Short Programs

Discrete Choice Analysis: Predicting Demand and Market Shares [14.61s]

Date: June 11-15, 2012 | Tuition: $3,900 | Continuing Education Units (CEUs): 2.7
Application Deadline »

Overview  |  Learning Objectives  |  Who Should Attend  |  Outline of the Program  |  Schedule  | 
Included in Tuition  |  Participants' Comments  |  About the Presenters  |  Scholarships  |  Location  |  Updates

Overview

This one-week program undertakes an in-depth study of discrete choice models (logit, nested logit, generalized extreme value, probit, logit mixtures), data collection, specification, estimation, statistical testing, forecasting, and application. The covered topics include analysis of revealed and stated preferences data, sampling, and simulation-based estimation, discrete panel data, Bayesian estimation, discrete-continuous models, menu choice, and models with latent variables. The course includes practical application sessions where participants will be provided with discrete choice software to learn how to estimate and test discrete choice models taught in lecture using real databases, and gain hands on experience in using new discrete choice techniques for practical applications. By examining actual case studies of discrete choice methods, students will be familiarized with problems of model formulation, testing, and forecasting.

Discrete choice models are widely used for the analysis of individual choice behavior and can be applied to choice problems in many fields such as economics, environmental management, urban planning, etc. For example, discrete choice modeling is used in marketing research to guide product positioning, pricing, product concept testing, and many other areas of strategic and tactical interest. Recent applications to predict changes in demand and market shares include areas such as choice of travel mode, coffee brand, telephone service, soft drinks and other foods, and choice of durables such as automobiles, air conditioners, and houses.

Learning Objectives

• Understand discrete choice models and their applications.
• Learn to apply new discrete choice techniques.
• Understand problems of data collection, model formulation, testing, and forecasting, as learned through case studies of discrete choice methods.
• Utilize commonly available software to estimate and test discrete choice models from real data bases.
• Evaluate theories of choice, random utility models, probabilistic choice models, alternative model formulations, statistical estimation procedures appropriate for alternative data sources, currently available computer software, tests of validity, and forecasting procedures.

Who Should Attend

This program is intended for academics and professionals interested in learning new discrete choice techniques and how to predict choice and forecast demand. They will gain hands-on experience in applying discrete choice software in real-world case studies. Participants need only have a basic working knowledge of statistical methods.

Outline of the Program

The course consists of a series of lectures and labs that develop discrete choice concepts and techniques, and demonstrate their applications. The labs offer hands-on experience in applying the material covered in the lectures using discrete choice software and real-world data sets.

Topics Covered Include the Following:

Theories of choice, random utility models, probabilistic choice models, alternative model formulations, statistical estimation procedures appropriate for alternative data sources, currently available computer software, tests of validity, forecasting procedures, and examples of empirical applications.

More specifically, the following subjects will be addressed during the course:

• Choice Behavior
• Binary Choice Models
• Stated Preference Surveys
• Multinomial Choice Models: Properties of Probit, Logit and Discriminant Analysis
• Specification and Estimation of Discrete Choice Models
• Statistical Tests of Discrete Choice Models
• Forecasting and Micro-Simulation
• Nested Logit Models
• Multi-variate (Generalized) Extreme Value Models
• Mixture Models, such as Logit Kernel or Mixed Logit
• Simulation-based Estimation
• Bayesian Estimation
• Discrete Panel Data
• Combining Revealed and Stated Preferences
• Sampling Strategies for Discrete Choice Analysis
• Joint Discrete / Continuous Models
• Choice from a Menu
• Choice Models with Latent Variables

Course schedule, registration times and Special Events

View (tentative) 2012 Course Schedule

Class runs 9:30 am - 5:00 pm every day.

Registration is on Monday morning from 8:45 - 9:15 am.

Special events include a reception for course participants and faculty on Monday night and a dinner on Thursday evening. All evening activities are included in tuition.

Please note that laptops are required for this course.

Included in Tuition

On the first day of class, participants will receive a copy of Discrete Choice Analysis by Moshe Ben-Akiva and Steven R. Lerman (MIT Press) and a set of additional reading materials covering recent developments and new applications.

Participants' Comments

"Discrete choice analysis is one of the most valuable new tools available to marketers interested in understanding and predicting consumers' choices ... It goes beyond current textbook treatments of discrete choice analysis with discussions of state-of-the-art developments in the area and experimental applications." Colleen Collins-Dodd, Assistant Professor of Marketing, Simon Fraser University, British Columbia

"The MIT Summer Session in Discrete Choice Analysis is without doubt the most valuable course on the topic offered anywhere. Being taught by the very persons who have developed and implemented the field, the course strikes a perfect balance between theory and practice. To me, it was an intellectual thrill." Lasse Fridstrom, Research Economist, Institute of Transport Economics, Norway

"Discrete choice analysis is an effective way to evaluate a new service and forecast future demand. I have applied it to the estimation of userÆs preference ... in order to develop new telecommunication services." Akiya Inoue, Senior Research Engineer, NTT Telecommunications Networks Laboratories

"This course is an excellent review and introduction to discrete analysis theory, ... provides opportunities for hands-on applications. Further, the instructors are major contributors to this area. I highly recommend it." Terence J. La Du, Member of Technical Staff, Bellcore

About the Presenters

The course director and principal lecturer is Moshe Ben-Akiva. Lab sessions are led by Dr. Carmine Gioia (Research Scholar/Visiting Scientist, MIT & CEO Choice Science).

Scholarships

One full-tuition scholarship will be awarded to an outstanding doctoral student. Half-tuition scholarships are also available for doctoral students. Scholarship applications must include a CV and a letter stating the relevance of the course to the applicant’s research.

Please contact Katie Rosa at krosa@mit.edu with any questions. The deadline to apply is
May 7, 2012.

Discounts for Faculty
In addition, a limited number of partial-tuition scholarships are available for teaching faculty, rank of instructor or higher, at other educational institutions. You may submit a scholarship request by filling out a Scholarship Request Form after your application to the course has been submitted. Please note that these scholarships are only for tuition and do not cover travel, lodging, or other expenses associated with the course.

If you have any questions please contact the Short Programs office.

Location

This course takes place on the MIT campus in Cambridge, Massachusetts. We can also offer this course for groups of employees at your location. Please contact the Short Programs office for further details.

Updates

Please note that laptops are required for the afternoon lab sessions.