Clearinghouse Standards of Evidence on the Transparency, Openness, and Reproducibility of Intervention Evaluations

Clearinghouses are influential repositories of information on the effectiveness of social interventions. To identify which interventions are “evidence-based”, clearinghouses evaluate empirical research using published standards of evidence that focus on study design features. Study designs that support causal inferences are necessary but insufficient for intervention evaluations to produce true results. The use of open science practices can improve the probability that evaluations produce true results and increase trust in research. In this study, we examined the degree to which the policies, procedures, and practices of 10 federal evidence clearinghouses consider the transparency, openness, and reproducibility of intervention evaluations. We found that seven clearinghouses consider at least one open science practice: replication (6 of 10 clearinghouses), public availability of results (6), investigator conflicts of interest (3), design and analysis transparency (3), study registration (2), and protocol sharing (1). We did not identify any policies, procedures, or practices related to analysis plan registration, data sharing, code sharing, materials sharing, and citation standards. Clearinghouse processes and standards could be updated to promote research transparency and reproducibility by reporting whether evaluations used open science practices, incorporating open science practices in their standards for receiving “evidence-based” designations, and verifying that evaluations used open science practices. Doing so could improve research quality, increase trustworthiness of evidence used for policy making, and support the evidence ecosystem to adopt open science practices.