A practical guide to study preregistration [ResBaz]

Many research literatures, across disciplines, are full of findings that cannot be replicated. A major contributor to this “replication crisis” is the use of questionable research practices that inflate statistical significance, meaning that many published findings are not “real”. Preregistration is one solution to this problem. A preregistration is a time-stamped document outlining a study’s hypotheses, sample size, methodology, and analytical plan before data collection begins. It therefore reduces researcher flexibility and increases the credibility of research findings. It is increasingly required or recommended by journals and funders, and is a valuable practice for all researchers. In this practical session I’ll introduce the options for where to post your preregistration and walk through some of the trickier decisions that must be made, including sample size determination, stopping rules, data exclusions, violations of assumptions, and exploratory analyses.