Statistics Colloquium by Brandon Paguada '23

Violating a Statistician’s Favorite Assumption: Using Generalized Estimating Equations to Model Correlated, Repeated, or Clustered Multinomial Response Data by Brandon Paguada ’23, Statistics Colloquium, Wednesday, March 1, 1:10 – 1:50 pm, North Science Building 114, Wachenheim.

Abstract:  How do we model multinomial response data when our observations are correlated, repeated, or clustered?  We will explore General Estimating Equation approaches to modeling ordinal and nominal multinomial responses.  We will discuss how to reparametrize within-cluster associations using a correlated, global-odds ratio, and a local-odds ratio approach.  Finally, we will apply the model to explore a dataset of car accidents in the United States.