Arrived in my pile

by on January 3, 2014 at 3:40 pm in Books | Permalink

1. Lars Peter Hansen and Thomas Sargent, Recursive Models of Dynamic Linear Economies.

2. Alberto Simpser, Why Governments and Parties Manipulate Elections.

3. Glenn Reynolds, The New School: How the Information Age Will Save American Education from Itself.

mike January 3, 2014 at 4:28 pm

Don’t bother reading #3 if it starts from the premise that American primary and secondary schools are bad, as Bob Somerby and Steve Sailer have shown pretty conclusively that they’re pretty much the best in the world.

albert magnus January 3, 2014 at 4:34 pm

“Best in the world” is maybe a little bit much. Pretty good at educating our mostly not very promising population is probably more accurate. Charter school, school choice and whatever Reynolds is suggesting is probably not going to change things a lot one way or another.

fffffff January 3, 2014 at 5:09 pm

In case anyone is wondering what exactly mike means, it’s that after controlling for race, American students do very well on the PISA. (The PISA is the international test whose results you are hearing about whenever you read yet another news story about horrible American test scores relative to the rest of the developed world.) And if you’re skeptical of anything Steve Sailer says, feel free to check out the PISA data yourself. (Steve Sailer does a good job looking through it and you can easily replicate the investigation he performed, straight from the publicly available PISA data.)

john personna January 3, 2014 at 7:52 pm

The problem with “controlling for race” is that you also, unavoidably, “control for income.” You also, unintentionally, control for “parent’s educational attainment.” Thus the claim becomes one that America can educate its well off, its children of college graduates, as well as other countries educate their “everyone.”

It becomes a just-so story that children of poor and less educated parents must not be expected to do well.

fffffff January 3, 2014 at 8:01 pm

Controlling for race is hardly the same thing as controlling for income. For instance, whites from families earning less than $10k per year outscore blacks from families earning over $100k per year on the SAT.

http://www.jbhe.com/features/49_college_admissions-test.html

mike January 3, 2014 at 8:06 pm

Even if it was just a way of controlling for income, it would still be good methodology. Controlling for extrinsic factors is a good thing, unless you’re a shitlib who doesn’t want to know the truth. But yeah, it’s not about income or parents parental attainment, it’s about genes. Given the state of the research, this can no longer be disputed in good faith.

mike January 3, 2014 at 8:07 pm

*parents’ education achievement, obvi

mike January 3, 2014 at 8:07 pm

* you know what i mean

john personna January 3, 2014 at 8:33 pm

Hanging a lot on 10 points (out of ~900) there. Ooh boy, outperformed by about 1%! (The approx equal income comparison showed a 61 point difference, or about 7%)

Not a lot to hang a racial theory on. Certainly other small things (culture, etc.) might produce a small 7% delta.

mike January 3, 2014 at 9:08 pm

“Ooh boy, outperformed by about 1%! ”

Said the guy who claimed the black-white difference was due to family income, when presented with data showing that the poorest of poor whites outperformed rich blacks

john personna January 3, 2014 at 9:15 pm

Said the guy who knows statistics. A 1% “difference” is the same as equality.

mike January 4, 2014 at 12:14 pm

“A 1% “difference” is the same as equality.”

This is an embarrassingly ignorant statement.

byomtov January 4, 2014 at 1:59 pm

The SAT is a very poor measure for these purposes, because of enormous differences in who takes it.

fffffff January 3, 2014 at 8:03 pm

I believe there is also the statistic that whites whose parents either did not even start college outscore blacks with a parent who has a graduate degree. If I find the source I will post it a bit later.

john personna January 3, 2014 at 8:40 pm

Why? What a way to even live. If say, when you did spend your “time adjusting for race,” you found a 7% difference, that would mean what in practice?

Certainly you’d know nothing about the next person who walks through the door (to be classified as one of four or so races, by your eye).

The whole effort seeks to be self-justifying, and only self-justifying.

mike January 3, 2014 at 8:59 pm

Nobody’s saying anything about inferring things about the next person who walks in the door, just explaining the group differences which we’re constantly being hammered over the head with. Do you know even the slightest thing about statistics, btw, or are you just flying by the seat of your pants?

john personna January 3, 2014 at 9:06 pm

Would someone who “knew anything about statistics” leap to believe there was something significant in a 7% difference, in averages?

john personna January 3, 2014 at 9:07 pm

I mean, give me 50%, give me 70%, and then maybe you’ve got something. 7%? You need a better hobby-horse.

mike January 3, 2014 at 9:08 pm

Ok so the answer is “no”, then.

