>>13409146Let me explain to you why the Mega Test isn't a good IQ test.
IQ exists as a concept to measure "g" or general intelligence. If IQ ceases to measure g under certain circumstances, than IQ is useless under those circumstances; it's equivalent to just attaching a number to a test and calling it your smartness level.
Now, IQ is normally distributed. This is true because the designers of IQ tests made it so. When a bunch of people take an IQ test, the computer ranks all their raw scores and inevitably assigns them IQ's such that the distribution of IQ scores is normal.
But g itself (which actually determines how smart you are) is NOT normally distributed. In the middle, and about 4 standard deviations out either way, it very closely resembles a normal distribution (because of the Central Limit Theorem) so you can get away with treating it normally in most cases. But when you are specifically trying to measure people 5, 6, or more standard deviations out, the difference between the distribution of g and the distribution of IQ becomes relevant.
So you have the Mega Test, which takes people's scores on a test and, instead of trying to normally distribute them, tries to arrange them to correspond better with g, which they do by making high scores more common to be in line with high g being more common. What the Mega society calls "one in a million" is not nearly that rare. They use a normal distribution to calculate how rare a score is, but use a more "realistic" method to actually assign the scores. If you do the math there shouldn't be a single person alive with an IQ over 195, and that's without excluding the inaccessible geniuses who are currently rotting away in some poor country.
Basically the problem with the Mega test is that it vastly exaggerates scores in order to feed the egos of people taking it.
