McArdle is certainly right that tuition costs have been far outpacing the rate of inflation for other goods. Here's how the cost of college tuition and fees have increased relative to the cost of all goods, according to the Bureau of Labor Statistics:
Prices overall have grown by a factor of about 3.6 since 1978, but college tuition has grown by a factor of 12. All told, in the 2010-11 school year, the average cost of college was an astounding $27,435, according to the National Center for Education Statistics.
But buried in McArdle's piece is a curious admission. "Experts tend to agree that for the average student, college is still worth it today," she writes. "But they also agree that the rapid increase in price is eating up more and more of the potential return." So college is still worth it! The best numbers on this come from the Brookings Institution's Hamilton Project, which calculates the return on investment for spending on college. Hamilton's Michael Greenstone and Adam Looney estimate that the return on investment for an associate's degree is about 20 percent, and the return for a bachelor's is about 15 percent (mostly because bachelor's degrees are much more expensive). That's enormous compared to returns from the stock market, bonds, or (for you Ron Paul fans) gold:
They also estimate that the lifetime value of a bachelor's degree is $570,000 relative to a high school diploma, at a cost of only about $109,740 (four times the $27,435 average cost figure). That's a huge, huge return on investment.
But McArdle's central claim is not that this return on investment doesn't exist. It's that it's going away, especially for "borderline students" who have not been adequately prepared for college by their high school education. To figure out whether this is the case, I calculated* how the value in dollars of getting a bachelor's relative to getting a high school diploma has changed between 1991 and 2010, looking at full-time workers from age 22 and up.
*For the nerds, I'm calculating the net present value using 5 percent discount rate, using the census data (P-32 here) on mean income by household attainment (which I recalculated to be the mean for all workers, not men and women separately), and using World Bank life expectancy estimates. It's the same method that Brookings' Greenstone and Looney used.