We are very happy to announce the next talk in the Recent Trends in Language Acquisition Colloquium, which will take place on Tuesday, February 12, 2 – 4 pm in IG 3.301.
Tom Fritzsche will present “Data analysis with linear mixed models: A practical example from language acquisition”.
Abstract: Linear mixed-effects (LME) models have become the standard in analysing psycholinguistic and psychological data. Compared to t-tests and ANOVAs LME models offer numerous advantages but also require additional attention for specifying a model as each can (and needs to) be tailored to the structure (and the amount) of the data. Advantages and challenges will be discussed using a specific data set: Szendrői et al. (2017). The paper along with the code is available online (links below). The purpose of this presentation is to go through the analysis of this study in R and address questions regarding:
– model specification
– contrast coding
– fixed & random components
– factors & continuous predictors
– model evaluation and selection
– limits and problems
You are encouraged to download the data and code and run the analysis during the presentation. For this, it is necessary to have R installed (R Studio is optional) on a notebook including the R packages “reshape2”, “ggplot2”, “lme4”, “ez” and “RePsychLing”.
Paper:
Szendrői, Kriszta, Schumacher, Rebecca, Fritzsche, Tom, & Höhle, Barbara (2017). Acquisition of quantifier raising of a universal across an existential: Evidence from German. Glossa: a journal of general linguistics, 2(1), 46. doi:10.5334/gjgl.261 https://www.glossa-journal.org/articles/abstract/10.5334/gjgl.261/
Data & Code (Datavers repository):
https://doi.org/10.7910/DVN/D6XN9S
R & RStudio:
https://cran.r-project.org/
https://www.rstudio.com/
You are all cordially invited.