[off-topic comments will only be published in the open thread]
So in part 1 of this series we learned that Americans who take the new GRE have the following scores: Verbal 152.7 (standard deviation [SD] = 7.6), Quant 150.2 (SD = 7.8), and Combined 302.9 (SD = 13.6)
In part 2 we learned that Americans who take the new GRE have the following IQs: Verbal 117 (SD = 13.5), Math 117 (SD = 13.3) and Combined 119 (SD = 13.2), on a scale where all Americans score 100 (SD = 15).
Mapping the verbal, quant and combined GRE scores to their IQ counterparts, new GRE scores can be equated to IQ as follows:
Verbal IQ = 1.776(new GRE Verbal) – 154.243
Quant IQ = 1.705(new GRE Math) – 139.11
Combined IQ = 0.971(new GRE Combined) – 174.991
All IQs are calculated based on U.S. norms which is standard on official IQ tests like the WAIS-IV. To convert to U.S. white norms which are sometimes used in international research, see here.
Pumpkin Person 
So maximum measured IQ would be 150. And most elite graduate programs even in humanities would have average IQ of 130-145 with an average around 138.
I believe the math part is very crystallized intelligence loaded. If you take only the verbal part, the picture is better. You get score from 120-145 with an average around 132. Physics and math being the most selective.
The 4 best programs in physics would have an average of 146. Almosft the same for math. The 4 best in engendering and computer science would be at 132.
Then for better prediction you could use math score to deflate high verbal score in case math score is under 1sd (159). So Psy instead of 132 would go down to 125.
Then even for verbal that way, data are probably 3 points too high … but that’s a good proxy I guess.
*https://www.mbacrystalball.com/gre/gre-scores-ms-usa-universities
Hey Pumpkin, have you seen the GRE to LSAT converter? Really fascinating how the different subsections of the GRE seem to be weighted.
https://www.ets.org/gre/institutions/about/law/comparison_tool?WT.ac=gre_law_comparison_180302&utm_source=general&utm_medium=general&utm_campaign=gre&utm_content=law_comparison
Also really great article below on the conversion tool. They suggest that the ostensibly higher weighting of verbal may have to do more with their similar normalizations.
https://blog.powerscore.com/gre/gre-to-lsat-conversion
GRE and LSAT are such different tests, and my experience they test entirely different things. The GRE verbal is mostly a vocabulary, while the math portion is mostly a test of rudimentary algebra and geometry concepts. The LSAT (which I found to be much, MUCH more difficult) seems to tax working memory much more–knowing 4-syllable vocab words won’t help at all here. Mostly testing your ability to keep multiple sentences and clauses in your mind at one time and then asking you how the clauses interact. A lot of verbally ordering and grouping variables, if then conditional statements, causation vs. correlation distinction, identifying premises and conclusions in arguments, analogical reasoning, comprehending conjunctive and disjunctive sratements, informal logic, etc.
Yes the LSAT seems to test the verbal + working memory indexes of the WAIS-IV while the SAT tests the verbal + fluid indexes of the WISC-V
I said SAT not GRE because the GRE also has/had an analytical section which tests working memory
But it makes sense that you’d do better on the SAT than the LSAT because on the WAIS-IV working memory was not your strongest area but i still think the WAIS-IV underestimated you overall because you got robbed on similarities
Makes sense that the LSAT should emphasize working memory given the complex contacts lawyers have to interpret
I don’t think the GRE has had the old analytical with logic games and arguments for a while (maybe since the 80s?). The new analytical section is just an essay.
But the LSAT I believe is a much more comprehensive measure of verbal ability than the GRE verbal. Even though the ETS study showed that people who taken both tests usually scored similarly on both in terms of percentile. the GRE verbal just seems easier to beat with rote memorization (same is true of much of the math on the GRE– memorize some formulas and procedures, and you’re good to go. LSAT improvement requires a more complete overhaul of your thought processes, in a way.)
I’ll also probably start working as a LSAT tutor soon.
