3 ms·

I can't find the study right now, but economists are not stupid, and there are experiment design to combat sampling bias. One study I remember followed student

by sanxiyn 12y ago

I can't find the study right now, but economists are not stupid, and there are experiment design to combat sampling bias. One study I remember followed students who barely made the admission cut, and who barely didn't make the cut. To illustrate, let's say people who go to college average 1800 in SAT score, and general population average 1200. This indeed causes sampling bias. So you compare two groups with average 1610 and 1590 instead. The first group went to the college, the second group didn't. This should minimize sampling bias. You get the idea.

socrates1998 12y ago
I agree that they should do this, but I don't think SAT scores would be a good measure. And I don't think they could find a large enough sample of the population that wouldn't be biased in some other way. Why? Because the only reason a smart and motivated person doesn't go to college is because they either have a great opportunity that would delay it (like a highly ranked chess player) or they some other outlying condition. This would really be biased because it would show how highly talented people really don't need college at all. I just don't see how you can get around this problem because it is so heavily ingrained in our culture that you have to go to college, so almost everyone that can, does.
- sanxiyn 12y ago
  "... the only reason a smart and motivated person doesn't go to college ..." You seem to think being smart and motivated is binary, which of course they aren't. They are continuous. You compare people who are barely smart enough and barely motivated enough to go to college, against people who are barely not smart enough and barely not motivated enough to go to college, to reduce bias. On the other hand, this does raise an interesting issue of inhomogeneous return to education. It is possible (even probable) that return to education is not same for groups with different abilities. It may be the case going to college is a very good deal if you are not very smart, but not a good deal if you are very smart. I think economists are actively researching this area.

by sanxiyn 12y ago

socrates1998 12y ago

I agree that they should do this, but I don't think SAT scores would be a good measure. And I don't think they could find a large enough sample of the population that wouldn't be biased in some other way. Why? Because the only reason a smart and motivated person doesn't go to college is because they either have a great opportunity that would delay it (like a highly ranked chess player) or they some other outlying condition. This would really be biased because it would show how highly talented people really don't need college at all. I just don't see how you can get around this problem because it is so heavily ingrained in our culture that you have to go to college, so almost everyone that can, does.

sanxiyn 12y ago
"... the only reason a smart and motivated person doesn't go to college ..." You seem to think being smart and motivated is binary, which of course they aren't. They are continuous. You compare people who are barely smart enough and barely motivated enough to go to college, against people who are barely not smart enough and barely not motivated enough to go to college, to reduce bias. On the other hand, this does raise an interesting issue of inhomogeneous return to education. It is possible (even probable) that return to education is not same for groups with different abilities. It may be the case going to college is a very good deal if you are not very smart, but not a good deal if you are very smart. I think economists are actively researching this area.