William Greene
Stern School of Business
New York University
Discrete Choice
Modeling
University of Southern Denmark
December 16-18, 2013
Professor
Professor William. Greene
e-mail: wgreene@stern.nyu.edu
Home Page: http://people.stern.nyu.edu/wgreene
Abstract
This course will survey
techniques used in discrete choice modeling in cross section, panel and stated
choice data. Discrete choice models have become an essential tool in modeling
individual behavior. The techniques are used in all social sciences, health
economics, medical research, marketing research, transport research, and in a
constellation of other disciplines. This course will examine a large number of
models and techniques used in these studies. We will begin with a brief review
of regression modeling concepts, then turn to the fundamental building block in
discrete choice modeling, the binary choice model. The remainder of the course
will be devoted to multinomial choice models of the sort used, e.g., in
modeling brand choice in marketing, travel mode choice in transport, and a huge
variety of applications in the social and behavioral sciences.
The course will include lectures
that develop the relevant theory and extensive practical, laboratory
applications. Emphasis in the laboratory sessions will be on estimation of
discrete choice models and using them to describe behavior and to predict
discrete outcomes. Course participants will apply the techniques on their own
computers using the NLOGIT computer
program and several real data sets that have been used in applications already
in the literature.
Prerequisites
Prior knowledge
is assumed to include calculus at the level assumed in the first year of a
Ph.D. program in economics and a course in econometrics at the beginning Ph.D.
level out of a textbook such as Greene, W., Econometric
Analysis, 7th edition. Familiarity with NLOGIT will be helpful, but is not necessary. A useful reference
book for the course is the primer Applied
Choice Analysis by David Hensher, John Rose and William Greene (Cambridge
University Press, 2005).
Students in this
course will obtain background in both the theory and methods of estimation for
discrete choice modeling. This course will provide a gateway to the
professional literature as well as practical applications of the methods at the
level of the contemporary research in the field. Emphasis in the course is on
applications of methods in health economics. Derivations of, e.g., asymptotic
properties of estimators, and theoretical fine points, such as the implications
of different types of independence assumptions in panel data models are left
for more advanced treatments.
Course Outline Left click to open. Right click to download. (This is a PDF
file.)
This is a course in econometric
analysis of discrete data. The course will consist of discussions and
laboratory sessions which will apply the techniques to live data sets and some
time devoted to topics, discussions and laboratory work on student projects.
Discussions will cover the topics listed below. Lab sessions will apply the
methods discussed in the preceding sessions. Practicals will consist of
directed exercises and student assignments to be completed singly or in groups.
No specific textbook is assigned
for the course. Some of the presentation will be based on Econometric Analysis, 7th ed., by Greene, W. (Prentice
Hall, 2012). 7 chapters are included with the course materials:
Left click to open. Right click to download. (These are PDF
files.)
Greene 11-Panel data methods
Greene 12-Estimation methods
Greene 14-Maximum likelihood estimation
Greene 15-Simulation based estimation and inference
Greene 17-Discrete choice models
Greene
18-Count data models
Greene 19-Censoring and truncation
The received literature on
discrete choice models is vast - one could easily compose a list of thousands
of articles. Your course materials include a small handful of articles. The
first set are methodological papers that focus on particular techniques. The
second set are recent applications in health economics that use the techniques
we will discuss in our course:
Methodology Left click to open. Right click to
download. (These are PDF files.)
Economic Choices.
Mixed Logit Models.
Fixed Effects Models.
Discrete Choice Models.
Methods Left click to open. Right click to
download. (These are PDF files.)
Correcting estimated standard errors in the presence
of clustering.
Modeling endogeneity in nonlinear
models.
Interaction effects in nonlinear models.
Modeling dynamic effects in nonlinear models.
Health Economics Applications Left click to open. Right click to
download. (These are PDF files.)
Bago d Uva and Jones Latent Class Health Care
Models
Finkelstein et al. Oregon Health
Insurance Experiment
Scott et al. Recursive Bivariate Probit
Analysis of Quality of Diabetes Care
Lagarde Latent Class Logit
Analysis of Infant Care
Johnston, Schurer and Shields
Dynamic Ordered Choice Model
Jones and Schurer Dynamic
Ordered Choice Model
Riphahn, Wambach, Million Mixed Poisson
Models for Health Care Utilization
Contoyannis, Rice,
Jones Dynamic Ordered Choice Model of Health Satisfaction
Winkelmann Econometric Exploration
of Count Models of Health Care
Van Ophem Extension of Winkelmann Hurdle Model
Laporte Quantile Regression
Gannon, Dynamic
Probit Model
Christensen and Kallestrup-Lamb
The received literature on
discrete choice models is vast - one could easily compose a list of thousands
of articles. The following are two classics and an interesting survey:
Economic Choices, American Economic Review, McFadden, D. (2001). McFadden�s Nobel
Prize lecture.
Mixed MNL Models for Discrete Response,
McFadden, D. and Train, K., Journal of
Applied Econometrics, 2000.
A
Control Function Approach to Endogeneity in Consumer Choice Models, Petrin, A
and K. Train, Journal of Marketing
Research, 2009.
