Ignore the Media: Choosing a College Major Edition
I’m always conflicted about the debate about whether students should be nudged into more quantitative and “practical” college degrees. Partially, it’s a question of who should be doing the nudging – parents? Popular culture? Universities? Or the actual government? But mostly because it strikes me as a very odd question shaped by some very poorly-thought-out assumptions on all sides. But the most vexing is when people, whether they’re pushing applied fields or liberal arts, bringing to bear income data on various college majors.
Jordan Weissman brings this critique to a recent survey purporting to compare STEM majors and humanities majors, but doesn’t go far enough. As he points out, there is self-selection here – students majoring in English are more likely to be focused on career goals other than salary. But there are so many other confounding factors – for example, family money. Perhaps students from wealthier families are simply freer to take less practical majors, since they are more secure in their economic lives (and have the connections to get better jobs). Furthermore, business students don’t earn more than humanities students – but there are a LOT more business and pre-professional students in America than there are humanities students. There are many more business majors coming out of community colleges than there are history students.
This whole comparison is fatally flawed, because it doesn’t answer the key question: if I’m a student entering college, what should I major in if I want to make a good income? It would be ideal to solve this once and for all by applying proper randomization. But unfortunately, in America in 2014 it’s not really practical to randomly assign students a major. One could try a crude approximation of a matched-pair study by looking at twins who choose different majors – though that still doesn’t solve the self-selection issue. But even without a sound experiment, there’s one very way to make this study less hopeless, which is to stratify.
The relevant issue, for a student making the decision, is the peer set. Knowing what these graphs look like for someone from your background, or for the type/caliber of school you attend, is vastly more valuable than aggregate figures. These graphs probably look extremely different for top-tier 4-year colleges and community college associate programs. Not to mention that by only counting the wages of the employed, you’re doing unnecessary violence to the statistics. A student cares about the expected value of education – if English students earn just fine but are grossly underemployed*, that’s extremely relevant information.
Anyway, the point is that the data is inherently terrible but that people seeking to make points are only making it worse. Actual students shouldn’t invest a whole lot of effort trying to figure out who is correct, they should do what they’re interested in. Though they should probably be aware that “architect” and “writer” are less appealing careers than they might sound at first blush.
*: They’re doing fine, incidentally.