October 31, 2013

How to fix the SAT

In response to yesterday's post on whether or not new College Board supremo David Coleman knows what he's doing in his still mysterious plans to revamp the SAT, one commenter responded:
Charles Murray said... 
Actually, I'm for the approach Coleman appears to be using. Even went on record saying so. http://www.aei.org/article/society-and-culture/abolish-the-sat/. The data finally prevailed and forced me to accept that a lot of my thinking about the differential effects of the SAT and the achievement tests doesn't pan out in practice (i.e., the SAT doesn't identify diamonds in the rough any better than the achievement tests. ...
I'm inclined to think Coleman knows what he's doing. Which is to say, he seems to agree with me. http://american.com/archive/2007/july-august-magazine-contents/abolish-the-sat
Charles M.  

Charles' 2007 plan is to get rid of the SAT I (the traditional aptitude test) in favor of the SAT II (a.k.a., SAT Subject tests: the less widely required subject achievement tests).

The SAT II achievement tests are miniature advanced placement subject tests, but heavily multiple choice and mostly without essay questions. So, moving from the SAT I test, which was originally intended to be more or less of an IQ-like aptitude test to achievement tests might have the advantage that if you are going to spend a thousand hours test-prepping, you might as well test-prep on real subjects like American History and Biology.

The University of California used to require three SAT II Achievement tests, but then started demanding expansion of the SAT I to include the demanded writing test. Having gotten its way on the flagship SAT I, the U. of Cal then logically announced plans to pare back the SAT II. 

What happened next?

As I wrote in VDARE in 2009:
But this simplification of the application process would likely hurt Asian high schoolers because their parents are more likely than other students' parents to get them signed up in time to take all these superfluous tests. 
Figuring out the various ideal points in your child's high school career at which he or she should take each of the three SAT Subject tests is the kind of complicated strategizing that Asian parents are most likely to obsess over. 
The California Asian Pacific Islander Legislative Caucus (who seem, significantly, to be all Democratic state legislators—the Asian Pacific Islander Legislative Caucus is part of the California Democrats website) wrote an angry letter to the chairman of the UC Board of Regents denouncing the reforms.

So, switching from the SAT I to your choice of the SAT II Subject tests is not going to lessen test-prepping. Tiger Moms love the SAT IIs.

Something that nobody ever talks about but could make a lot of sense is trying to get the SAT I back to its original intention of being a Scholastic Aptitude Test: i.e., more like an IQ test. Over the decades the SAT I has been repeatedly diluted in the name of "fairness," which, ironically, just plays into the hands of the Tiger Mothers. For example, each year the College Board now releases the list of the 2,000 words from which vocabulary questions will be drawn.

One way to find the IQ subtests that are hardest to prep is to look at James Flynn's list of the Wechsler subtests that have had the smallest Flynn Effect. Raw scores on Information (e.g., What continent is Argentina on?), Vocabulary, and Arithmetic subtests have barely bumped up over the decades, while raw scores on tasks more like programming your VCR have skyrocketed.

Indeed, when I look at what I'm still good at as a middle-aged man relative to the young guns who have enjoyed decades more of the Flynn Effect, Information, Vocabulary, and Arithmetic stand out (along with Comprehension, which is next in least Flynn Effect). Since what every pundit who expresses an opinion on the SAT really wants is for the rest of humanity to Be More Like Me, I think we should make the SAT a tool for producing millions of Steve Sailers. Granted, every single electronic device in America would soon be flashing "12:00," but that's a small price to pay for everybody to possess a more nuanced view of historical trends in golf course architecture.

So, why not merge Murray's proposal to emphasize the SAT Subject Tests with a downsizing of the SAT I away its current bloated (3 hour and 45 minute) state back to an aptitude test? They could cut down on the effects of test prep by emphasizing the elements of IQ tests with the least Flynn Effect, such as general information. Stop releasing ahead of time what's going to be on the test and just draw questions from huge pools. (For example, awhile ago hbd chick linked to a vocabulary test that tried to estimate your total vocabulary. I vaguely recall that mine was somewhere around, maybe 40,000 words. The exact magnitude isn't important, but the ratio to the College Board's annual list of 2,000 words to study is.)

Granted, the the IQ subtests that have the least Flynn Effect are among the most culture-loaded. But, life is pretty culture-loaded, too.

24 comments:

Veracitor said...

