View Text Only Version

Applied Statistics

Categorical Data Analysis, STA 507

Contact Us  

Applied Statistics

Address:
25 University Ave
West Chester, PA 19383


Phone: 610-436-2440
Email: Randy Reiger

Introduction to Categorical Data Analysis, STA 507

Course Objectives
This class will provide data-driven introduction to statistical techniques for analysis of categorical data. Each topic will be introduced from the foundations. However, the main focus of the class will be how to apply the various techniques to actual data and how to interpret subsequent results. All techniques will be demonstrated using SAS. While students will be provided with all appropriate syntax specific to new material, it is expected that all students will have completed STA511 or have a working knowledge of SAS.

Topics
Topics covered may include measures of association and statistical tests for (2 x 2) contingency tables, Mantel-Haenszel methods for sets of (2 x 2) contingency tables, measures of association for (s x r) contingency tables, methods for ordinal response data, logistic regression, methods for correlated categorical data including McNemar's Test, methods for describing survival data including actuarial estimates, Poisson regression, observer agreement studies, and, time permitting, generalized estimating equation techniques, log-linear models and non-parametric techniques.

Read the Course Descriptions

Example Syllabus

STA507: Introduction to Categorical Data Analysis FALL 2007

Instructor:Dr. Randall H. Rieger

Office:(Anderson 328)

Email Address: smcclintoc@wcupa.edu

Office Telephone: 610-436-2893

Office Hours:Monday, 3:30-5:30; Tuesday, 1:30-3:00; Wednesday 4:00-5:30

Classroom: Anderson 120

Campus Emergencies
For campus emergencies call WCU's Department of Public Safety at (610)436-3311

Require Text
Stokes, Davis and Koch, Categorical Data Analysis Using the SAS System.

Suggested Supplemental Text
Agresti,Categorical Data Analysis

Fleiss, Statistical Methods for Rates and Proportions

Hosmer and Lemeshow, Applied Logistic Regression

Statistics Texts

In order to review very basic statistical topics used in this class, you may want to have an introductory (non-mathematical) statistics textbook to use for reference, such as:

Brase and Brase, Understanding Basic Statistics, Houghton Mifflin,

Alternatively, any introductory mathematical statistics text book will serve this purpose on a more theoretical level. The following are recommended:

Hogg and Tanis, Probability and Statistical Inference, Prentice Hall

OR

Wackerly, Mendenhall and Sheaffer,Mathematical Statistics with Applications

Duxbury

It is also expected that you have some familiarity with linear regression and ANOVA methods and SAS statistical software (see below).

SAS Software

The SAS Statistical software package is available free to any West Chester University student. The software can be loaded from a CD onto a student's PC. The method of distributing the CDs to interested students will be discussed on the first day of class. As part of the licensing agreement, upon graduating from or leaving WCU, the software will expire and no longer be available for the student's use.

In addition, the SAS software (and other statistical packages) will be available on all computers in Anderson Hall, including the Applied Statistics Laboratory.

Graduate Assistants

As an additional resource, three graduate assistants in the Applied Statistics MS Program will be available at various times to assist students with questions, assignments, and for additional practice. Please bear in mind that these students are also taking this class, so they should only be expected to provide collaborative discussion, not tutoring or teaching. Their hours and location will be provided on the first day of classes.

Day Time Location
Sunday 7-10 PM Anderson 119
Monday TBA Mitchell 407
Tuesday TBA Mitchell 407
Wednesday TBA Mitchell 407
Thursday TBA Mitchell 407

Course Objectives
This class will provide data-driven introduction to statistical techniques for analysis of categorical data. Each topic will be introduced from the foundations. However, the main focus of the class will be how to apply the various techniques to actual data and how to interpret subsequent results. All techniques will be demonstrated using SAS. While students will be provided with all appropriate syntax specific to new material, it is expected that all students will have completed STA511 or have a working knowledge of SAS.

Topics
Topics covered may include measures of association and statistical tests for (2 x 2) contingency tables, Mantel-Haenszel methods for sets of (2 x 2) contingency tables, measures of association for (s x r) contingency tables, methods for ordinal response data, logistic regression, methods for correlated categorical data including McNemar's Test, methods for describing survival data including actuarial estimates, Poisson regression, observer agreement studies, and, time permitting, generalized estimating equation techniques, log-linear models and non-parametric techniques.

