Syllabus -- PS 602

Advanced Quantitative Methods
Fall 2007
(Revised 8/21/07)

Robert D. Duval

Office:

Class:
Bob.Duval@mail.wvu.edu

301A Woodburn

G15 Woodburn

Phone: 293-3811 x5299
or
599-8913

Hrs: MTWThF: 11:30-12:30 Hrs: TTh 1:00-2:15

Course Description

General

This course is designed to introduce the graduate student to the principle method of empirical inquiry in the social sciences - regression analysis. The overwhelming majority of studies which test hypotheses, empirically fit models, produce predictions, or estimate policy impacts are based upon some form of linear least squares analysis. This course will cover the range of these basic linear models. The level of mathematical treatment is somewhat more advanced than PS601, but will still be moderate. The course will not dwell upon derivations. We will discuss the mathematic treatment of the topics, but memorization of complex formulae and the ability to reproduce the mathematical treatment is not part of the course objectives. The conceptual understanding of the topics covered is, however, critical to succeeding in the class. While it is desirable to have had some prior coursework in regression analysis, this course begins with the basics.

The basic organizing concepts of the course are:

Course Requirements

The major course requirements are an in-class open-book examination just after midsemester, a take-home final, and a seminar paper due the next to last week of  class. Each will count 30%. The paper topic must be arranged with me, with a preliminary outline/design approved by the 8th week of class. Need I add that it must use regression analysis? In addition, there will be a number of computer assignments. It is expected that the student is at least moderately familiar with a statistical package such as Stata, NCSS for Windows, SPSS, or SAS, and as such, the teaching of computer skills is not an objective of this course. The assignments are for illustrative purposes, and in exchange for not formally grading them, I expect you to complete them. Thus the remaining 10% will be class participation/computer exercises. I also ask that each person obtain some data of interest, including at least one time series of at least 30 observations. This data may overlap with your paper, and may prove useful during the course.
   
 

Texts for the Course 

Basic Econometrics 4/e Damodar N. Gujarati McGraw-Hill (2003) Order
Regression Diagnostics John Fox Sage QASS #79 (1991) Order
Statistics with STATA (Suggested!) Lawrence C. Hamilton Thomson (2004) (Updated for Stata 9) Order
Additional Features

A number of additional articles or books will be placed on an electronic reserve. In addition, I am beginning to assemble a methods bibliography that may prove useful for additional or supplemental material.

Course Notes

My class notes (the Powerpoint slides) are available to you. I recommend that you print about 12-18 slides ahead of the class lecture, rather than print the entire file at the beginning of the semester. [Instructions on saving paper when printing Powerpoint.] These slides will change as the course goes on. In fact, the best time to print them is about 5 minutes before class, since I will likely be revising them up until then!

Students with Special Needs


Social Justice Statement 

Course Outline with Readings

This is an initial reading list. I will likely add a few more as the semester goes on.

Week 1: Aug 21-23 Introduction, Ordinary Least Squares 
  • Gujarati, Basic Econometrics pp. 1-57
  • Edward Tufte. Data Analysis for Politics and Policy. (Recommended)
  • Haslett Statistics Made Simple. The Chapter(s) on Regression (Your text from PS601)
 
Week 2: Aug 28- Aug 30 Least Squares Estimation and Multiple Regression
 
Week 3: Sept 4-6 Assumptions of the Model: A closer look at the error terms
 
Week 4: Sept 11-13 Hypothesis Testing & Properties of Estimators
 
Week 5: Sept 18-20 Dummy Variables
   
Week 6: Sept 25-27  Multicollinearity
 
  • Gujarati, Basic Econometrics. pp. 341-379
  • Farrar, D.E. and  Glauber, R.R. (1967) "Multicollinearity in Regression Analysis: The Problem Revisited," Review of Economics and Statistics
  • pp. 92-107.  (the classic piece)
  • Examining for Multicollinearity using STATA
  • Exercise 5 Multicollinearity Exercise
Week 7:  Oct 2-4 Heteroskedasticity
 
Week 8:  Oct 9-11 Autocorrelation and Trend: I
 
Week 9: Oct 16-18 Autocorrelation and Trend: II
   
Week 10: Oct 23-25 Model Specification and Stepwise Regression
 
Week 11: Oct 30 - Nov 1 Curvilinear and Non-Linear Regression
  • Gujarati, Basic Econometrics. pp. 164-192, 563-574
  • (1) Curvilinear Regression 
    • Wonnacott & Wonnacott. Econometrics. pp 116-134.
  • (2) Intractable Non-Linear Regression and Logistic Regression
  • Sample tests: 602TST.F04.doc
  • Exercise 9: & Example

Test Review Session - Wednesday, Nov 2nd in 306E WDB (Seminar Room) at 7:00pm

Test 1 - Nov 1st

 
Week 12: Nov 6-8 Regression Diagnostics
  • Gujarati, Basic Econometrics. pp. 540-548
  • Fox, John. (1991)  Regression Diagnostics.  Sage
  • Williams, James G. (1982) "Internal Exploration of Regression Data" Political Methodology. pp: 107-123.
  • Examining regression Diagnostics Using Stata
Week 13: Nov 13-15 Dichotomous Dependent Variables - Logit & Probit (a quick look)
  • Gujarati, Basic Econometrics. pp.580-625
    Exercise 10:
Week 14: Nov 20-22 Thanksgiving Recess
 
  • Catch up!
Week 15: Nov 27- Nov 29 Simultaneous Equations & Causal Modeling
  • Gujarati, Basic Econometrics. pp. 717-758
  • Asher, Herbert. Causal Modeling (Recommended)
  • Domke, William K., Richard C. Eichenberg; and Catherine M. Kelleher. (1983). "The Illusion of Choice" American Political Science Review. Vol. 77, No. 1, (March). pp. 19-35.

Papers Due Dec 4th !
   
Week 16 Dec 4-6 Other Topics
   
Week 17 Final Exam