Principles of statistical analyses: old and new tools
Franziska Kretzschmar, Phillip M. Alday · 2020
The present chapter provides an overview of old and new statistical tools to analyze electrophysiological data on language processing. We will first introduce the very basic tenets of experimental designs and their intimate links to statistical design. Based on this, we introduce the analysis of variance (ANOVA) approach which has been the classical statistical tool to analyze event-related potentials (ERPs) in language research. After discussing the merits and disadvantages of the approach, we focus on introducing mixed-effects regression models as a viable alternative to traditional ANOVA analyses. We close with an overview offuture directions that mixed-effects modeling opens up for language researchers using ERPs or other electrophysiological measures.