SPECIAL SESSION: Replicability in Plant Pathology: Do we have a problem? - Panel Discussion
Replicability in science: concept and status
Felipe Dalla Lana - The Ohio State University.
Obtaining consistent results in independent studies is one of the goals of science. A study is called replicable if the results are consistent in independent repetitions of an experiment. Although studies that failed to be replicated are not necessarily wrong, this is far from desirable. There is ongoing concern in multiple fields about the “replicability crisis in science”, which is the common problem of the increased number of studies that have failed to be replicated (when attempted). Fields such as psychology, statistics, medicine, and social sciences have an engaged community of scholars who are dedicated to the investigation of the replicability crisis, with proposals for solutions. Such proposals include better-informed interpretation and use of P-values and other statistics, improved experimental approaches, research-method transparency, and proactive attempted repetition of published studies (with reports on the negative and positive findings). As with any multidisciplinary debate, communication can be challenging because terms frequently have different meanings in different disciplines or with different authors, such as replicability and reproducibility, sometimes confusingly used as synonymous. In this presentation, I will address the concepts and terms related to replicability, and give an overview of the status of the debate in other fields of science.