|
|
Pols 602-600: Spring 2004 |
||
|
B.
Dan Wood |
Time: |
|
|
|
Office: 2098 |
Room: 2064 |
|
|
|
Phone: 845-1610 |
Office Hours: |
|
|
Purpose-
This is a course in doing quantitative empirical political science
research. Good quantitative research is
an essential element of theory building for the political scientist.
The class will teach basic quantitative skills
emphasizing the regression framework.
Students should have some prior background, including descriptive
statistics, probability, inference, and hypothesis testing. Nevertheless, we will review basic statistics
for about the first three weeks, with the remainder of the semester devoted to
the regression model. Methodological and
statistical theory will be discussed, but the emphasis will be on doing.
We will use the statistical package R as the primary
software for the course. Note that this is the public domain version of S-Plus
and can be freely downloaded to your machines. I have also established links to
the textbook exercises worked through in STATA by the UCLA statistical support
group. I encourage students to use both programs, but it is expected that any
exam or homework problems requiring software will be completed in R.
Homework includes problems from the primary text, as
well as computer exercises from the companion statistics guide. Problem sets
will normally be assigned one week and due the next. You are encouraged to work together in doing
the problems. However, you should not
simply copy the work of others or you will find yourselves in trouble on the
examinations. Examinations will be of the
take home type to enable sufficient time to demonstrate competency on the
associated problems.
Course Grade- The final grade will be based on four components. Weekly
homework assignments will count one fourth of the grade. Homework must be
handed in on time or no credit will be given. A midsemester
and end of semester examination will each count one fourth of the grade. The
examinations will test your skill in doing and understanding statistical
methods. The remaining one fourth will be based on an empirical paper which
utilizes one or more of the methods taught in this course. You should discuss
with me the paper topic sometime before the midsemester
examination. The paper is due on the last class day before the final
examination.
Prerequisites- Prior to
entering the course you should have reviewed the basic principles of
probability, and also gained some background in basic linear algebra and
calculus. Good sources for these materials are the following.
Recommended texts-
Dowling, Edward T. 2001. Introduction
to Mathematical Economics, Third Edition.
Spiegel,
Required texts-
Fox, John. 1997. Applied
Regression Analysis, Linear Models, and Related Methods. Sage Publications.
Fox, John. 2002. An
R and S-Plus Companion to Applied Regression. Sage Publications.
Fox (1997) is referred to
below as Applied Regression.
Fox (2002) is referred to
below as Companion.
Go to the Applied Regression textbook website by
clicking here.
Go to the Companion textbook website by clicking here.
Get the textbook data in ascii zipped format by clicking here.
A brief tutorial on STATA is here.
Get instructions to download
the STATA data for the textbook exercises here.
Topics,
Following is the order of the subjects taught in this course.
Note that there are only 12 headings, which implies that some may be given
multiple week treatments, while others may receive less than a week. I do not
attach dates to allow flexibility in timing.
1. Introduction- Applied Regression,
chapters 1 and 2; STATA 2;
Companion, chapters 1 and 2; R Ch1-script.txt; R Ch2-script.txt.
2. Review of basic
statistical concepts- Applied Regression,
chapters 3 and 4 and Appendix D
; STATA 3,. STATA 4; Companion, chapter 3; R Ch3-script.txt.
3. Linear Least Squares Regression- Applied
Regression, chapter 5; STATA 5; Companion, chapter 4.1; R Ch4-script.txt.
4. Statistical Inference,
5. Dummy Variable Regression- Applied
Regression, chapter 7; STATA
7; Companion, chapter 4.2.
6. Analysis of Variance- Applied
Regression, chapter 8; STATA
8; Companion, chapters 4.3-4.4.
7. Diagnosing Unusual and Influential Data- Applied
Regression, chapter 11; STATA
11; Companion, chapter 6.1; R Ch6-script.txt.
8. Diagnosing Nonlinearity, Nonconstant Error
Variance, and Non-Normality- Applied
Regression, chapter 12; STATA 12; Companion, chapters 6.2-6.4.
9. Diagnosing Collinearity and Variable Selection- Applied Regression, chapter 13; STATA 13; Companion, chapter 6.5.
10. Diagnosing Serial Correlation- Applied
Regression, chapters 14.1; STATA 14;
Web
Appendix-Time-Series Regression and Generalized Least Squares ; R
Script for Web Appendix .
11. Logit and Probit
Analysis for Dichotomous Variables- Applied
Regression, chapter 15.1; STATA 15; Companion, chapter 5; R Ch5-script.txt
12. Logit and Probit
Analysis for Polytomous Variables- Applied Regression, chapter 15.2.