The predominant factor in black deaths by police is more crimes commited - not racism

[maxparrish]

Except those agencies cover most of the population:

In 2007, law enforcement agencies active in the UCR Program represented more than 285 million United States inhabitants—94.6 percent of the total population.

Looks like whoever does not participate must be nit-picking small irrelevances.

http://www.fbi.gov/about-us/cjis/ucr/frequently-asked-questions/ucr_faqs

Here's what the UCR tracks:

The UCR Program collects offense information for murder and nonnegligent manslaughter, forcible rape, robbery, aggravated assault, burglary, larceny-theft, motor vehicle theft, and arson. These are Part I offenses and are serious crimes by nature and/or volume. Not all crimes, such as embezzlement, are readily brought to the attention of the police. Also, some serious crimes, such as kidnapping, occur infrequently. Therefore, the UCR Program limits the reporting of offenses known to the eight selected crime classifications because they are the crimes most likely to be reported and most likely to occur with sufficient frequency to provide an adequate basis for comparison.

This caveat is also specifically spelled out and bolded on the UCR page you linked (I underlined and italicized what I thought were other important parts):

Each year when Crime in the United States is published, some entities use reported figures to compile rankings of cities and counties. These rough rankings provide no insight into the numerous variables that mold crime in a particular town, city, county, state, or region. Consequently, they lead to simplistic and/or incomplete analyses that often create misleading perceptions adversely affecting communities and their residents. Valid assessments are possible only with careful study and analysis of the range of unique conditions affecting each local law enforcement jurisdiction. The data user is, therefore, cautioned against comparing statistical data of individual reporting units from cities, metropolitan areas, states, or colleges or universities solely on the basis of their population coverage or student enrollment. “Variables Affecting Crime” in Crime in the United States has more information on this topic.

And yet people wanting to blame black people for the amount of police violence visited upon them freely use this incomplete analysis to blame an entire demographic of people.

Whoops, I meant to add in this section that the UCR doesn't track the amount of times police officers shoot people so why should we expect any relevant information on how often and why police officers resort to deadly force in the UCR. So it doesn't surprise me that the UCR gets way more voluntary responses than the ARD does.

No one need "blame" people for being black, nor more than one should "blame" the innocent young black male for the unfortunate fact that so many of his racial and age group peers are disproportionately criminally prone. The fact remains, however, more blacks will have more run-ins with the law because more of them (proportionally) commit crimes of violence. And being "profiled" for being in the wrong place at the wrong time is a natural consequence and necessary.

Lets just look at the facts we DO (and don't) know:

1. UCR covers 95% of the population. Of those "between 85% and 90% of all homicides reported in the UCR summary data also have a corresponding SHR form." The SHR form is the one that provides information on total number of (police) justifiable homicides, and racial breakdowns. It is unclear what percentage of the covered population, are not covered for justifiable homicides and racial breakdowns for those homicides.

2. The three official sources of data provide similar consistent results. The FBI UCR, the CDC NVSS (based on death certificates), and the DOJ Statistics report around 400 to 450 police homicides per year. http://www.motherjones.com/politics/2014/08/police-shootings-ferguson-race-data

3. A separate study using a different methodology, the NVDRS states (16 of them), reported 125 legal-intervention deaths. "... 56.9% were non-Hispanic whites and 29.2% were non-Hispanic blacks. With respect to location, 44.6% of legal-intervention deaths occurred in a house or apartment, 29.2% on a street or highway, and 7.7% in a motor vehicle (Table 35). The majority of decedents were aged 30–54 years (Table 36). Of the 92.3% of decedents from legal-intervention deaths who were tested for alcohol, 40.1% were positive for alcohol, and 79.6% of these decedents had a BAC of ≥0.08 g/dL (Table 37). The percentage of victims tested for other substances varied (range: 47.7%–80.0%). The presence of other drugs for which tests were positive also varied: 17.7% of those tested for marijuana, 15.5% of those tested for antidepressants, 10.8% of those tested for opiates, 10.6% of decedents tested for cocaine, and 4.0% of those tested for amphetamines were positive for these substances (Table 37)."

These four sources use different methods, and while each has its weaknesses, the overall impression is pretty consistent (note: the NVDRS study does not state the total population of the 16 states). For example, a chart of the first three sources:

4. There are reasons to suspect significant under counting. They are:

Undercounting:

- Not all agencies report SHR statistics in their UCR. Exactly what proportion of the US population that is covered by SHR is unknown.
- Some number of 'unjustified homicides' by police might be left out.
- Some coroners, issuing death certificates, may leave out police as a cause of death.

