University of Oregon

Political Science 446/546

 

Methods for Political Analysis II

 

Professor Genie Baker

 

Class

Office

Spring 2008

TR 12-1:20

Hours: TR 1:45-3:15

     & by appt.

905 PLC

920 PLC, 346-4623

 

genie@uoregon.edu

 

 

Course Objectives: This course is more aptly titled “Statistical Methods for Political Analysis II.”  It is a continuation of PS 445/545, and will introduce the theory, practice, application, and some extensions of linear regression analysis.   The goal is to provide students with the background necessary to design and implement studies involving regression, to read and evaluate literature that uses regression methods, and to pursue more advanced training.

 

The course will balance statistical theory with applied methodology.  A central theme will be the integration of statistical analysis and substantive context.  It is hoped that students move beyond cookbook thinking and toward careful thought about the connections between formal statistical concepts and their implications in applied contexts.

 

Course prerequisites: Students are assumed to have a working knowledge of the basic statistical concepts and techniques covered in PS 445/545: descriptive statistics, elementary probability theory, sampling distributions, statistical inference and classical hypothesis testing, and bivariate regression analysis.

 

Course requirements and grading

 

·        Homework assignments, 30%: Homework assignments will include both problems to be discussed in class and homework sets to be turned in for grades.  The latter will include pencil-and-paper exercises and computer-based data exercises.  Students are welcome to think through their homework together, but should write up their own assignments, showing all work.  In my experience, students who do not keep up with the homework, learn the statistical software, etc., meet with unpleasant fates on the exams.

 

·        In-class midterm, 30%: Tuesday, May 6th.

 

·        Take-home final, 40%: Due date TBA (distributed last day of class, due during exam week).

 

Readings:

 

Recommended Texts:

 

David Freedman, Robert Pisani & Roger Purves (1998) Statistics, 3rd Ed.  New York: Norton

 

Damodar N. Gujarati (2003) Basic Econometrics (4th Edition).  New York: McGraw-Hill

 

Jack Johnston and John DiNardo (1997) Econometric Methods (Fourth Edition).  New York: McGraw-Hill.

 

Peter Kennedy (1998) A Guide to Econometrics (Fourth Edition).  Cambridge, Mass.: MIT Press.

 

G.S. Maddala (2001) Introduction to Econometrics, 3rd ed.  NY: Wiley.

 

Larry D. Schroeder, David L. Sjoquist and Paula E. Stephan (1986) Understanding Regression Analysis: An Introductory Guide. Beverly Hills, CA: Sage. 

 

   Gujarati provides the most accessible introduction to most of the material we will be covering.  All of these books cover much of the same material, though, and different readers tend to have different preferences.  As the books are very expensive, I would be prepared to look through them at the library and decide which you find most helpful.

 

Additional readings will be made available on Blackboard.

 

Software:  Homework assignments will require the use of statistical software.  I will provide a demonstration and examples using Stata, but students are free to use another package if they can do so independently and are willing to accept the risks involved.  There is a $20 lab fee for access to the Social Sciences Instructional Lab (SSIL).

 

TOPICS AND READING ASSIGNMENTS

 

Week 1.  Introduction

A. Topics:

Why we’re here

Descriptive vs. Inferential statistics

Types of Variables

Describing single variables

Summation Operators

 

Readings:

Freedman, Pisani & Purves: Chapter 4

Gujarati: Introduction (pp. 1-14), Appendix A1

Maddala, pp. 15-17

 

 

B.  Introduction to Stata, No readings.

 

 

Week 2.  Review: Describing Relationships & Basic Probability Theory

 

A.  Topics:

Describing relationships between variables:

Difference of means

Covariance, correlation & regression

Analyzing tables

Confounding

 

Readings:

Freedman, Pisani & Purves: Chapters 8-12 (Correlation and Regression)

Schroeder et. al., pp. 11-29 (Regression Analysis)

Christopher H. Achen (2000) “Warren Miller and the Future of Political Data Analysis.”  Political Analysis 8,2: 142-146  (SKIM -- The lost art of analyzing tables)

