Presented By: Department of Psychology
Psychology Methods Hour: Context effects in Hierarchical Linear Modeling (HLM) – obvious and no-so-obvious issues using a simple data example
Kai Cortina, Professor of Psychology, University of Michigan
Context effects are a common element in testing hypotheses involving nested data structures: Do students learn better if they are surrounded by high achieving students? Is the association between unemployment and depression stronger in affluent neighborhoods? Unfortunately, it is not always clear how to specify a context effect correctly in hierarchical linear modeling (HLM). Dr. Cortina will demonstrate that the different options of centering predictor variables can be confusing and often leads to inconsistent statistical conclusions. While there are special cases that require more complex models, he argues that most empirical studies in psychological research follow a straightforward definition of context effects.
Co-Sponsored By
Explore Similar Events
-
Loading Similar Events...