I. Class Notes:
Course description and overview. Notes for use during the class
sessions.
Left click to open. Right click to download. (These are
.pptx files)
Notes
0. Outline
and Introduction
Notes
1. Methodology and Binary Choice
Notes
2. Multinomial Choice Models
Notes
3. Nested Logit and Multinomial Probit
Notes
4. Modeling Heterogeneity
Notes
5. Latent Class Models
Notes
6. Mixed Logit Models
Notes
7. Stated Preference
Notes
8. Hybrid
Choice Models
II. Scripted
Exercises: These are scripted
applications that illustrate using NLOGIT to carry out some of the
computations discuss in class. There are two parts to each,
the .pdf file for the exercise and the .lim file that contains
the NLOGIT commands.
You can first load the indicated dat sets, open the command files, then execute
the script
and follow along as the computations are completed.
Left click to open. Right click to download. (These are .pdf
files)
Scripted Exercise 1: Basic Regression, Binary Choice (NLOGIT Commands for Exercise 1)
Scripted Exercise 2: Binary Choice: Estimation and
Testing, Panel Data (NLOGIT Commands for Exercise 2)
Scripted Exercise 3: Binary Choice Modeling with
Heterogeneity (NLOGIT Commands for Exercise 3)
Scripted Exercise 4: Ordered Choice and Count Data Models
(NLOGIT Commands for Exercise 4)
Scripted Exercise 5: Multinomial Choice Models, Stated
Preferences (NLOGIT Commands for Exercise 5)
III. Assignments:
These are exercises for
the student to do on their own.
Left click to open. Right click to download. (These are .pdf
files)
Exercise 1: Basic Regression, Binary Choice
Exercise 2: Binary Choice with Panel
Data; Delta Method, Bootstrapping
Exercise 3: Count Data;
Modeling with Heterogeneity, Latent Class and Mixed Models
Exercise 4: Multinomial Choice
Multinomial
Choice Exercises
1.
Binary Choice Models: (Assignment-pdf),
(Commands-lim),
(Data-lpj)
2.
Multinomial Choice: (Assignment-pdf),
(Commands-Part1-lim)
(Commands-Part2-lim)
(Data-lpj)
3.
Advanced Multinomial Choice (Assignment-pdf),
(Commands-lim)
(Data-lpj)
(Description
of Data)
4.
Stated Preference Multinomial Choice (Assignment-pdf),
(Commands-lim)
(Data-lpj)
IV. Lab Notes: These
are Powerpoint slide presentations (pptx) that show how to use NLOGIT.
Left click to open. Right click to download. These are pptx
files
Getting started tutorial
Lab
1: Basic Operation
Lab
2: Binary Choice
Lab
3: Testing Hypotheses
Lab 4:
Panel Data
Lab
5: Useful Tools: Simulation, Partial Effects, Bootstrapping
Lab
6: Random Parameters and Latent Class Models
Lab
7: Multinomial Choice Models
Lab
8: Mixed Logit Models
V. Data Sets:
These datasets are provided in two forms, NLOGIT project (.lpj)
files and portable CSV files.
Left click to open. Right click to download. CSV files will open
Excel.
Small
Demonstration Income/Education Data, 14 observations (Filename=IncomeData) (lpj) (csv)
Combined
Travel Mode and Brand Choices Data, 12800 observations (Filename=mnc) (lpj) (csv)
Health
Economics, Small Subset of GSOEP Data, 2039 observations (Filename=HealthData)
(lpj) (csv)
Manufacturing
Innovation Data, 6350 observations (Filename=panelprobit) (lpj) (csv)
Multinomial
Choice Stated Preference Experiment, 9408 observations (Filename=sprp) (lpj) (csv)
Labor
Supply Data, 753 observations (Filename=labor) (lpj) (csv)
Health
Care Panel Data 27326 observations (Filename=healthcare) (lpj) (csv)
Southern
California Fishing Data (long form), 4728 observations
(Filename=fishing_data(NL)) (lpj) (csv)
Rand
Study Data, 19339 observations (Filename=rand_data_2012) (lpj) (csv)
General
Practitioner Visits Data, 342 observations
(Filename=gpvisits_panel_shp_data_2011) (lpj) (csv)
Kenneth
Train California Public Utility Survey, 17232 rows, 4308 obs.
(Filename=TrainCalUtilitySurvey) (lpj)
(csv)
VI. NLOGIT Software: This
section contains a brief introduction and two manuals: The short introduction
is a getting started guide. The LIMDEP manual explains the basics of
using LIMDEP and NLOGIT. (LIMDEP is
embedded in NLOGIT). The NLOGIT manual contains descriptions of
how to use the special features for discrete choice modeling with NLOGIT (as well as some additional material
on other discrete choice models that are also contained in LIMDEP). The setup file contains an installation kit for installing
a copy of NLOGIT made specially for
this course on your own computer. You should download the setup file to your
own computer and execute it there, rather than launching it from your web
browser.
Left click to open. Right click to download.
Quickstart Introduction to NLOGIT (Command script file to use with Quickstart)
LIMDEP Student User Manual
NLOGIT Student User Manual
NLOGIT Software Setup for Installing NLOGIT
on Your Computer