SoylentNews is people
SoylentNews
from the maybe-they-only-surveyed-the-nimnobs dept.
Why can we talk about PISA results, comparing the performance of students in school, but we are not allowed to talk about differences in IQ? Bring this subject up, and you are immediately accused of racism. And yet. And yet, if there are substantial differences in intellectual capability, might this not explain some of the world's problems?
An update of a massive "study of studies" is underway; this article summarizes the work to date, and provides links to the work in progress. A quick summary of the answers to the questions no one dares ask:
- Eastern Asia (Japan, China): IQ around 105
- Europe/North America: IQ around 98
- Middle East: IQ around 85
- Africa: IQ around 70
In the first instance, it doesn't even matter why there are differences. They may be genetic, or disease related, or nutrition related, or something else. If these differences are real (and the evidence is pretty strong that they are), then we need to deal with them. Imagine if the low IQs in Africa turn out to be fixable - what would the impact be, if we could raise the IQ of an entire continent by 30 points?!
Sticking our collective heads in the sand, because the topic is not PC, is not going to solve any problems.
(Score: 2) by jcross on Sunday November 05 2017, @11:06PM
I think your numeric series example is excellent because it highlights the extent to which a test is unintentionally what Schelling calls a "coordination problem", in that when finding the "right" answer to the series, you have to model the thinking of the writer of the test and figure out which answer they'd be most likely intend. I suspect almost any such short series would look somewhat ambiguous to a number theorist, who might for example discard the simple answer in favor of a more interesting solution, and be graded wrong for it by the simpleminded test-maker. Anyway, any kind of coordination problem is going to work better the better the two parties understand each other, so I think tests will tend to lose their predictive ability when crossing cultural divides, even without any intentionally biased formulation. It might be possible to overcome this, but it would take a kind of care and deep thought that I doubt is generally spent on IQ tests, given that test writers seem satisfied if their test just correlates well with previous tests that are assumed to be good.