In a typically superb post, Tim Burke explains the problem with trying to apply a rational-choice model to the decisions college students make about their education:

There is a lot of information that you could acquire about courses or about colleges that you could reasonably use to assemble a decision matrix. What size is the class or the college? Do you have a good reason for thinking that you flourish in small or large classes or institutions? What do you think you need in terms of knowledge or training? What kinds of environments and teaching styles do you enjoy or find stimulating? And so on — this often information you could have, and sometimes, I agree, information that is hard to come by that shouldn’t be so hard to get.  

But then think on all the things that make a difference in a class or a university that you cannot possibly know about no matter how much information you have. The friends you will make. The people you will love. The mentors who will strike a chord with you. The class that will surprise you and change your views of everything. The chance experience you have that will transform you. You can find an environment that is rich in people, in time, in resources, in the unexpected (and some colleges and classes are impoverished in all or most of those). But you can’t determine any of the specifics with all the information in the world and yet it is these specifics that create the most “added value”. Perhaps even more importantly, it’s not the person who just had the experience who will value the commodity most (or rue it most dearly), it’s the person you will be in the years to come. That person is not you in so many ways: you are today very very bad at predicting what that person needed or wanted and you always will be bad at it. If we could sue our younger selves, many of us probably would.

So the people who look and say, “Oh, just make sure there’s more information and people will make the choices that economists think they ought to make” are doomed to disappointment. Which would be fine if they would consent to just being disappointed but policy wonks tend to think that when the outcome that the models predicted doesn’t happen, the answer is to make people behave like the models said they would.

So wise. Please read it all.