I’ve discussed this topic before, but in light of my revised formula for converting new SAT scores to IQ equivalents, some of my old conclusions require revision.

The average Harvard student has a post-1995 SAT score of 1490, which according to my latest estimate, equates to an IQ of 145 (U.S. norms), or 143 (U.S. white norms).

However according to my analysis of a Harvard study, Harvard undergrads have an average Wechsler IQ of 124 (U.S. norms) or 122 (U.S. white norms). This might be an underestimate because of sampling error or ceiling bumping, but taking the figure at face value, if students selected largely by the SAT to have an IQ equivalent 45 points above the U.S. mean on the SAT, average only 24 points above the U.S. mean on the Wechsler, a 0.53 correlation between SATs and Wechsler IQ in the general U.S. population is crudely implied: 24/45 = 0.53.

There is reason to believe this result is not anomalous.

Scientists Meredith C. Frey and Douglas K. Detterman came up with a formula for predicting how one would score on an official IQ test from their performance on the post-1995 SAT. The formula was presumably derived from regression predicting IQs on the Raven Advanced Progressive Matrices (RAPM) from scores on the SAT (see scatter-plot C below) of 104 students at a selective private university.

Such formulas have value, because if you ever hit your head, a psychologist might administer an IQ test to see if you have brain damage. But if you score low, it doesn’t necessarily prove any damage; it could just be that you’re naturally dumb. Thus knowing your statistically expected performance on an official IQ test from your past performance on tests like the SAT is diagnostically useful.

The formula is as follows:

X’IQ = (0.095 * SAT-M) + (-0.003 * SAT-V) + 50.241

It is interesting to apply this formula to the average Harvard student who scored 1490 on the SAT (reading + math). Assuming the typical Harvard undergrad scored 745 on both the reading and the math section, the formula predicts they will score ~~123~~ 119 on the Raven IQ test (the Raven was normed in lily white Iowa so ~~123~~ 119 reflects U.S. white norms, which are equivalent to ~~125~~ 121 in U.S. norms). Remarkably close to the actual IQ of Harvard students on the Wechsler!

So people who have an IQ equivalent 45 points above the U.S. mean on the new SAT, scoring only about ~~25~~ 21 points above the U.S. mean on the Raven, once again implying a post-1995 SAT-IQ correlation ~~that is only in the mid 0.50s: ~~around 0.5:

~~25/45 = 0.56~~ 21/45= 0.47

One possibility is that this formula underestimates the expected intelligence of high SAT people because the scatter-plot clearly shows a pile up of scores at the ceiling of the RAPM, however this pileup only seems to consist of 12% of the sample (not enough to significantly flatten the regression line, especially since some of them would have scored the same even if the test had more hard items).

In addition, I personally looked at the scatter plot carefully and did my best to write down the RAPM IQs of every single participant with an SAT score from 1400-1600. This was an admittedly subjective and imprecise exercise given how small the graph is, but I counted 38 top SAT performers and these were their approximate RAPM IQs: 95, 102, 105, 108, 108, 110, 110, 113, 113, 113, 113, 113, 117, 117, 117, 117, 117, 120, 120, 120, 122, 122, 128, 128, 128, 128, 134, 134, 134, 134, 134, 134, 134, 134, 134, 134, 134, 134.

While the mean RAPM IQ of the top SAT performers was likely reduced by ceiling bumping, the median wold not have been, and that median was 120, suggesting the regression prediction of ~~123~~ 119 was not too low.

If we assume that the correlation between the SAT and Raven is entirely caused by *g* (general intelligence)…a somewhat reasonable assumption given they seem at least superficially to have little in common but reasoning.. and if we further assume that the RAPM has a *g* loading of about 0.8, then we can deduce that the new SAT has a *g* loading of ~~0.7 (0.56/0.8)~~ 0.59 (0.47/0.8). That’s hardly bad, but not as good as the Wechsler IQ tests which are thought to have g loadings of 0.9-0.95. I guess there’s a reason beyond political correctness why Mensa stopped accepting SAT scores for admission.

chartreuse

said:selected largely by the SATYET MORE LIES FROM PEEPEE-TARD.chartreuse

said:THE SAT IS AN IQ TEST, SO IT CAN’T “CORRELATE WITH IQ” FUCKTARD…ALL THAT CAN BE SAID IS IT CORRELATES WITH A PARTICULAR IQ TEST.

AND FUCKTARDED PEEPEE AS USUAL SHOWS HER MATHEMATICAL IMBECILITY BY NOT TAKING INTO ACCOUNT RESTRICTION OF RANGE.chartreuse

said:average only 24 points above the U.S. mean on the Wechsler, a 0.53 correlation between SATs and Wechsler IQ in the general U.S. population is crudely implied: 24/45 = 0.53.FALSE!MATHEMATICALLY RETARDED PEEPEE.chartreuse

said:if the conditional distribution is the same for the subpopulation one would expect that:

1. the mean to be the full range correlation multiplied by the mean of the subpopulation.

2. the correlatuion to be greatly attenuated.

peepee is using only #1.she needs to look at both.chartreuse

said:If we assume that the correlation between the SAT and Raven is entirely caused by g (general intelligence)…a somewhat reasonable assumption given they seem at least superficially to have little in common but reasoning.. and if we further assume that the RAPM has a g loading of about 0.8, then we can deduce that the new SAT has a g loading of 0.7 (0.56/0.8). That’s hardly bad, but not as good as the Wechsler IQ tests which are thought to have g loadings of 0.9-0.95. I guess there’s a reason beyond political correctness why Mensa stopped accepting SAT scores for admission.100% RETARDED GIBBERISH.THE G LOAD IS IN YOUR PANTS PEEPEE.

