Commenter “Tenn” asked me to estimate the IQ of several people, most notably Ronald Reagan, and what better day to do so than President’s Day?
In 2005, Reagan was elected the Greatest American of all time. He was unbelievably influential because not only was he President for eight years, but his Vice President became President for four years, and his Vice President’s son became President for another eight years. The Democratic party was so shell-shocked by Reagan’s success, that they moved economically to the right during the Clinton administration.
Reagan was influential, not only in changing the trajectory of American politics for decades, but was instrumental in making America the World’s sole super power.
Historians can debate whether his influence was positive or negative, but anyone who changed the World as much as Reagan did, would likely have an IQ way above 100.
On the other hand, as far as I know, Reagan was the only U.S. President to have been mentally impaired during old age. Thus by definition, Reagan’s IQ during old age, was in the bottom 2.27% of American Presidents. This might be an underestimate because some U.S. Presidents have not (yet) lived to old age, on the other hand it might be an overestimate because Reagan might still have been the only one to become impaired, even if there had been 88 Presidents. But assuming it’s roughly correct, it implies that if all 44 U.S. Presidents took an IQ test when they were very old, Reagan would have likely scored two standard deviations below the average U.S. President.
Of course, the relevant question is not how Reagan would have scored when he had Alzheimer’s, but how he would have scored at his peak.
In one of the most fascinating studies in the history of psychometrics, Ian Deary and a team of other scholars, tracked down 101 people who took an IQ test at age 11, and tested them again at age 77, on the exact same test!
They found a 0.63 correlation between IQ measured at 11 (by age 10, IQ more or less stabilizes within measurement error) and IQ tested 66 years later. Because the sample was somewhat restricted (the standard deviation was only 77% of the national SD), the 0.63 correlation underestimates the relationship in a representative sample, however since the IQ variance of U.S. Presidents is similarly restricted (they seem to have a mean IQ of about 130 with an SD of 12, compared to the national white average of about 100 with an SD of about 15), I will use the 0.63 correlation.
So if Reagan’s elderly IQ was two standard deviations below the elderly IQ of the average U.S. President, his peak IQ was likely 2 SD(0.63) = 1.26 SD below the peak IQ of the average U.S. President. As mentioned above, U.S. Presidents have peak IQs (U.S. white norms) of about 130 with a restricted SD of 12.
Thus:
Reagan’s likely peak IQ = -1.26(12) + 130
Reagan’s likely peak IQ = 115
The standard error of the estimate is 9.32, so you might say with 95% confidence that Reagan’s peak IQ was anywhere from 96 to 134.
Conservatives might argue he had a towering IQ of 134, since he was a supremely important President who wrote eloquent love letters to wife Nancy, but arrogant liberal professors might argue he was closer to 96 since in their eyes, he did enormous damage to the most vulnerable in society.
9.32 points might seem like a large standard error, but it’s actually smaller than the standard error you would get if you tried to predict someone’s Wechsler IQ from their SAT scores, or if you tried to predict someone’s Raven IQ from their Stanford Binet IQ. Psychometrics is an inexact science.
It would be interesting to know the SD of this blogs readership. My estimate of a SD of 19 is only a guess. I do not know how many people read this blog but the mean IQ is 129. If restricted samples lower the SD compared to the total population (white mean 100) my guess of 19 is way off.
The SD is probably quite large here.
I have a question for chartreuse… It’s the “philosophy paradox”. Philosophy majors have some of the highest test scores, yet so much of contemporary seems, erm, stupid, to say the least. I didn’t study much contemporary philosophy at school, but what little I did study seemed pretentious and tedious. Especially the ethical philosophy. Is this because is Spearman’s law? If so, then why was philosophy so much better a hundred years ago?
Actually, more like 150 years ago
Why do offspring of very bright people regress to the mean? I remember one post where you estimated the expected range of the iq of the offspring of Langan and his wife’s, the upper limit was lower than Langan’s. Yes the wife’s iq was lower but it seemed more than I would have thought. Also, doesn’t this go against the flynn effect? Because to me at least, it seems like the bright half of the population converges to the mean over a long time and so does the dull half (roughly), so do you think the flynn effect supports the notion that the dull half just converges faster (due to nutrition, whatever). I haven’t really considered the effect of two bright parents however.
Do you think that looking at the entire US population, the amount of “underachievers” with a high achieving parent are more likely to just underachieve because of lower iq or a more comfortable upbringing that leads to complacency?
Why do offspring of very bright people regress to the mean?
Because the parent off-spring correlation is imperfect. By definition, an imperfect correlation means the slope of the standardized line of best fit predicting Y from X is less than 1, so the average value of Y for a given X, will be less extreme than X.
And it works both ways. The parents of extremely bright people also regress to the mean, as do the parents and offspring of extremely dull people.
Perfect. Thanks for the definition.
Wow I was going to protest on the grounds that the correlation between younger and old age IQ would be reduced by the onset of alzheimer’s, but the correlation between childhood verbal IQ and risk of developing alzheimer’s is amazingly strong:
http://www1.cs.columbia.edu/~julia/papers/snodonetal06.pdf
Yes I’ve posted about that study., There is a high negative correlation between early life idea density and Alzheimer’s in old age, although the correlation between idea density and conventional measures of intelligence might not be that high, since the nun study found virtually no correlation between the grades of the nuns and their IDs measured contemporaneously. Strangely, ID was a better measure of their cognitive function 60 years later.
I like the fact that grammatical complexity, i.e. complexity of sentences is not as important as idea density. There is a trend in certain academic circles to expend an enormous amount of words to say very little, which I think is connected to a fear of expressing things clearly enough that they may be falsifiable. Here’s an amusing essay on that topic: http://people.ucsc.edu/~nuclear/econ184/hotnews/dancingwithprofessors.htm
You should estimate Ralph Ellison’s IQ, perhaps even using some of those methods if you have a handle on them.
I’ll add him to the waiting list. I did estimate Oprah’s IQ using idea density
https://brainsize.wordpress.com/2014/07/20/oprahs-idea-density/
Nice job. Here is a possible material source, though it might be too small: http://mentalfloss.com/article/30937/famous-novelists-symbolism-their-work-and-whether-it-was-intentional
Also I asked after being impressed with Ellison’s answers so there is probably selection bias. No shortage of writing from writers anyway.
Apparently, he was tested with an IQ of 105.
Never heard of him taking an IQ test