Spring 2005
Tuesday 11:10-12:25, Claxton Addition 238
Instructor:
Russell Zaretzki
328 Stokely Management Center
rzaretzk[remove]@utk.edu
974-8326
SOMS 578 is a three credit masters level course in categorical data
analysis. Topics will include basic goodness of fit measures such
as Pearson's $X^2$ Statistic, contingency tables, log-linear
analysis of multidimensional contingency tables, and logistic regression.
1 yr graduate-level statistics, regression analysis and analysis of
variance, or consent of instructor. Some basic knowledge of SAS or
SPSS is also required.
Office Hours
TBD or by appointment.
- Siminoff, Jeffrey S.(2003). Analyzing Categorical Data. Springer.
- This is the basic textbook for the course. Below we give an outline
with sections covered. A web page for the book is contains data sets and computer code used in the book -
Siminoff Textbook Home.
- Stokes, M. ,Davis, C., Koch, G. (2001). Categorical Data Analysis Using The SAS System
- Because SAS is such an important tool throughout the statistical
world, we will regularly use it to analyze categorical data and
explore the options available within appropriate procedures. The website for this text contains all of the datasets used throughout the book-SAS Textbook Home.
.
Grading
Grades will be based on homework assignments, midterm and final exam or project. Homework will be assigned approximately every two weeks. Because this is a practical course,
assignments will largely focus on learning to execute and interpret analyses using SAS or SPSS. Answers to specific questions and summary of program results and performance should be well organized and clearly presented. Program code must be well documented.
Final Exam
The final exam is scheduled for May 4 between 10:15am-12:15pm. As a class we can decide whether a final project or final exam is preferred. In the event that the final exam is selected no alternative time will be offered for the exam.
Programming Languages
Examples and sample code demonstrated or discussed within class will
be written in the SAS language. Students may find it more
appropriate to use SPSS for homework assignments. Homework done
with SPSS will be accepted but the instructor will not be available
for assistance in using this package.
Participation
You are strongly encouraged to participate in class discussions.
Please ask questions when you have them. This is the first time
that I am teaching such a course and I will need feedback to
understand which elements of the presentation are unclear. Personal
experiences and interests related with this material are also
appreciated.
Detailed Course Outline (Tentative)
- Jan 13. Discussion of syllabus, a taxonomy of data types, techniques for analyzing categorical data.
- Read Siminoff Chapter 1., SAS chapter 1.
- Jan 18. The binomial distribution and 2-by-2 tables.
Using the SAS procedure Freq.
- Read SAS chapter 2. Siminoff chapter 6.
- Jan 20. 2-by-2 tables cont. Discuss odds ratios and relative risk, exact tests, McNemar's test.
- Read SAS chapter 2.
- Jan 25.Higher way tables and Mantel-Hanzel type tests
- Read SAS chapter 3.
- Jan 27. Further analysis of higher way tables
- Browse SAS chapter 4-5.
- Feb 1.The Linear Regression Model.
- Read Siminoff Chapters 1 & 2
- Feb 3. Model Selection for Linear Regression.
- Read Siminoff Sec 3.4-3.6
- Feb 8.Multinomial and Poisson random variables. Goodness of
fit testing.
- Siminoff Chapter 4
- Feb 10.
- Feb 15. Regression Models for Count Data I.
- Siminoff Chapter 5
- Feb 17. Regression Models for Count Data II.
- Siminoff Chapter 5
- Feb 22. Regression Models for Count Data III.
- Siminoff Chapter 5
- Feb 24. Loglinear Models I.
- Siminoff section 6.2
- Mar 1. Loglinear Models II.
- Siminoff chapter 7
- Mar 3. Loglinear Models III.
- Siminoff chapter 8
- Mar 8. Introduction to Logistic Regression and GLM
- Mar 10. Logistic Regression II.
- Mar 15. Logistic Regression III. Analysis of Deviance and model building strategy.
- Mar 17.Model Building and Diagnostics for Logistic Regression
- Mar 22 & Mar 24 . Spring Break.
- Mar 29. Regression Models for Mulitnomial Data.
- Mar 31. SAS procedure CATMOD
- Apr 5.Regression models for Ordinal data
- Apr 7.Regression modeling for Ordinal data cont.
- Apr 12.Case Studies 1.
- Apr 14.Case Studies 2.
- Apr 19.Case Studies 3.
- Apr 21.Project Presentations
- Apr 26.Course Review
- May 4. Final Exam 10:15am-12:15pm.
Links to Related Courses with Online Notes Available
- Penn State Statistics Department
- Slides for a year 2000 course on categorical data done at a slightly more mathematical level
- Penn State Statistics Department
- Slides for a year 2004 course on categorical data done at a slightly more mathematical level. Links seem to be gone but old homework is still there.
- Online Notes from UMASS
- Slides discuss standard analyses of 2-by-2 tables and various related distributions.
- Sociology Course from UMD
- Powerpoint slides along with data sets giving examples in the sociology context.
- Generalized Linear Models.
- Excellent Sets of Notes for a course on Generalized Linear Models i.e. Regression Models for Categorical Data from a professor at Princeton's Woodrow Wilson School.
- Generalized Linear Models.
- More Notes from a course in Denmark.
- Generalized Linear Models.
- Splus-R notes on GLM's.
- Generalized Linear Mixed Models.
- Combining Generalized Linear Models with random effects.
- Generalized Linear Models.
- Extensions in the area of robotics.
SAS Resources
- OIT SAS Course
- The office of information technology offers training in SAS as well as other languages.
- SAS resources on the Web
- A list of resources from a guy in North Carolina?
- UCLA Help Page
-
- Internet and Web Resourses for SAS
-
- Online SAS Users Guide
-
- Harvard SAS Homepage
- Contains links to SAS online documentation.
- Harvard SAS Notes
- A thirty page introduction to SAS.
- SAS Notes
- Links from York University in Canada.
- SAS Notes
- University of Indiana.
Other Useful Internet References
- The Siminoff Textbook Home The textbook home again
- Categorical Data with Graphics
- This author has written a book about using graphical methods for categorical data
- Categorical Data for Social Scientists
-
- University of Notre Dame
- A downloadable document discussing discussing analysis of contingency tables
- Link to a syllabus from Australia
-
- Course Homepage with Syllabus
- From Cornell University Dept. of Rural Sociology
- Oxford University Syllabus
-
- Ohio State
- Extended and detailed syllabus from a Political Science 818 course
- Duke University Economics
-
- Intro Statistics from Texas A & M
-
- Homepage for a course from South Carolina
-