G IS NOT A THING FUCKTARD.

IT DEPENDS ENTIRELY ON THE SET OF TESTS USED.

THE RPM HAS THE WORST CORRELATION WITH OTHER SELF-NAMED IQ TESTS. MUCH WORSE THAN THE SAT, ETC.chartreuse

said:THE RPM ALSO HAS A SHITTY HERITABILITY…ONLY .55 IN THE MINNESOTA STUDY FOR MZ TWINS…ADULTS AVERAGE AGE 41.chartreuse

said:IN FACT AS ONE MIGHT EXPECT…IF THE SD OF THE POPULATION IS HALF THAT OF THE ENTIRE POPULATION, A CORRELATION OF .53 CORRESPONDS TO A FULL RANGE CORRELATION OF…

.9!

PEEPEE IS HOIST BY HER OWNRETARD.pumpkinperson

said:LOL!

You’re so stupid.

You are correcting for range restriction which is only appropriate when an actual correlation was CALCULATED in a restricted group.

That’s not what I did.

What I did was report the Wechsler IQ and Raven IQ of people with SAT IQs of 145.

Since people with SAT IQs 45 points above average have Wechsler IQs and Raven IQs that are only 53-56% as many points above 100 (IQ 124 and IQ 125 respectively), I estimated the REGRESSION SLOPE in the general population predicting official IQ from SAT IQ is 0.53-0.56

Since in a standardized bivariate normal distribution, the regression slope equals the correlation, I estimated the correlation in the general population to be 0.53-0.56 since that is the apparent slope of the regression line.

WHAT PART OF THAT DO YOU NOT UNDERSTAND??????My analysis was not perfect.

There are criticisms that might be made

But you just blathering on about range restriction indicates a serious conceptual problem on your part.

chartreuse

said:apparently you CANNOT read.read what i wrote peepee-TARD.

YOU ARE A COMPULSIVE LIAR.chartreuse

said:AND YOU’RE SO MATHEMATICALLY ILLITERATE YOU HAVE ASSUMED THAT THERE IS THE SAME RELATIONSHIP AT THE TAILS AS AT THE CENTER…WHEN THIS HAS BEEN SHOWN TO BE…FALSE!PEEPEE IS:

1. MATHEMATICALLY ILLITERATE

2. A COMPULSIVE LIAR

GIVE THE FUCKING CORRELATION TOO PEEPEE-TARD, NOT JUST THE SAMPLE MEAN OVER THE SAMPLE MEAN.CAN YOU?

DO YOU KNOW HOW?

FREY AND DETTERMAN CORRECTED FOR RESTRICTION OF RANGE IN THEIR OWN WAY AND FOUND A CORRELATION OF…

.7!

CAN YOU STOP LYING FOR EVEN A SECOND RETARD?

OR IS IT PART OF YOUR LOW IQ THAT YOU HAVE TO LIE CONSTANTLY?chartreuse

said:LOL PEEPEE LAUGHING AT HER OWN LIES.

APPARENTLY SHE CAN’T TELL THE DIFFERENCE BETWEEN FACTS AND LIES.

SO MAYBE SHE

REALLYBELIEVES THAT SHE’S BETTER AT MATH THAN MISDREAVS OR ME.OR MAYBE SHE’S BLOCKED THE TERM “MISDREAVS” BECAUSE SHE KNOWS THE TRUTH.

So people who have an IQ equivalent 45 points above the U.S. mean on the new SAT, scoring only about 25 points above the U.S. mean on the Raven, once again implying a post-1995 SAT-IQ correlation that is only in the mid 0.50s:25/45 = 0.56

…

Study 2 investigated the correlation between revised and recentered SAT scores and scores on the Raven’s Advanced Progressive Matrices among 104 undergraduates. The resulting correlation was .483 (.72 corrected for restricted range). http://pss.sagepub.com/content/15/6/373.shortIF I WERE PEEPEE I’D HANG MYSELF FOR HAVING SUCH A LOW IQ AND BEING A COMPULSIVE LIAR.chartreuse

said:it WAS FUCKTARD…THOSE WHO TAKE THE SAT ARE A restricted group.

u

have

a

LOW

IQ.pumpkinperson

said:it WAS FUCKTARD…THOSE WHO TAKE THE SAT ARE A restricted group.

So much stupidity. So little time to correct it. Those who take the SAT are a restricted group, but I did not calculate the correlation among those who took the SAT.I estimated the correlation AMONG ALL AMERICANS OLD ENOUGH TO HAVE TAKEN THE NEW SAT!!!

Those who took the SAT and were 45 IQ points above the U.S. population average (NOT THE SAT taker average) on the SAT, were only 25 points above the U.S. average on the Raven.

That gives 25/45 = 0.56 regression slope in the GENERAL POPULATION, not the SAT POPULATION!!!!!!!!!!!!!!!!!

Your IQ is too low to do statistical research in the social sciences.

Your Wechsler scores were so incredibly right and so were all the universities who rejected youchartreuse

said:i UNDERSTOOD all that peepee-tard.but because you’re SOCIALLY RETARDED in addition to being retarded in every other way

you didn’t understand that i understood that

misdreavs and i know

YOU’RE A MORON.

