Understanding Mixed-Effects Models through Simulation
- Location
- 52 Pritchatts Road 412, Zoom
- Dates
- Friday 12 July 2024 (12:00-13:00)
Join Dr Julian Quandt, postdoctoral researcher at WU Vienna University of Economics and Business, for his hybrid seminar entitled 'Understanding Mixed-Effects Models through Simulation' from 12 - 1pm, Friday 12 July 2024.
"In this talk, we will delve into the increasingly popular mixed-effects models, frequently employed in experimental fields such as psycholinguistics where data are nested within participants and items. These models can often appear opaque, making it challenging to grasp what specific parameters signify and how they are reflected in the data.
Through the power of simulation, we aim to demystify these models by constructing and manipulating their parameters from scratch. This approach not only aids in a deeper comprehension of the models but also serves as an invaluable tool for conducting power analysis and refining experimental predictions.
Key areas we will explore include:
- Simulating data and see whether the model estimates mirror our expectations.
- Understanding random effects and how to simulate them.
- Exploring how simulation techniques can be used for power analysis.
This session is designed to be useful for a wide audience but of course familiarity with the lme4 package and R will help. If you would like to familiarize yourself with data simulation, you can have a look at my blog posts about this topic."
About the Speaker
Julian Quandt, currently a postdoctoral researcher at WU Vienna University of Economics and Business, completed his PhD studies in Psychology at Radboud University, Nijmegen, focusing on metacognition in value-based decision making. During his doctoral studies, faced with the absence of a standard tool for power analysis for Mixed-Effects Models, he used data simulation instead. Through this process, he discovered the power of data simulation as a tool for learning and enhancing his own understanding of statistical models. Julian authored four blog posts on data simulation, covering both simple and complex statistical models. His intention was to share his intuition that simulation can serve as a valuable tool for comprehending these models, in addition to enabling us to conduct power analysis for (nearly) any statistical model.