In the past having a university degree was an impressive achievement. There are two reasons for this: 1) university educated people were less common, and 2) the correlation between IQ and years of education was much higher.
In the 1950s, the correlation between Wechsler global IQ and years of education among American adults was a potent 0.7 (roughly as high as the correlation between two different IQ tests) but by the late 1970s it had sunk to 0.57, where it remained through the 1990s and presumably today.
By 2006, roughly 26% of American adults, aged 25+ had a bachelor’s degree or more. That means that the median university graduate, is in the top 13% of education. If you’re in the top 13% of IQ, you’d have an IQ of 117 (U.S. norms), in other words, 17 points above the U.S. mean of 100. But since the correlation between IQ and education is only 0.57, the expected IQ of university grads would be 0.57(17) + 100 = 109.69.
How close does this prediction come to the actual data? According to a source provided to me by commenter C, the actual IQ of U.S. university grads (age 20-90) tested in the WAIS-IV 2006 norming was 110.77.
Very close to the predicted value.
In 2006, roughly 17% of American adults, aged 25+ lacked a high school diploma or equivalent. That means that the median high school dropout was in the bottom 8.5% of education. If you’re in the bottom 8.5% of IQ, you’d have an IQ of 80 (U.S. norms), or 20 points below the U.S. mean of 100. But since the correlation between IQ and education is only 0.57, we’d expect high school dropouts to have an average IQ of 0.57(-20) + 100 = 88.6.
The actual average IQ of Americans with only a 9th to 11th grade education (age 20-90) tested in the WAIS-IV 2006 norming was 88.77.
Again, very close to the predicted value.
Pumpkin Person 
PP are you able show what the average IQ of children or old people is compared to a mid-20s person?
The age based IQ sucks, I would rather know how smart an average 10 year old is compared to a 25 year old. Or how would a 150 IQ kid score if it were held to the same standard as an adult.
Here’s Paul Cooijmans’ take: http://paulcooijmans.com/intelligence/iq_development_with_age_modelled.html
I generally take Cooijmans as gospel truth but I dunno how Peepee feels.
There’s no simple answer to that question. It depends on the test. Some tests stop increasing with age around 16, others keep increasing with age until middle age. Some tests decline sharply with old age. Other tests hold up.
This is excellent work PP, thank you. Particularly, thank you for the tie-in to the economic utility of a post-secondary education via IQ signalling. This can also be viewed in light of The Bell Curve‘s Chapter 2 treatment of IQ as a predictor of job success.
Particularly, success in high-IQ professions, which is more strongly dependent on IQ. The ceiling for marginal return on IQ increases as job complexity increases, i.e. as job complexity increases, the return on IQ above prerequisite gets higher.
So for instance, say the baseline to be a decent programmer is IQ 110 (with exceptions), and the baseline to be a decent theoretical physicist is IQ 150 (with exceptions, e.g. Feynman). A theoretical physicist benefits more (does more and better work) if he’s thirty-points above the baseline.
According to https://en.wikipedia.org/wiki/Educational_attainment_in_the_United_States#General_attainment_of_degrees.2Fdiplomas , 1.77% of American adults, aged 25+ had a Doctorate. That means that the median Doctor, is in the top 0.885% of education. If you’re in the top 0.885% of IQ, you’d have an IQ of 136 (U.S. norms), in other words, 36 points above the U.S. mean of 100. But since the correlation between IQ and education is only 0.57, the expected IQ of Doctors would be 0.57(36) + 100 = 120.52. That’s a few points below the commonly cited 125.
Did I get it right?
You did the calculations correctly, and the predicted value is indeed a few points below the often cited average of 125. However it could be that the often cited 125 is too high, and is perhaps based on crystallized tests (i.e. ASVAB) that perhaps over-correlate with schooling. I haven’t seen any data on how PhDs perform on any of the recent editions of the WAIS. They may indeed have dropped in IQ.
Pumpkin I recall you making a post where you cited research showing that the higher the IQ differential between two people, the greater the “intellectual dominance” of the higher IQ person, and the higher the percentage of arguments or debates between them that he or she won; and that above a 2S.D. difference, people have difficulty maintaining any meaningful intellectual relationship.
I am sure I got quite a few details wrong but would you happen to recall this post of yours? Cheers.
I don’t recall a single post which contained all those ideas, though it’s just common sense that high IQ people will tend to dominate lower IQ people in debates, or any other cognitive task.
The notion that beyond a 30 point IQ gap, communication breaks down is a very old idea, which I think originated with Leta Hollingworth
D.K. Simonton argued that the optimum IQ for a leader is about 18 points higher than his followers. A big enough gap to gain respect and dominate most of the competition, but not so big that folks can’t understand you.
I don’t endorse that theory, but it’s out there.
Perhaps you’re thinking of this?:
http://garthzietsman.blogspot.com/2012/03/chess-intelligence-and-winning.html
That’s the one!
Thanks a bunch (and to PP).
You’re very welcome. Impressive that black national black merit finalist ruhkukah found the precise link. Shows his incredible WAIS-IV General Knowledge score even predicts knowledge of the HBD-o-sphere! His brain is like a sponge. It just absorbs information in whatever environment he’s placed in.
No problem! I had it bookmarked.
If the average college-educated student is lower in intelligence than in previous decades, wouldn’t the same logic pertain to the average high school student that is college bound (i.e. a student that takes the SAT/ACT)?