Monday, October 16, 2017

g - a Statistical Myth

A nice essay from 2007.

One of the best parts: "How to make 2766 independent abilities look like one g factor".

Comment (1)

Loading... Logging you in...
  • Logged in as
Stupid American's avatar

Stupid American · 377 weeks ago

I wonder if this bit of obfuscation got the author the tenure he was begging for? He seems to have
a couple of points, neither of which seems very relevant to me. (1)If you have consistent correlations among
measurements, factor analysis can extract a largest principal component and (2)One factor does not completely
describe performance on mental tests. The first is a tautology and the second has been known almost forever
(again, from factor analysis.)

Neither point does anything to cast doubt on the usefulness of IQ as a concept or as a technique for judging
potential capabilities. Yet another pointless screed on IQ.

Post a new comment

Comments by