Hacker News new | past | comments | ask | show | jobs | submit login

Removing the penalty for wrong answers seems to be an interesting choice to me. I thought a strong part of the test was the emphasis on carefully considering the questions so that the answers given were deliberate.

Without a penalty, guessing is just a chance for free points. Not really a measure of any skill aside from test taking (of which the SAT already suffers from enough). But I guess if all of your tests at a university are similar to the SAT, it might be a good measure of your potential success.




Penalties for wrong answers are the ultimate test of test-taking abilities (actually, to an extent, an ability to implement probabilities). Your strategy changes based on how many answers you can eliminate, rather than on if you know the correct answer.


Being able to eliminate wrong answers is pretty much the same skill as being able to notice that you made a mistake when you look over your work. It's valuable for more than test taking.


I disagree. Checking your work and realizing you forgot a semi colon or that you made an addition mistake is a very different skill than being able to eliminate an answer since you know it is incorrect. Further, it's an even more distinct skill to know when you have eliminated enough answers for your expected correct answer value >= your expected wrong answer value.

That's not saying that they are not valuable skills on their own, but at age 16, they are far more associated with test prep than what the SAT purports to measure.


No, if you're actually checking your work properly, you're going in with the assumption that there are mistakes to be found, and you're looking for reasons why your answer might be wrong - the same as when you're trying to eliminate a candidate answer in a multiple choice test (especially if "none of the above" is one of the choices). The gaming the odds aspect of the SAT only comes into play when you've run out of time or run out of knowledge to use to support or contradict possible answers.


Checking my work with the assumption I made mistakes is distinct from knowing that 2 of the 5 answers are wrong, and therefore my expected value of a guess is positive vs. when i cannot eliminate any. Saying that checking your work and knowing some answers are wrong but not which one is right are the same thing is simply not true.


Way to ignore the last sentence of my comment. I consider the guess to be separate from the process of elimination that precedes it and sometimes obviates it.


If you know all the answers though, you won't skip any or get any wrong. I didn't get a perfect score by guessing on the hard questions.


So no penalty for wrong answers increases the chances of an undeserved perfect score but decreases the test-taking skills advantage in understanding when to guess and when not to guess. The latter is much more common than the former.

EDIT: Said the penalty, meant no penalty




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: