"When I was six years old, I was given the Stanford Binet test and obtained a score that was equivalent to that attained by a typical 14-year old. Based on this, an IQ of just over 220 could be inferred by simple division, but the test is extremely noisy at these scales, and with realistic error bars a more accurate estimate would be “greater than 175”. This was documented in the book “Exceptionally gifted children” by Miraca Gross, where I was given the pseudonym “Adrian Seng”; a relevant excerpt can be found here.
But there is no reason to expect that this ratio would continue in my later years; I did take some other cognitive tests at age eight (see the article of Clements linked to above), but have not taken such tests since."
He's definitely math tilted based on that - difficult to make an inference with these distributions as there isn't a good sample of comparable children who took the SATV.
Math tests may have a g-loading of 0.77 in the general population, but isn't there some extra regression towards the mean for someone's whose strength is math? If you judge everyone on what they do best on everyone will over-perform their IQs.
"When I was six years old, I was given the Stanford Binet test and obtained a score that was equivalent to that attained by a typical 14-year old. Based on this, an IQ of just over 220 could be inferred by simple division, but the test is extremely noisy at these scales, and with realistic error bars a more accurate estimate would be “greater than 175”. This was documented in the book “Exceptionally gifted children” by Miraca Gross, where I was given the pseudonym “Adrian Seng”; a relevant excerpt can be found here.
But there is no reason to expect that this ratio would continue in my later years; I did take some other cognitive tests at age eight (see the article of Clements linked to above), but have not taken such tests since."
From his blog: https://terrytao.wordpress.com/career-advice/advice-on-gifted-education/
He's definitely math tilted based on that - difficult to make an inference with these distributions as there isn't a good sample of comparable children who took the SATV.
Math tests may have a g-loading of 0.77 in the general population, but isn't there some extra regression towards the mean for someone's whose strength is math? If you judge everyone on what they do best on everyone will over-perform their IQs.
> isn't there some extra regression towards the mean for someone's whose strength is math?
Probably, but it's hard to judge exactly how much regression to the mean there is.