Monday, October 5, 2015

A Short Defense of Generalizations

Some people have a problem with generalizations. But there is no logical fallacy called, "generalization", because in fact we can make true generalizations from particulars. All of social science is built on generalizations from particulars because we can't sample the entire population. But through the magic of statistics, especially the law of large numbers, we can make true generalizations, even about people, and still be on solid ground.

Tuesday, September 8, 2015

Are the people who disagree with you Hypocrites and Bigots?

Consider this meme. Is it a valid point?

Bigots are intolerant of other beliefs. Hypocrites do things in conflict with their own beliefs. These definitions sound right to me.

It's not bigotry if they're asked for nothing more than tolerance. Tolerance only obliges inaction. Their job obliges action. In neither situation is the individual asked to tolerate something.

It's not hypocrisy if there are relevant differences in the rationale of the supporter. There are plenty of possible differences in the rationale of the supporter. Most obviously, Kim refused to do a small fraction of her job (marry gay couples). The Muslim refused to do a very large fraction of her job (marry all those without hijabs).

Maybe one only supports people who refuse to do their job because they have the right beliefs. And don't support people who refuse to do their jobs for wrong beliefs. That too is not inconsistent or hypocritical.

The charge of hypocrisy is over-levied because we often think in terms only as specific as we need to in order to maintain the charge.

I think of the "hypocrisy" of pro life pro-capital punishment, or the other way around if you prefer. It seems that anyone who gives a yes-yes or no-no answer would be able to site relevant differences (a fetus isn't a real life / a criminal is a guilty life), but their opponents are thinking in terms only specific enough to maintain a charge of hypocrisy.

And then when their beliefs are under scrutiny, OF COURSE there are relevant differences!

Friday, September 4, 2015

Education Spending in the U.S. doesn't seem to be doing any good

I was caught up in a debate recently over whether school funding provides better student achievement. My tentative view is that extra spending have very trivial impact beyond a relatively low point.

When I take a few gulps  of the research on the subject, I find that the general trend is on my side with a few exceptions. I take away that schools with more money have higher student achievement, but schools very similar to each other, and with similar student body compositions but very different budgets, don't get these results.

The remarkable finding from combining the 377 estimates across 90 separate published works is that neither variation in teacher-pupil ratio nor variations in teacher educations are systematically related to student performance...
Teacher experience stands out in that about 30 percent of the parameter estimates are statistically significant in a positive direction. whereas 7 percent are statistically significant in a negative direction. But, again, this is far from unqualified support for the efficacy of employing more experienced teachers, since 71 percent are statistically insignificant or negative. Since these three inputs are combined to indicate variations in instructional spending per pupil, the results lead to the conclusion that there is no strong or systematic relationship between spending and student performance.
(Does Money Matter? Gary T. Burtless)

Why do schools with more money have higher achievement? The alternative story is that successful parents have successful kids regardless of where the kids go to school. It just so happens that successful parents also buy their kids sexy educations, if not by affording private school, then by buying an expensive house in a good school district. The expensive education isn't an investment, but a consumption good. The kid's experience is better, but it doesn't change where the kid is going.

The research also meshes with my experience. In the town where I grew up I knew lots of kids who didn't go anywhere because they had a "I don't care", "I live for today" mindset. No amount of spending was going to change that. The district, by the way, spent $2,000 more per student than average. You wouldn't have thought that if you went to school there.

There is little doubt that on a national scale education spending isn't the magic that some think it is. The U.S. governments spend about $ 1 Trillion on education now after ramping up the funding over a period when student achievement has only gone downhill. Most of it is spent on K-12. The U.S. spends  more than almost every other country on earth, beat only by Switzerland. All that increased spending didn't all go do the richest districts, although they were preferred they're also too few to consume all that extra spending.

I believe that many impressions people have of how dry the education budgets are in the U.S. come two places. One is a lot of political zealots manipulating data by citing only federal expenditure when states and local governments are the main spenders. The seconds is the anecdotal stereotypical "budget cuts" that are immediately accepted without so much as a whiff of skepticism. How do you know that there are budget cuts? Who did you hear this from? Does "budget cut" mean what we think it means? (The term is often used to describe the difference between what the district asked for and how much they got, rather than the difference between what they got and what they got last year. Bullshit, I know)

Tuesday, August 18, 2015

Saturday, August 15, 2015

SlatStar on My Id on Defensiveness

Here's another great post by the best blogger in town (high praise from both David Friedman and Bryan Caplan) Scott Alexander:

I like discussion, debate, and reasoned criticism. But a lot of arguments aren’t any of those things. They’re the style I describe as ethnic tension, where you try to associate something you don’t like with negative affect so that other people have an instinctive disgust reaction to it.

Thursday, August 13, 2015

No Universals?

Though twilight may be long, there is a difference between night and day.

Wednesday, August 12, 2015

Higher class correlates with charity?

A Large Scale Test of the Effect of Social Class on Prosocial Behavior

Across eight studies with large and representative international samples, we predominantly found positive effects of social class on prosociality: Higher class individuals were more likely to make a charitable donation and contribute a higher percentage of their family income to charity (32,090 ≥ N ≥ 3,957; Studies 1–3), were more likely to volunteer (37,136 ≥N ≥ 3,964; Studies 4–6), were more helpful (N = 3,902; Study 7), and were more trusting and trustworthy in an economic game when interacting with a stranger (N = 1,421; Study 8) than lower social class individuals.

If this finding is basically right, it might seem natural to say that money is causing this good behavior. I doubt it for the same reasons I doubt that lack of money causes many of the poor's systematic bad behaviors (like divorce, alcoholism, and single parenthood). Rather, bad behavior generally leads to failure, and good behavior lead leads to success.