Averaging the two groups is not very meaningful. After all, it only takes a few smart dudes to get to an IPO or win a Nobel prize, although a critical mass of smart people often proves essential.
Averaging the two groups is not very meaningful. After all, it only takes a few smart dudes to get to an IPO or win a Nobel prize, although a critical mass of smart people often proves essential.
I am inclined to agree with you that this study is just as worthless as the ones comparing American students to foreign ones, but for different reasons. It's just about what is being tested -- kids may do worse on some measures, but on average, in the age of Google, everyone is a great deal smarter than 30 years ago.
In any case, we're talking about the exceptional students, so racial averages are irrelevant to the conversation.
Nonsense. Assume intellect is normally distributed, with equal sigma to simplify things (1). Assume a racial difference in mean of 0.5 sigma, then 2.2% of the higher intellect group will be 2sigma beyond the (high intellect) mean, while 0.6% of the lower intellect group will reach this level.
This difference grows exponentially when we consider more exceptional people (2).
(1) Between males and females, sigma is not equal. Men have higher sigma, as Larry Summers famously hypothesized.
(2) To see this, write out the CDF for a normal distribution and use the asymptotic expansion for erf. The rate of growth is (smarter group's mean - dumber group's mean)/sigma^2.
Statistics?
I see your point, but one must intellectually curious to make use of the vast pool of resources that is available. Google also allows mis-information to be more readily disseminated.