HiQ

Tyler Cowen points us to Sam Knight of the Financial Times writing on high IQs. The featured player is Marilyn vos Savant, owner of the Guinness World Records highest IQ, now and forever, of 228.

Savant – the surname is real, it was her mother’s maiden name – has had a unique claim to fame since the mid-1980s. It was then, almost 30 years after she took a test as a schoolgirl in downtown St Louis, Missouri, that her IQ came to light. In 1985, Guinness World Records accepted that she had answered every question correctly on an adult Stanford-Binet IQ test at the age of just 10, a result that gave her a corresponding mental age of 22 years and 11 months, and an unearthly IQ of 228.

I say now and forever because Guinness has dropped that category.

If we were to try to interpret that IQ in the conventional fashion for adults, at the usual rate of 15 points per standard deviation, that number comes to just about eight and one half standard deviations and a probability of less than 10^(-17) or one in 100 quadrillion. That's not plausible, of course, but as a ten year old she supposedly did as well as the average 23 year old on an IQ test.

In any case, she was obviously a smart kid and has become a very clever and knowlegible adult. The FT article is a winner in my book, touching a number of points that fascinate me: the high IQ societies, the superstitious awe in which society holds IQ, and a few hints of the relationship to achievement (modest but real) - not to mention the good old Monty Hall problem, probably Marilyn's most famous controversy, but one that she did get right.

So what do IQ tests test? One of the beauties of the IQ theory is that once you have convinced yourself that a bunch of skills are correlated, testing almost any of them seems equally plausible. Because correlations are imperfect, most IQ tests present a grab bag, but three key components are pattern recognition, speed and complexity of mental processing.

The Wechsler Adult Inteligence Test is one of the most popular ones. They test four general areas, each of which has sub areas:

Verbal Comprehension Index (VCI)
Perceptual Reasoning Index (PRI)
Working Memory Index (WMI)
Processing Speed Index (PSI)

The well-substantiated claim for IQ tests is that they predicts academic performance. Looking at the individual skills certainly makes that claim seem plausible to me. It's also claimed that it predicts performance in all sorts of other areas. There is little obvious correlation of very high IQ to exceptional intellectual accomplishment, however. Geniuses are smart, but might have IQs only a couple of standard deviations above the norm. Many of the so-called super IQs are extremely modest in accomplishment. vos Savan, for example, was distinctly ordinary until she got the idea of claiming the world's highest IQ - two failed marriages (the first at age 16) and employment in the family business A very high IQ friend once explained to me the reason why he left academia - he found that the thing he was actually good at was taking tests.

A claim is made in the article for what I would call the Sheldon Cooper phenomenon - that very high IQ individuals tend to be socially incompetent and often suffer from Asberger's syndrome.

Comments

Popular posts from this blog

Anti-Libertarian: re-post

Uneasy Lies The Head

We Call it Soccer