Thursday, May 06, 2010

Taking Out the Trash

by Daniel L. Bennett

Ben Miller has a great suggestion over at The Quick and the ED blog in which he calls for 3 outdated higher ed statistics to be put to rest. He believes that the following 3 arguments need to be taken out with the trash:
Each Year 170,000 qualified students will not attend a college or university–often because of monetary concerns.

Around 28 percent of all freshmen, and 42 percent of freshmen at two-year institutions require remediation. Only 30 percent of students who completed a remedial reading course, and 42 percent of those who completed two or fewer remedial math courses, earned any sort of degree within eight years of enrolling

The United States is falling behind other countries in postsecondary attainment.
Miller also asks if there are any other higher ed stats that deserve the same treatment. I have one:

Stat: College graduates will earn a million dollars more than non-graduates over their lifetime because of their degree

Original source: The US Census Bureau

Why it needs to be retired: The figure is severely flawed. Mark Schneider of AEI has suggested that it fails to account for factors such as school selectivity, the cost of a college education, foregone wages and the net present value stream of income. He estimates that the actual figure is closer to $280k. Furthermore, the figure is based on statistical averages of non-continuous income groups over time and fails to account for the earnings increases of actual individuals over time. It also fails to account for other factors that may have contributed to the increased earnings of individuals, such as innate ability, on-the-job training and outright luck.

What we need to talk about instead: Developing longitudinal datasets that track individuals from K-Retirement that will allow us to track individuals throughout their education and work experience in order to provide us with ample evidence to assess and evaluate the outcomes and value-added of our education system.


Steve said...

I agree that that statistic should be laid to rest, but I think you're mistaken about what we need.

We don't need longitudinal data, because that still leaves us with endogeneity issues. What we need is either credible instruments or experimental evidence on the return to schooling?

All the evidence we have from the US, esp. from studies by Josh Angrist, is that the return to going to college is indeed in the hundreds of thousands for most people not going.

Daniel L. Bennett said...

Thanks for the comment Steve. I think that we can agree that what we need is more outcome-based information.

Richard Vedder has on several occasions proposed the idea of linking college outcomes to IRS income records, data that is already collected by the federal government, in an effort to assess value added, or as you say, returns to schooling.

Another important factor that is often neglected in such assessments is the return of non-schooling human capital, such as on-the-job training.