Racism against blacks in police use of fatal force

Raw disparity
On average, black people are killed by police from around 2.5 to 3.5 times more than whites according to Edwards, Lee, and Esposito 2019 and Ross 2015Despite making up around 13 percent of the population, blacks makes up around 30 percent of all police killings.

For the data below, I am using data from the FBI, Killedbypolice.net, the CDC, and the Washington Post.


 * Harvard economist Sendhil Mullainathan estimates around 32 percent of those killed by police were black after analyzing data from the F.B.I.’s Supplementary Homicide Report
 * Sociologist Peter Moskos analyzed data from the website Killedbypolice.net, which claims to be “The most accurate, most comprehensive and always up-to-date list of people killed by U.S. law enforcement officers.” Using this data set, Moskos found that 30 percent of those killed by police in 2013-2015 were Black
 * The CDC’s Compressed Mortality Database measures the amount of people who are being killed by law enforcement (excluding legal execution). Using this data, we can see that the CDC estimates Blacks to have been 27 percent of those killed by police between 1999 and 2014
 * Washington Post's comprehensive analysis on police shootings show that around 26 percent of those killed by police from 2015 to 2019 are black

Resisting Arrest
It should be expected that there should be a higher disparity between blacks and whites if blacks on average are more likely to resist arrest


 * A five-year study of non-felony arrests in San Francisco found that blacks were almost 10 times more likely than whites and Hispanics to be charged with resisting arrest.
 * In Chicago, from 2014 to 2015, blacks accounted for 77 percent of arrests for obstruction of justice and resisting arrest (pg 4), meaning that blacks were almost 7 times more likely than non-blacks to be arrested on these charges.

Raw arrest rates for various violent crimes compared with police fatal force
This brings up a question, does disparities inherently mean racism? Not necessarily

This graph includes murderers, violent criminals, and criminals in general. The Uniform Crime Report state that blacks are about 28 percent of arrested people and 34 percent of all violent criminals. Therefore: we should expect a fatality proportion of around 28 percent to 34 percent. I additionally added murder to compare.

Cesario et al 2018
When adjusting for crime, Cesario et al 2018 finds no systematic evidence of anti-Black disparities in fatal shootings, fatal shootings of unarmed citizens, or fatal shootings involving misidentification of harmless objects.


 * This study analyzes the Guardian’s online database (The Counted, 2016). This database is more complete than official federal databases; police departments under-report to the federal government by *50%

Results:



Tregle et al 2018
This analysis states that "using violent crime arrests or weapons offense arrests, we observe that black citizens appear less likely to be fatally shot by police officers."

Streeter 2019

This analysis found that the race of someone being shot by police could not be predicted by a detailed list of 120 factors concerning the incidents. This implies that black and white civilians are shot by police under the same sorts of circumstances. This is contrary to popular narratives which would suggest that factors like “no witnesses present” or “suspect unarmed” should increase the chance that the person being shot was non-white.

Moreover, Streeter’s analysis implies that black people are killed more often than white people by police only because they more often find themselves in circumstances in which police are likely to kill someone regardless of their race.

Patty and Hanson 2020
This analysis does not find evidence for anti-black bias in police homicides

Weisburst 2019
This local analysis found that black people were not over-represented among those killed by the Dallas police department relative to their arrest rates in Dallas.

Mentch 2020
Simple population level summary statistics fail to take into account fundamental local characteristics such as county-level racial demography, local arrest demography, and law enforcement density. Using data on fatal police shootings between January 2015 and July 2016, Mentch 2020 implements a number of straightforward resampling procedures designed to carefully examine how unlikely the victim totals from each race are with respect to these local population characteristics if no racial bias were present in the decision to shoot by police.

The study found that black and Hispanic Americans were not over-represented among those killed by police relative to what you would expect given their local arrest rates.

Johnson et al 2020
This study conducted a county level analysis showing that minorities are less likely to by shot by police than are whites using a crime benchmark but (in the case of blacks, but not Hispanics) more likely using a population size benchmark.

Racial disparity in shooting threshold and reaction time before the use of fatal force
According to a meta-analysis by Mekawi and Bresin 2015, Black victims are shot more quickly than Whites victims, and police have a more liberal threshold for shooting Black victims.

