Is anyone aware of research studying the correlation between the results of IQ tests given to young school-age children (ages 5-7) and those kids’ subsequent performance on standardized college admissions tests?
Nothing to back it up, but I’d assume there would be a positive correlation (A is higher and B is higher) since they both test more or less the same thing more or less. IQ tests are intended to quantify the lowest level of cognition (that is to say operational level, not level of intelligence) but they usually rely on the same forehand knowledge that a test would.
Based on that I’d guess that any child that was able to score at a certain level in an exam would also be able to get a decent score on an IQ test.
According to this, there’s a correlation of .82 between I.Q. and the SAT test:
Now you’re going to ask what .82 means. I would have to explain how one measures correlation coefficients, and that would take a while. Would it suffice to say that a correlation coefficient of .82 means that it’s a good predictor but not a perfect one?
If you add that 1 is perfect correlation, 0 is no correlation, and -1 is perfect anti-correlation.
Actually, this page has some nice pictures for a qualitative feel.
I find it’s easier to explain it as “a correlation of 82%.” Good enough for the layman and the guy who forgot everything from stats class (that’s me.)
This is purely anecdotal, but also relevant.
I was I.Q. tested fairly early on, and the same relative score has come up even decades after the first one. To me that seems to justify that it’s at least relatively precise, “accurate” is still up for debate I suppose.
My High School required 3 years of math in order to graduate, most students for some reason were either not aware of this, or made better decisions than I in HS. I opted for 3 years, because math is by far my weakest field.
Consequently, when I went to take my ACT and SAT tests, I found that I had absolutely no idea how to answer several of the math questions having zero introduction to even pre-calc. I did well enough, based on my other strengths, but I should have taken that other math course in order to have a better overall picture of how I compared to others that had received the same instruction.
An interesting thing happened when I took my placement test for math at Ohio State, it seems I may tend to either test well, or get lucky. I placed into a course that was practically Greek to me. The whole idea of calculus in college kinda made me mad. I knew I would never, ever, have to use that sort of math in real life, because I would avoid any job that required it. Yet, a B.S. in anything requires it, oh well.
I wonder if the correlation between SAT and IQ still holds as strongly as it once did. The problem is in the last 20 years, the SAT has been gamed to help people increase their scores. This not only includes the various third party tutoring services, but also schools themselves as they mutate their teachings to teach towards the SAT exam.
What is an interesting correlation is between IQ, the SAT, and the Marshmallow Test conducted by Walter Mischel about 1968 or so.
Walter Mischel took young kids and put a marshmallow in front of them. He told them they could eat the marshmallow now, or if they wait until the researcher got back, they would get a second marshmallow. Although he never wanted to study the long term consequences, he realized when he asked his daughter about her classmates who took place in the study, that the ones who were more patient did better in school and in life.
I don’t think either of your points are relevant, qazwart. The SAT has been recentered, not gamed. When the SAT test was first given (in the 1940’s, I think), it was only taken by a small set of the best prospective college students in the U.S. It was only meant for those students applying to a small set of highly selective colleges. The students who applied to other colleges (and even then those were the majority of all students) didn’t have to take any such test. Gradually from the 1940’s to the 1990’s, more and more colleges began insisting that prospective students take the SAT. (Also, the ACT was required at other colleges, but that’s not relevant either.) By the 1990’s, a hundred times as many students were taking the SAT test each year as in the 1940’s.
The SAT’s scoring was originally fixed so that the first set of students who took it had an average score of 500. Each successive year’s test’s scoring was fixed so that a student who scored a given score on the previous year’s test would score approximately the same on this year’s test. Not surprisingly, this meant that the average score dropped a little bit each year till by the 1990’s the average score was in the low 400’s. This did not mean that the average students in American colleges were less prepared for college. It meant that if at one point you only give the test to the best college students and then at another point you give the test to every college student, you can expect the average to drop.
Then in the 1990’s the SAT test scoring was changed so that every year now the average score will be 500. This, in effect, only differs from the previous scoring by adding the same number to each score. If you correlate either score to the I.Q., you will get the same correlation coefficient.
The marshmallow test isn’t particularly relevant to the points made so far either. Nobody has ever said that I.Q. is the only thing that affects SAT scores. That’s why the correlation between I.Q. and SAT scores is good but not perfect. Obviously things like patience affect how well one learns.
I should add that my answer of a correlation of .82 doesn’t quite answer the question that the OP asked. Hotel Cass wanted to know the correlation between the I.Q. at ages 5-7 and the SAT test. I believe that the correlation of .82 is between the I.Q. on a test given around the same time as the SAT and the SAT score.
However, IQ does seem to correlate pretty strongly from age to age, so I wouldn’t expect the number the OP is asking for to be too far off from .82 .