YOU HAVE A VERY LOW IQ.

YOU HATE IT.

SO YOU’VE CONVINCED YOURSELF YOU HAVE A NORMAL OR HIGHER IQ.

SO I’M STILL WONDERING:

1. DOES PEEPEE EVEN KNOW SHE’S LYING?

2. CAN PEOPLE WITH IQs OF 60 OR LOWER LIKE PEEPEE TELL THE DIFFERENCE BETWEEN FACTS AND LIES?chartreuse

said:AND PEEPEE KEEPS LYING ABOUT MY WISC SCORES…PEEPEE IS A VERY STRANGE PERSON.

1. SHE HAS A VERY LOW IQ.

2. SHE BELIEVES SHE HAS A HIGH IQ.

3. ALMOST EVERYTHING OUT OF HER MOUTH IS A LIE.chartreuse

said:ONE ASPECT OF PEEPEE’S MO IS TO1. LIE ABOUT WHAT OTHER PEOPLE SAY.

2. CONDEMN THEM FOR WHAT THEY NEVER SAID.Animekitty

said:Ravens IQ 125 = SAT IQ of 145

If I score 125 on Wechsler and take it twice the range between the two tests is 5 points. I could score on the second test between 127 or 122 because of the correlation of 0.9

If I take the SAT test then my score will vary between the it and the Wechsler test by 50 points. If I score 125 on the Wechsler I could score between 100 and 145 on the SAT. So the correlation is 0.56

5 points variation is smaller than 50 points variation.

These numbers are not exact but rough estimates.

Animekitty

said:Sorry I messed up:

Wechsler IQ 125 = SAT IQ of 145

Animekitty

said:50 point is the wrong number but not hugely wrong. Just less precise. I know I am bad at math.

Animekitty

said:Yes jean I am like 22 points bellow pumpkin person. And I know a guy with an IQ of 90 so the difference is obvious to me. I do know that a test correlated with itself will mean that the score may vary when you take the test again. Maybe that does not apply here?

Stetind

said:C’mon, be nic to AK, Jean. He’s the nicest person on here. He’s never insulted anyone.

chartreuse

said:and what are SAT 2 AND SAT 3?combining SAT, SAT 2, SAT 3 gives an almost perfect correlation.Animekitty

said:X’IQ = (0.095 * SAT-M) + (-0.003 * SAT-V) + 50.241

Paul Allen got 1600

So on the Ravens he would get 123.841

Previously I thought he would be 170 because previously 1600 = 170.

Now 1600 = 123.841 if you are correct.

Paul Allen founded Microsoft with Bill gates.

His net worth is 17.4 billion but the SAT by your formula predicts 123.841 on the Ravens at the highest score. And you said the average self made billionaire would be 130 or so.

Tomorrow I am going to have an MRI scan and an EEG. I think the reason I have low processing speed is because of physical and mental trauma. I have a friend who plays playstation four. He says he has fighter jet pilot reflexes. If my friend has such reflexes the neuronal signals would be less fuzzy than mine are so can store more complex patterns. I am curious about how Chronometric scores would work out for a person with a real IQ of 170. I would like to understand how Neuroplasticity produces high (g). The patterns must be very complex and highly coordinated.

Gypsyman

said:X’IQ = (0.095 * SAT-M) + (-0.003 * SAT-V) + 50.241

Paul Allen got 1600

So on the Ravens he would get 123.841″

No Animekitty, this is a formula for POST 1995 SAT, Paul Allen did the PRE 1995 SAT.

His Real IQ would have been much, much higher.

Animekitty

said:I just assumed he took it in college. College students are 18-20 so I based that on his age and when he would likely have taken it. He was 18 before 1995.

Animekitty

said:Forget what I said

you said post not pre

chartreuse

said:PEEPEE IS SO FUCKING RETARDED I’M SURPRISED SHE CAN BREATH AND BREATH AT THE SAME TIME.So people who have an IQ equivalent 45 points above the U.S. mean on the new SAT, scoring only about 25 points above the U.S. mean on the Raven, once again implying a post-1995 SAT-IQ correlation that is only in the mid 0.50s:25/45 = 0.56Study 2 investigated the correlation between revised and recentered SAT scores and scores on the Raven’s Advanced Progressive Matrices among 104 undergraduates. The resulting correlation was .483 (.72 corrected for restricted range).http://pss.sagepub.com/content/15/6/373.shortIF I WERE PEEPEE I’D HANG MYSELF FOR HAVING SUCH A LOW IQ AND BEING A COMPULSIVE LIAR.chartreuse

said:…OR RATHER BREATHE AND BREATHE… IF YOU LIKE THAT SPELLING.NOTICE THAT .72 IS A LOT HIGHER THAN .56.

PEEPEE CAUGHT IN A LIE ONCE AGAIN.

AND GIVEN THE REALITY OF SLDR THE FULL RANGE CORRELATION IS PROBABLY MORE LIKE .9.pumpkinperson

said:Study 2 investigated the correlation between revised and recentered SAT scores and scores on the Raven’s Advanced Progressive Matrices among 104 undergraduates. The resulting correlation was .483 (.72 corrected for restricted range). http://pss.sagepub.com/content/15/6/373.shortFrey and Detterman found a 0.483 correlation between the Raven and the post-1995 SAT in a restricted sample. They jacked the correlation up to be 0.72 by correcting for range restriction, but if the correlation really were 0.72 in the general population, then assuming a bivariate normal distribution, people with SAT IQs that are 45 points above average should be 45(0.72) = 32 points above average on the Raven, on average.