…part-time of course
Btw Gman have you ever thought about obsessively practicing the LSAT? Even if working memory is not your strongest area, performance on any one type of test can be dramatically improved through practice.
But what probably can’t be much improved is overall cognition
As a black with a sky high LSAT you could get into Harvard law & really rake in the coin
I gave this same advice in details to G last year.
I did practiced obsessively (almost 400 hours) for the 2018 February LSAT and got a 172. 🙂 Incredibly intense but worth it.
I didn’t even apply to Harvard even though everyone said I should (including people on your blog and others, and everyone I know in real life). Am I an idiot for this? I don’t know…I’m perfectly happy going to the University of Miami. Everyone says that your chances of getting a job is much better at the former but I have no desire for Big Law.
OMG i had no idea you scored so incredibly high!
You’re probably one of the top 7 blacks of the entire year on the LSAT
You’re insane not to apply to Harvard & Yale
I also recently received an upgrade to full-ride for University of Florida…should I consider this over Miami? I don’t know, Miami is more fun 🙂 Amazingly, I was rejected by Arizona State (ranked 25) and UCLA (ranked 16) after a long hold….I guess I’m the one genius black male they don’t like???
Amazingly, I was rejected by Arizona State (ranked 25) and UCLA (ranked 16) after a long hold
That’s bizarre. Do you have a low GPA? Or maybe you took the LSAT multiple times and scored lower the first time and they know?
Top 7? The data I’ve seen put roughly 20-40 blacks in the 170+ range in a given year. But black females are probably over-represented almost 2-1 in this range.
According to this source, it’s far fewer:
I’ll be using the data for African-Americans from the 2009-10 testing years. The key numbers are as follows:
Total number of test-takers: 14,585
Mean score: 142.04
Standard Deviation: 8.74
1SD Above Mean: 150.78 (151)
2 SD Above Mean: 159.52 (160)
2.5SD Above Mean: 163.89(164)
3 SD Above Mean: 168.26 (168)
Now with this data in hand, we can use this standard deviation curve (LinkRemoved) to get an idea of approximately how many AA test takers (what percentage, roughly) fall at or above each of the standard deviation milestones I listed above.
Number of Black Test Takers At or Above 2SD (159.5): 335.455 (335-336)
At or Above 2.5SD (164): 87.51 (87-88)
At or above 3SD (168): 14.585 (14-15)
Before any conclusions are made, it is important to note a couple of major caveats.
Firstly, numbers vary by year. 14,584 blacks took the test in 2009-10. This number could shift a little year by year (it was 13,205 in 2007-08), as can the mean and standard deviation. Granted, these shifts won’t really make much of a difference, particularly at the high ends of the scoring spectrum (2SD and above) we’re focusing on, but they should be noted anyway.
Secondly, your GPA is obviously quite important, and could in theory enhance your competitiveness to an extent even greater than your score would predict (even, in theory, if that score is south of 160). The most recent data I could find on the mean UGPA for black applicants was from Fall 2008, when it was a 2.96. This figure had been increasing in years past, though that increase seems to have leveled off.
If you’ve got a GPA north of 3.0, you should help yourself out quite a bit here by narrowing your competition within your scoring range (and possibly allowing yourself to compete more evenly with higher scoring “splitters” who have better LSATs but lower GPA’s). When you look at the number of AA’s scoring in the ranges I’ve indicated above, you should expect that the number who manage to do so with an above average (3.0+) GPA is probably even smaller.
Finally, my analysis doesn’t account for folks who take the test and don’t go to law school, among other anomalies. The pool of test takers who are seriously going for law school may be smaller than I’ve assumed for a host of reasons, though there’s no way to find out exactly how much smaller. This is the best I could come up with.
Alright, so onto some conclusions:
1. A 159 should, in theory, put you approximately near the top 360 or so among black test takers (about 2 full Standard Deviations above the mean), and a 160 should put you more comfortably and confidently within that range. The T-14 will accept approximately 331 AA’s in a given year, so you’ll contend with these scores, maybe some money. If the “next three” (Vandy, UCLA, UT-Austin) are also on your list, you should be in an even better shot at getting an acceptance. Overall, this isn’t a bad place to be. If you’re an AA law school hopeful who wants a T-14, I’d recommend a 160 as the minimum standard to shoot for assuming an average (or better) GPA.
