An ordinal regression workshop with Guilherme D. Garcia
In the fall of 2022, I conducted a workshop at the Hispanic Linguistics Symposium together with Guilherme. This workshop arose after meeting at GASLA earlier that year and having a discussion about how statistical modelling practices could be improved in linguistics. I brought up that many people at the conference were treating ordinal data inappropriately, and we started discussions about a possible paper.
The paper talks ended when I realized that João Veríssimo had already written the paper we were talking about (Veríssimo 2021). We decided that our part in promoting the use of ordinal models would be to run a workshop as part of a conference where many linguists would attend. This workshop outlined the conceptual problem with treating ordinal data as numerical, and then gave concrete examples of how to run and interpret the proportional odds models using package ordinal. Our slides and the workshop materials are available below.