Database of police who shoot citizens reveals who’s most likely to shoot | Hacker News Reader

Database of police who shoot citizens reveals who’s most likely to shoot

theconversation.com

71 pointssohkamyung16 comments

Here is the original paper: https://doi.org/10.1073/pnas.1903856116

They use multinomial regression analysis with several county-level socioeconomic factors. It looks like a solid study that shows no statisticly valid correlation between victim and shooter ethnicity can be established.

If use county-level data how do they account for more granular neighborhood demographics? Maricopa County encompasses all of Phoenix, Glendale, Mesa, etc. Cook County has a huge amount of the Chicago metro area. I expect this effect plays out throughout the US.

Perhaps from crime reporting aggregation like how heyjackass[0] does it.

[0]: https://heyjackass.com/

Profiling stability of both the police and the potential victims should be able to predicatively show who is danger of being shot. Making that into a public health survey could be beneficial.

So it sounds like the entire national narrative about racism and law enforcement we've been having for years now is built on sand.

Please don't take HN threads on generic ideological tangents or further into ideological flamewar.

https://news.ycombinator.com/newsguidelines.html

No, because that narrative has focused on systemic racism, which has nothing to do with the race of the individual officers, but rather with policing policies.

e.g. "Black Lives Matter (BLM) […] campaigns against violence and systemic racism towards black people." From [1]; emphasis mine.

[1] https://en.wikipedia.org/wiki/Black_Lives_Matter

Statistical disparities aren't evidence of racism.

And let's not pretend we haven't been talking about a supposed epidemic of white cops shooting innocent black people for no reason for years and years now.

I don't understand the point of this comment. It does not bolster your comment I replied to, nor does it argue against my own.

Are you starting a new, different argument, and now claiming that systemic racism doesn't exist, and that large bodies of research about it are unfounded, and supporting your claim with a single pithy statement about statistics?

If so, sorry, I don't feel like being dragged into an off-topic argument about whether the concept of systemic racism is valid.

If there is systemic racism then it would exist through the approval of laws that disproportionately affect minorities, not through the actions of individuals who are merely guided by the system.

Because the solutions are obvious but unacceptable (even to those who are discriminated against) I don't think we will make any progress on this front.

>> And let's not pretend we haven't been talking about a supposed epidemic of white cops shooting innocent black people for no reason for years and years now

Double negatives and sarcastoc political signaling of a POV that ignores statistics in the same statement?

A successful broadcast of talking points and your agenda, but not really participation in a discussion.

What statistics am I ignoring? I'm talking about the ones from the OP's post.

It looks like a lot of data and some poor statistics.

The blog post does not include the original analysis. You can find the full study here: https://doi.org/10.1073/pnas.1903856116

PNAS is a highly regarded and high impact journal with stringent peer review. Scepticism is always a must in sociology. However, it is poor form to outright dismiss the findings as "poor statistics" without argument or reference to the source.

I've seen some pretty iffy articles on molecular biology/micro biology etc published in nature (not scientific reports) science cell. They were studies that were technically OK, but made some excessively bad claims and questionable methodological choices (eg left out performing a validation step that's cheap and the next step in the research, curiously resigning that step for anyone who would try and replicate the study). Prestige of the paper only says standards are higher, not that any given paper is great, and especially not that the conclusions are great.

Reply on news.ycombinator.com