27.05.2016: afex - Analysis of Factorial Experiments in R (Dr. Henrik Singmann)

afex - Analysis of Factorial Experiments in R

Abstract: 

Factorial experiments play an important role in experimental psychology. In my talk I give an overview of the functionality of R package afex which provides user-friendly functions allowing to perform the two most common analyses for data originating from such designs: analysis of variance (ANOVA) and mixed models. For ANOVAs, any number and combination of independent-samples (i.e., between-subjects) and repeated-measures (i.e., within-subjects) factors are supported via the same interface. Mixed models are estimated via lme4::lmer and p-values for terms/effects are calculated either via the Kenward-Roger approximation, parametric bootstrap, or likelihood-ratio tests. For both, ANOVAs and mixed models, the obtained results can be directly passed to lsmeans for any type of follow-up/post-hoc contrast test. In its default settings afex imitates commercial statistical packages by employing Type III sums-of-squares and sum-to-zero contrast making it particularly suitable for new R users.

Slides: afex - Analysis of Factorial EXperiments in R

Dr. Henrik Singmann is a postdoc in the Cognitive Psychology lab of Klaus Oberauer at the University of Zurich (Switzerland). After studying for his Diploma at the University of Hamburg, he obtained his PhD in 2014 at the University of Freiburg (Germany) under the supervision of Christoph Klauer. He is mainly interested in formal models of reasoning and recognition memory, but also in the development of tools for cognitive and statistical modeling in the statistical programming language R.