News for hackers. Stuff that matters.
Re:Too many prisons make people worse by rayiner on Saturday May 27, 2017 @11:31AM (#14430492)
I agree with that in general, but it's a bit oversimplified. America is also a dangerous and violent society--our per capita murder rate has averaged 5-10x higher than the U.K. for over a century (long before private prisons and modern mass media).

I agree that it's a moral failing if the people that we don't deal with that problem in a more graceful and humane way.

Re:Too many prisons make people worse by gragas on Saturday May 27, 2017 @11:54AM (#14430585)
.
Re:Too many prisons make people worse by vitaminbandit on Saturday May 27, 2017 @12:01PM (#14430610)
Explain how Simpson's Paradox works here. Explain how it can be mathematically possible for each state/city/etc to have a lower murder per capita than the national murder per capita. Use concrete numbers as examples.

Spoiler: it isn't mathematically possible.

Re:Too many prisons make people worse by gragas on Saturday May 27, 2017 @12:46PM (#14430792)
.
Re:Too many prisons make people worse by vitaminbandit on Saturday May 27, 2017 @12:59PM (#14430844)
> That's a logical fallacy (Simpson's Paradox). It could be the case that when you condition on each neighborhood, the per capita murder rate is lower than Britain's, but the overall per capita murder rate is higher.

One, phrase your statement properly. You intended to compare neighborhood to corresponding neighborhood between the two countries. Not "Britain's".

Two, as others have mentioned. You have no basis for invoking Simpson's Paradox.

> Explain how Simpson's Paradox works here. Explain how it can be mathematically possible for each state/city/etc to have a lower murder per capita than the national murder per capita. Use concrete numbers as examples.

Three, how does your example satisfy my comment?

Four, I know Simpson's Paradox. I also know the difference between knowing something and knowing the name of something.

Re:Too many prisons make people worse by gragas on Saturday May 27, 2017 @01:02PM (#14430856)
You're right. I'm an idiot. Glad I got that sorted out, though.
Re:Too many prisons make people worse by fjdlwlv on Saturday May 27, 2017 @12:05PM (#14430629)
Your math is wrong, rayiner didn't mention neighborhood variation at all, and Simpson paradox is relevant to changes over time or resampling, not simple rate comparisons
Re:Too many prisons make people worse by gragas on Saturday May 27, 2017 @12:57PM (#14430835)
>Simpson paradox is relevant to changes over time or resampling, not simple rate comparisons

Not true. It absolutely works with simple rate comparisons.

Re:Too many prisons make people worse by dasil003 on Saturday May 27, 2017 @12:19PM (#14430698)
What's your evidence that this is Simpson's Paradox? In order for that to be the case, you'd have to have a significantly higher proportion of the US population living in the most dangerous neighborhoods as compared to the UK. That may be the case, but I'd be reluctant to assume that.
Re:Too many prisons make people worse by gragas on Saturday May 27, 2017 @12:51PM (#14430815)
Yeah, I don't think the muder rate in the US vs. the UK actually exhibits that.

I was just pointing out that it's not okay to say "the unconditional rate of murder in A is greater than the unconditional rate of murder in B, therefore A is more dangerous than B". While it may be true that A is more dangerous than B, the implication is wrong.

Re:Too many prisons make people worse by rayiner on Saturday May 27, 2017 @12:00PM (#14430605)
I don't follow how Simpson's paradox is relevant here.