<p>I've been reading quite a bit about the old SAT and how it had such a high ceiling. Also, I've heard that it used to be accepted by High-IQ societies for membership...was the old SAT much harder and more like an IQ test? If so, are there any reasons why it was changed?</p>
<p>It was marginally harder. But, as far as I can tell, the main difference was the recentering--- meaning for the same number of questions wrong, suddenly after 1995 that gave you a higher score.</p>
<p>For example, the score I received when I took it in the late 80s would have been about 60 points higher if I had gotten the same number of questions wrong and took it in 1995.</p>
<p>The thing was recentered because ETS always wanted a score of "500" to be average--- that is, around the 50%. 500 was average when the test was orginallly started decades ago, but by 1996 the average score had slipped to the mid 400s... I cant recall what exactly, something like a 450 was the 50%. They never discussed why this was--- maybe people were getting stupider, or the test was getting harder. Who knows? </p>
<p>So, what they did was just shift everyone up by like 50 points. This was tough luck for the Ivies and other top schools, because while in 1994 you needed to get every single question right to get a 1600 (perfect) suddenly in 1995 you could miss three or four questions and still get a 1600 score. </p>
<p>That is, after 1995 you could no longer distinguish very well between folks at the top end, ie above the 95%. </p>
<p>I'm sure this recentering made average test takers feel a lot better, but it no longer became as useful for the high IQ societies and some top schools.</p>
<p>Thanks for that explanation, GregB77! :)</p>
<p>great writeup</p>
<p>Just to add a bit to Greg's explanation, the scoring scale/curve/whatever you want to call it was hugely inconsistent at the top of the range (like, one or two wrong would drop you 40+ points), whereas in the middle, someone who got a 500 might have missed 2 more than someone else who got a 500. The scale also did not seem to be based on any sort of logical mathematical principles (i.e., they seemed to be making up the scale along the way). </p>
<p>However, you still can't miss 4 and get a 1600/1600. My 1580/1600 is two wrong, and I would have needed only 1 wrong to get a 1600...you can hardly ever, if ever, score an 800 on the math section with even one wrong. There is a little more leeway with a tough verbal section, but I haven't personally seen someone who got an 800 with more than 2 wrong. Sooo...a 1600 is still pretty meaningful!</p>
<p>2400 > old 1600</p>
<p>I really don't think a 1600 (M+V) is really actually that meaningful today. </p>
<p>When I took the test in 1988, in the ENTIRE UNITED STATES there were 16 people who scored a 1600. That means they got every question correct. (I wasn't one of them, by the way!)</p>
<p>According to the College Board itself (see the link) 984 people last year got a 1600 (M+V only). That is about 60 TIMES the number of people that got a 1600 the year I took it. 968 of those people would have gotten a 1580 or 1550 or even lower pre-recentering.</p>
<p>It just doesn't mean as much as it used to, especially to the folks who care (ie. highly selective colleges, high IQ societies.) </p>
<p>I guess you could argue "well, it's pretty easy to miss one or two questions, so it's better to have that lee-way." I suppose. But is they were so easy, then you should have missed NONE of them, right? </p>
<p>Think about it this way: If the whole test was 100 questions asking you to add single-digit numbers (4+6, 3+7 etc), you woud expect to get NONE wrong, even on a bad day. Well, you could make the same argument about getting a few wrong on the SAT. For some people, the SAT difficulty is like adding one digit numbers. Those people are probably smarter than you. But you can't tell anymore with new recentering, since their "none wrong" gives the same score as your "two or three wrong."</p>
<p>^^AS meaningful...no, I guess not. Of course, the stats you quoted don't include how many more people took the SAT last year as opposed to in 1988...I am sure that it is less than 60 times as many, but that info is needed to make a fair assessment as well. I just think that you are being a little too harsh about how much "easier" recentering has made the SAT. First of all the SAT NEEDED to be recentered--the old way was not mathematically sound. It's not a matter of making people feel better--it's a matter of having a scaling system that makes since. Secondly, it's still not easy or common to score a 1600. Finally, you just can't convince me that someone who missed one or two is appreciably less intelligent than someone who missed none. They MIGHT be, but I think that that would be difficult (impossible) to prove with the SAT alone, recentering or not.</p>
<p>you wrote:</p>
<p>"Finally, you just can't convince me that someone who missed one or two is appreciably less intelligent than someone who missed none. They MIGHT be, but I think that that would be difficult (impossible) to prove with the SAT alone, recentering or not."</p>
<p>I would agree with you on a case by case basis. That is, "he got an 800 and missed no questions and I got a 780 and missed one question. That means he is DEFINITELY smarter than me." Not necessarily. Not in that one case. I agree.</p>
<p>HOWEVER, this is EXACTLY what it means when we talk about statistical distributions. In fact, this is the VERY reason the high IQ societies decided to accept the SAT score as useful for admissions. They spent a lot of time and money correlating SAT scores with Stanford-Binet, Wechsler, and other IQ test scores, and they found that, in fact, the SAT score (of the non-recentered test) correlated QUITE STRONGLY with IQ as measured by Stanford-Binet--even, and especially, at the high end. That's why the test was useful to them!!</p>