5. The total number of police homicides may be much higher. The WSJ survey showed a 50 percent under-count from the sampled agencies. http://graphics.wsj.com/justifiable-homicides-by-police/. At least one crowd source site suggests in might be three times the reported number.

6. Whatever the total number of police killings each year, be it 400, 600 or 1200 the range of black deaths seems to hover from 30 to 43 percent as a proportion. In other words, the proportion you might expect for the crimes and demographics involved.

 

[Axulus]

I think the point that an unarmed suspect is much less likely to engender a reasonable fear of death than an armed suspsect was obvious. So your grasping at straws indicates that you recognize the paucity of your position.

You realize you are making my case, not yours with those examples.

How so? Killings of unarmed individuals by police is rare in comparison to killing suspects that are armed. The cases get all the attention because they tend to have more unanswered questions (was he really in fear for his life?), but they are rare. Exactly as we'd predict based on the rarity of having reasonable fear for life against an unarmed individual. Especially when we remove the cases that strongly indicate suicide by cop. What was your case again?

- - - Updated - - -

What can't be waved away though is that the police kill a lot of civilians . . . a lot. And in a modern, democratic society that should be cause for concern.

Around 1 per million for whites and 3 per million for blacks. Let's say that the numbers are actually underreported by a full 50% - that makes 1.5 per million for whites and 4.5 per million for blacks. We can improve on those numbers, and we should, but I'd like to know what your definition of "a lot" is?

 

[laughing dog]

You are the only one inventing a story to dismiss the data. You are assuming systematic bias in a particular direction, despite having not a shred of basis to assume this....blah blah blah...

Nope. You are the one making assumptions about the missing data, not me. The data is incomplete. Hell even on the sources (the FBI) cautions users against drawing the very types of conclusions you are defending.

Honest and rational dialogue does require that you not cherry pick the tiny portions of counter arguments that you feel you can invent a b.s. response to. But your are correct, that honest and rational dialogue is not required.

The data is incomplete. Anyone truly interested in honest and rational dialogue would acknowledge that fact instead of literally making stuff up to support it.

The BJS report includes any and all forms of death for any reason. Deaths are categorized as "homicide" both "by officer" and by "other person", "suicide", "intoxication", "accidental", and "natural causes". The majority are attributed as "homicide by officer" with no indication about it being "justified" or not.
In addition, the CDC uses death certificates and emergency room records collected without any requiring reporting by police agencies and they reach similar numbers. Whatever limits the varied data sets have differ with the differing methods used. Convergence of results among the differing methods shows that those limitations have minimal impact upon the estimates, otherwise the different methods would yield highly varying results. It is called convergent validity, look it up.

I noticed none of that addressed "What about the fact that we have no idea if the police in those reporting agencies use the same criteria for responding to, observing, categorizing and reporting incidents? ". According to you, that means you cannot find a flaw in that criticism. Hmmm.

Your continued refusal to address these points is evidence of your incapacity to show any flaw with them.

That is one logical possibility. Another is that once the data is shown to be incomplete there is no need to address any other point concerning the data. Another is that I am simply bored with your spin. A poster who is truly interested in honest and rational dialogue would enumerate those possibilities (and perhaps add more).

Wrong. Here is the caution in the BJS report
[P]
Arrest-related deaths are under-reported. BJS did not
attempt to estimate for partial or non-responding
jurisdictions. Data are more representative of the nature
of arrest-related deaths than the volume at which they
occur. [/P]


The underlined part is key. Variables like race represent "the nature of arrest-related deaths" that the BJS states are represented by the data. They include all those variables related to who, when, and how deaths occur because they think they are reasonably valid estimates of how those deaths covary with other factors, such as race. Their caution refers to the total "volume" (aka raw number) of deaths being lower than the true number, but that is true across all levels of the other variables, so it doesn't impact the relation with those variables. IOW, the absolute raw number of deaths is an under-estimate but that is separate from relative comparisons and % of deaths of different types for various groups.

Nice story. Too bad it does not follow from what the DOJ quote.

I am defending rational thought and honest use of relevant evidence.

Too bad your responses do not reflect such noble ideals.

No, you're argument has the logical problem in assuming that national level stats on shootings provide evidence that unarmed blacks are shot due to racism......