Angus Campbell, Philip E. Converse, Warren E. Miller and Donald E. Stokes (1964) The American Voter, abridged edition.  New York: Wiley.  Chapter 5: The Impact of Party Identification, pp. 67-86 (Analyzing tables, confounding)

 

 

B. Topics:

Basic Probability Theory

Mathematical Expectation

Histograms & Probability Distributions

Start inferential statistics: inferences about a single mean

 

Readings:

Maddala, 2.1-2.11

Freedman, Pisani & Purves: Chapters 3, 13, 14-18

Gujarati: Introduction (pp. 1-14), Appendix A1-A5

 

 

 

Week 3.  Inferential Statistics: Relationships Between Variables

A.  More Inferential Statistics

Topics:

The Central Limit Theorem

Difference of means tests

Chi-squared test

Tests related to regression analysis


 

Readings:

Freedman, Pisani & Purves, ch. 16-18

(Brilliant, deceptively simple presentation of the Central Limit Theorem)

Neil Weiss (2002) Introductory Statistics, 6th ed.  Boston: Addison-Wesley, chapter 10.  (Difference of Means Tests)

Freedman, Pisani & Purves, ch. 28 (Chi-Squared Tests)

Gujarati,  Appendix A6-A8 (Inference)

Gujarati,  Chapter 3 (Regression Inference)

Johnston and DiNardo: Appendix B.2-B.4

 

B.  Regression & Inference

Topics:

More Regression: Inference, Prediction & Presentation

 

Readings:

Maddala, 5.5 & 5.6

Schroeder et. al.: pp. 29-31

Gujarati: 7.8

Johnston and DiNardo: 3.2

Kennedy: 5.5

Christopher H. Achen (1977) “Measuring Representation: Perils of the Correlation Coefficient.”  American Journal of Political Science 21: 805-815.

Gary King, Michael Tomz, and Jason Wittenberg (2000) “Making the Most of Statistical Analyses: Improving Interpretation and Presentation.” American Journal of Political Science 44, 2 (April): 347-61.

 

 

Week 4.  Regression Diagnostics, Multiple Regression, Begin Matrix Algebra

A.  Regression Diagnostics

Topics:

Outliers

Nonlinearities

 

Readings:

Maddala, sections 3.5-3.9

Gujarati: Chapters 4-6 (skip or skim 4.4, skip 5.9 for now)

Johnston and DiNardo: 1.5-1.8

John Fox (1991) Regression Diagnostics.   NY: Sage,  pp. 21-40  (Analysis of outliers, influence diagnostics.)

John Fox (1991) Regression Diagnostics.   NY: Sage,  pp. 53-60 (Nonlinearities)

 

Some Applications

Robert W. Jackman ( 1974) “Political Democracy and Social Equality: A Comparative Analysis.”  American Sociological Review 39,1:29-45.

 

B. Multiple Regression, Matrix Algebra

 

Topics:

Introduction to Multiple Regression

Introduction to Matrix Algebra

 

Readings:

Maddala, 4.1-4.4

Schroeder et. al.: pp. 29-31, 32-35

Gujarati, 7.1-7.2

Gary King (1986) “How Not to Lie with Statistics: Avoiding Common Mistakes in Quantitative Political Science.”  American Journal of Political Science 30: 666-687.

Maddala, Appendix to Ch. 2

Gujarati, Appendix B

 

Week 5.  Regression in Matrix Form; Multicollinearity and Omitted Variables

 

A.  Multiple Regression in Matrix  Form

Readings:

Maddala, Appendix to Chapter 4

Gujarati, Appendix C.1-C.10

Johnston and DiNardo: 3.1, 3.3-3.5

Kennedy: 50-53, 4.1-4.4

 

B.  Specification Error: Multicollinearity and Variable Selection

 

Readings:

Maddala 4.9, ch. 7 (skim 7.5, 7.6, 7.8)

Schroeder et. al: pp. 65-72

Gujarati: Chapter 7.7, 10, 13.1-13.4

Johnston and DiNardo:4.1

Edward E. Leamer (1983) “Let’s Take the Con Out of Econometrics.”  American Economic Review 73,1: 31-43.

