"We find no evidence of anti-Black or anti-Hispanic disparities across shootings, and White officers are no more likely to shoot minority civilians than non-White officers."
An excellent study that examines, in detail, the data behind the data. There are a number of press and other outlets that urge the reader to see systemic police bias against blacks, using a simple method of counting, for instance, that "23% of those shot and killed by police are black, while blacks make up only 13% of the population." This study's conclusions are in contradiction with such simple statistics and their inferred bias. And this is what got the authors into hot water with their employer.
The Full Study - Click image.
David Johnson, co-author, found “no evidence of anti-Black or anti-Hispanic disparities across shootings, and White officers are not more likely to shoot minority civilians than non-White officers."
The Retraction - click image.
In his request for retraction he stood by the central conclusion of the study that "We found there was no relationship between the race of officers and the civilians they fatally shot when controlling for county racial demographics and crime rates."
Removed from MSU web site - click image.
"As we should have made more clear in the introduction to our article, the reason we did not calculate
P(shot|race) is that many researchers have already calculated approximations to this quantity. Indeed,
benchmarking approaches (in which the number of Black and White civilians shot is compared to the
proportion of Black and White civilians of some relevant pool) have been widely used for decades."
What was the cause of the retraction?
The authors responded to pressure from MSU university. The faculty sponsor of the study was removed from his administrative position.
The reason for their retraction
"We did not calculate the probability that a person is shot given their race... we should have made more clear in the introduction to our article, the reason we did not calculate
(this) is that many researchers have already calculated approximations to this quantity."
Interpretation: Under pressure from Michigan State University and others, the authors state that they did not include the raw statistic: That 23% of those shot and killed by police are black, while blacks make up only 13% of the population.
This is the raw statistic, repeated by the press and others, in order to support the belief that police are biased against blacks. The study helps to counter this common statistical error in logic, one that finds a correlation where none exists.
The retraction affirms the study: "We found there was no relationship between the race of officers and the civilians they fatally shot when
controlling for county racial demographics and crime rates."