New Course for Fall 2006     (MWF 9am) 

STA 380        Categorical Data Analysis (3)

Not all data is continuous and methods based on the normal distribution are not appropriate in many situations. Questionnaire data and small counts are typical examples.  This course will present some of the most common and most practical methods for analyzing categorical and count data.

 SAS will be the main software used but other software may be introduced.

 Course prereq: STA 216 or instructor permission

 Course Topics will include:

•  Probability & sampling models for categorical/count data
•  Odds ratios & related analysis methods
•  Methods for analysis of 2 and 3 way contingency tables
•  Generalized linear models (GLMs) and their applications
  • Poisson Regression, Logistic Regression & other models
•  Building and applying loglinear models
  • 2 way and higher dimensions, ordinal associations
•  Multinomial models, logit models for ordinal responses  
•  Selected topics (eg. interrater agreement, sparse data)