Towards a better understanding and communication of uncertainty in empirical research by Dr Lonni Besançon

About the speaker

Lonni Besançon is a postdoctoral fellow at Monash University, Australia. He received his PhD in computer science at University Paris Saclay, France. His thesis “An interaction continuum for 3D dataset visualization” received the second price of the prix de these GDR-IGRV. He is particularly interested in interactive visualization techniques for 3D spatial data relying on new input paradigms and his recent work focuses on the visualization and understanding of uncertainty in empirical results in computer science.

About the talk

Statistical analysis is at the heart of empirical research work. Yet current reporting styles and guidelines on interpretations have been found to be prone to dichotomous thinking and eventually lead to issues such as the file-drawer effect or even the replication crisis. In my work, I have first explored my research field to find evidence of dichotomous reporting. Based on the threats of a replication crisis I have observed in empirical computer science, I have explored alternative visual presentations of statistical results that could help to avoid binary interpretations of results.