim, or was he responding to a complaint? Was it sunny and clear, or dark of night? What was the race of the officer?

In these emotionally-charged times, please do not think I am trying to excuse unwarranted killings by police; they should be investigated and justice should be served. But faulty analysis - especially when presented as self-evident "data" - doesn't help the situation, it can only lead to the wrong solutions.

Flawed. You need to show the crime rate in each category.

So, you are comparing people killed by cops in 2015 to population census data from 2014. Why not compare data from the same year? Are you sure there are more hispanic/latin americans than blacks in the US? I want to see the raw numbers, not just race/ethnicity proportions (i.e. percentages) of US population.

This data is meaningless without including statistics to address the percentage of each race committing crimes. If a particular race is

I looked at the article, but did not read the whole thing. Did I miss the stats on the percentage of violent crime committed by each of the listed ethnic groups compared to their percentage of the population?

now, show data about crime rate by race...you will get an answer for why the percentage is so high for black deaths...they simply cause more crime.

Nice work.

Why are you comparing The Guardian's data to general demographics instead of those who have been approached by the police?

There are statistics on a vast array of scenarios here, but I was curious as to whether statistics are kept concerning the killing of police by people of different ethnic backgrounds. Could you steer me in the right direction?

Certain racial groups are charged with more crimes by our justice system and have frequent police encounters. How does this not weigh into this "exhaustive" study.

Unless, I'm missing something, if the object is to prove whether race matters in a lethal encounter with the police, the important data point is not the percentage of each ethnic/racial group in the the total population, but their representation in encounters with the police. That is, how many Whites, Blacks, Hispanics, etc. were stopped by police and not killed. I doubt that data is easy to come by. I'm pretty sure blacks have proportionally more encounters with police than Whites in the US (which may well indicate structural and direct racism among other factors), so this may account for at least some of the disproportion with respect to the total US population. No?

I disagree with the author's implication that the differing shooting rates is largely a result of police racial bias (as claimed by Black Lives Matter). Blacks, due to socio-economic reasons, have much higher rates of police interaction, due to the high crime areas so many find themselves living. (stemming from destructive gov programs IMO). This much higher interaction will lead to higher rates of violent encounters that result in police use of force. The perception of police oppression can further be expected to create a more hostile environment and result in further escalation of resistance. In fact if you consider arrest and incarceration rates by race, you would expect the death rate among blacks to be much higher.
Note again that nearly twice as many whites were killed as blacks.

By limiting the data to race relative to percentage of population, this study ignores factors that may lead to quite different conclusions. The study shows that, "compared to their portion of the overall population, blacks are disproportionally the victims of lethal force by law enforcement officers, even when they are unarmed." What if the population figures were limited to the areas where most killings take place? If most law enforcement killings take place in inner cities and the population there is predominately black, then this might account for the increased percentage of black deaths.

What about population percentages relative to crimes? Are there more police shootings in areas with very high crime? Are blacks disproportionately represented in high crime areas, and therefore more likely to be victims of those shootings? And what percentage of the black killings were by black officers (and vice versa)?

By focusing solely on race versus general population, this study implies that police shootings are racially motivated. If other factors are considered a different picture may be presented. And if a distorted or incorrect cause for these killings is reached and accepted by the public and those in government, steps that are taken to correct this problem will not only be ineffective, they may worsen the problem.

Something has to be done, but basing policy and law on incomplete and potentially incorrect studies is not going to solve anything. The authors of this study should be looking at all the potential possibilities, rather than settling for the easy one.

I would hope that truth was more important.

Now, please, do a study on race and crime. The two go hand in hand. It is clear blacks are disproportionately the victims of lethal force but someone needs to answer the question of WHY. The answer is far more complicated than racism. The answer lies somewhere in a muddled concoction of history, fading racism, poverty, and crime. Before we can EFFECTIVELY address and resolve the problem, we as a nation need to understand the cause.

Thanks for sharing this disturbing (but important) analysis), Todd. I had a read a similar piece recently (sorry, can no longer locate) that suggested the underlying racism if our society was more to blame (people of color were more likely to be living in high-crime areas, for example) then racism of police. Thoughts?

Where in your article did you mention that blacks comment a higher disproportionate amount of violent crime? Does that not explain why the police are many more times likely to confront Blacks than whites?

Wouldn't it be more accurate to compare the percentage killed against the number of interactions with police instead of overall population? The general population totals include ages 0->100, disabled or otherwise incapacitated, etc. A large portion of the population does not interact with law enforcement at all or only in situations of a benign nature. I believe it would be more accurate to look at the numbers involving situations where police became involved due to the probability of laws being broken, threats of violence, etc. What is the percentage of police involved situations that turn violent and/or deadly? What about the race of the officers, not just the suspects involved? These are the numbers that would truly reflect a problem with race by law enforcement.

Excellent article. Being in a technical field we always look to root causes to slice problems. I would expect Dr. Beer to dig deeper in to the causes of disparity found by race and not simply assume racism as the cause. Compare these numbers against the race of police officers involved; Perhaps a review of crime statistics by race as published by the FBI and marry the 2 data sources to build evidence based hypothesis before jumping to a conclusion. I have done this and have a hypothesis that police are not 'racist' because of society but because of personal experience on the job and have been conditioned subconsciously to have biases tied to the offenders race. Such that race becomes an intermediate marker as to the causal. In the high level data this presents itself as apparent racism by police , regardless of the race of the officer. Police have a very difficult role and should be given training to help individualize thier tactical response. On the other hand addresses causes of crime in specific populations at a root cause level would address the other side of this issue. Oversimplification doesn't help solve the issue.
Regards,

"If we lived in a society where one’s race had no impact, then we would expect to see the portion of each racial/ethnic group killed by police equal to that of the portion in the general population. "
The thesis is flawed. There are two critical stats that are missing in order for the thesis to hold:
1) the proportional rate of crime per group (more crime = more run ins with police)
2) the race of the shooting police officer.
The authors given data processing the old expression, "there are three kinds of lies: lies, damn lies and statistics.

