Abstract

Preregistration has been lauded as one of the key solutions to the replication crisis in the social sciences, mainly because it has the potential to prevent p-hacking. However, empirical evidence about the effectiveness of preregistration is scarce. In this symposium, four speakers will shed light on this meta-scientific dark spot by presenting the first empirical studies that systematically assessed the effectiveness of preregistration.

First, Marcel van Assen (Utrecht University / Tilburg University) presents a study examining the quality of hypotheses in preregistrations. Van Assen and colleagues defined four criteria for a hypothesis to be valid, and assessed these criteria for a large sample of preregistered hypotheses. The results of this study yield important recommendations to improve how we formulate hypotheses in preregistrations. Second, Olmo van den Akker (Tilburg University) and George Ofosu (London School of Economics) present work in which published studies were compared to their corresponding preregistration. Van den Akker will discuss preregistrations in psychology, while Ofosu will discuss preregistrations in economics and political science. This multidisciplinary perspective allows a comparison of these fields’ preregistration practices and the effectiveness of these practices. Finally, Sarahanne Field (Groningen University) presents a study that assesses whether preregistration increases perceptions of trust in science. This is important information because the uptake of preregistration hinges on researcher’s perception that it is a useful tool to improve the validity of research.

Together, the results from these studies provide the first large-scale empirical evidence on the impact of preregistration on the replicability of the social sciences.