Re:Measuring 'Return on Investment' of Various College Degrees
by rahimnathwani on Wednesday August 18, 2021 @10:59AM
(#28221959)
Imagine there are two groups of 30-year-olds:
A) Those who were offered university places, but decided not to go.
B) Those who were not offered university places.
I would expect group A to have a significant 'earnings premium' over group B. Even though neither group attended university, group A are more likely to be smart, hard working, lucky, or whatever drives career success.
If you want to estimate the earnings premium of a university degree, you have to compare the folks who went to university with other folks just like them, i.e. with folks in group A. Not with a mixture of groups A and B.
> Is there another thing you would attribute it to?
Yes, some of the earnings premium is due to the students pre-university chances of success.
> It's not hard to simply ask "of this profession,
> what is the salary of those w/ a degree, and those w/o" at the national level
This wouldn't measure the thing we're trying to measure, because it suffers from selection bias.
A) Those who were offered university places, but decided not to go.
B) Those who were not offered university places.
I would expect group A to have a significant 'earnings premium' over group B. Even though neither group attended university, group A are more likely to be smart, hard working, lucky, or whatever drives career success.
If you want to estimate the earnings premium of a university degree, you have to compare the folks who went to university with other folks just like them, i.e. with folks in group A. Not with a mixture of groups A and B.
> Is there another thing you would attribute it to?
Yes, some of the earnings premium is due to the students pre-university chances of success.
> It's not hard to simply ask "of this profession, > what is the salary of those w/ a degree, and those w/o" at the national level
This wouldn't measure the thing we're trying to measure, because it suffers from selection bias.