3 ms·

> It has been possible for a century to self-study to the equivalent of a bachelor's degree using textbooks but it is very difficult to verify in a reasonable a

> It has been possible for a century to self-study to the equivalent of a bachelor's degree using textbooks but it is very difficult to verify in a reasonable amount of time whether someone has done this I challenge this. In fact I've wondered why people don't devise a test that capitalizes on this. Some edtech company should make a test - let's call it "the Genius test" - that's so hard it would challenge Terence Tao. I mean, why not? You have the whole world of possible test takers, anyone with access to a computer. If you make the incentives good enough, you can attract a million people to take your test. So this is your class population; now weed them out. Include International Math Olympiad questions on there. Partial differential equations. Algorithms. Quantum mechanics questions. Every hard question you can imagine should be on this test. Then make the reward something like a full ride to an Ivy, or a strong state school, or a group of them. You'll probably only have a hundred people with decent scores on it; and it will be so immediately obvious that they cheated if you call them up and ask them to explain a single answer, most fraudsters won't try. Seems like a good model to me. I'm surprise it hasn't been tried.

TeMPOraL 4y ago
Look at this from the perspective of the participants of your test: you're asking them to play in a lottery, whose odds are one to million, and the price of admission is years of long, hard work. No thanks, anyone capable of solving even the easiest of your tests will know they're better off playing the actual lottery, and spending the rest of the time on any kind of job or social activity. You'd have to offer something truly out of this world, like guaranteed UBI for the winner and their whole family, two generations back and two generation ahead, for it to be worth the risk. But then, with incentives that strong, you'll probably have only fraudsters getting close to the finish line. Unfortunately, the deck is already stacked against higher education here: as long as your long-term economic prospects hinge directly on your educational performance, the most motivated and effective crowd will be cheaters. If you weed out people based on performance, you'll end up selecting for cheaters, and against honest students.
MagnumOpus 4y ago
> I'm surprise it hasn't been tried. It has? You are describing every centralised college admissions exam, from the Chinese NCEE ("gaokao") to the Indian JEE-A or the French concours for admission to the better universities...

TeMPOraL 4y ago

Look at this from the perspective of the participants of your test: you're asking them to play in a lottery, whose odds are one to million, and the price of admission is years of long, hard work. No thanks, anyone capable of solving even the easiest of your tests will know they're better off playing the actual lottery, and spending the rest of the time on any kind of job or social activity. You'd have to offer something truly out of this world, like guaranteed UBI for the winner and their whole family, two generations back and two generation ahead, for it to be worth the risk. But then, with incentives that strong, you'll probably have only fraudsters getting close to the finish line. Unfortunately, the deck is already stacked against higher education here: as long as your long-term economic prospects hinge directly on your educational performance, the most motivated and effective crowd will be cheaters. If you weed out people based on performance, you'll end up selecting for cheaters, and against honest students.

MagnumOpus 4y ago

> I'm surprise it hasn't been tried. It has? You are describing every centralised college admissions exam, from the Chinese NCEE ("gaokao") to the Indian JEE-A or the French concours for admission to the better universities...