Thursday, August 30, 2007

A revolutionary discovery

Ideal for learning about the importance of studying would be a random experiment in which two groups of students that are identical in all respects at the beginning of school are forced to study different amounts during school, but continue to behave identically in all other ways (class attendance, sleeping, drinking, study efficiency, paid employment etc.) that could influence the outcome of interest. In this paper we examine the effect of studying on college grade performance by using an Instrumental Variable (IV) approach that takes advantage of a real-world situation which we find closely resembles this ideal experiment. ...

In this case, we can learn about the causal effect of studying by comparing average grade outcomes between the two groups. Linking our survey data to administrative data, we find that grades are significantly lower, on average, for the group that studies less, on average, and we estimate that studying has an important effect on grade performance.
--Todd Stinebrickner and Ralph Stinebrickner, NBER Working Paper 13341, documenting the stunning fact that studying does raise your grades. I'm all for being careful about conflating correlation with causality, but this is really pushing it to the extreme! This would be a great nominee for an IgNobel Prize. Or an Onion article.


Jess Austin said...

For those who don't care to google through the cliffhanger abstract: the IV is whether a student's roommate brought a video game to school. Which is hilarious, for anyone who attended college in the last 15 years.

James Choi said...

Marathon forever, baby!