Young, minority community college students in California perform better in courses taught by minority instructors, concludes a new NBER working paper.

Economists Robert Fairlie, Florian Hoffman and Philip Oreopoulos estimate that when minority students are assigned to instructors of the same background, the usual gaps in outcomes, such as grades they earn and the rates courses are dropped or failed, are cut nearly in half.

And here's something even more interesting: the authors were able to posit a "direction" for the phenomenon, the direction being from students to teachers, not from teachers to students. In other words, it is not that teachers of a particular race or ethnicity teach differently depending upon whom they are teaching, but that it is the students who behave differently based on who is teaching them.

How do we know this?

First, minority students drop classes at disproportionate rates when the instructor is a different race or ethnicity before receiving any feedback in the form of a grade.

Second, the researchers didn't find any differences in outcomes when they looked only at older students. The findings were significant only for young students, those under 22 years old. If instructors were changing their behavior based on the race or ethnicity of the students, one might expect the data to show significant gaps across all age groups.

It's this second finding that points to another tactic regarding the achievement gap (in addition to recruiting, training, and retaining more minority instructors), namely, to provide instruction and experiences that accelerate the process that makes students less susceptible to stereotype threat as they age.