Steve, you're my hero. As soon as I saw Charles Murray's comment I wanted to post a link to your previous discussion of Murray's suggestion but you beat me to it. I have great respect for Murray but can't really agree with him about the best future direction for the SAT.

Beliavsky said...

Anyone who advocated replacing the SAT Reasoning test with SAT subject tests should read the chapter "On Substituting Achievement Tests for Aptitude Tests in College Admissions" in the book "Uneducated Guesses: Using Evidence to Uncover Misguided Education Policies" by Howard Wainer, who worked for 21 years as a statistician for the Educational Testing Service.

Wainer's key point is that the fact that takers of SAT subject tests also take the SAT Reasoning tests allows the ETS to calibrate the scales of the SAT subject tests. Otherwise, there would be no way to make a 700 on the SAT physics test comparable to a 700 on the SAT U.S. history test.

Anonymous said...

Kinduva stoopid question, but how does handing over the responsibility for overseeing the future of the entire American edumakashunal establishment, to but one single solitary man, differ from, say, handing over the responsibility for overseeing the future of the entire American medical complex to but one single solitary woman?

Aren't we Paleocons supposed to be deeply viscerally innately suspicious of this sort of thing?

And equally cynical about the motivations of the people who would seek to amass that sort of power for themselves?

But don't worry - if only we can wait until the next Five Year Plan, then surely our betters will correct all of the problems with our current Five Year Plan?

sunbeam said...

"The SAT II achievement tests are miniature advanced placement subject tests, but heavily multiple choice and mostly without essay questions. So, moving from the SAT I test, which was originally intended to be more or less of an IQ-like aptitude test to achievement tests might have the advantage that if you are going to spend a thousand hours test-prepping, you might as well test-prep on real subjects like American History and Biology."

I guess I don't understand things too well.

How could a change like this not fail to favor the usual suspects?

And to be blunt, out of the usual suspects I think this is going to set up another interesting Jewish/Asian dynamic. You'd think this change would heavily favor Jews, with their higher verbal intelligence.

But while I wouldn't call Jews lazy,... well let's face it compared to Asians and their Tiger Moms they pretty much are. My world view tells me I can't get a Jewish kid to practice Piano like the original Tiger Mom did, unless the kid actually liked the piano.

Another thing that gets me, is that I imagine their are a lot of high schools along the Northeast Seaboard, the Pacific Coast in parts, and a select few other areas of the country that have good high school biology classes, the kind where the kids win Westinghouse Awards (I never even heard of this, or Math Olympiads, or anything of the sort till I got to college).

You think Valdosta High School is going to have one? Usually you are scrounging to find a teacher who actually understands algebra, trig, and geometry. Someone who knows Biology, Physics, or Chemistry, in more than a basic way is far harder to come by.

Then:

"The data finally prevailed and forced me to accept that a lot of my thinking about the differential effects of the SAT and the achievement tests doesn't pan out in practice (i.e., the SAT doesn't identify diamonds in the rough any better than the achievement tests."

What's Murray smoking? He wants to take a test that he thinks doesn't work at identifying "diamonds in the rough," and replace it with one that is guaranteed not to?

Or has he tossed in the towel on Fishtown?

I'd really like it if someone could correlate AP, ACT, and SAT test scores with student demographics.

I'll eat my hat if this doesn't favor "the usual suspects" more than the SAT did.

With the SAT, super smart Joe Stoner has a shot at the big leagues. Won't his single mom in the apartment with the hellhole school district they moved to when Dad ran off with his secretary be surprised.

Or Sally Striver from Butte who fanatically works at academics might have a shot. But unless Sally is willing to go far afield on her own, she isn't going to compete with a kid from an elite suburban high school in Boston, based on what she was exposed to in her classes. I give her a better shot on the SAT than the ACT or this proposed test, even if she is a workaholic, not a talent.

Anonymous said...

As the son of a poor, rural divorced mom who was busy with three jobs and two other kids, I forgot to take the SAT 2s because nobody told me they were necessary and our school had no college counseling to speak of. Luckily, my SAT 1 scores were above 1400 and the UC admissions people knew enough of my hard-knock story to cut me a break and let me in anyway.

Of course this was back in the 1980s. I doubt that in 2013 there'd be room for a kid like me, wedged between underrepresented minorities on one side and Asian/white suburban test preppers on the other.

Anonymous said...

the son of a poor, rural divorced mom

What do you know about your father?

Anonymous said...

