3 ms·
That simply isn’t true, eg see https://www.washingtonpost.com/news/answer-sheet/wp/2017/05/09/can-coaching-truly-boost-sat-scores-for-years-the-college-board-sa
by seanmcdirmid 7y ago
That simply isn’t true, eg see https://www.washingtonpost.com/news/answer-sheet/wp/2017/05/09/can-coaching-truly-boost-sat-scores-for-years-the-college-board-said-no-now-it-says-yes/ https://www.washingtonpost.com/news/answer-sheet/wp/2017/05/...
The College Board can’t really have it both ways. And of course: https://www.reuters.com/investigates/special-report/college-sat-two/ https://www.reuters.com/investigates/special-report/college-...
Which is just one example of pulling out all the stops.
- thaumasiotes 7y agoYes, no one disputes that memorizing an answer key can give you a higher score. But by the same token, no one believes that the influence of cheating is of interest when studying potential gains from "test prep".
- seanmcdirmid 7y agoThat isn’t just it though: they have access to old tests, they have access to test takers who memorize questions, given the importance and popularity of any test, these options will arise and people with means will have access to them. There is no avoiding Goodhart’s law.
- thaumasiotes 7y agoYou can make a case for a much weaker form of Goodhart's Law, that with the SAT as a well-known target, SAT scores are less informative than they would be otherwise. But "there is no avoiding Goodhart’s law" is obviously wrong if you interpret Goodhart's Law as it's actually phrased; the SAT is doing a fine job of avoiding it today. The SAT hasn't ceased to be a good measure. It's still very informative.