Time: 3:30-4:45 p.m., MW, Bryan 105

Instructor: Dr. Scott Richter

Office: Bryan 389

E-mail: sjricht2@uncg.eduPhone: 256-1123

Office Hours: MW 2:30-3:30; TTh 2:00-3:30. Other times are available by appointment.

Web Page: http://www.uncg.edu/~sjricht2

Required Text: Agresti, An Introduction to Categorical Data Analysis, John Wiley & Sons, 1996.

Course Description: Methods for analyzing dichotomous, multinomial and ordinal responses. Measures of association; inference for proportions and contingency tables; generalized linear models including logistic regression and loglinear models. Computations will be illustrated using SAS and possibly other statistical software.

Prerequisites: STA 662 or permission of instructor. The student should have some exposure to regression analysis and some practice using statistical software, since most of this course deals with extensions of regression to handling categorical response variables.

Homework: Assignments consisting of exercises from the text and occasionally supplementary exercises will be assigned regularly, collected and graded. Due dates will be announced in class. These will be designed to provide practice and to help synthesize readings, class discussions, and lectures. Students are encouraged to work together in teams to help each other in understanding the course material and completing the homework problems, but each student should write up their own solutions.
*Late assignments will not generally be accepted without prior arrangement, and will receive a score of zero.

Exams/Project: There will three take-home exams. The test due dates will be announced in class. Students have the option to complete a project instead of the third exam.
*Late exams will not be accepted.

Project: The project will consist of a written report. This report should present a statistical analysis based on modeling a data set containing a categorical response variable. If you have any questions about what to do for your project, please talk to me about it (if you do not have ready access to a dataset, you might create one from the General Social Survey at the web site, http://www.icpsr.umich.edu/gss/, or from other online sources of data). The report should include sections addressing: (1) description of the data and their source; (2) statement of questions to be addressed; (3) specification of application of models to data and model-checking; (5) interpretations of results of model fitting; (6) summary and conclusions. Include, in a separate appendix, copies of relevant parts of your computer printouts and (if possible) the data. The project is due on the same day as the third exam.

Determination of course grade:

Assignments: 40% of course grade.
Exam 1: 20% of course grade.
Exam 2: 20% of course grade.
Exam 3/Project: 20% of course grade.

Grading scale:
 

Overall average	Grade
90 or above	A
80-89	B
70-79	C
Below 70	F

 

Academic Integrity:
Students are encouraged to discuss solutions to assignments, but each student is expected to write up his or her solutions independently. Copying other people's work is plagiarism and is an Honor Code violation. You are responsible for knowing and abiding by the UNCG Honor Code.