Tuesday, August 3, 2010

Standards for Regression Discontinuity Designs

I had the great pleasure of being on the panel that helped to develop quality standards to be used in the grading of studies for the What Works Clearinghouse funded by the U.S. Department of Education and operated by Mathematica Policy Research. The WWC is a compendium of evidence on educational interventions.

The standards are now publicly available on the WWC website.

In some ways designing standards for RD studies is easier than for other types of studies, as the nature of the design focuses analyses on a small set of issues. On the other hand, we were aiming at a moving target, as the literature on things like standard errors and bandwidth selection is still evolving. We assumed, or at least I assumed, that there would be a revisiting of the standards after a few years to update them in the light of more recent research.

No comments: