On each major subscale of the SAT, scores range from 200 to 800. Most sources converting SAT scores to IQ are based on one’s combined score on the reading and math subscales and thus range from 400 to 1600. But current forms of the SAT actually have three subscales (reading + math + writing) so it’s useful to know how the combined score on all three subscales converts to IQ, since many sources only provide total scores from all three subscales. These ranges from 600 to 2400.
In 2013, the average college bound high school senior had an SAT score of 1499 out of 2400 (reading + math + writing) and a score of 1008 out of 1600 (reading + math). An equation I created shows that since 1995, a 1008 out of 1600 on the SAT is somewhat equivalent to an IQ of 105; thus a score of 1499 out of 2400 is also roughly equal to IQ 105. Now of the 1,660,047 people who took the SAT in 2013, only 494 scored a perfect 2400. That’s one out of every 3,360 people. However only about 35% of American 17 year olds take the SAT, and scholars Charles Murray and Ron Hoeflin have argued that the higher the ability, the greater the odds of taking the test. Thus among 17 year-old Americans capable of scoring 2400 on the SAT, virtually 100% actually do so. Thus one out of 3,360 becomes one out of 9,600, because virtually none of the people who didn’t take the test were capable of scoring 2400.
If the IQs of Americans are forced to fit a bell curve with a mean of 100 and a standard deviation of 15, a score at the one in 9,600 level is assigned an IQ of about 156. Thus we have two data points:
SAT 2400 out of 2400 = IQ 156
SAT 1499 out of 2400 = IQ 105
From here we can create the following linear equation converting SAT scores to IQ:
IQ = 0.0566(SAT score) + 20.15094
Of course it should be noted that these IQs are on a scale where the American mean is set at 100 and the SD set at 15. Increasingly in the technical literature, IQs are scored relative to the white distribution, not the American distribution.
To convert from American IQ scores to white IQ scores apply this formula:
White IQ = [(American IQ – 103)/14.5](15) + 100
The advantage of using white norms is they provide a more stable reference for anchoring IQ that is much less influenced by demographic changes.
[This article was slightly revised on January 3, 2019]
professor shoe is getting uppity again.
now he’s posted a paper correlating brain images with IQ…for 164 subjects.
let’s take the same formula and apply it to subjects in another country.
what would be found?
that the previous results are bollocks.
What struck me was how high the correlation was…0.7…about the same as two IQ tests correlate with one another.
I think the formula would likely work in other countries, you’d probably just have to add or subtract a constant.
Jensen was struck by the fact that the correlation between brain-size and intelligence test scores was almost identical in mice as it is in people (0.48 vs 0.4)…so cross-cultural replication is nothing.
and that there are, if there are, physical brain correlates of IQ says nothing about the effect of genes on IQ even if such correlates are highly heritable.
there’s little more damage one can do to his brain than WK, yet all of the evidence is that WK has no effect on IQ.
Shouldn’t you factor in that the SAT only correlates at .8 or so with IQ?
No, because different IQ tests only correlate 0.8 with one another
Is the correlation that low? Because most full-scale IQ tests have correlations with g above .95. I’m pretty sure Wechsler correlates highly with other full-battery IQ test like Woodcock-Johnson, more so than the SAT does.
I think my next post will be devoted to this…
Yea good topic
” Increasingly in the technical literature, IQs are scored relative to the white distribution, not the American distribution.” – do you have a link to that?
Also interesting because there was that recent kerfuffle in your neck of the woods about how Canadian IQ’s suddenly rose when the used the US, rather than the Canadian standard. To me it seemed obvious that this was because Canada only has about 10% immigrants with sub-100 average IQ, whereas in the USA it’s more like 30%, who drag down the average, making it much easier to get 100. But you say that the US norms (“increasingly”) are based on the white population in the US, so maybe that’s not the case.
I think the idea originated with Richard Lynn (he discusses it in his 2006 book) but it can be found everywhere in articles about international IQ comparisons. It also seems to be known as the British scale or Greenwich scale:
The Canadian standardization sample of the WAIS IV obtained a Full Scale IQ 104.5 in relation to 100 for the American standardization sample, giving Canada a British (Greenwich) IQ of 102.5.
http://openpsych.net/ODP/2014/08/the-canadian-iq-calculated-from-the-standardization-of-the-wais-iv/
what is the IQ of Chinese Americans
East Asian Americans score about 5 points higher than white Americans, I believe.
Pingback: WE ARE ALL WITH STUPID… | Have Coffee Will Write
Could you convert my SAT in middle school to an IQ score? 1780/2400 at 14 years 1 month, when I was in 8th grade.
Well if you had taken it as a 17-year-old, 1780 would equate to IQ 123 (U.S. norms) by my formula. The question is how much of an age bonus you should get for taking it at 14.
Traditionally IQ was calculated as the ratio of mental age to chronological age, with adult mental age being set at 16, so a 14 year old who performs at adult level is 16/14= 1.14 times smarter, so your 123 would become 123(1.14) = 140!
But that’s a sloppy way of estimating your IQ. On the WISC-R IQ test, the closest subtest they have to the SAT is vocabulary and a 14.08 year old who scores equivalent to the mid 120s with reference to the oldest age group (16.9), scores 135 with respect to his own age. So 135 would be my best estimate for your IQ (as imperfectly measured by the SAT)
I know this is about a year late but I just came across this. I took the SAT in 2014 and had a composite of ~1920 (math was weakest). What would that convert to?
Pingback: Gossip Girl (2007 to 2012) | Yellowos
I wonder about factoring in test prep. Naturally, as a vain person, I want as high an IQ as possible. I scored 2190/2400 (about 14 years ago). I did not study for the SAT for even a minute (Note: I suspect in the ensuing 14 years that I have experienced normal cognitive decline). I am not asking for any hard numbers, but more for some indication as to how test prep affects score and thereby IQ measurements. If some average could be attained for test prep’s duration and for its effect on scoring, couldn’t a “no-test-prep” score of X be reasonably compared to a “test-prep” score of X?