[SomeStupidPoint]

> people are willing and able to make judgements from circumstantial evidence without rigorous quantitative work- you don't demand a formal Bayesian analysis of the day's weather before you decide to take an umbrella with you

But I do do a Bayesian analysis rather than a naive one: it's cloudy without raining here too often to carry an umbrella just because it's cloudy. That would be a maladaptive response, even if naive analysis is "clouds -> rain; prepare!"

Similarly, social issues tend to have enough complexity that you can't just resort to "clouds -> rain!" In this particular case, I'm just arguing that you need to control for background statistics rather than saying police shootings are randomly distributed across the entire population without any correlations.

No different than thinking a little about how often it's cloudy without rain before grabbing an umbrella. And I dislike that people like you are just screaming "DON'T THINK! CLOUDS! GET UMBRELLA!"

[maxander]

And I'm arguing that, in any other issue, you would think "controlling for background statistics" in a formal sense is unnecessary. I'm not telling you not to think, quite the opposite. Think of this:

In a country with a long history of terrible racial oppression and documented contemporary racist sentiment, a large body of people are convinced by their lived experiences [1] that police violence is racially biased, and this conviction is further supported by a massive number of known individual cases. Among the huge space of possible injustices that could be claimed, this one particular form has received tremendous attention and accumulated consistent material evidence across the entire nation.

We as yet lack the huge amounts of data to settle the issue in a way that would satisfy a physicist. But if you think about the issue with an open mind, you're still going to feel pretty damn confident about what's actually going on.

[1] Taking this as distinct from the large body of people who are convinced at a safe distance via their TVs.

[SomeStupidPoint]

> a large body of people are convinced by their lived experiences [1] that police violence is racially biased, and this conviction is further supported by a massive number of known individual cases

Okay, we know that there's an effect. No one is arguing that black people not only feel that way, but interact with (and are killed by) police at a higher-than-white per capita rate. That certainly should be addressed, and is a major social concern.

I think we agree on how to interpret the data at least this far -- that there's an effect, originally identified anecdotally by black people but supported by data, where black people are killed by police more often per capita. (Numbers I've seen are like 2.5x per capita -- but order of magnitude, I think we can agree on somewhere between 2x and 5x.)

Where we disagree, and I think we're just going to have to disagree, is that this is evidence that there's a lot of systemic racism going on. (Think "70 cent" wage gap, not "95 cent" wage gap.)

I think we're seeing a little bit of active racism, a lot of lingering economic effects of historic racism, a little bit of active cultural maladaptation, and a lot of general police violence. (That's me "thinking about the issue with an open mind" -- that distribution is basically the prior on social issue breakdown.)

You seem to suggest that the data suggests just "a lot of racism", which is where I disagree: I don't think the evidence is anywhere near moving the needle from the prior of "complex weave of the usual issues" to "outright, ongoing, systemic racism".

> But if you think about the issue with an open mind, you're still going to feel pretty damn confident about what's actually going on.

Finally, I just want to say, that this argument supports literally anything that sounds appealing, regardless of how likely it is to be true. There's a long history of creating new problems while attempting to solve problems by adopting solutions that make no sense upon detailed analysis, but sound good or appeal to our emotions in some way.