The first requirement for research findings to be credible is that they are push-button replicable, meaning a third party can access the original data and authors’ programming code and use them to reproduce the findings. This study finds that only 27 of 109 sampled development impact evaluations from both the health sciences and social sciences published in 2014 meet this requirement. In many cases, authors are noncompliant with journal policies regarding replication files.
Push button replication: Is impact evaluation evidence for international development verifiable?
Written By
Wood BD, Muller R, Brown AN
