College Guide


January 11, 2011 3:44 PM Mismeasuring Success

By Daniel Luzer

At least one college is going to have to start measuring its output. This could be a good—or very, very bad—thing. According to a piece in the Wall Street Journal:

Texas A&M is starting to ask such basic questions as: Is that psychology or engineering professor worth his $125,000 salary?
The school is trying to answer this question by applying a cost-benefit analysis of how much each professor earns in salary per student taught. The school also uses such metrics of value added as research dollars brought in by a professor and student evaluations of how well a teacher performs in a classroom. For high-achieving professors, the new pay-for-performance standards offer bonuses of up to $10,000 a year.

Why is Texas A&M doing this? Well according to the article it’s because,

Since 1978 college costs have risen by more than tenfold, about three times the rate of inflation, according to an American Enterprise Institute study. Four years of college now cost as much as $200,000 at some private institutions, making this perhaps the only industry in America that has recorded negative productivity gains. In 2009 tuitions rose by 6%, four times overall prices. With rising tuition comes rising indebtedness, and for the first time student loan debt of $850 billion now exceeds credit card debt of $830 billion. State subsidies keep rising but are swallowed up in higher university costs and thus haven’t lowered tuitions.
Professors’ salaries and benefits make up about 60% to 70% of university noncapital costs.

A&M professors are protesting this but in theory it’s a very good idea. For decades American universities have been resisting any efforts to measure their outputs. This is wrong. Grading, metrics, and measures of success are good. The more information we have about how universities operate the better.

Just be careful what you measure. Because once someone measures something in education, it becomes the only thing people pay attention to.

Unfortunately Texas A&M seems to be interested in measuring the wrong things. The most relevant cost-benefit analysis for a university to perform is probably not “how much each professor earns in salary per student taught.” Added research dollars and student evaluations are also not necessarily the most crucial measures of success. The most important thing, arguably, is how much students actually learn in college.

And the rise in student tuition is here only the excuse for this concern about professors’ expenses; it’s not actually the reason. Professors don’t set tuition, the university board of regents does.

And if academic salaries and benefits make up 60 percent of noncapital costs, that means there’s a good 40 percent of college costs that are eaten up by things like running the campus, building buildings, administrators’ salaries (the president of Texas A&M earns $425,001 a year), and sports (the athletic director earns $690,000 a year). And while tuition is rising, academic salaries actually aren’t, adjusting for inflation they’re about the same as they were in 1984.

So if A&M (and other schools) really wants to perform an honest cost-benefit analysis, it needs to look at some other costs, and some different benefits.

Daniel Luzer is the news editor at Governing Magazine and former web editor of the Washington Monthly. Find him on Twitter: @Daniel_Luzer


  • PQuincy on January 11, 2011 11:08 PM:

    Thanks for the sensible review of a silly policy.

    Of course, assessing the value of college education is important, and the costs of providing that value. The hard part, though, is exactly measuring the value added. College education is a long-term good, one with multiplier effects that may take decades to emerge, and also one that is highly modulated by branding effects in various non-linear ways. (A Harvard BA will always be an asset to its holder, probably, even if that holder's circumstances are far removed from undergraduate education; a Washington State degree, in contrast, has much shorter brand value, even though the quality of the education and the long-term value added for the student's life may equal exceed that of a Harvard grad.

    Administrators, however (particularly in Texas, one suspects), much prefer to measure things easily measured, whether or not they have much connection to what one would like to asses. This is the great flaw in most current efforts to measure "learning outcomes": they are caught between the fact that there are things easy to measure that probably bear a very indirect relationship to the actual thing we want to know, and the fact that there are things very difficult to measure (and especially to quantify) that are close to what we want to know.