It is simple enough to construct straw men. I find it fascinating that anyone would think that the rationales for shooting armed suspects would be identical in nature and frequency as the rationales for shooting unarmed suspects.

Honest and rational dialogues compels me to observe that your conclusions are due to the use of incomplete data, illogic and reading comprehension.

 

[Davka]

i think i found the real cause of all those shootings:

 

[laughing dog]

How so? Killings of unarmed individuals by police is rare in comparison to killing suspects that are armed. The cases get all the attention because they tend to have more unanswered questions (was he really in fear for his life?), but they are rare. Exactly as we'd predict based on the rarity of having reasonable fear for life against an unarmed individual. Especially when we remove the cases that strongly indicate suicide by cop. What was your case again?

Ah, the well worn out "suicide by cop" excuse. Keep grasping at those straws.

[

 

[Davka]

...although apparently those Brits aren't as dangerous as eating mozzarella:

 

[Axulus]

How so? Killings of unarmed individuals by police is rare in comparison to killing suspects that are armed. The cases get all the attention because they tend to have more unanswered questions (was he really in fear for his life?), but they are rare. Exactly as we'd predict based on the rarity of having reasonable fear for life against an unarmed individual. Especially when we remove the cases that strongly indicate suicide by cop. What was your case again?

Ah, the well worn out "suicide by cop" excuse. Keep grasping at those straws.

[

What was your case? I have no idea at this point. That killings of unarmed individuals by cops should be rare due to the fewer possible scenarios where such an unarmed individual could create a situation to put the cop in fear of his life? They are.

 

[ksen]

How so? Killings of unarmed individuals by police is rare in comparison to killing suspects that are armed. The cases get all the attention because they tend to have more unanswered questions (was he really in fear for his life?), but they are rare. Exactly as we'd predict based on the rarity of having reasonable fear for life against an unarmed individual. Especially when we remove the cases that strongly indicate suicide by cop. What was your case again?

Remind me, how do you know the killing of unarmed individuals is rare in comparison to the killing of armed suspects?

What can't be waved away though is that the police kill a lot of civilians . . . a lot. And in a modern, democratic society that should be cause for concern.

Around 1 per million for whites and 3 per million for blacks. Let's say that the numbers are actually underreported by a full 50% - that makes 1.5 per million for whites and 4.5 per million for blacks. We can improve on those numbers, and we should, but I'd like to know what your definition of "a lot" is?

50% underreporting is probably too low it's probably more like 200% underreported.

My definition of "a lot" is more than is absolutely necessary.

- - - Updated - - -

...although apparently those Brits aren't as dangerous as eating mozzarella:

View attachment 1897

keep 'em coming buddy!

 

[Jimmy Higgins]

How so? Killings of unarmed individuals by police is rare in comparison to killing suspects that are armed.

Rare? Can you cite that please.

The cases get all the attention because they tend to have more unanswered questions (was he really in fear for his life?), but they are rare.

They typically get attention because the shooting officers rarely face any repercussions for the shooting.

What can't be waved away though is that the police kill a lot of civilians . . . a lot. And in a modern, democratic society that should be cause for concern.

Around 1 per million for whites and 3 per million for blacks. Let's say that the numbers are actually underreported by a full 50% - that makes 1.5 per million for whites and 4.5 per million for blacks. We can improve on those numbers, and we should, but I'd like to know what your definition of "a lot" is?

Over 1000 can be considered a lot. That is 3 people in the nation a day killed by the people who are supposed to protect them. It is also about 1000 (2000?) times the number of ebola deaths in the US this year.

 

[ksen]

And to add a little perspective:

http://nypost.com/2013/12/30/fewest-police-deaths-by-firearms-in-2013-since-1887/

The annual report from the nonprofit National Law Enforcement Officers Memorial Fund also found that deaths in the line of duty generally fell by 8 percent and were the fewest since 1959.

According to the report, 111 federal, state, local, tribal and territorial officers were killed in the line of duty nationwide this past year, compared to 121 in 2012.

Forty-six officers were killed in traffic related accidents, and 33 were killed by firearms.

The number of firearms deaths fell 33 percent in 2013 and was the lowest since 1887.

They kill over a thousand civilians a year while only 33 of them were killed by firearms in 2013 the lowest since 1887.

But they fear for their lives.

