IQ, Creativity and the Twisted Pear,
Why the Sidekick Gets theGirl

Grady M. Towers

Reprinted from Lucid Vol. VIII, No. 4/5 (#42/43), Aug./Oct. 1988.
Lucid is the newsletter of the Mensa “Truth SIG.” This issue was edited by
Dale Adams. Some of Grady’s essays were printed in more than one place;
if this has been published elsewhere, a further citation will appear in Noesis.

It has been repeated again and again until it has passed into psychometric folklore that above a threshold IQ of about 120, there is no relationship between measured intelligence and creative accomplishment. This bit of common wisdom, like many other myths, is a misinterpretation of the facts and is only half true. The facts are these: that the correlation between IQ and creativity is a twisted pear correlation, and that in a classic twisted pear correlation there is no systematic relationship between individual measurements on one variable and individual measurements on a second variable above a given threshold. There is, however, a definite relationship between measured intelligence and creative accomplishment for groups of people. There is a definite, empirically-observable, optimum IQ for creative accomplishment in intellectually demanding fields, that lies well below the maximum IQ, but also well above the 120 IQ limit.

Anyone familiar with elementary statistics has been introduced to the Pearson product moment correlation coefficient. In most cases he will also have been introduced to the scatter diagram. This is where score pairs are plotted on a graph. He will usually have been taught that score pairs almost always form an ellipse, and that if the ellipse is narrow, the correlation is high, but if the ellipse is circular, the correlation is zero: i.e., there is no systematic relationship between variables.

pear1.gif (2462 bytes)

pear2.gif (2117 bytes)

Moderate Positive Correlation

No Correlation

There is, however, a compromise between these two extremes. It is found when there is a systematic relationship between two variables below a threshold, and no systematic relationship is found above the threshold. Its scatter diagram is a union of an ellipse with a circle. It looks like this:

pear3.gif (5080 bytes)

It should now be obvious how the twisted pear correlation got its name; its scatter diagram looks exactly like a twisted pear. If we let the horizontal dimension on the graph represent IQ, and the vertical dimension represent creativity, then point A represents the IQ/creativity threshold. Below this 120 IQ level, measured intelligence and creativity have a positive correlation. Above point A there is no systematic relationship between IQ measurements and creativity for individuals.

Point B on the graph represents the very highest level of measured intelligence. This is the group with the very highest IQs. As the reader can see, they tend to be well above average in creativity, but also well below the highest grades of real-world creative accomplishment.

Point C on the graph represents the group with the very highest level of creative achievement in intellectually demanding fields. Although this level of accomplishment is found well below the very highest IQ level, it is still very much higher than the 120 IQ threshold. We now know what that optimum IQ is.

The answer to this question has been provided for us by Doctor Anne Roe, whose study in the 1950’s of 64 of America’s most eminent scientists and scholars remains perhaps the most important study of creativity ever made. Being the wife of Doctor Gaylord Simpson, the widely respected paleontologist and evolutionary biologist, gave her access to the kind of proven creative performer that few researchers have ever had. At least one of her sample had won the Nobel prize, and, without doubt, all of them were world-class achievers in their various fields.

Doctor Roe made a very detailed study of each of these men from a wide variety of scientific and scholarly fields. Her sample included physicists, chemists, biochemists, geneticists, psychologists, anthropologists and others. She obtained detailed personal histories, examined family backgrounds, educational records, marriages and so on. She also examined their emotional lives through Rorschach tests and other methods. Included in her mental examination were three IQ tests: a verbal, a spatial, and a mathematical test. As the intelligence of these men went through the ceilings of ordinary IQ tests, Doctor Roe had the Educational Testing Service, the developers of the SAT, construct especially difficult tests for her subjects. The IQ equivalent earned by her subjects on the verbal test—the test which is probably the closest equivalent to a conventional IQ test—was a median of 166, a score that is about the same as that earned by the average member of the Prometheus Society.1

Creativity research has produced some of the most defective studies in psychometric literature. Tests were constructed that purported to measure this elusive quality, but turned out to have no correlation with real-world achievement in any field. Subjects were often school children, sometimes elementary-school children, as though the “creativity” of a ten-year-old could be compared to that of a Newton or a Goethe. Claims about the relationship between IQ and creativity were often based on mixtures of tests that measured different functions, had inadequate ceilings, and were of uneven reliability, rather than being based on the results from one good test. But the biggest flaw in these research designs was that the creativity being studied often wasn’t creativity at all, in any meaningful sense. Finding a hundred uses for a brick is in no way comparable to discovering a new scientific principle, or inventing a new experiment. Doctor Roe’s research avoided all of those pitfalls. There is absolutely no question about her subjects’ intelligence, or their creativity. And although their verbal IQs ranged from 121 to 177, only six of her 64 subjects scored below 148. There is simply no question that creativity at the very highest levels, in the most intellectually demanding fields, is heavily dependent on the same kinds of abilities as those sampled by verbal IQ tests.

The twisted pear correlation also has something to teach us about Prometheus Society members. It tells us that almost all of them show some degree of creativity. It tells us that the average member of the society is about as creative as those with the very highest IQs—those with IQs of 180, or 190, or 200. But above all, the twisted pear tells us that the greatest real-world achievements will come from people very much like ourselves—those with IQs below the very highest levels. It tells us why the sidekick gets the girl.


1. Anne Roe, The Making of a Scientist (New York: Greenwood, 1953), p. 164.