Return to Dorms Bleg

I would think this would remain uninvestigated as research because almost everyone who would do such research has already learned that dorm life sucks, especially if you want to do something like study.

I wonder if you could exploit distance of the parental home to commuter school location where the student is in attendance to instrument for the role of commute time in GPA.

This doesn’t speak directly to your query and the sample size of one offers zero degrees of freedom, but:

One strong reason I didn’t even bother applying to the Tuck School of Business at Dartmouth was precisely that it was only one hundred yards from my old fraternity from my undergraduate days. So for that observation the value of the binary variable would have been, um, not one.

Score one for self-knowledge, though.

Bryan: There is a thriving peer effects literature that compares the effects of different types of roommate relationships on a variety of outcomes, but very little to my knowledge that compares people who live on campus to those who don’t. A well identified study on the effects on racial attitudes of interracial roommate pairings: http://www.povertyactionlab.org/evaluation/peer-effects-diversity-and-college-roommates and here’s one on socializing and GPA http://www.uh.edu/~adkugler/Sacerdote.pdf – but I think you’re right that there’s nothing comparing returns (income? happiness?) to people who lived on campus to those who didn’t.

But I’m not sure how you would obviate concerns about underlying dissimilarities between the two groups — commuters are likely much poorer, on average. Perhaps an RD design of people at either side of a scholarship that covered dorms?

Lots of colleges claim that it does, right? Lots of schools require that freshman live on campus, for example, so that they can better integrate into “campus life”? Might be interesting to ask those colleges for evidence that that actually leads to better educational outcomes. :->

Scott Carrell does a bunch of education production function papers and he has access to random assignment data from the military schools. Might be something there.

I wonder if on or off campus is the correct split.

I recall from long ago that there is evidence that the social ties formed in college may be more important than the knowledge gained — at least if we’re looking at career advancement.

So I would expect the most advantaged group to be those who lived off campus in an active fraternity or sorority. Then lived on campus in a dorm. Then lived off campus in an apartment building with lots of other students. Then probably lived at home (with whatever support that gives for one’s social life). Then finally, at the rear, the off campus social isolates.