 

[maxparrish]

Wrong. Those flaws mean nothing other than that the raw # of deaths is under-reported. But given the different methods and different reasons for overall under reporting, it is implausible that the relative relationships with race would be similar in the various reports if those estimated relationships were far off in either or both report. It is a basic principle in sciences that whether data validly reflect how a variable covaries with other factors is completely distinct from whether the total frequency with which an outcome occurs is accurately estimated in absolute terms. The data showing how a disease covaries with other factors does not require accuracy in the total frequency with which the disease occurs in the population.

And yet again, even if the relative difference in rates of death were more than double what any of the reports show that would be perfectly in line with the 8:1 ratio in the commission of the most violent crimes. So, what we have are multiple independent estimates that all under report absolute frequency but converge on a 2:1 to 4:1 estimate in relative frequencies which given the different methods and equal probability of over and under estimation is a highly unlikely convergence if the true ratio were something over 10:1, which would be the extremity required to suggest a disparity out of line with disparities in violent crime and of shooting at the police.

Ksen (and others) are glossing over the pivotal question with a bit of handwaving distraction. Data need not be "complete", and rarely is. Most empirical data come from samples and, if random, they can be used as representative of the population being studied. As such, whether the reported data sampled is as much as a 100 percent of all police homicides, or at minimum 33 percent, is not important. The sample size is more than sufficient.

Of course, the data sample(s) are not fully random (for example smaller rural agencies are less likely to report, and thus understate the number of police killings of whites in comparison to blacks) you have a bit of a problem. But the fact that three or four different methods of collection has come to similar conclusions on numbers and composition is a strong inference that the proportions reported are plausible.

 

[laughing dog]

And to add a little perspective:

http://nypost.com/2013/12/30/fewest-police-deaths-by-firearms-in-2013-since-1887/

The annual report from the nonprofit National Law Enforcement Officers Memorial Fund also found that deaths in the line of duty generally fell by 8 percent and were the fewest since 1959.

According to the report, 111 federal, state, local, tribal and territorial officers were killed in the line of duty nationwide this past year, compared to 121 in 2012.

Forty-six officers were killed in traffic related accidents, and 33 were killed by firearms.

The number of firearms deaths fell 33 percent in 2013 and was the lowest since 1887.

They kill over a thousand civilians a year while only 33 of them were killed by firearms in 2013 the lowest since 1887.

But they fear for their lives.

They are obviously safer with those extra 1000 thugs per year gone.

 

[ksen]

Ksen (and others) are glossing over the pivotal question with a bit of handwaving distraction. Data need not be "complete", and rarely is. Most empirical data come from samples and, if random, they can be used as representative of the population being studied. As such, whether the reported data sampled is as much as a 100 percent of all police homicides, or at minimum 33 percent, is not important. The sample size is more than sufficient.

Of course, the data sample(s) are not fully random (for example smaller rural agencies are less likely to report, and thus understate the number of police killings of whites in comparison to blacks) you have a bit of a problem. But the fact that three or four different methods of collection has come to similar conclusions on numbers and composition is a strong inference that the proportions reported are plausible.

Thank you for admitting me (and others) aren't glossing over anything since you admit that the data is not random.

The data is what the reporting police departments want us to know.

 

[Axulus]

Ksen (and others) are glossing over the pivotal question with a bit of handwaving distraction. Data need not be "complete", and rarely is. Most empirical data come from samples and, if random, they can be used as representative of the population being studied. As such, whether the reported data sampled is as much as a 100 percent of all police homicides, or at minimum 33 percent, is not important. The sample size is more than sufficient.

Of course, the data sample(s) are not fully random (for example smaller rural agencies are less likely to report, and thus understate the number of police killings of whites in comparison to blacks) you have a bit of a problem. But the fact that three or four different methods of collection has come to similar conclusions on numbers and composition is a strong inference that the proportions reported are plausible.

Thank you for admitting me (and others) aren't glossing over anything since you admit that the data is not random.

The data is what the reporting police departments want us to know.

 

[Nice Squirrel]

And to add a little perspective:

http://nypost.com/2013/12/30/fewest-police-deaths-by-firearms-in-2013-since-1887/

The annual report from the nonprofit National Law Enforcement Officers Memorial Fund also found that deaths in the line of duty generally fell by 8 percent and were the fewest since 1959.

According to the report, 111 federal, state, local, tribal and territorial officers were killed in the line of duty nationwide this past year, compared to 121 in 2012.

Forty-six officers were killed in traffic related accidents, and 33 were killed by firearms.

The number of firearms deaths fell 33 percent in 2013 and was the lowest since 1887.

