Measuring the return on investment of attending college (Part 2)

In my last post, I described the challenges in obtaining accurate and comprehensive data on the salaries earned by graduates of individual colleges around the nation. While there is much interest in using data on the earnings of college graduates as a means for ranking colleges, there is no single source of data on all colleges in the country (or even a majority of them) that allows us to compare institutions in this manner. Even if we did have accurate data, is the salary of recent graduates the primary criterion we should use in judging higher education institutions?

One attempt at collecting better data on earnings is underway at the CollegeMeasures.org website, which I also mentioned in my earlier post. One part of the website uses the inaccurate PayScale.com data on earnings to calculate a return on investment for hundreds of colleges around the country. But a series of reports on that website, which have received less attention, use state unemployment insurance data matched up with student records from public and private institutions in a small handful of states to calculate earnings by major at individual colleges and universities in those states. And what these reports demonstrate is that any variation in average wages of graduates of different institutions is generally overshadowed by differences in earnings of those majors within institutions.

The College Measures state reports are very upfront about the limitations of the earnings data used in their studies. For example, because they have access only to those individuals at employers in the state who participate in the unemployment insurance program, they exclude graduates who take jobs outside the state. If one believes that “better” students have more labor market options, including taking jobs out of state, then using data on graduates who stay in the state may provide downwardly-biased earnings estimates. The reports only use earnings for the first year after college graduation, so you cannot tell if the graduates of some institutions – or with different majors – have varying career earnings trajectories. Nevertheless, there are some interesting points that can be gleaned from these reports.

One of the states included in College Measures’ database is Virginia. You can use the database to look at first-year earnings of graduates for each institution. For the University of Virginia, the state’s flagship institution and one of the most well-respected public universities in the country, the average first-year earnings of bachelor degree recipients is $39,648. The range at UVa goes from a high of $60,300 for graduates of the systems engineering program, to a low of $27,209 for those with a degree in biology. At George Mason University, a less selective public university in the state, the average first-year earnings are actually higher than at UVa – $41,153, or about 4 percent higher. And the range at GMU is from $58,924 for graduates in computer engineering to $28,648 in athletic training.

Interested in majoring in anthropology, a degree that Florida Governor Rick Scott has said is not worth the state investing in? You’d be better off at George Mason, where the first-year earnings of $31,554 are 13 percent higher than at UVa. How about business administration? Go to UVa, where the earnings of $53,222 are 20 percent higher than at GMU. Electrical engineering? Back to GMU, where the average earnings of $57,239 are 7 percent higher than at UVa.

I think you probably get the point. If you are interested in earnings as a measure of institutional performance – and remember, the data I’ve shown here from the College Measures website is for first-year graduates only – then you need to take into account what the mix of majors is in a given institution. This is why colleges such as MIT, Cal Tech, and Harvey Mudd fare so well when you look at the earnings of their graduates – they are all heavily weighted toward engineering and the sciences.

And for any single student who is thinking about using earnings data to determine where to go to college? Don’t waste your time. You are better off worrying about what you will study, rather than where you will study it. Recognizing that a large percentage of students end up majoring in something different than what they were thinking about when applying to colleges, it makes little sense to choose a postsecondary institution based solely on its earnings profiles.

Leave a Reply