Why not just give them the Wonderlic?

Anonymous said...

What if we had a world where the point of 12+ years of education at no small cost to all...prepared you for adulthood instead of making you a debt slave for the bloated educational establishment? With our masters skimming off the top slots they leave open for talented peasants?

As to all the usual suspects, if you are willing to have alien masters and all you will do is bitch, then you are fated to have Masters.

We might improve matters by having our own as masters, but that would require you know work [not studying, work] and risk.

Scholars cannot rule. The lesson is ancient. In particular scholars who wish to control your minds and souls - the core of Progressvism.

And what they really want.

Whiskey said...

Steve, both yourself and Murray are wrong, and you hit on why with your post on NASCAR being more open (including to Black drivers) generations ago.

SAT functions as winnower to help Ivies and the rest in hierarchical fashion take "draft picks" on the most likely to succeed. And this is total bunk.

It is bunk because it involves making predictions about 18 year olds that are often wrong in both directions ("high draft picks" fail and "low draft picks" succeed) but emphasizes hereditary class differences that advantage Tiger Mothers and Helicopter Moms.

Instead, lets emphasize cheap, fast, efficient MOOCs and certification. A college degree for anyone who wants one, at a cost of around $5-6k, and three years at most. Excepting a few degrees requiring hands-on labs and such. Even those can be rented out by certifying Universities like Distance Stanford and the like.

Say you are AT&T. You want to hire an electrical engineer who is just out of college. Do you want to interview a whole bunch of people many of whom have heavy debt and require big salaries to pay it off, or people Stanford have certified have passed their program? And can work for cheaper?

We need to lower barriers to entry for most jobs by allowing certification that applicants know their stuff cheaply. Not diddling around with Class Division Enhancements like SATs and ACTs and the tests the Japanese, Koreans, and Chinese use to determine the fate of 12 year olds.

Getting rid of SATs and ACTs for all but the Harvard and Yales and Stanford in-person social club with university attached will retain the BEST attributes of the old America you love: free wheeling, social mobility, low barriers to entry, lack of credentialism/professionalism, informality, lower nepotistic networks.

bdoran said...

Here's how Germany succeeds.

http://www.bbc.co.uk/news/business-18868704

Among other things they don't have an Ivy League elite running everything to maximize their own hedge funds, they wouldn't think of stocking it with non-Germans to make some more money for themselves..and the German Government seems to actually like Germans.

Anonymous said...

I would they rather started grading it on a 100 to 900 scale, or introduce a second harder test for the top decile of SAT population if it's impossible to accomplish the former without lengthening it to soporific levels. Reducing the chaff that takes it won't hurt either, though that would mean less monies.
Are SAT given state-wide to every student the same way ACT is done in some states?

momo said...

What if SAT doesn't identify diamonds in the rough very well because there are fewer of them? In other words, what with all the social changes Murray and Herrnstein discussed in The Bell Curve, most of the diamonds are in nice neighborhoods with good school districts already? Steve: I suspect I'm wrong about this, but do you know why?

Steve Sailer said...

Dr. Hoxby of Stanford finds that most of the overlooked high school students are in Red State America.

In the big cities there is lots of charity money to find inner city diamonds in the rough. (I was an advisor to one such charity in the early 1990s in Chicago.) In South Dakota, not so much.

Anonymous said...

We have identified three inconsistencies between this composition explanation and the evidence.

The first
is that SAT verbal scores should have declined during the 1952—1963 period because that was when the
composition of the test-takers was changing most (from about 5 to over 50% of the senior cohort). To the
contrary, ETS reported that scores fluctuated within a narrow range—between 472 and 478. While the
average test-taker became less elite (academically) and more diverse in class, race and ethnic background,
mean verbal scores did not fall.



The second inconsistency concerns the prediction that when the changes in test-taker composition
slowed, in the 1963-1979 period, mean verbal scores should have leveled off and remained relatively sta-
ble. Throughout this period, the fraction of the senior cohort taking this test was stable at just over 50%. it
was in this period that verbal scores fell from a high of 478 to 424 (just 2 points above its lowest level ever,
422, recorded in 1991). That is, virtually the entire change in post-World War II American verbal achieve-
ment levels — from an initial high plateau of about 475 in the 1950-early 1960s to a much lower plateau of
scores from 1979 through 1994 — occurred within this single 16-year period.


