[sollewitt]

Data on lifetime likelihood of death by cop in the US by ethnicity. https://www.pnas.org/content/116/34/16793

It's very evident the US has strong implicit biases just from the Justice system. We could also look at incarceration rates, or sentencing disparities by ethnicity.

At some point saying "hey, we can't just _assume_ this is racist" _appears_ less like good faith debate and more like bad faith denialism.

[belorn]

Statistics do not prove intent, unless you want us to use the same data to prove that the Justice system is an order of magnitude more sexist than racist because 2000% more men are killed then women. The numbers for incarceration rates and sentencing disparities are similar damning.

If we go by the data in the study, the likelihood of death by cop in the US is determined by age first, gender second, race third. If we look at other studies we can add social economical factors, geographic location, geopolitical and cultural factors, all which has correlation impact with age, gender and race.

[sollewitt]

Agreed, I think one of the ways the two camps talk past each other is using "racist" to refer to both the motive of individual actors and the systemic and cultural biases.

Also yes, the judicial system is also sexist - police don't see women as being threats demanding lethal force, juries are more willing to see women as victims including of their circumstances. That's not the only reason for discrepancies, sociobiology looks into links between biological factors and criminality too, but the cultural issues are enough that defendants are coached to behave differently based on gender.

I see your final point, but my hunch is if you dig in to the correlation between age, gender and race with economics and geography, the eigenvalues will be ordered the other way (race, gender, age) - and if so I think that would indicate the Justice system is only a reflection of the larger society, rather than negate the idea of a racist justice system?

[belorn]

My hunch is leaning towards the opposite. Social economic status is determine from primarily income, education and occupation, and each three is correlated to age groups and gender to a point where one can make pretty decent probability curves for each. There are not many 20 years old that share income of 55 years old. Naturally there aren't many phd's or people working senior positions at a very young age.

If we look at age groups, 20-29 is the most common age for criminals to be found guilty by the justice system which matches the age when people exit from schools and have the lowest amount of income and highest rate of unemployment. High crime areas also tend to have a lower median age than low crime areas. For being shot by a cop, this age jumps a bit to 30-40, but I don't know why.

The general problem with this kind of statistics is that you will always end up with correlations that goes both way. If a country has a massive wave of refugees of a certain age, refugees with a lower social economic status that the existing population, you get an obvious correlation with social economic status, age and refugee status. At which point some statistician try to normalize values and those numbers suddenly becomes a discussion about which factors are considered and which ones aren't.

Is the justice system ageist, sexist, racist, and -phobic? When it tries to do a risk assessment based on little else than demographic data then yes.

[dbrueck]

I'm all for looking at incarceration rates and sentencing disparities, etc. - if there are actual imbalances after controlling for relevant factors, then it makes sense to root them out and fix them.

That said, studies like the one you're citing appear to be flawed because they fail to account for rates of crime (especially violent crime) as well as socioeconomic factors. Correlation is not causation and all that.

(see e.g. https://www.wsj.com/articles/the-myth-of-systemic-police-rac...)