EPPS 6316.

502: Advanced Regression Analysis in the Social Sciences

Spring 2011, Monday, 7:00pm to 9:45pm, SLC2.202
Preliminary Syllabus, December 3, 2010

Professor: Paul A. Jargowsky, GR 3.202; 972-883-2992; jargo@utdallas.edu

Office Hours: Thursday, 3:00pm to 6:00pm, and by appointment.

TA: TBA

Overview. This is a course on empirical methods that are useful in the investigation of
hypotheses in the social sciences and the analysis of public policies and programs. The bulk of
the course is a detailed examination of the bivariate and multiple regression models estimated
using Ordinary Least Squares (OLS), with an emphasis on constructing regression models to test
social and economic hypothesis. Several special topics in regression analysis are addressed as
well, including violations of OLS assumptions, the use of dummy variables, fixed effects models,
and path analysis. Throughout, examples are drawn from the literature so students can see the
models and methods in action.

Prerequisite. POEC 6313 or 7313, or the equivalent knowledge. In other words, you are
expected to be familiar with descriptive and inferential statistics and basic concepts of
probability. You need not have any previous exposure to regression analysis. If you wish to
review probability and statistics before or during the class, I recommend Mendenhall, Beaver,
and Beaver, Introduction to Probability and Statistics, 11th ed. (chapters 4-6 on probability and
chapters 7-10, especially 7 and 9, on sampling theory and hypothesis testing). A solid
foundation in basic mathematics and college algebra is essential, but you do not need to know
calculus or matrix algebra.

Textbooks. Damodar M. Gujarati and Dawn C. Porter, Basic Econometrics, 5th ed. New York:
McGraw Hill Irwin, 2009. Previous editions are acceptable, but you are responsible to identify
the corresponding chapters and pages.

Course Web Site. The course web site on eLearning serves several purposes. Mainly, it is a
place to download lecture notes, assignments, practice questions, etc. Second, there is a
discussion board where you can post questions and/or comments. Third, you can use it to send
email to some or all of the other students in the class, arrange study groups, etc.

Requirements and Grading. Grades will be determined as follows:

8 of 10 Problem Sets 30 percent (two lowest scores dropped)
Midterm 30 percent
Final Exam 30 percent
Empirical Paper 10 percent

We will often discuss the exercises in class after they are handed in. For that reason, late
1
assignments can not be accepted. There will be a short problem set due almost every week.
However, the lowest two problem set scores will be dropped, so you can miss a few with no
grading consequences. If you are going to miss class, you still need to submit your assignment
(as an email attachment, by FAX, pony express, or whatever) before 4:00 pm on the day it is due.

Students are encouraged to form study groups and to collaborate on the exercises, particularly the
estimation of models on the computer. However, each and every student should write up his or
her own answer to ensure they have internalized the group discussions.

Attendance. Attendance is entirely optional. Having said that, the material covered in the
lectures is essential to passing the course. You are responsible for all material covered in the
lectures whether or not it is in the textbook. You are also responsible for all announcements
made in class, whether or not you attend. Thus, the best policy is to attend every class. If you
have to miss a class, due to illness or other unavoidable conflict, try to borrow a fellow student’s
notes.

Academic Honesty Notice. There can be absolutely no collaboration of any kind

on examinations. This includes communication of any kind, sharing of books or
notes, or sharing of calculators. The penalty I request in cases of academic
dishonesty is a grade of zero (0) for the examination or assignment in question,
although the final penalty is determined by the responsible university official. A
grade of zero on an exam could lead to failing the course. In addition, the
incident is reported to the Dean of Students for disposition and becomes a part of
the student’s record.

Computing. The official software for the course is Stata, Version 11. Stata 9 and 10 will work
but not as well. Stata 11 is available for student use in the Social Sciences Computer Lab (GR
3.206. However, if you wish to purchase a copy for your own computer, you may purchase the
software at an educational discount through the UTD Social Sciences Grad Plan.

Schedule and Readings. The schedule and readings are posted on the course website in e-
Learning. The tentative schedule is shown below, but always consult the online version for
updates. Note the following dates:

March 7 – Midterm Exam

May 9 – Final Exam

Early or late examinations cannot be given except in extreme circumstances, so plan your
schedule accordingly!

