Johann-Friedrich Salzmann receives StanCon 2026 Poster Award
At the StanCon 2026 Joff presented his work on short-term changes in electoral preferences with his poster „Modelling Short-Term Voting Propensity Shifts as Trajectories on the Simplex: A Bayesian State-Space Imputation Model Using High-Frequency Panels“. His poster presentation emphasised the potential of modelling categorical panel data as hierarchical Dirichlet-distributed trajectories on the simplex, with convex combinations powered by Aitchison operators, superseding conventional log transformations. Using this method, short-term transitions in representative individuals' party preferences are estimated from a high-frequency panel to shed light on the sudden surge in Die Linke's popularity right before the 2025 German federal election. We're excited to see this work further developed as part of Joff's PhD project.