Instead, they’re only about 25 points above average, suggesting they over-estimated the general population correlation OR that the correlation is not linear.

The problem is correcting for range restriction in test scores does not correct for all the other ways restricted samples are not representative (i..e. more motivated, more math training etc) and thus can overestimate the general population correlation (if all U.S. teens took the SAT)

My method circumvented that problem

Too bad you’re too stupid to see that

chartreuse

said:SHOULD BE PEEPEE.AND I SAID THAT BEFORE YOU DID…

ON THIS THREAD.

LEARN TO READ.chartreuse

said:YOUR RETARDED METHOD DOES NOT CIRCUMVENT THE PROBLEM FUCKTARD!IT’S JUST ANOTHER ESTIMATE.

WHEN THE GUYS WHO DID THE STUDY FIND THAT YOUR ESTIMATE IS SHIT…

IT’S BECAUSE IT IS SHIT.

NOT BECAUSE YOU’RE SMARTER THAN THEY ARE.chartreuse

said:FURTHER PEEPEE-TARD THINKS SHE CAN FIND THE G-LOADING OF THE SAT BY DIVIDING THE CORRELATION OF THE SAT WITH THE RPM BY THE CORRELATION OF THE RPM WITH G.TARDS GONNA TARD.

MISDREAVS WAS SOOOOOOOOOOOOOOOOO

RIGHT!chartreuse

said:then we can deduce that the new SAT has a g loading of 0.7 (0.56/0.8)no “we” can’t…you fucking subhuman garbage.chartreuse

said:following peepee’s subhuman un-logic if a test correlated with the RPM at .8 it would measure g itself.chartreuse

said:MORE LIES FROM PEEPEE-TARD…But school grades correlate better with SOME IQ tests than SOME IQ tests correlate with one another.FALSE!Subjective opinion of who is smart correlates better with SOME IQ tests than SOME IQ tests correlate with one another.FALSE!PEEPEE JUST MAKES SHIT UP WHEN IT SUITS HER…SHE LOVES TO LIE.

SHE’S A PERSON “OF THE LIE”.

OF COURSE PEEPEE FAILS TO MENTION THAT

THE ONLY THING IQ TESTS HAVE WHICH OTHER TESTS DON’T IS THAT THEY ARE MARKETED AS IQ TESTS…TO THE IDIOTS LIKE PEEPEE…THE CLAVER KNOW THAT THE SAT ETC. ARE IQ TESTS IN ALL BUT NAME…BUT PEEPEE KNOWS SHE’D SCORE ABOUT 600 ON HER COMBINED SAT SO SHE HAS TO BELIEVE THAT SHE’D SCORE HIGHER ON A SELF-DESCRIBED IQ TEST…SHE WOULDN’T.pumpkinperson

said:But school grades correlate better with SOME IQ tests than SOME IQ tests correlate with one another.FALSE!TRUE! Elementary school grades correlate 0.65 with IQ. Correlation between Draw a Person IQ test and the WISC? About 0.5.

pumpkinperson

said:Daniel Seligman cited the 0.65 figure but I guess it depends on the school and subject.

chartreuse

said:SHOW US YOUR SCORES PEEPEE.chartreuse

said:ELEMENTRAY SCHOOL GRADES ARE MEANINGLESS ACCORDING TO PEEPEE BECAUSE THEY’RE IN CHILDREN.

SHOW US YOUR SCORES PEEPEE.pumpkinperson

said:My scores:

https://pumpkinperson.com/2015/11/03/my-chronometric-scores/

chartreuse

said:SHOW US YOUR SCORES PEEPEE./NOT TETRUS SCORES./ruhkukah

said:No female is thist interested in iq

chartreuse

said:PEEPEE IS A FEMALE WHO “IDENTIFIES” AS MALE.

SHE’S A TRANS-SEXUAL.

chartreuse

said:SHOW US YOUR SCORES PEEPEE.chartreuse

said:ALSO.SHOW US YOUR DICK.

AND SHOW US YOUR OVERBITE AND ROUND EYES.chartreuse

said:SHOW US YOUR SCORES PEEPEE./LEGIT SCORES.

NOT TETRUS SCORES.chartreuse

said:STILL WAITING ON PEEPEE’S IQ…STILL WAITING.pumpkinperson

said:I gave the most legit scores there are.

Way more legit than the SAT and GRE

We don’t have SATs in Canada. I never went to graduate school so I didn’t take the GRE.

I probably would have scored 800+ on the math. Verbal, not so much.

pumpkinperson

said:I’ll look into it. I was even tempted to take the real SAT just for fun. I’m sure they offer it in canda for people interested in studying in the U.S.

chartreuse

said:YOU WOULD HAVE SCORED 200 ON THE MATH PEEPEE.

THE VERBAL MAYBE 400.

WAY LESS LEGIT.

YOU GAVE TETRIS SCORES PEEPEE.

STILL WAITING.

STILL………

WAITING……….

AND SHOW US YOUR DICK TOO.

AND YOUR ROUND EYES.

AND FAIR SKIN.

chartreuse

said:AND TAKE THE SoA EXAM IN PROBABILITY AND STATISTICS.

IT’S THE SAME TEST IN THE US AND CANADA, AND EVERY CANADIAN ACTUARY HAS TO TAKE IT.