2
 EPPS 6316 Preliminary Reading List
Draft, December 6, 2010
Always consult the live version on eLearning for the latest schedule!
Class/Date Due Major Topics, Assigned Readings, Events
Review as Mathematical Background: Probability and Statistics
necessary Gujarati and Porter, Basic Econometrics, Appendix A
prior to Jargowsky and Yang, Descriptive and Inferential Statistics
first class
Jan. 10 Introduction and Course Overview
Why is Regression Necessary?
Bivariate Regression: Theory and Estimation
Gujarati and Porter, Basic Econometrics, Chapters 1-3

Jan. 17 Martin Luther King Holiday – University Closed

Jan. 24 Problem Bivariate Regression: Inference

Set 1 Hypothesis Testing
Goodness of Fit
Gujarati and Porter, Basic Econometrics, Chapters 4-5
Jan. 31 Problem Alternative Functional Forms
Set 2 Gujarati and Porter, Basic Econometrics, Chapter 6
Feb. 7 Problem The Three Variable Regression Model
Set 3 The K Variable Regression Model
Polynomial Specifications
Gujarati and Porter, Basic Econometrics, Chapter 7
Omitted Variable Bias
Jargowsky, Omitted Variable Bias
Feb. 14 Problem Inference in the Multiple Regression Model
Set 4 Gujarati and Porter, Basic Econometrics, Chapters 8
Jargowsky, Using the F Statistic to Test Hypotheses
Feb. 21 Problem Indicator Variables (“Dummies”)
Set 5 Specification of Interactions
Comparing Effects Across Subgroups
Gujarati and Porter, Basic Econometrics, Chapter 9
Devaney, Barbara, Linda Bileimer, and Jennifer Schore. 1992. "Medicaid
Costs and Birth Outcomes: The Effect of Prenatal WIC Participation
and the Use of Prenatal Care." Journal of Policy Analysis and
Management 11:573-92.

3
Feb. 28 Problem Applications and Review for Midterm
Set 6 Jeanne Brooks-Gunn, Greg J. Duncan, Pamela Kato Klebanov and Naomi
Sealand. 1993. "Do Neighborhoods Influence Child and Adolescent
Development," American Journal of Sociology, 99: 353-395.
Alan B. Krueger, “How Computers Have Changed the Wage Structure:
Evidence from Microdata, 1984-1989,” Quarterly Journal of
Economics 108: 33-60.
Review for Midterm
Mar. 7 MIDTERM EXAM
Open note, open book, no laptop computers
Bring a calculator

Mar. 14 Spring Break – University Closed

Mar. 21 Review of Midterm

Multicollinearity
Gujarati and Porter, Basic Econometrics, Chapter 10
(Optional Topic) The OLS Model in Matrix Notation
Jargowsky, “Matrix Notation and OLS”
Gujarati and Porter, Basic Econometrics, Appendices B and C
Mar. 28 Problem Heteroskedasticity and Autocorrelation
Set 7 Weighted Least Squares
Generalized Least Squares
Gujarati and Porter, Basic Econometrics, Chapters 11 and 12
Apr. 4 Problem Instrumental Variables
Set 8 Jargowsky, Instrumental Variables.
Model Specification Issues
Gujarati and Porter, Basic Econometrics, Chapter 13
Path Analysis
Duane F. Alwin and Robert M. Hauser. “The Decomposition of Effects in
Path Analysis,” American Sociological Review 40: 37-47 (February
1975).
Apr. 11 Problem Qualitative Response Models: Logit and Probit
Set 9 Gujarati and Porter, Basic Econometrics, Chapters 15
(Optional Topic) Non-Linear Regression Models
Gujarati and Porter, Basic Econometrics, Chapters 14
Apr. 18 Problem Panel Data Regression Models
Set 10 Gujarati and Porter, Basic Econometrics, Chapter 16
Apr. 25 Simultaneous Equations and the Identification Problem
Gujarati and Porter, Basic Econometrics, Chapters 18-20
May 2 Short Student Presentations
Paper Review for Final
May 9 FINAL EXAM
Open note, open book, no laptop computers
Bring a calculator