2. A 164 will put you 2.5 SDs above the mean, a level that only approximately 90 AA’s should reach annually (88 or so did this in 2009-10), assuming a normally distributed sample. You should be very well placed for a shot at the top 14 with a score like this. A glance at my enrollment numbers above will also show that the holy trinity at the top (HYS) accepted 81 black applicants most recently. This indicates that a 164 may (depending on your GPA) give you an outside shot at contending for HYS, and should certainly put you in T6 contention.
Going further, I do not have exact approximations for where scores in between 2 and 2.5SDs (161-163) would put you, but it could probably be safely inferred that there aren’t many more than 130-160 AA’s at or above 162/163. A 163 might make shooting for HYS less realistic barring an above average GPA, but it should give you a shot at a T6 and should certainly set you up for a T-14 acceptance somewhere (perhaps with some good money). A 165/166 should place you firmly in HYS contention, again depending on where the GPA is.
3. A 168 will put you 3SDs above the mean, a level that less than 20 AA test takers (around 1-1000, to be more precise-only 15 or so likely pulled this off in 2009-10) will reach annually in a normally distributed sample. HYS is probably a given here barring a very poor GPA, and one should probably expect plenty of money.
So assuming black LSAT takers average 142.04 and have a Standard Deviation of 8.74, you’re 3.43 SD above their mean.
Assuming a bell cure, only one in 2,968 black LSAT takers should score this high.
Assuming 14,585 black LSAT takers a year, you should be in the top 5 for the year.
Of course the bell curve is an abstraction seldom perfectly observed in reality.
Athlone McGinnis used to be a commenter on the RooshV forum I believe.
I had a very low GPA in both undergrad and in graduate school. I don’t want to say what it was but I was nearly placed on probation both in undergrad and grad school. I spent more time on the Internet and going to buffets than in class. I’m just lazy as hell for some reason. That was part of the reason I got tested for ADHD a few years ago (when I got my IQ tested) and got Adderall more recently.
Also you should look here:
http://www.jbhe.com/news_views/51_graduate_admissions_test.html
It says in 2004 10,350 blacks tooks the LSAT and 29 got above a 170. So slightly different from the top law school data.
spent more time on the Internet and going to buffets than in class.
LOL! So pill was right about all his jokes about you overeating.
I’d also previously taken the LSAT a few times with much lower scores…
I’d also previously taken the LSAT a few times with much lower scores
That’s where you made your mistake. You should have just taken practice tests until you were ready to score 172. Because schools like Harvard view every score and average them.
But you still provide an interesting case study in how coachable these tests are. How much did you improve by studying? Herrnstein and Murray claimed college admission tests (at least the SAT) are very resistant to test prep but I don’t buy it.
Yup!!!…I mean, no, I was just trying to use a self-deprecating inside joke. I’ve been hitting the gym more lately, and I might take some jui-jitsu so I can kick Pill’s ass someday…
It used to be true that most law schools averaged the scores but the increasing emphasis on US News rankings has encouraged law schools to only look at the highest scores. At least this is the case at most non-T14 schools.
From my initial practice test (with a real LSAT exam) to my last official test score was a 16 point increase.
” I spent more time on the Internet and going to buffets than in class. ”
Nothing to be ashamed off Gman, we all have our vices.
Wow, thanks for being oh, so understanding Pill.🙄
Another thing: law school admissions is rolling and I didnt submit most of my applications until late March/early April. So that hurt too.
I recall someone who tutored LSAT said that the hardest part to improve scores was in reading comp:
– Analytical reasoning/logic games was most teachable => teach some basic diagramming and scores go up quick
– Logical reasoning was harder to improve, but people who put the time in could start groking the “test logic” and improve
– Reading comprehension was painful to improve, even though most students thought it would be the easiest. Many complained about time constraints, but even with no time limit on practice tests, they still would score the same. There were no quick hacks and most advice boils down to eliminating answers, being “engaged” and quirky marking/note strategies. It is a working memory thing?
