re: #731 Obdicut

Well shit, if all you’re saying is tests are somewhat predictive, sure. But there can easily be people who do badly on tests and perform very well in the real world.

Your original claim was ” The skills needed to answer a good test item are the exact same skills needed to use whatever is being tested in a real-world setting.”, which is amazingly hyperbolic and obviously untrue.

Hell, I mean, five seconds of thought can tell you this. Unless you just take this to the ridiculous level of saying that for some things a real world usage of the skill is the form the test would make, which would be hilarious.

A good test does its best, within the constraints of time and budget, to match the real world usage. A test of chess skill, for instance, could take the form of having the student play a computer a couple of games, with the computer level adjusted as one went along. A test of musical mastery? Play this piece.

A test of mathematical knowledge? Work this problem.

A test of literacy skills? All the essays we see at LGF are quick, impromptu demonstrations of literacy. Almost without exception, at a high level, for that matter.