STILL WAITING.

chartreuse

said:STILL WAITING…

pumpkinperson

said:I’ve already revealed a far more legit score than you have.

The only other legit scores I have are my childhood WISC which I do intend to do a post once I find the score report.

chartreuse

said:POST A SCORE FOR AGE >= 18.

CHILDHOOD SCORES ARE MEANINGLESS.

chartreuse

said:after you post someone else’s report…you’ve seen so many…

post a dick pic too.

and show us your round eyes and fair complexion.

STLL WAITNG.

pumpkinperson

said:Most Northeast Asians have whiter skin than you

Hal

said:Why do you keep responding to chartreuse, pumpkinperson? You are wasting your time with this person: he or she is either a troll, an idiot, or both. Their IQ is obviously not very high, too; probably no higher than 105–110, based on their lack of eloquence and the overall inanity of their observations. In fact, I seriously doubt that the 129 figure is even accurate (though perhaps it was back when chartreuse was a child), as cognitive decline during adulthood is common for people who suffer from mental illness, which seems to be the case for chartreuse.

Regarding the topic at hand, I am wholly unsurprised that the correlation between SAT scores and IQ scores is weak. Getting into an Ivy League school in this day and age generally only requires two characteristics: wealthy, elite parents; and ambition, which is a common personality trait among the children of rich, success-driven families.

One can have an IQ in the 115–120 range and still score exceptionally high on the SAT, much higher than a person with an IQ of 160 who lives in poverty and never had the opportunity to receive a good education. Someone with an IQ of 160 might be an autodidact as well, preferring to learn subjects that they personally deem more important than what the SAT tests for, and score poorly on the SAT as a result.

Furthermore, I have found that it is fairly easy to “hack” the SAT just by using the right textbooks, tutors, etc. This does not mean that anyone with the right resources will achieve a high score, just that a combination of the right resources, academic drive, and familial pressure is more indicative of success than a single IQ score.

The SAT measures knowledge more than it measures intelligence. Perhaps this was not always the case, but it certainly is now.

pumpkinperson

said:Why do you keep responding to chartreuse, pumpkinperson? You are wasting your time with this person: he or she is either a troll, an idiot, or both. Their IQ is obviously not very high, too; probably no higher than 105–110, based on their lack of eloquence and the overall inanity of their observations.You’re right. I should stop feeding the troll.

Hal

said:Not just a troll, but someone who has harassed and insulted you to an extent that is, in my opinion, very obsessive and creepy. Is there any way to block them completely?

(By the way, I have been lurking your blog for quite a while—I am terribly eager to see your analysis of Vladimir Nabokov! I was also wondering if you would consider analyzing the IQ of Bernie Sanders, because I find him extremely interesting.)

pumpkinperson

said:Yes Nabokov is coming soon, and Sanders sometime after that.

Hal

said:Excellent! Since I read this blog regularly, I’ll post my own rough estimates now and, in the near future, will be able to see how ours compare:

Bernie Sanders: 130–135 IQ

Vladimir Nabokov: 140–145 IQ

Hal

said:One more suggestion: Would you mind estimating the IQ of James Joyce at some point in the future? I find him incredibly fascinating.

pumpkinperson

said:Would you mind estimating the IQ of James Joyce at some point in the future?Good idea. I will.

pumpkinperson

said:people who suffer from mental illness, which seems to be the case for chartreuse.Understatement of the decade.

grey enlightenment

said:if everyone is coached then still smarter people among the coached group will still score higher than thier coached peers, so there is still a correlation.

ruhkukah

said:Interestingly SAT is one of the few tests where the non-g portion predicts grades above and beyond the g portion.

pumpkinperson

said:Makes sense. Reading and math are very useful skills in their own right,

Tenn

said:Isn’t there something wrong with the formula when having a high verbal score negatively impacts IQ? Why would having a CR of 200 correlate higher with a high IQ than would a CR of 800?

Tenn

said:I assume the formula is meant to be: X’IQ = (0.095 * SAT-M) + (0.003 * SAT-V) + 50.241.

This produces the expected Harvard result you mentioned. Just a typo then?

pumpkinperson

said:No, that’s the equation from the paper for predicting Raven IQ (see equation 2):

X’IQ = (0.095 * SAT-M) + (-0.003 * SAT-V) + 50.241

But it looks like I overlooked the negative sign when doing the calculation.

So if Harvard students average 1490 on the SAT (745 V, 745 M)

X’IQ = (0.095 * 745) + (-0.003 * 745) + 50.241

X’IQ = (70.775) + (-2.235) + 50.241

X’IQ = 68.54 + 50.241

X’IQ = 68.54 + 50.241

X’IQ = 119

In my post I said 123.

So it looks like the regression was even more severe than I thought, implying an even lower correlation between SAT and Ravens.

I should do an update.

Thanks for bringing this to my attention.

pumpkinperson

said:Okay, I crossed out the errors from the post.

pumpkinperson

said:I am sure the negative sign was an error. The idea that CR scores are negative correlates of IQ is preposterous.Maybe. The fact that the authors of the study didn’t even comment on such a strange result makes me think it might be an error.

