If intelligence is just defined as the g factor then it isn’t clear what grounds we have to dispute test results. Instead, what cases like Feynman’s suggest is that while a single variable account of intelligence is powerful predictor academic and job performance, income, etc society-wide, the cognitive abilities of individuals can be considerably disparate. This is why modern tests have multiple indexes. I took WAIS III when I was 18 and they test runners refused to actually give me my full IQ because the verbal IQ and the performance IQ were 30 points apart. (not that it wasn’t trivial to just average them if I wanted to).
I know I am several years late to this party, but I felt it appropriate to hop in.
Einstein, with an IQ of 160+, really has an unmeasurably high IQ—at that IQ level, you are outside of the bounds of statistical probability used to construct the test. 160 correlates to such a rare IQ that you cannot use IQ tests to measure differences between them in a very predictive fashion because there aren’t enough people with IQs that high.
Now, what about the case of Feynman?
Well, there are tons of possibilities here:
1) It was a bad IQ test.
2) He got extraordinarily unlucky (Remember, your variability on a modern test is +-5 points; if the older tests had higher margins of error, it could be he was significantly more intelligent than this)
3) IQ is not a perfect indicator of g. This is actually known. It strongly correlates with g, but it is not identical to g—it is entirely possible that he was smarter than the IQ test indicates because of this discrepency.
4) He did badly on the test for some external reason (he was tired, the test didn’t get graded properly, he got the wrong person’s score back… any number of possibilities that could theoretically lower his IQ).
5) He really DID have an IQ of 125 in high school, but via concerted effort increased his intelligence greatly over time. In other words, he may have had significant untapped potential. Did he take the IQ test before or after he went on his crazy math-learning spree? This is especially true given he was still in adolescence.
6) He may really have only had an IQ of about 125, maybe as high as the mid 130s, and simply made better use of it than most people.
If intelligence is just defined as the g factor then it isn’t clear what grounds we have to dispute test results. Instead, what cases like Feynman’s suggest is that while a single variable account of intelligence is powerful predictor academic and job performance, income, etc society-wide, the cognitive abilities of individuals can be considerably disparate. This is why modern tests have multiple indexes. I took WAIS III when I was 18 and they test runners refused to actually give me my full IQ because the verbal IQ and the performance IQ were 30 points apart. (not that it wasn’t trivial to just average them if I wanted to).
I know I am several years late to this party, but I felt it appropriate to hop in.
Einstein, with an IQ of 160+, really has an unmeasurably high IQ—at that IQ level, you are outside of the bounds of statistical probability used to construct the test. 160 correlates to such a rare IQ that you cannot use IQ tests to measure differences between them in a very predictive fashion because there aren’t enough people with IQs that high.
Now, what about the case of Feynman?
Well, there are tons of possibilities here:
1) It was a bad IQ test.
2) He got extraordinarily unlucky (Remember, your variability on a modern test is +-5 points; if the older tests had higher margins of error, it could be he was significantly more intelligent than this)
3) IQ is not a perfect indicator of g. This is actually known. It strongly correlates with g, but it is not identical to g—it is entirely possible that he was smarter than the IQ test indicates because of this discrepency.
4) He did badly on the test for some external reason (he was tired, the test didn’t get graded properly, he got the wrong person’s score back… any number of possibilities that could theoretically lower his IQ).
5) He really DID have an IQ of 125 in high school, but via concerted effort increased his intelligence greatly over time. In other words, he may have had significant untapped potential. Did he take the IQ test before or after he went on his crazy math-learning spree? This is especially true given he was still in adolescence.
6) He may really have only had an IQ of about 125, maybe as high as the mid 130s, and simply made better use of it than most people.
Some of which can be verified if you trace this widely repeated anecdote back to the source; see my old comment on the topic: http://news.ycombinator.com/item?id=1251164