—
Having examined both LSAT/SAT, A perfect GRE prob correlates to a ~172 +- 5, not the 178-180 +- 5 claimed by ETS.
—
Second I scored a 337 170V/167Q on the GRE.
I did study/practice a lot, but I don’t know how *hollow* that score is considering practice effects, as I studied intensely for a a few months (I worried I was/am dumb and wanted to prove myself).
Based on the formula my IQ would be around ~152, or 3+SD above the white American mean. Obv this is only a single data point, so Im probably not in the “Highly Gifted” (145 to 159 IQ) range.
Another data point: IRL I havent been that successful, as my salary is only about 30k USD and I’m not too dynamic (although I can read situations and know how to adapt, I always hesitate due to fear/anxiety). This is a huge demerit against my score, but that is probably more a case of a neurotic personality/lack of confidence versus lack of intellectual ability.
@pp Based on my scores, should I not be afraid to enter a more advanced/intellectually demanding field for a better salary? Should I be confident/proud of my abilities or was my score just a fluke?
—
@GM BTW, 172 LSAT is boss. Wish you luck in your law studies.
There was some study about the effect of speed on LSAT performance. I can’t remember the results, but it wouldn’t surprise me if it showed that Reading Comp performance hardly improved with more time given. On the flipside logic game performance is highly contingent on amount of time. I used the Powerscore and Manhattan logic games books, and once I learned the diagramming methods I could get close to 100% if given ~50 minutes per section, but under the 35 minutes timeline I was still only getting 60-70% correct. That is, until I started practicing more…
My GRE lined up exactly with my final LSAT score.
…according to that calculator I posted.
@pp Based on my scores, should I not be afraid to enter a more advanced/intellectually demanding field for a better salary? Should I be confident/proud of my abilities or was my score just a fluke?
Anytime someone gets an extreme score, it’s likely a fluke to some degree, but that’s true of everyone with high scores, not you only. I would be proud of your score and confident in your abilities until you see evidence to the contrary.
It makes sense that verbal comprehension would be the most resistant to practice, because we’ve all been practicing it all our lives (including right now in this discussion), so more practice offers diminishing returns.
You should try Harvard, Yale, Stanford, Columbia or Berkeley.
Are you 98% or 99% ?
There were a study where they said only 15 blacks were above 170 out of 10 000. So you must be at 99.95% among blacks.
Try to find among the highest average LSAT schools the one that have few blacks. You would get a double competitive advantage . You should definitely go for a top 14 school. You could end up clerking in the Supreme Court in 5 years .
172 used to be the lowest limit for 99% but now I think it’s like 98.7% or something.
Pumpkin, Gond,
1) I remembered those figures and the number of candidates dropped from 150K to 100K , that’s why I went from 15K blacks to 10K blacks. Notice there is a rebound named Trump effect on LSAT test takers this year, greatest increase ever !
https://magoosh.com/lsat/2016/many-people-take-lsat-year/
2) I think I predicted exactly a 172 score for G last year. Gond if you remember or can find the post … 🙂
Yes, I think I remember that post (I think). Maybe that sub-consciously influenced my final score.😂
And remember JD with low average LSAT can’t increase their SAT with one person. So they won’t want you. And if their smartest kid is a black guy, that’s even more weird. You can only safely go for the top. Even with shitty GPA. That won’t matter. They will take you because you have a business story. Find someone good at making application story. You need a story.
Well, if I wanted to put together a better application for Harvard/Yale I would have to wait another year and apply next cycle, which I’m not sure I want to do. But that would be “highest IQ” thing to do. 😅
Book a prep school like those https://www.manhattanprep.com
telling them your a black with 172 LSAT and low gpa with professional experience in real estate in Florida that would pay for a killer app. That would be worth some 2or3k.