On the other hand, in small data sets, where there’s a lot of multicollinearity, it’s not rare to have one statistically redundant positive predictor flip and become negative:

In statistics, multicollinearity (also collinearity) is a phenomenon in which two or more predictor variables in a multiple regression model are highly correlated, meaning that one can be linearly predicted from the others with a substantial degree of accuracy. In this situation the coefficient estimates of the multiple regression may change erratically in response to small changes in the model or the data. Multicollinearity does not reduce the predictive power or reliability of the model as a whole, at least within the sample data set; it only affects calculations regarding individual predictors. That is, a multiple regression model with correlated predictors can indicate how well the entire bundle of predictors predicts the outcome variable, but it may not give valid results about any individual predictor, or about which predictors are redundant with respect to others.https://en.wikipedia.org/wiki/Multicollinearity

pumpkinperson

said:Why would having a CR of 200 correlate higher with a high IQ than would a CR of 800?Yes that is strange. Perhaps doing well on the math section despite being a slow reader takes extra IQ? Perhaps a lot of the students with low CR scores were high IQ second language students. Who knows.

Because both reading and math are so highly correlated with each-other, there’s a great deal of redundancy in an equation that predicts Raven IQ from both variables, and that redundancy can lead to strange somewhat random results.

I wouldn’t over-interpret it. A regression equation based on a different sample might find the opposite.

Tenn

said:What’s particularly puzzling is that I’ve always seen the reading portion as far more g-loaded (seeing as it’s far less trainable). However, maybe the reason math skill correlates more closely with the Ravens is because both stress non-verbal abilities.

Still, it is undeniably baffling that any study could predict that a student with a score of 1000 (800-M, 200-CR) would be more likely to possess a high IQ than one with a perfect 1600. Utterly bizarre.

Tenn

said:On another note: when was the Raven normed? Would the 121/119 figure be inflated due to the Flynn Effect, or am I missing something?

pumpkinperson

said:On another note: when was the Raven normed? Would the 121/119 figure be inflated due to the Flynn Effect, or am I missing something?Good question. The RAPM had a norming in 1993 and the study was published circa 2003. At most there would have been a 2 point Flynn effect, but it might have even been zero over that period as I believe it supposedly was in Britain. I’ll look into it.

Stetind

said:Why do you think Jorge isn’t smart, pumpkin?

You usually accept Prometheans’ word as gospel. Jorge could join the Prometheus society. Look up their membership requirements. 1560 on the old SAT is the cutoff.

Gypsyman

said:Oooh, literally JUST on the cutoff.

But the Prometheus society also accepts a GRE of 1610 as acceptance for membership and unless I’m very much mistaken the GRE has a maximum score of 1600…

chartreuse

said:you’re assuming peepee is smart enough to appraise the smartness of those smarter than she is.the rule is as john cleese has put it:

Stetind

said:❤ Monty Python.

All joking aside, why are you so bitter, Jorge?

pumpkinperson

said:Why do you think Jorge isn’t smart, pumpkin?You usually accept Prometheans’ word as gospel. Jorge could join the Prometheus society. Look up their membership requirements. 1560 on the old SAT is the cutoff.

Prometheus accepts a pre-April 1995 SAT score of 1560. It’s unclear whether he took the SAT before or after April 1995.

The reason I doubt his IQ is because the Wechsler scales are much more comprehensive measures of intelligence than the SAT and his Wechsler scores were much less impressive than his SAT scores.

Also,the criticisms he’s made of my statistical arguments have all been either pedantic or confused. For example, he seemed to think that if X and Y correlate 0.4 in a bivariate normal distribution of the entire population then a subgroup that was 33% less variable in X should also be 33% less variable in Y. WRONG!

He was also slow to understand that the correlation I deduced from the degree of regression high SAT people show on the Wechsler and Raven did not need to be corrected for range restriction.

He also does not seem to understand that the correlation between two test scores is a product of their factor loadings and keeps reverting to the pedantic argument that knowing the correlation between A and B and between B and C does not tell you the correlation between A and C.

Perhaps he does or could understand all of these points but lacks the impulse control to argue rationally or read carefully, but the fact that he’s also declined to take my heritability test, despite this being his primary interest, really makes me wonder about his capacity for subtle mathematical thought:

https://pumpkinperson.com/2015/04/02/get-a-degree-in-heritability/

On the other hand, I think he’s incredibly knowledgeable which indicates a very good long-term memory. I think he has an extremely high verbal IQ. Pincher Martin thought he couldn’t write, but I think he’s a very powerful writer.

Stetind

said:Thank you for the excellent response, Pumpkin.

Gypsyman

said:I think it’s strange that the Verbal + Quantative accepted score is 1610, when it seems to me that the maximum score is 1600 on the GRE.

pumpkinperson

said:Before October 1, 1981, it was possible to score above 1600 on the V + Q composite of the GRE.

Gypsyman

said:How much further above and where can I find a source to read up on this?

Thankyou.

chartreuse

said:AN EXAMPLE OF PEEPEE’S RETARDATIONAlso,the criticisms he’s made of my statistical arguments have all been either pedantic or confused. LIE! For example, he seemed to think that if X and Y correlate 0.4 in a bivariate normal distribution of the entire population then a subgroup that was 33% less variable in X should also be 33% less variable in Y. WRONG!IT IS! RETARD USING YOUR ASSUMPTION THAT THE SUBPOPULATION HAS THE SAME CORRELATION WITH IQ AS THE POPULATION AS A WHOLE. BUT YOU’RE SO FUCKING RETARDED, YOU CONVENIENTLY FORGET ALL OF THE ASSUMPTIONS YOU’VE MADE, AND THEN ACCUSE ME OF NOT UNDERSTANDING.YOU’RE A FUCKING JOKE YOU’RE SO DUMB PEEPEE.

