Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Sure, I agree that in a general sense just pulling a number out of the air is crap.

Testing isn't wrong, it's just not as correlated to the things we would actually like to measure as we would like it to be. This is an open problem in the world, and solving it (besides potentially making the inventor fabulously wealthy) would dramatically change the world.

How do you decide if someone is qualified for a job, or will be successful at a university? In reality we don't have a very solid way to figure that out, so lots of people that would have done well are passed over and lots of people get into universities and hired for jobs and end up failing.

Testing is correlated to lots of things, and is quite nicely correlated to the quality of the training a student was given. I'm sure there is a better solution out there in the aether, but mankind hasn't discovered it yet.



Actually, we do have a good way to determine how someone will perform. The US military has used the ASVAB test for decades, and studies have shown that performance correlates with test results.

http://official-asvab.com/validity_res.htm


Took the pre-asvab in high school back in the 90s, got a perfect score. The recruiters called a lot. Took the full ASVAB at a recruiting station, still got a very high score. Too easy, how can that be a real test? My ACT/SAT scores were good but not perfect.




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

Search: