RISE Insight - The Evidence About What Works in Education: Graphs to Illustrate External Validity and Construct Validity

front page of RISE Insights

In the latest RISE Insight, Lant Pritchett (Harvard Kennedy School and the Center for Global Development) looks at how the bulk of the new empirical research on estimating the learning impact of various programmes and policies is inadequately theorised to be of immediate use in accelerating learning. 

Pritchett explains in this piece about the body of evidence used to identify what works for increasing learning, and more specifically, illustrates the two concepts of external validity and construct validity. In his view, one cannot understand the evidence about what works to improve learning from the systematic reviews of “rigorous” studies, without understanding the notions of design spaces, response surfaces, construct validity, and external validity.

Concluding, Pritchett explains the premise of RISE is to expand the body of rigorous research, but in a way that explicitly puts each of the collection of studies into a country context. Furthermore, he noted that RISE will evaluate reforms at scale, with a common system diagnostic, with specific, theory grounded hypotheses about what programme/policy/project design might (or might not) work. His idea is that this approach, which seeks to encompass the body of knowledge and contribute new knowledge, will move forward from the body of evidence to actionable, context-specific recommendations, including that “learning about learning” has to be embedded in projects/programmes/policies.

The Evidence” About “What Works” in Education: Graphs to Illustrate External Validity and Construct Validity