Perspectives on Scientific Error

University of Technology Eindhoven | Eurandom

In the last decade, several studies have raised concerns about a high prevalence of errors in published research

Department of Mathematics and Computer Science, 4th floor, meeting rooms MF11/12

Theoretical arguments have suggested the possibility that the number of false claims in the scientific literature is much higher than desired and reports across fields have subsequently shown that many findings that researchers believed were firmly established failed to replicate. These erroneous claims in the scientific literature raise important issues. From an empirical practitioner’s perspective, errors mislead and slow down research projects. From a philosophical perspective, scientific error raises questions about the right forms of scientific inference, scientific progress, and the reliability of science as a source of knowledge. From the perspective of the general public, scientific error undermines the epistemic authority of science and the degree to which policy-makers trust scientific experts.Following the success of the previous five editions, the sixth Perspectives on Scientific Error workshop will bring together an interdisciplinary group of researchers interested in statistical, philosophical, and meta-scientific issues of scientific error. Participants will share their views on how scientific errors can be detected and corrected, how they influence knowledge generation, and the perception of science as a whole.Hackathon: Reducing scientific error in practiceIn collaboration with the <a href="https://www.metamelb.org/" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">MetaMelb</a> research group we invite you to join a one day hackathon following the workshop. Join us for a Saturday of hackathons exploring how we can reduce scientific errors in practice, especially during pre- and post-publication peer review. We will break into small groups and develop tools, workflows, and guidelines that researchers can use to detect/reduce a variety of error types, such as non-reproducibility, statistical errors, non-reusable data, undisclosed preregistration deviations, and spin/hype. Everyone is welcome — you don’t need any special expertise to take part, just bring your brain (and maybe a laptop). You can register either for the workshop, or the hackathon, or both.