<p>What that means is that, if you take all the people that missed no questions on the SAT and averaged their IQ scores on the Stanford-Binet, and did the same for those who missed ONE question on the SAT, you would find that the average IQ of the former group IS HIGHER than the latter group, in a large and statistically significant way. That is, as a group, those who missed NO questions on the SAT are, in fact SMARTER than those who missed one question, as measured by IQ score.</p>
<p>That is a statistical FACT. </p>
<p>If you'd like to actually read the painful and very detailed discussion of this issue, I invite you to read the following report from the Prometheus Society. This is a high IQ soceity that takes on the top 1/30,000 people. That is, the top 99.99997% of the population based on IQ score. Think this is a joke? It isn't. I direct you to section 8.5 of the report, where they discuss SAT and its correlation to IQ.</p>
<p>So, as much as you'd like to believe that "missing one question" doesn't mean anything, it actually means quite a lot. </p>
<p>A lot of people spent a lot of time and a lot of statistics and a lot of data to prove that it does. </p>
<p>So, it was somewhat difficult to prove, as you suspected. However, it has, in fact, been proven.</p>
<p>My bad. Prometheus is actually the top 99.997%, not 99.99997%. Quite a large difference, actually!!</p>
<p>Okay...not trying to get into a huge argument over it. I'll take your word for it with the Prometheus test, for obvious reasons (too lazy and tired to read it! :)). I'm actually not too surprised, although my statement was confirmed...they had to introduce IQ tests to prove the correlation--SAT alone wasn't enough to PROVE it, although once it was proved you could make assumptions from the SAT score. </p>
<p>I'm actually fine with the conclusion...I can settle for saying that with a 1580/1600, I am pretty smart, but not a genius. Pretty smart is good enough for me! :)</p>
<p>No, with an SAT score of 1580, I think you would actually fall in the "genius" catagory. An IQ above 140 is considered "genius or near genius." An SAT score of 1580 correlates, roughly, to an IQ around 150. Even with statistical fluctuations, I think you can safely say your IQ is at least 140.</p>
<p>I'm actually not sure folks should simply call "IQ above 140" "genius." But that is a very common definition. I mean, Mozart was a genius, Einstein was a genius... and there are a heck of lot more people in the world with IQ s of 140 and who aren't Mozart. But, using that very loose defintion, you probably actually ARE a genius!</p>
<p>Lucky you!</p>
<p>Now, what are you going to do with all that horsepower?? It doesn't matter if your car has a supercharged engine if you aren't driving in the right direction!</p>
<p>-greg</p>
<p>I was lucky enough to take the SAT when one math question wrong was still an 800 (I would know!! Apparently I got one geometry question wrong or something.). But I know other people who got one wrong and then got 790 or whatever so it's not exactly fair. BUT if I got an 800 with one wrong, the theory is that if I took that other SAT where an 800 was none wrong, I would not have gotten one wrong. PS I took the SAT in 04 so I do not know the current state of new SAT math 1 wrong 800s.</p>
<p>I always thought it was weird when my mom said she got high 1300s and went to Penn Wharton undergrad but keep in mind she worked her way through college so she obviously did not have the money to take SAT prep courses or even the SAT more than once like people do now. With recentering you have to wonder how many people take the SAT only once with no prep and get those kinds of scores today. That probably explains the larger number of people getting perfect scores.</p>
<p>Oh, now I really feel flattered, Greg. Actually, when I was in MS, they had us take a standardized test that wasn't an IQ test but was scored out of 200 and had a bell curve exactly like an IQ test but they SWEAR it's not an IQ test...anyway, I believe I scored about 148 or 150 on that. So, it seems like the SAT really IS a good predictor of IQ...hmmm. </p>
<p>But that's true, I think of Mozart/Einstein/etc. when I think of genius, and I've pretty much decided that maybe really smart is a nicer thing to be than an Einstein genius. Of course, I'll never know, so that's a convienient conclusion. </p>
<p>Anyway, I think that the curve on the math has gotten harder (meaning the test has probably gotten easier...anyway), because when I took it last June, I got 1 wrong and that was a 780 :(. I don't know if you could even have gotten a 790...maybe if you just omitted one? The next time I took it (last October), two wrong was a 770, so the curve was a bit nicer. I was able to acheive an 800 CR with one wrong, but I do think that that CR section was pretty tough.</p>
<p>"I always thought it was weird when my mom said she got high 1300s and went to Penn Wharton undergrad"</p>
<p>Why did you think it was "weird"? A high 1300 (M+V) is a really good score, and was even better before 1995!! An old, pre 1995 1380 is about equal to a "new" 1430. I have no idea what universe you are living in to think a 1430 is "weird" for a Wharton student. </p>
<p>According to the following website, the average SAT(M+V) of entering Wharton undergrads is a 1448. Your mom's 1430 would certainly be in line with that!</p>
<p>Maybe that gives you some perspective. Plus, you must realize that if 1448 is the AVERAGE score, by definition HALF of the Wharton undergrads had SAT scores LOWER than 1448!!!!</p>