The Finch Formerly Known As Gold

10 October 2004

Half verbal, so to speak

I've always suspected that the SAT is important largely because ETS says it is, and I'd take these numbers with a grain of salt even if they didn't come with the disclaimer that "many of these scores are unverified."

For the record, I took it twice in high school, back in the Pleistocene era, and both times I scored between Scott McNealy and Rush Limbaugh.

(Via Bill Quick, who scored similarly.)

Posted at 3:12 PM to Almost Yogurt


Until at least 1990, the SAT scores correlated very closely to the first year GPA. Don't know what that means in terms of anything except just what it says. The higher your SAT the more likely you are to get high grades THE FIRST YEAR. Then it was dumbed down, or re-normed _and_ people really started getting their kids cram courses to up their scores. The combination appears (you shouldn't be surprised to know how closely stuff like this is guarded. I had to do some real back-door thinking to pry this out of the data I got.) to have a funny effect. The SAT may now be bi-modal. Apparently, dumbing down and heavy prep push those in the middle up several points while leaving those from the middle on down unchanged except for a slight upward movement in general based, probably on the dumbing down part.

I found this effect in the late 1990s, so it might be corrected for now, but I got it by getting hold of a bunch of raw SAT scores for my school with no identifying data included. I got about 15 years worth of all who had applied to a decent mid-level state school. There it was. Bi-modal. Evidently the school authorities (mostly IRB) didn't know their stat all that well. ;->=

Plus, I was part of an effort to test GRE scores against graduate performance. Again, GRE scores pretty well predict first year grad GPA. Nothing else. I was gratified, however, to find a positive correlation between the last four digits of the SS number and the likelihood of finishing a grad degree. Cool.

Posted by: JorgXMcKie at 4:39 PM on 10 October 2004

Hmmm. 1330 for me, which was I think right about what my PSAT scores predicted. In those days they expected people with relatively higher PSAT scores to get SAT scores slightly less than ten times the PSAT, and relatively lower scores slightly more than ten times.

The rationale was, if you scored poorly in the PSAT, you'd study harder to do better in the SAT, but if you did well, you'd coast in taking the SAT.

Of course, I don't think they took into account scores like a friend of mine got -- his total on the PSAT was 90, so he didn't bother taking the SAT. He's probably making more money now than I am.

Posted by: McGehee at 5:33 PM on 10 October 2004

I got about 11 times my PSAT score on the SAT. Not that I plan to be impressed by any of these statistics.

And besides, almost everyone is making more money than I am. Not that I care (much).

Posted by: CGHill at 8:15 PM on 10 October 2004

I'm amazed that anyone my age or thereabouts still remembers his score. Certainly the emphasis placed on it in the late 60's was nowhere near as profound as it is today. I barely remember taking the thing let alone what my score was.

Anybody have a clue if I would be able to find it through CollegeBoard.com or elsewhere before I start the process of attempting to locate it?

Posted by: Vickie at 7:06 AM on 11 October 2004

Thing is, Vickie, we're geeks -- and you're not.

Posted by: McGehee at 7:58 AM on 11 October 2004

Oh -- and I took my SAT in 1980.

Posted by: McGehee at 7:59 AM on 11 October 2004

Ooops,,Sorry Kevin. I aged you about 12 years...

Posted by: Vickie at 8:41 AM on 11 October 2004

Blogging has aged McGehee more than you can possibly imagine. :)

You can in fact order old scores from ETS, but they will cost you.

Posted by: CGHill at 9:10 AM on 11 October 2004

Are you calling me old?

Posted by: Vickie at 9:42 AM on 11 October 2004

Let's say I know you took this test before 1996 (which is the point at which they impose additional fees).

Beyond that, deponent saith not.

Posted by: CGHill at 10:01 AM on 11 October 2004