john personna January 3, 2014 at 9:10 pm

I think you just self-refuted your comment, mike. You didn’t explain to us why 7% difference in averages should matter at all, to any of us.

john personna January 3, 2014 at 9:12 pm

I forgot. “We believe in these small difference, because it allows us to keep looking for these small differences.”

mike January 3, 2014 at 9:15 pm

I’m just saying that if you were a statistically literate person you would be talking in terms of standard deviations, and not percentages out of some arbitrary total. I’m six foot five (77 inches). A seven foot tall man is only 10% taller than me (84 inches). Does that mean he’s not really substantially taller than I am? Go back to the kids table, man.

john personna January 3, 2014 at 9:17 pm

So why didn’t you bring in those distributions if you are so conversant? I responded to a guy who thought 10 points (POINTS) on a SAT result was a big deal.

mike January 3, 2014 at 9:17 pm

And just to make the analogy complete, if there was a race of one billion people that averaged 6’5″ and another race of a billion people that averaged 7’0″ would you be so stupid as to claim that there was no real difference in average height between the races?

mike January 3, 2014 at 9:19 pm

Oh sorry nevermind I just realized that since you don’t even understand the concepts of variance and standard deviations you won’t even realize how sample size affects the analysis

mike January 3, 2014 at 9:28 pm

BTW, you also demonstrate your logical illiteracy by harping on the 10 point difference between poor whites and rich blacks. You claimed that the black-white difference was due to income disparities. Therefore if your argument was correct, rich blacks should have scored much better than poor whites. But poor whites scored better than rich blacks, even if only by a small margin, proving that your hypothesis was sorely mistaken. But you can’t grasp this, because you’re a self-righteous moron.

Willitts January 4, 2014 at 12:12 am

That’s irrelevant. The purpose of including race is to avoid omitted variable bias on the variable we are interested in. If there is collinearity between race and other variables, it would increase the standard error of the estimate but not bias the estimators. If you reject the null hypothesis, you do so with a handicap.

You might have a point if someone drew inferences from the coefficient on race. Even if race has a negative coefficient, that doesn’t imply an inherent racial disadvantage, just one that is highly correlated with race. Problem is, people see race there and immediately think there is a racist agenda at play. To omit race is an eggregious statistical error. To do so to assuage hurt feelings is worse.

prior_approval January 4, 2014 at 3:55 am

‘Problem is, people see race there and immediately think there is a racist agenda at play. To omit race is an eggregious statistical error.’

Well, let’s use Sailer as a leading light in ‘seeing race’ without commenting on his agenda at all. So, using Sailer’s expertise in this area of classification, please let me know how we will avoid ‘eggregious statistical error’ if we don’t break out ‘ethnic Catholics’ into a group. Not to ‘assuage hurt feelings’ on the part of that ethnic group, but because Sailer’s very idea of ‘ethnic Catholics’ makes the concept of his idea of race particularly clear.

Coming up with such an euphemism shows the high level of though one expects – as if the new ‘ethnic Catholic’ immigrants belong to an identifiable group, while the older ‘ethnic Catholics’ from Germany or Ireland are not classified at all. It isn’t about feelings or statistics when one attempts to veil one’s overarching concerns about classification, leading to such ignorant statements as considering ‘ethnic Catholics’ an actual term, apart from something dredged from the author’s need to put things into groups he apparently needs to believe are valid.

The question about whether the Irish, the Germans, and the Poles are all the same race is left as a historical exercise – one demonstrating just how flexible the concept has been over the last century.

mike January 4, 2014 at 12:13 pm

If we had good statistics on the Irish, Polish, and Germans and they deviated significantly from the average then it would indeed be good to break them out. For example, we know Jews do well on these tests, so it’s possible that native American underperformance is slightly buoyed by having a larger population of Jews than most other countries. Note that the current analysis breaks out not only Blacks and Hispanics but also Asians, who perform better than Whites.

Ed January 3, 2014 at 8:44 pm

If you adjust American scores by omitting those of the visible minority group that happens to have the lowest scores, presumably to make the comparison accurate you would do the same with the scores for all the other countries?

mike January 3, 2014 at 8:57 pm

You’re certainly welcome to do so, but it would be laborious and most of the top countries that are said to be superior to the US have very few “visible minorities”, like an order of magnitude fewer, so it wouldn’t change much

mike January 3, 2014 at 9:00 pm

BTW, I don’t think any other countries expend the effort the USA does to sort out racial disparities (mostly because they have few non-natives and even those are extremely recent) so in most cases the data just isn’t even there

Rahul January 4, 2014 at 1:15 am

Let’s assume your racial thesis is correct & indeed white-only PISA scores are at par with the best of the world. So what?

