Title: Geometry of statistical models

Abstract:

As quantitative psychologists, we often use statistical models to measure psychological processes and explain behavior. Geometry and measurement are at least etymologically related, and geometry is deeply rooted in the measurement of quite a few physical quantities. However, from the author’s perspective, geometry has somewhat faded into the background of current psychological modeling. In this talk, I will not go into the reasons why this may the case, but I want to give an idea of how ideas from non-Euclidean geometry may be relevant for psychological modeling. We will discuss distance and curvature in specific applications.