Blinding and Randomization

Abstract

Most, if not all, guidelines, recommendations, and other texts on Good Research Practice emphasize the importance of blinding and randomization. There is, however, very limited specific guidance on when and how to apply blinding and randomization. This chapter aims to disambiguate these two terms by discussing what they mean, why they are applied, and how to conduct the acts of randomization and blinding. We discuss the use of blinding and randomization as the means against existing and potential risks of bias rather than a mandatory practice that is to be followed under all circumstances and at any cost. We argue that, in general, experiments should be blinded and randomized if (a) this is a confirmatory research that has a major impact on decision-making and that cannot be readily repeated (for ethical or resource-related reasons) and/or (b) no other measures can be applied to protect against existing and potential risks of bias.