The third inconsistency between the evidence and the composition explanation stems from its assump-
tion that as more lower scoring students took this test, the fraction scoring over 600 and 700 would neces-
sarily decline (so long as top students continued to take the test). The absolute numbers of students
scoring over 600 should have remained about the same.
7 The evidence is different: The entire distribution of verbal scores from top to bottom, shifted to lower levels.
There was not only a proportional decline in top scorers but an absolute decline in the number scoring
over 600. There are now 35% fewer scoring over 600. The number scoring over 700 fell from 17,500 in
1972 to just over 10,000 in 1993 — even as the number taking this test grew (Shea, 1993). Highly selective
colleges and universities report mean verbal declines on the order of 40 points.



Schoolbook Simplification and Its Relation to the
Decline in SAT-Verbal Scores

Anonymous said...

Otherwise, there would be no way to make a 700 on the SAT physics test comparable to a 700 on the SAT U.S. history test.

Could a 700 on the SAT Physics test ever be comparable to a 700 on the SAT US History test?

To assert that it could indicates a low IQ. It is like suggesting that some RAP trash is equivalent to Swan Lake.

TheLRC said...

Steve, exactly what you said at 4:48.

I grew up in Iowa just a few miles from the SoDak border, and although I was a classic 'diamond in the rough' as a student (blue collar parents; high test scores with absolutely no prep) I had no one helping me see there was an academic/intellectual world outside my home area: my high school guidance counselor's vision of the giddy heights of academia was attending a branch campus of the U of Minnesota!

I think things may have changed somewhat in red state America, but not that much. I visit my friends still living in the area, who now have college-aged kids of their own, and see them still pushing their kids, no matter how bright, to attend a small range of liberal arts colleges or nearby state universities.

There's also still a lot of home-town boosterism that's persistent and powerful. In other words, if there's a local scholarship your kid might qualify for, chances are it'll go to someone committed to attending the local liberal arts college, which is perceived (rightly) as a Very Good Thing that benefits the local economy.

On the other hand, kids who are seriously motivated now can find out lots about the wider academic world if they really want to 'get out' -- far more than I could discover 30 years ago.

And even if they don't end up at an Ivy or U of Chicago or the like, they will likely still get decent undergrad degrees (and then go on to grad school and end up as adjuncts anyway, but that's another story).

marco lalo said...

Steve, I don't quite understand how test prep is a kind of mild 'cheating.' All throughout school you were taught to study and prepare for exams, and the SAT is supposed to be different?

rob said...

Here's how Germany succeeds.

http://www.bbc.co.uk/news/business-18868704

Among other things they don't have an Ivy League elite running everything to maximize their own hedge funds, they wouldn't think of stocking it with non-Germans to make some more money for themselves..and the German Government seems to actually like Germans.


Well...Oh, what would get through 'komment kontrol'?

Philo said...

“[W]hen I look at what I'm still good at as a middle-aged man relative to the young guns who have enjoyed decades more of the Flynn Effect, Information, Vocabulary, and Arithmetic stand out (along with Comprehension . . .) . . . .” For a professional writer Information, Vocabulary, and Comprehension *should* be improved into late middle age, when brain deterioration starts to outweigh the accumulation of experiences. But as for Arithmetic, I’d bet on the young Steve Sailer as against the middle-aged version.

Anonymous said...

"Raw scores on Information (e.g., What continent is Argentina on?), Vocabulary..."

I don't think its a good idea to use tests that rely on vocabulary and general information to measure IQ. I know someone who is relatively smart and completed his master's degree (at a Dutch university) with no more difficulty than other students according to my impression, but he had absolutely no interest in geography and didn't like reading. So I don’t think measuring his vocabulary or geographic knowledge would measure his intelligence. He is from a working class background, so he is not representative of the average student.

Anonymous said...

I wonder if the Power Law which applies to networks also applies to society, IQ etc. as a whole.

http://www.shirky.com/writings/powerlaw_weblog.html

Anonymous said...

Well...Oh, what would get through 'komment kontrol'?

Not much more than that.

LOL'ed.

Man, you made my morning.


Anonymous said...

It makes sense that if you could do well on achievement tests, you will likely do well in college. You obviously have to have some aptitude to do well on these tests, but hard work also increases scores.

How do you get into one of the Oxford or Cambridge colleges? What tests do they use to evaluate candidates?

Anonymous said...

Ask the parents too.

http://www.youtube.com/watch?v=HJXpA919PZA