Thursday, March 27, 2014

Here's Why You Shouldn't Take Any One Study Too Seriously

Journalists often write about new studies as if they definitively prove or disprove something, when the study may not actually be all that great, and there have often been a lot of previous studies on the same topic that get little or no mention at all, and which may have conflicting results. Also, when a new paper gets posted online, it usually hasn't been vetted by very many people in the profession yet. Here is one more reason why you usually shouldn't take too much stock in any one new paper, courtesy of the 4th-year economics PhD students at the University of Michigan.

Every year, the Michigan economics department holds a skit night, and in the latest skit night two weeks ago, a 4th-year PhD student (who requested to remain anonymous) presented a supposedly typical job market paper. Mind you, this is all exaggerated and meant to be a joke. But I hope it can also serve as a reminder of how some studies aren't always the most robust, and people shouldn't be in awe of a piece of research just because it has numbers and jargon in it. There are differences in quality in academic research, just as in any other profession, and it's important not to take any one piece of research at face value.

Without further ado, here is the presentation, entitled "Every JMP [Job Market Paper] Ever":