Point is you are stuck with the raw material (I doubt genocide is an option…) you have so might as well try and do the best with it that you can! I don’t get this defeatist, nihilistic approach of “Oh! Students are black so we cannot make schools do any better!”

Further, go and ask some random parents of white kids what they think of their unionized school teachers. Your fantasy of all’s right with US schools isn’t true (at least in the parents’ eyes) even in almost fully white schools in Minnesota or Wisconsin. Yes, controlling for race might show better student performance but that’s more to the credit of parental effort than teacher or school performance.

21st Amendment Absolutist January 4, 2014 at 11:48 am

“I don’t get this defeatist, nihilistic approach of “Oh! Students are black so we cannot make schools do any better!”

Or maybe it means we (who is we exactly?) should focus elsewhere than in the schools like on “parental effort.” etc. But that might require some uncomfortable re-evaluation of victimhood and guilt amirite?

Rahul January 4, 2014 at 1:50 pm

Who is we? Policymakers. Politicians. School Boards. Parents.

Why don’t you spell it out? Where should we focus on.

mike January 4, 2014 at 12:08 pm

The point is (if you followed the actual argument and not the one in your head) that people like Reynolds, Duncan, etc are aggressively pursuing a radical reform effort based on the perception that American schools in general are terrible, based on the fact that our students on average score poorly on international testing. But when you break it out, we are actually just behind Finland and Korea by a tiny amount. Sure, try to improve the schools. Just don’t rely on falsehoods to do it. I mean, it’s the same thing with the gun grabbers. They want to blame Billy Bob out in the country and seize his guns, based on statistics caused by the epidemic of violence in Chicago, because they know nobody can name the real culprits. You can’t have an honest discussion built on a lie.

Rahul January 4, 2014 at 1:09 pm

Except that this perception that American schools are doing badly and need reform is shared by a large number of white parents. Reynolds and Duncan ( & Steve Sailer ) be damned.

This is a high stakes issue. No wonder everyone and his uncle is trying to sneak in his pet agenda and piggyback. ( You brought in gun control. )

mike January 4, 2014 at 1:18 pm

Except it’s not, really, except to the extent that people are misled by these test score statistics. Everyone in my area is extremely satisfied and proud of the school they send their own kids to. There are complaints of course, as with anything that’s not perfect. You think there’s any country anywhere in the world where parents don’t think the schools could be doing a better job? These education fads always seem great at the time, and virtually always turn out to be disasters. The best way to improve education right now is to have less of it.

Willitts January 4, 2014 at 6:35 pm

@Rahul:

Well, yes, if you determine that race is a significant explanatory variable, the policy might be, “let’s not waste any more money trying to turn lead into gold” or it could be, “we need to spend more on education or do it better because 1) it is our duty, 2) we are stuck with the problem, 3) they are human and deserve it.”

But reaching an honest policy discussion requires honesty in the underlying reasons and causes. Do our teachers suck? Is there not enough diversity among teachers? Is there an untreated mental health problem in a community fraught with violence? Are black parents deficient in parental guidance?

Nevertheless, the key point in this study is not yo determine the effect of race, per se, but to compare the quality of inputs. That there may be endogeneity between the quantity and quality of inputs and minority success is an empirical issue that needs to be dealt with. Including race, though, avoids omitted variable bias in the other regressors.

ww January 4, 2014 at 12:57 am

#3: Khan Academy. Nuff said.

Marie January 6, 2014 at 7:35 pm

Yeah, I liked it the first three videos, too.

Rob January 4, 2014 at 1:07 am

#2 Look quite interesting, but not 81 dollars interesting…

Alberto January 4, 2014 at 10:54 am

The first chapter, with an overview of the argument, is available for free on:

home.uchicago.edu/~asimpser

under “Research.” (Pricing is unfortunately set by Cambridge Press, but new copies are available on Amazon for $45)

Rahul January 4, 2014 at 1:53 pm

Out of curiosity, at a $81 price tag will all purchasers be libraries & institutions? Do you get many individual purchasers buy at this price tag?

Euripides January 4, 2014 at 9:14 am

#1. Tyler and Alex should recommend this text is adopted for GMU’s macro core. It’s the only way to do “serious” macro.

silivri evden eve nakliyat January 4, 2014 at 12:29 pm

sitenizi çok beğeniyoruz paylaşımlarınızı takip ediyoruz

Willitts January 4, 2014 at 6:50 pm

Tyler’s previous post seems to have attracted the interest of a shipping and moving company in Turkey. They obviously want to conduct business under European law. :)

Comments on this entry are closed.

Previous post:

Next post: