Björn Siepe

I am a postdoctoral researcher at Marburg University, Germany, where I recently finished my PhD in Psychological Methods. My research addresses two main questions: how much we can learn about psychological processes in daily life from intensive longitudinal data, and how we know which statistical methods work well.

For the first question, I develop and evaluate multilevel time series and dynamic network models, with a particular emphasis on quantifying and communicating uncertainty in model outputs. To help researchers applying these methods, I have (co-)developed multiple software packages for R and Python.

The second question belongs to the metascience of methodological research. I study how simulation studies and benchmarks are planned, preregistered, and reported. One result of this work was ADEMP-PreReg, a protocol for reporting and preregistering simulation studies. I’m currently (co-)chairing two working groups on quality standards for simulation studies, and I’m always open to chat about novel ideas to improve methodological research. The two questions meet in my current DFG project, in which I build an empirical benchmark for time series models in psychology.

I have always enjoyed teaching statistics and programming to students and researchers. Over the years, I have given multiple workshops on reproducible research and time series modelling, for which you can find the materials on my workshop page.

I’m committed to open science practices. Together with collaborators, I built openESM, an open-source database of 60+ experience sampling datasets. We are actively working on expanding the database and making it easier for researchers to use it. Updates on the project can be found on the openESM blog. I’m also a board member of the Open Science Initiative at Marburg University.