Advanced Statistical Workshop: Generalized Additive Mixed Models
Vincent Porretta (University of Alberta) & Aki-Juhani Kyrolainen (University of Turku)
This workshop provides a hands-on introduction in how to analyze eye-tracking data (reading, VWP, and Pupillometry) using Generalized Additive Mixed Modeling (GAMM) as implemented in the R package mgcv. GAMM is a nonlinear regression method, which can be used to analyze gaze data and is particularly suited to time course data. Over the one-day session, different topics will be addressed: 1) conceptual introduction to GAMMs with random effects; 2) basic implementation of GAMMs in R and visualization of effects; 3) Model fitting, model comparison, and interpretation of the model; and 4) Gaussian vs. Binomial models and discussion of autocorrelated residuals in time-series data. The workshop combines short lectures with practical hands-on sessions, so participants are advised to bring a laptop with R installed (or share a laptop with someone else during the practical sessions). While the required R code for the course will be provided, experience with R and linear modeling is definitely an advantage (though not strictly required).