CE 392R: Discrete Choice Theory and Modeling (3 credits)

Tuesdays and Thursdays, 3:30pm - 5:00pm

Econometric discrete choice analysis is an essential component of studying individual choice behavior and is used in many diverse fields to model consumer demand for commodities and services. Typical examples of the use of econometric discrete choice analysis include studying labor force participation, professional occupation, residential location, and house tenure status (owning versus renting) in the economic, geography, and regional science fields; choice of travel mode, destination and car ownership level in the travel demand field; purchase incidence and brand choice in the marketing field; and choice of marital status and number of children in sociology.

This course will provide the student with an understanding of the theory and models of individual choice behavior. The course builds on econometric modeling approaches to develop guidelines for the formulation and estimation of models of choice behavior and their use in service and product design, marketing and prediction. Practical problems (using the econometric software "LIMDEP" for estimation) will be assigned to give students familiarity with models discussed in class. These problems will focus on choice behavior in the context of travel demand analysis. However, the class instruction/discussion will be general and will focus on theory and modeling methodology for application to any discrete choice context.

Students are required to undertake a course project in choice model development and application. Students are free and will be encouraged to undertake a project that is closely related to their field of specialization.

Methods and statistics of model estimation with emphasis on maximum-likelihood estimation; Individual choice theory; binary choice models; unordered multinomial and multi-dimensional choice models; sampling theory and sample design; ordered multinomial models, aggregate prediction with choice models; introduction to advanced concepts such as accommodating unobserved population heterogeneity in choice behavior, joint stated preference and revealed preference modeling, and longitudinal choice analysis; review of state-of-the-art and future directions.

Familiarity with matrix algebra, statistical estimation and hypothesis testing, linear regression analysis, and basic differential calculus.

Classes will be conducted in lecture/discussion format. Students are expected to contribute to class discussions. Homework assignments will be given; these will provide a basis for some of the class discussion. Each student will undertake a course project in model development and/or application.

Grades will be based on homework assignments (60%), class participation (10%), and Course Project (30%).

A quantitative analysis project will be undertaken separately by each student (or pair of students, by permission). The project will consist of identifying a discrete choice analysis problem, defining the problem, and outlining an approach to the problem including issues of data acquisition, model formulation, estimation and interpretation of empirical results. The written part of the project should be clear, well-structured, and should achieve a quality consistent with that of a paper submitted for review in professional journals. The oral presentation should be clear and should succeed in sharing (with the rest of the class) the most important elements of the project in a brief period of time.

A brief abstract of the project (indicating the topic of the research, data source to be used, and methodology to be applied) and a draft outline of the project report will be due on March 12th. Instructor-student meetings will be scheduled on March 2nd and March 9th at which time students will discuss proposed projects with the instructor. The final project report is to be turned in by noon on May 12th. Student presentations are scheduled for April 28th, April 30th, May 5th and May 7th during the class period.

Web Site

The web site for the course is http://courses.utexas.edu/. Once you get to this site, log in with your UT EID and password and select CE392R from the list of courses. The web site will include the main text, course contents, course calendar, data sets to be used in the assignments, additional datasets that may be used in projects, the LIMDEP software, and several miscellaneous notes/links.

See “Calendar of Course Events” at the web site. Note that an additional class period will be held on February 2nd, February 9th, February 23rd, and March 23rd. No classes will be held on March 31st, April 16th, April 21st, and April 23rd (these class periods will be open for discussions regarding student projects). The front-end “loading” of the course is being done for several reasons. First, much background material will need to be covered in the beginning and additional classes early on will get us over this background “hump” early in the semester. Second, it will ensure a more uniformly-spaced distribution of the assignments. Third, the early coverage of material will allow students to start working on their projects early. Finally, the intent is also to provide students with more time toward the end of the semester to work on their projects.

Meeting Time and Location

Tuesdays and Thursdays, 3:30 pm - 5:00 pm, ECJ 7.208.

Office Hours:

Mondays 3:30 – 5:00 pm and Tuesdays 12:30 – 2:00 pm, ECJ 6.810.

Other General Information

The University of Texas at Austin provides, upon request, appropriate academic adjustments for qualified students with disabilities. Any student with a documented disability (physical or cognitive) who requires academic accommodations should contact the Services for Students with Disabilities area of the Office of the Dean of Students at 471-6259 as soon as possible to request an official letter outlining authorized accommodations. For more information, contact that Office, or TDD at 471-4641, or the College of Engineering Director of Students with Disabilities at 471-4321.

Anticipated Homework Assignments (with expected dates of distribution, submission and return)

2. Multinomial Logit Estimation and Interpretation

Distribution date: Feb. 23

Due date: March 5

Return date: March 10

3. Multinomial Logit Specification Refinement

Distribution date: March 10

Due date: March 23

Return date: April 2

4. Nested Logit Models - Specification and Estimation

Distribution date: April 2

Due date: April 9

Return date: April 14

5. Ordered Response Model Estimation

Distribution date: April 14

Due date: April 21

Return date: April 28

Readings are assigned to supplement lecture material. The main text for the course is A Self-Instructing Course (SIC) in Mode Choice Modeling: Multinomial and Nested Logit Models, prepared for U.S. Department of Transportation, Federal Transit Administration, by Frank S. Koppelman and Chanda Bhat, 2006. Readings will also be assigned in Discrete Choice Analysis: Theory and Application to Travel Demand by Moshe Ben-Akiva and Steven R. Lerman. The main text (SIC) will be made available at the web site for the course. Readings from Ben-Akiva and Lerman (BL) will be distributed in class. Occasionally, readings will include journal papers which will be provided in class. A detailed listing of topics and readings is provided below.