May
30th
2018
Wholesome Rage: Once more into the breach · 2:35pm May 30th, 2018
This one has an unwitting special guest at the end, it seems.
A look into why discrimination against black people and poor people blur together at the seams, and just how badly being low-income affects development.
Next week's article is up on Patreon already, and it's actually probably the first time I've cried while doing research for these. The next one hurt to write. After that I'm probably going to go back to doing some nerdier stuff for a while, history and engineering and the like. So if you haven't been a fan of these leftist-economics rants, I'll be giving you a bit of a respite for a while.
And if you have been a big fan of them, join my Discord server, I kind of do these a lot in real-time.
Thank you for doing what you’re doing.
Also can I join your discord server?
I feel like both you and Titanium Dragon appear to be making a small fundamental statistical error (or perhaps "oversimplification" is better than "error") here. Namely, just because X correlates with Y and Y correlates with Z, that doesn't necessarily say anything about whether X correlates with Z[1].
In the case of "are success differences racial or socioeconomic", this is the sort of problem observational studies are always prone to run into. Sure, you can establish that race relates to success, and socioeconomic status relates to success, AND EVEN that race relates to socioeconomic status. But it's not intrinsically clear that the part of success explained by race and the part of success explained by socioeconomic status are the same.
That's not to imply these are unanswerable questions. It's easy enough to model success differences based on both race and socioeconomic status and see how much they're explaining the same part of the variability. (I don't know the literature well here, but I'd tend to suspect they're essentially interchangeable and playing the same role.) But discussion of the mechanisms by which these things are acting really must address this sort of how-do-we-parcel-out-the-covariance issue. Since we know race doesn't completely determine SES, it's basically sufficient to just ask whether race has additional predictive validity for success criteria when SES is already known, or vice versa.
Based on the extracts, it looks like some of the studies you mention are zeroing in on this, with the need to control for other variables—but really, this should be getting looked at both ways, and the sources you're mentioning seem to be focusing on "does SES have an incremental effect when race and other variables are known" rather than "does race have an incremental effect when SES and other variables are known". Both of those are important questions, and at the heart of this issue. Certainly, most of us would like to hope that the first one is answered in the affirmative and the second one is answered in the negative, but logically we need to have answers to both.
The liberal conventional wisdom is that the answers will be affirmative and negative respectively. The rac(ial)ist conventional wisdom is, I think, that the answers will be negative and affirmative respectively. But there's a non-ignorable possibility that both answers could be affirmative, or both answers could be negative—which calls either conventional wisdom into question somewhat and means we'd probably need to re-evaluate some set of assumptions.
[1] Sort of. Depending on the size of the correlations, you can get a situation where X necessarily correlates with Z here. The easiest example is when cor(X,Y) and cor(Y,Z) are both greater than sqrt(0.5), or about 0.7. Those are VERY strong correlations, though.
:raises hand:
I was born to poor parents in the US, married rich for a bit, divorced, got a *great* job (I was good with numbers), and despite my physical disability, always told people it didn’t matter because I had my brain. (I should point out that all my aunts & uncles went to college, save one—who was young enough to be my brother & was always told he’d never amount to as much as his much older siblings. He ended up with more than any of them, may it give him cosmic enjoyment.)
Then a bad head injury, the subsequent moderate TBI, no more work (I can’t do office work with executive function & short term memory disorder, not to mention the memory losses). So here I am, poor again. Worse, actually. I’m on SSI, not SSDI. You don’t recover from genetics or brain injury, and the SSA expects your family to help with your finances (literally). But my uncle died and there’s no one else.
Less than $9k/yr. That's what the government thinks is proper for a fully disabled person, who raised kids (and so not enough work credit for SSDI) to live on. Including housing, utilities, dental (unless you only want extractions), medication (in my state), my SNAP this month was $41. Oh, and because of my disability? I *have* to have a landline & internet to make my assistive devices work. Those don’t count towards ‘utilities’ in expenses.
So, yes. Thank you for writing about these subjects. I tried to on DailyKos, several years ago, but not many were interested. Also, great editorial writing!