"These data lead us to conclude that there are no overall effects of shooting sensitivity or false alarm rate based on target race, but that there are small negative effects for reaction time for reaction time to armed targets and shooting threshold and a small positive effect for reaction time to unarmed targets. Relative to White targets, participants were quicker to shoot armed Black targets, slower to not shoot unarmed Black targets, and more likely to have a liberal shooting threshold for Black targets." -Mewaki and Bresin

Limitations from Mekawi and Bresin 2015:


 * There may be a possible bias for the selection of studies
 * In contingency (b) for inclusion in the meta-analysis, papers needed “a comparison of White male targets to Black male targets”. This is a hyper-specific contingency that will result in papers that more so favor their narrative, because the burden of proof for law enforcement officers shooting minorities more quickly is on left-wingers, not the inverse.
 * This study states that “467 participants are required. This is much larger than the average 85 participants in the studies included in this meta-analysis”
 * Regardless of the methodology, the effect size (d) between blacks and whites are small at best
 * Reaction time for gun trials: -0.13
 * Reaction time for no gun trials: 0.11
 * False alarms: -0.01
 * Shooting sensitivity: 0.07
 * Shooting threshold: -0.19

Shjarback et al 2020
This study analyzes racial disparities in the use of fatal force by police with violence directed toward law enforcement by white, black, and Hispanic suspects, respectively

Shjarback's justification for using this as a benchmark, "Perhaps a more straightforward benchmark, then, is the racial/ethnic distribution of violence directed toward police by citizens, despite the fact that such citizen behavior is ultimately conditioned on being encountered by police. If officers are biased either in the decision to initiate contact with a citizen or escalate the use of force, then it may be difficult to disentangle rates of contact with citizens across race/ethnicity, or the other contextual-level variables (e.g., violent crime; firearm availability; economic disadvantage), from the behavior of officers and citizens during those encounters. Essentially, we believe it is necessary to separate the decision to use deadly force during an interaction from the decision to initiate the interaction."



Shjarback found that using such a benchmark rendered the probability of a black American being shot by police roughly 40% lower than the probably of a white American being shot by police. For Hispanics, there was either no difference or evidence of an anti-Hispanic bias depending on whether the benchmark was the rate at which people killed police or the rate at which they assaulted police.

Analysis of racial disparities in the use of fatal force by police using simulation studies
====James 2012 and James 2016==== The studies found that, despite clear evidence of implicit bias against Black suspects, officers were slower to shoot armed Black suspects than armed White suspects, and they were less likely to shoot unarmed Black suspects than unarmed White suspects.


 * In these studies, officers were confronted with potentially armed suspects identical in all aspects, including body language and weapon, except for their race. The test subjects were not told the purpose of the research.



This simulation data is consistent with actual data from Worall et al 2020 who found that black suspects 33% less likely than white suspects to have a gun drawn on them by the Dallas police force.



Racial bias in police behavior when accounting for race of the police member
The reason why black suspects are more at risk from black cops is mainly geographic.

Crotty et al 2017
This study finds a non-linear relationship between the proportion of an area’s police that are black and the rate at which police kill black people. However, there was no evidence that police forces which were almost entirely black had significantly lower rates of police killing black people than did police forces that were almost entirely white.

Menifield 2018
This study also finds the same conclusion that white officers are no more likely to use lethal force against minorities than nonwhite officers. Black people account for 33% of those killed by non-white police officers compared to only 28% of those killed by white police officers.

Johnson et al 2019
This study finds that:


 * As the proportion of Black or Hispanic officers in a fatal officer-involved shootings (FOIS) increases, a person shot is more likely to be Black or Hispanic than White, a disparity explained by county demographics (Each group of police was likelier to shoot civilians of their own race. That’s likely true, the researchers say, because police tend to be drawn from the communities they work in and are thus more likely to have deadly encounters with civilians of the same race.)
 * race-specific county-level violent crime strongly predicts the race of the civilian shot
 * although we find no overall evidence of anti-Black or anti-Hispanic disparities in fatal shootings, when focusing on different subtypes of shootings (e.g., unarmed shootings or “suicide by cop”), data are too uncertain to draw firm conclusions. We highlight the need to enforce federal policies that record both officer and civilian information in FOIS.

Raw analysis of unarmed black people killed by police
According to Washington Post, in 2019, fourteen unarmed black men were killed by the police. The number of those unarmed and killed by police has generally been decreasing since 2015.



Read Later:
Schwartz et al 2020 https://www.sciencedirect.com/science/article/abs/pii/S0027968419301300?via%