View Single Post
Old 10-09-2018, 03:59 PM
iiandyiiii's Avatar
iiandyiiii is offline
Join Date: Aug 2010
Location: Arlington, VA
Posts: 36,142
Originally Posted by Damuri Ajashi View Post
Well at least you aren't disagreeing with the math. But apparently it doesn't matter to you that the sample size went from 1200 to a few dozen.

Why does it matter what I am, I have a degree in economics and I am familiar with economics and statistics even though I am not an economist or statistician. Just as I don't have to be a mathemetician to know that pi is the ratio between a circle's circumference and its diameter; I don't need to be a statistician to know that 37 deaths over 3 years is not enough to declare a racist epidemic as severe as you describe. Pro Publica is feeding you factoids to rile you up.

The only actual peer reviewed study published by a Harvard economist in a prominent journal says that cops don't really shoot blacks significantly more than whites. The differences to the extent they exist can be chalked up to statistical noise.

The one statistician that supported the pro publica statement has withdrawn his support. It is effectively P Hacking the data.

Police shoot and kill ~1000 people each year. You are saying that based on ~12 shooting in each of 3 years you can conclude something?

I'm not saying that the peer reviewed study is irrefutable but pro publica is a news organization. They are good at uncovering and reporting news. It is indeed an interesting factoid that in the population of 60+ kids 15-19 shot by police from 2015-2017, 37 of them were black. But this is a factoid because it is a cherry picked fact that distorts the truth. This factoid is not enough to reach the conclusion that cops shoot black kids at 21 times the rate that they shoot white kids.

Why is it that you think that a news article is the equivalent of a peer reviewed study?
I'm not fully convinced of anything by either study -- two studies, with different results, demand further investigation. All of your numbers here are uncited (and I found your earlier math explanation highly confusing), so I'm responding as if they are accurate for the purposes of this discussion, though I certainly would not vouch for them without further explanation. In my understanding, the ProPublica Risk Ratio calculation is valid and provides a reasonable level of confidence -- 95% that the real ratio is somewhere between 10 and 40, based on a solid exercise in statistical analysis.

I don't offer the ProPublica report as "You are WRONG!", but rather "here is conflicting data". And this data is not easy to come by -- most shootings aren't recorded in this manner and made available to analyze.

My overall opinion on law enforcement treatment of black people is based on many, many pieces of data -- the most prominent of which is the polling that shows that 50% of black people report personal mistreatment by law enforcement.