YOU HAVE NO MEMORY AT ALL, LET ALONE A LONG TERM MEMORY.

AND…

EVEN IF THE CORRELATION IS APPROPRIATELY ATTENUATED BY THE RESTRICTION OF RANGE OF THE SUBPOPULATION…

SO IS THE SUBPOPULATION’S SD IN VARIABLE Y, JUST NOT BY AS MUCH AS IF IT HAD THE SAME CORRELATION WITH Y AS THE WHOLE POPULATION. BUT YOU ASSUMED ZERO SUCH ATTENUATION.

AGAIN…

YOU’RE A FUCKING JOKE YOU’RE SO DUMB PEEPEE.

AND UTTERLY CONFUSED.pumpkinperson

said:I never assumed that the same IQ-income correlation existed within the subpopulation as existed within the general population. I assumed that the mean income of the subpopulation could be inferred from the mean IQ of the subpopulation using the IQ-income correlation within the general population. Here’s my original post:

https://pumpkinperson.com/2015/12/07/does-iq-explain-racial-differences-in-extreme-wealth/

I’m VERY concerned that you don’t have the cognitive ability to understand that subtle distinction.

chartreuse

said:STLL W8TING PEEPEE.DO YOU HAVE ANY SIGN OF MATHEMATICAL ABILITY?

NO!

DO I?

YES!

1. GRE 800 M AND SAT 800 M, 99th PERCENTILE ON GMAT QUANTITATIVE

2. MATH DEGREE

3. PASSED ALL ACTUARIAL EXAMS ON MY FIRST TRY.

4. HIGHEST SCORE IN NORTH AMERICA ON THE SoA EXAM 100 AND A PERFECT SCALED SCORE ON EXAM 110 (PROBABILITY AND STATISTICS)

YOU’RE DOING AN INCREDIBLE DISSERVICE BY TRYING TO DO ANY MATH AND THUS MISLEADING PEOPLE EVEN MORE INCOMPETENT THAN YOU.

CUT IT OUT PEEPEE!

PEEPEE IS A MATHERMATICAL MORON.

AND BOTH MISDREAVS AND SWANK REALIZED THIS.

MY MATH IQ IS AT LEAST 4 SDs ABOVE PEEPEE’S.

AND I HAVE PROOF.pumpkinperson

said:Swank never criticized my mathematical skills. Only Misdreavus did, and then I proceeded to BEAT THE LIVING SHIT OUT OF HIM!!!!:

https://pumpkinperson.com/2015/05/31/gay-black-hbder-attacks-me-on-twitter/

Your high math scores are extremely impressive but you haven’t displayed that level of talent here, perhaps because the math on this blog is more inutitive, subtle, and non-linear, and you can’t think laterally.

I on the other hand scored well into the top 1% on a 100% culture fair, 100% physiological measure of pure unadulterated biological intelligence, and thus I’m able to think intuitively and non-linearly:

https://pumpkinperson.com/2015/11/03/my-chronometric-scores/

Paper and pencil tests are so 20th century. Chronometrics is the future.

chartreuse

said:I KNOW that you don’t have the cognitive ability to understand that VERY UN-subtle distinction.

you made it AGAIN in your most recent post! FUCKTARD!

here’s an example pee-tard.

it’s easiest to see for strong correlations.

if a subpopulation has an SD 1/2 that of the whole population, and the population level correlation with variable Y is .8.

what is the SD of the subpopulation in variable Y relative to that of the entire population?

IT’S NOT 1!

ANY GUESSES PEEPEE-TARD?

IT’S ONLY 72% THE SD OF THE ENTIRE POPULATION IN VARIABLE Y.

NOW AS TO THE .4 AND 2/3ds EXAMPLE. THE SUBPOPULATION’S SD IS ALSO NOT 1!

IT’S < .95.

AND THIS IS IMPORTANT WHEN CALCULATING THE NUMBER IN THE 400.

AND YOUR ASSUMPTIONS LEAD TO THERE BEING MORE THAN 400 PEOPLE IN THE 400.

GIVE UP ALREADY PEEPEE.

1. GRE 800 M AND SAT 800 M, 99th PERCENTILE ON GMAT QUANTITATIVE

2. MATH DEGREE

3. PASSED ALL ACTUARIAL EXAMS ON MY FIRST TRY.

4. HIGHEST SCORE IN NORTH AMERICA ON THE SoA EXAM 100 AND A PERFECT SCALED SCORE ON EXAM 110 (PROBABILITY AND STATISTICS)

AS LONG AS YOU CAN'T ADMIT BEING STUPID YOU'LL NEVER BE SMART.

pumpkinperson

said:Except you foolishly thought it was 2/3rds. After I pointed out the absurdity, you are now coming back with a more tenable criticism. Too bad you don’t have integrity to admit you were originally quite confused.

chartreuse

said:AND WALCOTT KOed MARCIANO.AND I HAVE DISPLAYED THAT TALENT HERE.

YOU’RE JUST SO UTTERLY CONFUSED AND STUPID YOU CAN’T SEE IT.