I did a mock test at Kaplan some years ago and the preparator who has my phone and was in this kind of prep proposed me plenty of money to do that job (allegedly). But I would hab’ve need a 800 GMAT. But I got a 790 at Kapan first try that are more difficult than real GMAT.
which contemporary american entrance exam has the highest ceiling?
the GMAT, LSAT, GRE, MCAT, a GRE subject test for some subject, SAT, ACT,…?
MCAT is the most difficult to ace. Then it’s GMAT. There are only 2 or 3 perfect scores a year and it’s often professional test takers (coach on preparation). Most Business school doens’t have perfect scores on GMAT. They pretend to refuse perfect scores but I believe it’s fake because I believe there must be some years without any perfect score candidates (if you take out the coach and the quant guy who do a master in quant finance instead of an MBA). That’s what I was told by this Indian who had a 790 and was trying to find a 800 for his prep.
https://poetsandquants.com/2018/02/28/highest-lowest-gmat-scores-at-the-top-25-schools/2/
I’m not sure if the average # of perfect GMAT scores per year is that low…but I do know way more people take the GMAT (the test can be taken almost any day of the year, unlike 5 times a year for the LSAT (which just changed recently from 4) and the test has a much more international scope–I think almost just as many people take it in India as they do in the US. 95% of LSAT takers are in the US). I don’t know of anyone personally who’s gotten a perfect GMAT, or even a 750, but I’ve met a couple 175+ LSAT takers.
OK, I know that anecdotal sample is kinda silly but even scouring the Internet I can’t find any GMAT perfect scores, but there’s a number of LSAT tutors with 180s.
What’s really odd is that the median GMAT score is more easily attainable than the median LSAT score.
It is because GMAT candidate are not as good as LSAT candidates academically and, by correlation, in their average IQ. Plus, there are more foreigners and non English native speakers in GMAT (lots of people with GMAY score 1 sd under the average).
Thus average LSAT for law school is higher than a average GMAT (the gap is reducing).
The test are the same up to 99.9 (178 and 780). But for GMAT, they made it very difficult to go 790 and 800. Like 99.98 and 99.999 instead of 99.92 and 99.96). Thus a perfect score in GMAT is 40 times rarer than for LSAT. If all the people taking GMAT took the LSAT, maybe it would be only 4 times more difficult only 99.99).
For MCAT, it’s the same but worse . Lots of foreigner take it to study medicine in the USA. They drag down average . And the ceiling is next to impossible to reach (at least in the 45 points format, but the 528 must be the same). It’s mlre scholastic than IQ because the pure scholastic part is only the verbal. But then it’s more purely g loaded because crystalllized intelligence for science is more g loaded than math lsat le gmat le sat who is a bit stupid )
Thanks Pumpkin. Great work!
At what IQ does the disparity between males begin to grow? I’m guessing this would change the expected IQ of kids.
I’m curious as to what would be the expected IQ of our(yours and mine) babies. 🙂
Here are my stats if you have time for family planning.
ACT 34
GRE 155/161 Math/Verbal
European, 5’7
Head size unknown (sorry!)
“I’m curious as to what would be the expected IQ of our(yours and mine) babies.”
There is the “breeder’s equation”, but if you Google “breeder’s equation IQ”, you weirdly only get blogs and not academic papers… So as far as I can tell, the “breeder’s equation” has not been “tested” in humans.
Of course, my stance is that IQ tests are really “middle-class tests” and your score is your knowledge of the middle class.
Stop to repeat the same thing… everyone here is hellly boring about your stupid and rigid nstances…
Don’t read or respond to my comments. It’s that simple.
Where is this 119 average coming from…. I’d guess the average GRE taker would score 110 on the wais and no higher.
Nvm I see it. 119 still seems pretty high, considering college grads are at around 110.
I was surprised by the high number too, but:
1) college grads probably have a higher IQ as measured by the GRE/SAT than by the WAIS because the former are a more direct measure of academic skills
2) people who take the GRE may be smarter than most college grads because they are typically graduate school material
It would be interesting if you made a post ranking college/graduate admission test (GRE, GMAT, SAT, ACT, MCAT etc.) by g-loading by correlation with the WAIS or by the type of questions that are asked on those tests (emphasis working memory vs pattern recognition etc.)