Evaluation

Assignment % of Grade
Exam 1 (in-class) 24%
Exam 2 (computer, take home) 24%
Final Exam (format TBD) 26%
Homework 18%
Participation/Lab 8%

Class Format

  • Most classes will consist of
    1. a lecture/presentation introducing new material and, possibly a short "LAB" assignment to be completed in class.
    2. a bi-weekly homework assignment to be worked on outside of class. While you are strongly encouraged to work on problems independently, you are encouraged to discuss aspects of problems with classmates. However, the write-up of all assignments MUST be done independently.
  • All homework assignments, lecture notes, and LAB assignments will be made available prior to class on the class web page.
  • No late assignments will be accepted without a written medical note or equivalent.
  • On take-home examination assignments,students must work independently on these assignments, under the guidelines of the Honor Code. If you have questions about the exam, Dr. Rieger will be glad to help you. No late assignments will be accepted without a written medical note or equivalent.
  • EXAM I will be a written, in-class examination. EXAM II will be a one-week computer-based, take-home examination.
  • The format of the Final Exam will be determined at a later date and will be announced several weeks prior to the final exam period.
  • Missed exams can only be made up with a valid, verifiable, written university-approved excuse and must be made up within a week of the originally-scheduled exam.

Class Rules

Students engaging in disruptive behavior will be dealt with according to university policy. Students are encouraged to consult the undergraduate catalog for details of this policy.

  • If there is an emergency or unsafe situation while you are in class or on campus, please call 610-436-3311 immediately.
  • Academic dishonesty will not be tolerated in this class. Any cases of academic dishonesty will be dealt with according to university policy, and I will recommend the maximum possible penalty.
  • Students with disabilities are encouraged to make their needs known to the instructor and the Office of Services for Students with Disabilities early in the semester.
  • Please make use of office hours and other department and university resources if extra help is needed.
  • In the event that I am unable to meet a class, I will a) notify you in person at a prior time or b) an official class cancellation notification on the stationary of the Department of Mathematics, signed and date stamped by the Department Secretary (Barbara Maleno) will be posted on the classroom door. All other postings announcing the cancellation of this class are to be considered unofficial and are to be ignored.
  • Students should bring a storage device to each class to save any relevant lab work or data sets.
  • Consistent lateness will affect the participation portion of a students' grade.

Class Philosophy

This should be a fun class where you will learn valuable skills. By the end of this class, you will be proficient in working with categorical data. In order to truly learn the skills taught in this class, you must give yourselves time to "play around" with assignments, procedures, and computer manuals. If you wait until the night before class to do your work, it will be hard to be successful in this class.

Students are encouraged to give me feedback about the class. If you have any suggestions about any aspect of the class, please let me know in person or via e-mail or a note in my box. I will do what I can to foster student learning. However, I need to know what you are thinking!

"The fastest way to achieve success is to increase your rate of failure." Sometimes more can be learned by struggling through a problem and allowing yourselves several iterations at solving it. Please have the confidence and resolve to not give up on a problem or idea until giving it maximum effort. I guarantee that all of you will have moments of great frustration while trying to solve a problem. But, learning to persevere and think critically under stress is one of the skills I hope to develop in this course.

Do not hesitate to ask questions in class or in office hours.

Tentative Schedule

Week Topics
1 Introduction to Categorical Data, Rates and Proportions, Discrete Probability Distributions
2 Working With Rates and Proportions; The (2x2) Table
3 The (2 x 2) Table ; Sets of (2x2) Tables
4 Sets of (2 x r) and (s x 2) Tables
5 The (s x r) Table; Sets of (s x r) Tables
6 EXAM I
7 Logistic Regression
8 Logistic Regression (cont'd)
9 Logisitic Regression for Polytomous Response
10 Logisitic Regression for Polytomous Response (cont'd), Conditional Logisitc Regression
11 Poisson Regression
12 Loglinear Models
13 Introduction to Survival Analysis, Mantel-Cox Test, Categorized Time to Event Data, Life Table Estimation of Survival Rates
14 Non-parametric Methods or Generalized Estimating Equations

* This schedule is perhaps overly ambitious. Depending upon how quickly we move through the first sets of topics, there is a strong possibility that we will not be able to cover some or all of the topics listed in Weeks 13-14.

September 3: No Class (Labor Day)

October 15: No Class (Fall Break)

December 11-15: Final Exams

Back to top of page.