They kill over a thousand civilians a year while only 33 of them were killed by firearms in 2013 the lowest since 1887.

But they fear for their lives.

How many were killed by donuts?

 

[maxparrish]

Ksen (and others) are glossing over the pivotal question with a bit of handwaving distraction. Data need not be "complete", and rarely is. Most empirical data come from samples and, if random, they can be used as representative of the population being studied. As such, whether the reported data sampled is as much as a 100 percent of all police homicides, or at minimum 33 percent, is not important. The sample size is more than sufficient.

Of course, the data sample(s) are not fully random (for example smaller rural agencies are less likely to report, and thus understate the number of police killings of whites in comparison to blacks) you have a bit of a problem. But the fact that three or four different methods of collection has come to similar conclusions on numbers and composition is a strong inference that the proportions reported are plausible.

Thank you for admitting me (and others) aren't glossing over anything since you admit that the data is not random.

The data is what the reporting police departments want us to know.

And, in my experience, the statistic keeping section of local police departments strive to report accurately. Having assisted them in records management (as an IT person) I have never encountered this dark "conspiratorial" view of local police statistics. Granted, you can speculate anything you like but pure speculation without a parsley snip of evidence is not an argument.

From many data sources we have inferences that are contrary to your faith; until such time as you can show them to sufficiently flawed in their reporting of proportional racial makeup, your gut feelings are not relevant.

 

[laughing dog]

Thank you for admitting me (and others) aren't glossing over anything since you admit that the data is not random.

The data is what the reporting police departments want us to know.

And, in my experience, the statistic keeping section of local police departments strive to report accurately. Having assisted them in records management (as an IT person) I have never encountered this dark "conspiratorial" view of local police statistics. Granted, you can speculate anything you like but pure speculation without a parsley snip of evidence is not an argument.

From many data sources we have inferences that are contrary to your faith; until such time as you can show them to sufficiently flawed in their reporting of proportional racial makeup, your gut feelings are not relevant.

Neither are your anecdotal dismissals. It is up to the users to show the data is complete. To date, that has not occurred.

 

[George S]

Using the bad CDC numbers:

View attachment 1895

The biggest disparity is in 2003 and 2011. The rate for blacks is only double the "all." (ETA: oops the first one is triple.)

As I understand it the % of blacks involved with violent crime is more than double. So the expected rate given that race is involved is percentagewise lower for blacks. There is no other-than-white-but-not-black line. It would be above the blue line.

Do I have the facts right? The participation rate in violent crime among the black population is more than double (triple) that of others? Is the above graph accurate?

It shows favoritism to blacks. Racism.

 

[AthenaAwakened]

Tell me,

Was there ever a time in US history when racism factored into the treatment of black people by the police? If there was such a time, what the racism wide spread or limited? And if there was such a time, when did the racism lessen and or stop?

Yes, widespread in the south, lesser extent in the north,

what makes you think it was lesser in the north? What data do you have that police department were racist in the south? What stats were kept in say 1880?

began lessening in the late 1960's, but have no idea how much it has lessened since because apparently no one is aboe to provide any data to show it.

Well, without data, how do you know it has lessened at all? We do have data that the incarceration rate for AAs began to increase around the same time the civil rights movement began to gain victories.

It appears to be a small enough factor in police killings today that it is not readily apparent in the data that is available.

And as has been pointed out, the data available is incomplete and quite possibly biased. And even when it isn't, it is often used and interpreted incorrectly.

Do you have any such data that demonstrates the racism is readily apparent?

You won't read the book but

The New Jim Crow by Michelle Alexander

And yes you do need to read the whole book since you have demonstrated a severe deficit in knowledge of the US criminal justice system and policing tactics.

 

[maxparrish]

And, in my experience, the statistic keeping section of local police departments strive to report accurately. Having assisted them in records management (as an IT person) I have never encountered this dark "conspiratorial" view of local police statistics. Granted, you can speculate anything you like but pure speculation without a parsley snip of evidence is not an argument.

From many data sources we have inferences that are contrary to your faith; until such time as you can show them to sufficiently flawed in their reporting of proportional racial makeup, your gut feelings are not relevant.

Neither are your anecdotal dismissals. It is up to the users to show the data is complete. To date, that has not occurred.

Dismissal of groundless speculation is always warranted, with or without having had personal experience in police statistic compilation. It is not up to those of us using the data to show it "complete", because being "complete" is not necessary.