If you’re trying to keep abreast of the latest behavioural economics research on education, it’s worth popping in every so often at the Social Science Research Network (SSRN) to check out the latest from the National Bureau of Economic Research (NBER). It’s mostly about K-12, but when it does tackle higher education, it’s unfailingly interesting.

Maybe the most interesting piece published recently is called The Effects of Student Coaching in College, by Rachel Baker and Eric Bettinger (who is, IMHO, a genius). Over a period of two years, students at a number of U.S. institutions were assigned by lottery to a program run by a company called Inside Track in which they received various forms of personal and academic coaching. (The authors have posted a fulltext version of the paper for free here.)

The results were striking: one year of coaching created an immediate increase of 12 percent in year-on-year persistence, which did not shrink in subsequent years. Coaching is a pretty intensive (and expensive) enterprise, but 12 percent is an enormous return, and compares very favourably to the results achieved by increasing student aid.

A second great paper is A Community College Instructor Like Me: Race and Ethnicity Interactions in the Classroom by Robert Fairlie, Florian Hoffman and Philip Oreopoulos. Using data from a community college where low-achieving students are quasi-randomly assigned to instructors, the authors try to work out whether minority students taught by members of their own ethnic group do better than those taught by members of other ethnic groups.

As it turns out, they do – or, at least the younger ones do (there was no role-model effect among older students). When taught by a member of their own ethnic group, non-white students closed roughly half the educational gap with white students, and the effect was even greater among black students.

It’s great research, but unlike the Bettinger piece, the policy implications are less clear-cut as the political acceptability of greater classroom segregation seems limited, even backed by results like these. And hiring more instructors of one ethnicity may lead to more classroom sorting, which could have other knock-on effects.

Both papers are great, but if you can only read one, read Baker and Bettinger – it’s a result that has the potential to seriously change the way we look at retention.