As so many have said in replies, although not as directly as I will, this analysis of the data is misleading and is confounding our understanding of the situation. As a result, it is making everything worse. To repeat a point that was stated by a few, the prevalence of males in the list of "victims," by this same simple logic, would "prove" that the police hate men. The fact that the data cannot be normalized by several important but unrecorded factors is not justification for ignoring these factors. The author unwittingly escalates matters by encouraging all people of color to identify with violent criminals who confront police. This simple treatment combines with anecdotes of truly poor judgment by police to reaffirm a preconception that makes no sense: police everywhere hate black people.

Would be good if you also added the perspective from a police officer. A Police Officer makes up 806k of the population and had 130 death while on the job. It certainaly is another perspective.

It only makes sense to use the number of interactions with police as the denominator in this case, IF you assume that would be an unbiased sample. Which it's not. Profiling begins far before a person is killed. Profiling is part of the reason that blacks are killed at higher rates - because law enforcement believes that they are more dangerous from the start.

By the way, this is also the reason that men are killed more often than women. Law enforcement officers think men are more dangerous from the start - similar to the way they think black people are more dangerous. Obviously there's a bias there too. That doesn't invalidate the original claim - it's a separate but related issue.

So many of the responses above show that people still want to believe that those who were shot deserved it. In short, their presupposition, while accusing the article of having an underlying bias (which I think it does) is that more blacks are shot because more blacks (per percentage of the population) are involved in crimes and violence. I think we still really, really want to excuse a systematic problem because it doesn't line up with our view of what America is about. This is called denial. I think (I work in mental health) that there is such a thing as an addiction to power (think domestic violence), and that too many people are drawn to law enforcement with its guns and power, who have this addiction. The over use power because the use of power is a drug. This group needs to be weeded out. If they are not, they endanger all the good people who are part of the police community. And they endanger all of us. I think the disparity is significant. Some deviation in % can be explained away, but when the rate is double? However the most startling statistic is simply the raw number of American's killed by those charge to serve and protect. Open carry is making this number rise (a policeman will shoot quick with less provocation if the other has a gun, legally or not). But we must address the number of people killed, white, black, asian, hispanic...all

The article does not take in to account that crime rates affect police

Given the history of the US I'm sure there is racial bias in the way the law is enforced. But comparing deaths vs. % of general general population is simply insufficient to draw firm conclusions about racial bias. What about other contributing factors to the deaths? How many attempted arrests involved a struggle? How many deaths were related to reports of violence by the suspect? How many acts of violence against police are committed by each race? It's a complex topic, but to really figure out solutions and prevent people from dismissing the data, you have to conduct a more thorough analysis.

I see many are trying to edge in the criminality by race. I guess then, maybe it will go into educational and economical disparities and inequalities since we're still seeking to vilify blacks on the basis of supposed higher criminality. Nice try. Amass ALL the data since we got off the boats in stolen America and you'll still come out having to swallow the pill you run from.

Thank you for putting together this page, super comprehensive and much-needed. The only tweak I would make is to the column in the first table labeled "percentage of THOSE killed by police." It's not wrong, but it could be more clear to distinguish "those" since it can construed as "the percentage of people of ___ race that were killed by police with respect to the total number of people killed by police." -Or- "the percentage of ___ race that were killed by police." Obviously it's the first and not the latter because 50.1% of all white people weren't killed by police...However, the second is an interesting statistic as well that I was looking for, specifically, and the title of that column in that chart threw me off upon first glance...a minor critique for an otherwise excellent statistical study.

I must beg to differ with the professor. His data is skewed.
Where his numbers are flawed is he is comparing incidents with the general population. He fails to understand it must relate to the criminal population in contact with the police.
Until they do that these numbers will be flawed.

Where in your article did you mention that blacks comment a higher disproportionate amount of violent crime? Does that not explain why the police are many more times likely to confront Blacks than whites?

Where in your article did you mention that blacks comment a higher disproportionate amount of violent crime? Does that not explain why the police are many more times likely to confront Blacks than whites?

Where in your article did you mention that blacks comment a higher disproportionate amount of violent crime? Does that not explain why the police are many more times likely to confront Blacks than whites?

– FACT: Despite making up just 13% of the population, blacks commit around half of homicides in the United States. DOJ statistics show that between 1980 and 2008, blacks committed 52% of homicides, compared to 45% of homicides committed by whites.
More up to date FBI statistics tell a similar story. In 2013, black criminals carried out 38% of murders, compared to 31.1% for whites, again despite the fact that there are five times more white people in the U.S.
– FACT: From 2011 to 2013, 38.5 per cent of people arrested for murder, manslaughter, rape, robbery, and aggravated assault were black. This figure is three times higher than the 13% black population figure. When you account for the fact that black males aged 15-34, who account for around 3% of the population, are responsible for the vast majority of these crimes, the figures are even more staggering.
– FACT: Despite the fact that black people commit an equal or greater number of violent crimes than whites, whites are almost TWICE as likely to be killed by police officers.