Administrative data is increasingly available, accelerated by the Evidence-Based Policymaking Act, which has encouraged government agencies to make those data available for researchers inside and outside of government. And of course, the availability of state data in education—which is a direct result of IES’s national longitudinal data system funding—is certainly new within the past decade. We have new approaches to artificial intelligence, new approaches to data scraping, new approaches to the use of big data that are improving all the time, and we need both research that uses those techniques, and we need research on the techniques themselves.
IES has for a long time recognized the importance of understanding, not just what works, but what works for whom and under what circumstance. But the way the research proceeds is first, the what works question is asked, and then the heterogeneity questions—what works for whom, where, under what circumstances—get added on later. And our committee is recommending that that attention to variation be built in from the start, so that it’s not an afterthought, but rather the heart of the problem.

The committee is also very interested in that question, but we do not have the answer. We don’t have the data. IES has released a little bit of data on who gets funded, but we don’t have data on who applies so we are not able to discern at what point in the pipeline the inequities are being introduced. That’s why we call for [a] comprehensive [grant funding] review.

Some of the recommendations could be implemented with few additional resources. However, all of them will require staff time, and that is a scarce resource at IES. We have recommended that Congress re-examine IES’s budget, recognizing both that it is modest compared to other scientific research agencies and that it is not enough to allow all the recommendations of this report to be implemented. Indeed, staffing resources are essential for implementing these recommendations.