View Single Post
Old 10-09-2018, 02:03 PM
Damuri Ajashi is offline
Join Date: Jul 2006
Posts: 20,623
Originally Posted by iiandyiiii View Post
Okay, assuming your calculations are valid, now show me the statistical analysis that says that this sample size is too small to be valid and provide useful information for the populations we're talking about. I've taken a statistics class, but it's been a long time. If you don't think that ProPublica's statistical analysis is valid, you'll have to prove it if you want to convince anyone.
No, I don't. People can look at the anemic sample size and reach the conclusion that the sample size is too small to be able to say that 21::1 is a valid statistic. It might very well be the result of a violent turf war in Chicago or couple of violent drug/gang encounters in LA. The only ACTUAL study done has said the exact opposite of what you are saying. The only ACTUAL study done has said that cops don't shoot blacks at higher rates than white after adjusting for reasonable variables. What pro publica did was take a few very small subset of numbers and extrapolated something about society as a whole.

Based on this methodology, I could say that cops kill asians age 14-19 three times as frequently as they shoot whites.

From my reading, their analysis is good math and in line with good statistical and probabilistic practices - they even give their confidence levels (that it's really between ten and forty times different). If you feel differently, show me why.
Its bad math. The numbers are small and you are cherrypicking the data to reach those statistics. The fact remains that cops shoot blacks generally 3 times ore frequently than whites generally. P Hacking has been pretty prominent in the news recently and this sort of cherrypicking is a close relative. You get a big data set and look for anomalies that help prove your point.