What’s odd is that when I did the mock test at Kaplan there were another guy who scores 720 in first try. Indian said it was really rare but was obsessed with my 790. And my English was poorer than the other guy English.
But the Indian informed me he got a 800, went to Stanford business school (this year Stanford had a 800!), and now he is a Partner at Goldman Sachs.
His parents are very elite (like court of appel judge for father and Supreme Court for’mother). Parents look eggheads. they come from aristocratic background , Russian orthodox for the father and tradionalist catholic (like Mel Gibson or me ) for the mother .
This is a bit late but may be of some interest. Based on a factor analysis I performed on the GRE’s correlation matrix, I obtained that the verbal score and the quant score have correlations of .78 and .85 respectively with a common latent factor (f). I then obtained the multiple regression formula for f = .56V + .67Q (all variables standardized). If we treat the common factor f to be representative of g, which is reasonable upon some thought and conservatively assume an average IQ of 110 and SD of 15 among the current GRE’s test-taking population then a revised formula for
IQ = [.56*(V-150.8)/8.4 + .67*(Q-152)/8.8]*15 + 110 (V,Q and IQ no longer standardized).
Additional (somewhat) justification for treating f as the g factor. A factor analysis I conducted on the old 1995 GRE which included V,Q and IQ had f correlated with FIQ by .93, which is about the reliability of retakes of the WAIS.
The g loading of college admission tests depends a lot on who is taking them. Among academically homogenous samples the g loading might exceed 0.9, but if all American adults took it, the g loading might be as low as 0.6
Good point about correlations between variables depending on the reference population. The population for the old GRE factor analysis was 30 graduate students, so it should be fairly representative of the population of test takers. The reason I brought it up is that although it’s reasonable to claim f = g for the analysis done on the newer test, the claim is vulnerable to questioning. But without the additional support for f = g, I think that if we assume that on the whole test takers are equally prepared for the test then the primary cause for common variance of V and Q must be g.
My LSAT score was 172 and my GRE score was 333 (169 V, 164 Q). Based on my GRE score, the ETS converter predicts a 173 LSAT. I studied very hard for the LSAT while studying comparatively little for the GRE. No verbal study, fairly modest quant study. The correlation is probably less than the converter predicts. Speculating, I think my GRE score should predict a 171. The GRE is still a decent measure of IQ, but the LSAT is king among admissions tests on this metric.
Update: 169 V, 167 Q ‘=’ 174 LSAT. Still very modest study. The ETS GRE/LSAT sample could have prepared less for the GRE, I suppose.
Also, interesting that holding my Q score constant, bumping 1 pt. to a 170 V bumps the LSAT 2 pts. to 176; holding V constant, bumping Q 1 pt. to 168 bumps the LSAT 2 pts. to 176. Bumping Q to 169 causes no change, whereas bumping Q to 170 bumps the LSAT another 2 pts. to 178.
Hey pumpkinperson, another question, while looking at the conversion equations for the gmat and the gre I noticed a big discrepancy, because following your conclusions, I have according to the gre an iq of ~131 (160 verbal, 157 math) but when i decided to convert my gre scores to gmat scores on the ETS site I came across 610 which according to you corresponds to an iq of 117,a staggering 1 SD of difference.
I never did the gmat but I was absolutely shocked when i saw the conversion, what’s up with that? Why such a big difference? On verbal+math iq test I always have above 125,in iq tests like the raven’s apm set 2 I had 133, my spatial is about 115-120 can you elucidate me, thanks.
Interesting. Obviously someone made an error but i’m not sure if it’s you, me or ETS
I hope i don’t come off as hubristic but it’s probably the ets, because i did the same experiment with the lsat and then went with as much celerity as i possibly could to your post about converting the lsat into iq and the result was very similar to the gre about ~128.
Regards from Portugal.