Speaker
Description
Statistical models are essential tools for understanding data and drawing scientific conclusions. Despite the adoption of modern statistical methods (hierarchical models, structural equation models, Bayesian inference, …) in many disciplines, using these tools responsibly can be challenging. Classical issues, such as model assumptions, uncertainty, and experimental design remain and intersect with new developments, including unprecedented data volumes, complex model structures, and the temptation to replace classical statistical methodology with AI-based methods.
We invite participants to reflect on old and emerging statistical challenges to scientific understanding. Which challenges are the most urgent in everyday research practice? Which ones could be addressed through, for example, better statistical training, more rigorous analyses, improved study design, or increased computational power? Topics may include model generality versus model complexity, correlation versus causation, or statistical inference versus forecasting.
About:
Benjamin is a Computational Ecologist with a PhD in Applied Mathematics, based in iDiv’s ‘Theory in Biodiversity Science’ group. Here, his research focuses on statistical methods for population dynamics, species interactions and food webs. Through his work as a statistical advisor, he has been involved in many biodiversity-related projects. He regularly teaches Bayesian statistics and also offers statistical consulting for the whole iDiv community.
