The Benefits of Preregistration and Registered Reports

From a methodological falsificationist philosophy of science it is desirable that scientific claims are severely tested. We review evidence about the self-reported prevalence of research practices that reduce the severity of tests. Such biasing selection effects are relatively common, although the extent to which they lead to biased claims is impossible to empirically establish. These practices are amplified by publication bias, where desired results are more likely to end up in the published literature. We explain how, from an error-statistical epistemology, the goal of preregistration is to allow peers to transparently evaluate how severely a claim has been tested, and how the goal of Registered Reports is to reduce publication bias. A review of metascientific research on preregistration and Registered Reports confirms both practices achieve these goals. The literature also reveals researchers perceive additional benefits of preregistration, mainly concerning the improved quality of their studies. We review criticisms on preregistration, and find conflicting viewpoints, several misunderstandings, and a general lack of empirical support. An exception is the valid concern that the mere presence of a preregistration will mindlessly be used as a proxy for high quality. We conclude with general recommendations to increase the quality and practice of preregistration.