Correlation between Iq and SAT scores....

<br>


<br>

<p>I was just objecting to the sweeping language you used; if your goal is to make sure that individual kids don't start to believe that they're doomed to fail because of risk factors, of course I agree. Society-wide correlations can never give the final word about one person's potential.</p>

<br>


<br>

<p>That's true, but it's a whole lot easier to do when your parents are pushing you to do it and your whole peer group thinks that's the cool thing to do. I certainly didn't have the wherewithal at 17 to be that responsible without a lot of outside encouragement. Kids who have the kind of maturity you're talking about are likely to go far in life whether they succeed in bringing up their SAT scores or not.</p>

<p>"Let's suppose it is shown that an IQ score of 130 on a particular brand of IQ test indicates a rank order at the 98th percentile of the general population (which is approximately correct for most current IQ tests). Does that mean that a person who scored 130 on an IQ test can be sure to score at the 98th percentile level on the SAT? Well, no"</p>

<p>Thats true of different IQ tests some people score higher and lower on others why should SAT be any different.</p>

<p>Okay, so I hope everyone agrees with me that the prediction </p>

<p>
[quote]
This means my 144 iq will only get me a 1490 sat score.

[/quote]
</p>

<p>just can't be counted on. The test-taker might be pleasantly surprised. It's a lot easier to get a surprisingly high SAT score by going into the preparation process believing that a pleasant surprise is possible, which it is for many more people than know it is possible.</p>

<p>I happen to have training in psychological testing, and I feel I owe it to this forum to add my $.02.</p>

<p>The IQ test is not a test of "intelligence". Most IQ tests, such as the Weschler Assessment of Adult Intelligence, or Stanfor Binet, are a battery of about 12-17 separate tests measuring various specific abilities. This includes vocabulary (big surprise), arithematic (big surprise), but also pattern recognition tasks, compying tasks, picking missing details, etc. These tests test precisely what they are testing for (very specific abilities whose real world meaning is far from obvious). The idea that by creating an index from them and call this index "intelligence" is very controversial. In fact, there is no agreement about what intelligence is. Ask a neuropsychologist.</p>

<p>As it happens the SAT is modeled on IQ tests, and is made up of alot of subtests which are combined in two main indexes (Verbal and Math). And there is a strong correlation between IQ and SAT scores (especially the Verbal SAT). But it says almost nothing about intelligence. It instead says alot about the similarity of what is being measured, whatever that is.</p>

<p>To give you an example of how weak the SAT really is, knowing a person's SAT gives you about a 10% advantage of guessing their first year college grades beyond knowing their GPA. That means 90% is explained by other factors. Not exactly earth shattering. Think about that when you find people viewing the SAT as some powerful measure of who someone is or what they are capable of doing. It's not surprising schools like UC Berkeley are getting rid of using the SAT. </p>

<p>Maybe some day we'll have a real understanding of intelligence, whatever that is, and a way of measuring it, whatever that may be. But it ain't the IQ tests of today or the SAT.</p>

<p>I have to say I agree.</p>

<p>
[quote]
To give you an example of how weak the SAT really is, knowing a person's SAT gives you about a 10% advantage of guessing their first year college grades beyond knowing their GPA.

[/quote]
</p>

<p>But I think that is partly a problem of restriction of range. In the entering classes of the University of California (where such a study of SAT scores predicting college grades was done), there is not as much variance in SAT scores as there is in the general high school test-taking population including high school students who don't get admitted to UC. If you consider the students whose SAT scores are so low that they never get admitted, it's a reasonably safe GENERAL prediction that they would get worse grades in UC classes than the applicants who were admitted. </p>

