Geoffrey MillerVerified account

My challenge to folks who are skeptical about IQ tests such as the Stanford-Binet or WAIS-R: Develop your own measure of general cognitive ability and show that it has higher reliability & validity in predicting real-world success. I'll wait.

The lack of an alternative measure doesn’t make a flawed measure any more valid.

Its imperfect, just like knowing the temperature outside alone doesnt tell you how to dress, but it isn't meaningless.

For the vast majority of uses, temperature satisfies monotonicity and transitivity, necessary of any coherent measure. Outside the left tail, IQ fails to satisfy these for predicting nonacademic performance. https://medium.com/incerto/iq-is-largely-a-pseudoscientific-swindle-f131c101ba39 …

You only care because of what it implies. Most IQ discussions are about group averages. Think of playing chess against 100 people.. 50 from group A have an avg of 85, 50 from group B have an avg of 115. Who are you gonna do better against?

While there are convexity effects in chess, they’re light. Still too much structure. Not comparable to real world. This is ludic fallacy.