 

 

Week 6.   Midterm Exam, Dummy Variables and Interaction Terms

 

A.  Midterm Exam


 

B.     Specification Error: Dummy variables, interaction terms and multiplicative relationships

 

Topics:

As above

Aside on t-tests and ANOVA

 

Readings:

Maddala, 8.1-8.5

Schroeder: pp. 53-59

Gujarati: Chapter 9

Johnston and DiNardo: 4.6

Thomas Brambor, William Roberts Clark and Matt Golder (2005) “Understanding Interaction Models: Improving Empirical Analyses.”  Political Analysis, forthcoming.

Tse-min Lin and Timothy Fackler (1995) “Political Corruption and Presidential Elections.”  Journal of Politics 57,4: 971-993.

 

Week 7.  Catch-up, Heteroscedasticity

 

Readings:

Maddala, 5.1-5.4

Schroeder et. al.: pp. 75-77

Gujarati: Chapter 11

Johnston and  DiNardo: 6.1-6.3

Kennedy: 8.1-8.3

George W. Downs and David M Rocke (1979) “Interpreting Heteroscedasticity.”  American Journal of Political Science 23,4: 816-828. 

Peter Lemieux (1976) “Heteroscedasticity and Causal Inference in Political Research.”  Political Methodology 3: 287-316.

 

 

Week 8.  Autocorrelation and Autoregressive Distributed Lag Models

 

A.  Autocorrelation

Readings:

Maddala, 6.1-6.6

Schroeder et. al.: pp. 72-75

Gujarati: Chapter 12

Johnston and DiNardo: 6.4-6.8, skim 6.9

 

B.  Autoregressive and Distributed Lag Models

Readings:

Maddala, 6.7-6.10

Kennedy: pp. 263-270

Gujarati: Chapter 17 (MORE BELOW)

Johnston and DiNardo: Chapter 8

Christopher H. Achen (2000) “Why Lagged Dependent Variables Can Suppress the Explanatory Power of Other Independent Variables.”  Paper presented at the Annual Meeting of the Political Methodology Section of the American Political Science Association, UCLA, July 20-22.

Regina M. Baker (2008) “Lagged Dependent Variables and Reality: Did You Specify that Autocorrelation A Priori?” 

 

Week 9.  Simultaneous Equations Models,

Intro to Dichotomous Dependent Variables

 

A.      Simultaneous Equations Models

Readings:

Maddala, ch. 18.  Skim ch. 19, 20.

Schroeder et. al.: pp. 77-79

Gujarati: Chapters 18-20

Johnston and DiNardo: 9.4-9.6

John Bound, David A Jaeger and Regina M. Baker (1995) “Problems with Instrumental Variables Estimation when the Correlation Between the Instruments and the Endogenous Explanatory Variable is Weak.”  Journal of the American Statistical Association 90,430: 443-450.

Applications:

William K. Domke, Richard C. Eichenberg and Catherine M. Kelleher (1983) “The Illusion of Choice: Defense in Advanced Industrial Democracies.”  American Political Science Review 77,1: 19-35.

Gary C. Jacobson (1978) “The Effects of Campaign Spending in Congressional Elections.”  American Political Science Review 72, 2: 469-491.

 

B.     Dichotomous Dependent Variables: Logit and Probit

Readings:

Schroeder et. al.: pp. 79-80

Maddala 8.7-8.10

David W. Hosmer and Stanley Lemeshow (2000) Applied Logistic Regression, 2nd ed.   NY: Wiley, chapters 3 & 5.

Kennedy: pp. 233-237

Gujarati: 15.1-15.10

Johnston and DiNardo: 13.1-13.7

Michael W. Doyle and Nicholas Sambanis (2000) “International Peacebuilding: A Theoretical and Quantitative Analysis.”  American Political Science Review 94,4:779-801.

 

Week 10.  Finish Dichotomous Dependent Variables,

Introduction to Maximum Likelihood Estimation

 

Reading:

 

Jan Kmenta (1971) Elements of Econometrics.  NY: Macmillan, pp. 174-182.