<p>But I agree with you that the construct estimated by an IQ test score is best not characterized as "general intelligence," but rather as something like "scholastic aptitude," which is the label I have seen for that construct in an old book on educational testing by two eminent psychologists. No doubt the UC campuses would be very different places if they only admitted applicants with SAT scores in the lowest two deciles, rather than mostly applicants in the highest two deciles. But there are plenty of low-SAT-scorers, and low-IQ-scorers, who have a great deal of real-world problem-solving ability and life success that could be characterized as "intelligence" just as fairly as the ability to score high on those tests.</p>

<p>Tokenadult raises a good question. Very low SAT or IQ scores suggest that the person taking the test has at least several areas of significant impairment on a number of the subtests (by the way you could still be great at some things but it would be hidden by the indexing).</p>

<p>However, what I was saying was that if you already know the GPA then adding the SAT as additional info only gives you 10% added information. So a high GPA plus low SAT combination is a tough call, since ON AVERAGE the SAT only is 10% better than a random guess on how that person will do. It may suggest some unusual problems with specific subtests, difficulties testing, being sick at the time of test, having inflated GPA, whatever. You just don't know. If adcoms could use the SAT like a real IQ test they would be able to look at the various subscore factors and that might help them, but they aren't trained in how to do that. </p>

<p>I never even mentioned the coaching factor. Which skews results so much that it may make the 10% advantage no longer apply.</p>

<p>What I really worry about is all these elite colleges with very high SAT expectations. There is very little predictive value between a 1400 versus 1600 SAT but don't tell that to the parents and youth (or Prep course companies)! It sends a bad message about false exactness and an even worse push towards perfectionism that may cause more harm than good. Of course, someone will ask "what's the alternative". Fortunately, a few institutions are trying to figure that one out right now.</p>

<p>I read somewhere that Mensa accepts SAT scores for membership in lieu of Standford-Binet, Cattell, or other formal IQ tests. The SAT doesn't have (shape) is to (shape) as (etc.) analogies like most IQ tests do. Otherwise, it is roughly similar, assuming that the test takers have all developed some vocabulary through reading and developed some rudimentary math skills.</p>

<p>And guys, colleges have ALWAYS considered the SAT as roughly equivalent to an intelligence test. It's just that no one will admit to it for image and marketing reasons. It is equivalent to the political manipulation motives illustrated by the Bush administration's term "ownership society" - which translated, can be rendered as ("we own, you suffer" + "you're on your own, ciao.") [flames ok]</p>

<p>"What I really worry about is all these elite colleges with very high SAT expectations. There is very little predictive value between a 1400 versus 1600 SAT but don't tell that to the parents and youth (or Prep course companies)! It sends a bad message about false exactness and an even worse push towards perfectionism that may cause more harm than good. Of course, someone will ask "what's the alternative". Fortunately, a few institutions are trying to figure that one out right now."</p>

<p>Thats what worries me too but there is not much people can do there is money to be made in keeping the status quo</p>

<p>I took the SAT in 1994, and I found a table for the IQ correlation to my SAT score. </p>

<p>Surprisingly, it matched perfectly to my IQ from a Stanford Binet test many years ago. I find that the SAT has gotten to be too easy over the years. The ceiling has been lowered, and it has made a 1500-1600 score too crowded and undistinguishable. My converted score from 1994 would be 50 points above a 1600 on the last version of the SAT.</p>

<p>I find that college grades are more based on willpower and grade inflation/deflation more than anything. Many students who spent their entire high school lives studying find the urge to socialize and have fun in their first years in college away from home.</p>

<p>We need hard workers in society to be creative, build buildings, and do a lot of things that doesn’t require a high IQ score. </p>

<p>They can make as much or more money than a high IQ scientist. But you cannot deny the basic intelligence levels as being an important factor for those jobs that require high intelligence. Just like you cannot deny the value of hard work for jobs that require hard work as its most important factor.</p>

<p>Adam Smith said the best economy is one where you have different skill sets amongst the population. I think one of the main problems with our economy today is that we emphasize too much on hard work alone, and not on intelligence. We need intelligent people to make the hard work of others relevant for a super power economy.</p>