HERE’S YOU MAKING EXACTLY THE MISTAKE YOU ACCUSE ME OF.it is interesting to apply this formula to the average Harvard student who scored 1490 on the SAT (reading + math). Assuming the typical Harvard undergrad scored 745 on both the reading and the math section, the formula predicts they will score 123 119 on the Raven IQ test (the Raven was normed in lily white Iowa so 123 119 reflects U.S. white norms, which are equivalent to 125 121 in U.S. norms). Remarkably close to the actual IQ of Harvard students on the Wechsler!So people who have an IQ equivalent 45 points above the U.S. mean on the new SAT, scoring only about 25 21 points above the U.S. mean on the Raven, once again implying a post-1995 SAT-IQ correlation that is only in the mid 0.50s: around 0.5

RANGE RESTRICTED:25/45 = 0.56 21/45= 0.47RANGE RESTRICTEDTHAT FORMULA WAS RANGE RESTRICTED.

YET YOU’VE USED IT TO COME UP WITH A POPULATION LEVEL CORRELATION.

YOU CAN’T DO THAT!

IT’S NOT A MEANS OF CORRECTING FOR RESTRICTION OF RANGE, IT JUST GIVES YOU THE SLOPE IN Z-SCORES OF THE RANGE RESTRICTED REGRESSION LINE…WHICH IS MUCH LESS THAN THE FULL RANGE REGRESSION LINE’S SLOPE.

THE AUTHORS’ CORRECTION TO THIS PROBLEM OF RANGE RESTRICTION WAS THE CORRELATION .72! NOT .56 OR .53 OR WHATEVER.

BUT IT DEPENDS EXACTLY HOW THE RANGE WAS RESTRICTED.

READ THE FOLLOWING:

http://www.amazon.com/Introduction-Mathematical-Statistics-7th-Edition/dp/0321795431

pumpkinperson

said:It doesn’t matter that the formula was for a range restricted sample. It still reveals the mean Raven IQ of high SAT IQ people. If one assumes those high SAT people are cognitively representative of high SAT people nationwide, then:

(Raven IQ of high SAT people – general population average)/(SAT IQ of high SAT people – general population average) = estimated correlation between Raven and SAT in general population

pumpkinperson

said:You seem to think I estimated the slope of the standardized regression line of the people in the SAT-Raven study. I did not. I estimated the slope of the standardized regression line if ALL American young adults had been in the study. Had I wanted to estimate the slope of those in the study, I would not have used IQs which are a general population rank, but rather Z scores based on the SAT & Raven means and SDs of only the study sample.

chartreuse

said:JUST MORE LIES.I NEVER ASSUMED IT WAS 2/3ds EXCEPT WHEN THE SUBPOPULATION HAD THE SAME CORRELATION WITH Y AS THE ENTIRE POPULATION.pumpkinperson

said:Which was never implied.

Anonymous

said:As abrasive as Jorge is, I’ve actually kind of grown to like him in an endearing way. I’m guessing PP thinks similarly. I kind of like the thing he does where he says one important word on its own

LINE!

But he could stand to post a little less.

Carl Churchill

said:THE SAT HAS NO CORRELATION WITH INCOME

Tenn

said:Hi Carl — The study you cite is utterly irrelevant to the point you’re trying to make. This is for two reasons.

First, the study draws a distinction from between aptitude tests (i.e. SAT) and achievement tests. It delineates that its focus is the latter (and in fact, specifically says it’s NOT about the SAT).

Second, the study doesn’t ever say there is no correlation between income and success on tests. From the conclusion: “We conclude that average income can have some impact on standardized test scores, yet it may

not be the main influencing factor.”

^That doesn’t exactly scream “NO CORRELATION” does it?

ruhkukah

said:LOL!

Carl Churchill

said:I mistyped. I meant that income has no impact on SAT scores, so claiming that it gives an advantage to rich test-takers is false.

Deeru

said:Random Question:

Since the pre-1995 sat has quite a high g-loading and there are old official tests out there, could you derive or provide a reasonable estimate of someone’s IQ if they were to take it today?

If this is the case it’s a cheap way for someone to find out their IQ. I was thinking if someone who took the old test fell into the age group of the majority of SAT takers (16-18 yrs), then maybe you could apply the Flynn effect by shifting the table -9 IQ points if the test was from 1985 for example. However, that’s assuming test takers get linearly better at the verbal and the math sections, which some data out there suggests otherwise.

I don’t know where I saw this off the top of my head, but there is data that shows the converted IQ gains that students make on the verbal and math sections over the years, so you could extrapolate from that what the adjusted table is for someone to take an old test today, given they are 16-18 yrs old. If they are outside of the age group, it probably is more complicated. What are your thoughts about this?

pumpkinperson

said:The Flynn effect is much smaller on the SAT than on conventional IQ tests.

marshamurphy1

said:The SAT is the only IQ test that matters to me, because it’s the one IQ test virtually every smart American takes. Few smart people take the WAIS or the GRE or the LSAT.

Virtually every American above 135 IQ takes the SAT.

It’s the one test where every smart American can be judged by the same standard.

But then I’m biased because I got a 1590 at only 16, back in 1994, the year before the test was recentered.

But then I was raised by New York psychologists who were obsessed with turning my brother and I into prodigies.

jorge videla

said:and the jews want to portray their enemies as dumb, prole, trailer trash,…

just look at the full on assault of trump…even though trump deserves it.

hitler was right…to a large extent…at least.

but two psychologists raising very high scoring kids is much more likely than two CPAs raising very high scoring kids.

even though it takes a much higher IQ to pass the uniform CPA exam than to get a masters or phd in psychology.

environment strikes again.btw…

peepee has yet to mention the greatest speech in american history…

oh believe me…

the donald’s acceptance speech.

the longest ever.