Political Science 552
Professor Herbert Kritzer
II, 2003-04

NOTE: This syllabus is subject to minor changes both before and during the semester.


This course is an intermediate level statistics course concentrating on linear regression. The first six weeks will be devoted to bivariate regression (including matrix algebra); the balance of the semester covers multiple regression and other applications of the general linear model, including logistic regression, dummy variables, and causal models.

The class is scheduled to meet from 9-10:45 am Tuesday and Thursdays. Approximately 30 minutes of each session is intended as time for discussing problem set issues and math-related topics.

The prerequisite for the course is introductory statistics (e.g., Political Science 551). In addition students will need to have or obtain some familiarity with computer based data analysis using a statistical package like Stata, SPSS or SAS; students are strongly encouraged to also enroll in Political Science 553, a one credit course meeting Fridays 11-1 that will teach students how to use Stata. Students will be able access Stata (or SPSS) either from the machines in 313 North Hall or at the the Social Science Micro Lab (SSML) in the Social Sciences Building (3rd floor). All students will be set up with accounts on the Political Science Server.

The course requirements include a research design (7-10 pages), due March 25, a research paper (15-25 pages), midterm and final examinations, and six to eight problem sets.

The research paper may be done in conjunction with another course, but must be a piece of original research; it should in some way apply the material from this course. In format and substance, the paper should be modeled after a empirical research article in a general political science journal (e.g., APSR, JOP, AJPS, etc.) or a specialized journal (e.g., ISQ, ASQ, LSQ, LSR, etc.). The paper will be evaluated in terms of the overall research (25%), the statistical analysis (25%), the presentation of the research and analysis (25%), and writing (25%).

The following texts are required:

Many students find it helpful to have a second reading for the material. In the assignments below, I have included "alternative" readings from two other regression texts:

I have also included a number of required and "recommended" readings drawn from the the Sage "Quantitative Applications" series. Because these short monographs are excessively overpriced, I have put them on reserve at Helen C. White:

Lastly, a number of the assignments include examples of published research that employs the techniques that we will be studying. These are all from journals that should be available in the Dean Room (Room 313, North Hall). As noted above, I will arrange to have them available through the electronic reserves system, and a reading packet will be available from the Social Science Copy Center.


Jan. 20

Review of Inferential Statistics, and Principles of Estimation

    Required reading:
  • Kutner, Appendix A
Problem Set 1
Jan. 27 Bivariate Regression: The Model and Estimation
    Alternate reading:
  • Wonnacott & Wonnacott, 1-29, 152-162
  • Hilton, Ch. 2
Problem Set 2
Feb. 3

Bivariate Regression: Inference, Correlation and Analysis of Variance

    Alternate reading:
  • Wonnacott & Wonnacott, 29-74, 163-173
  • Hilton, Ch. 3
Feb. 10 Bivariate Regression: Diagnosing and Resolving Problems
    Alternate reading:
  • Wonnacott & Wonnacott, 208-238
    Recommended reading:
  • Hilton, Ch. 4 (section on heteroscedasticity)
Problem Set 3

Feb. 17


Bivariate Regression: Diagnosing and Resolving Problems, continued
Matrix Algebra

    Alternate reading:
  • Hilton, Ch. 5
Feb. 24 Bivariate Regression: The Matrix Algebra Approach and Introduction to Multiple Regression
    Required reading:
  • Kutner, Ch. 5-6
    Alternate reading:
  • Hilton, Ch. 6
  • Wonnacott & Wonnacott, pp. 75-103
Problem Set 4
Mar. 2 MIDTERM WEEK (sample exam)  
Mar. 9 Multiple Regression
    Alternate reading:
  • Wonnacott & Wonnacott, pp. 179-193
  • Hilton, Ch. 7

    Recommended reading (great at the beach, no kidding!):
  • Andrew Abbott, Flatland
Problem Set 5
Mar. 23 Special Problems with Preditors: Dummy Variables, Nonlinear Predictors, and Interactions
    Alternate reading:
  • Wonnacott & Wonnacott, 104-119
  • Hilton, 186-215
Problem Set 6
Mar. 30 Dichotomous Dependent Variables (Logistic Regression [example]) and Other Nonnormal Problems (Poisson Regression )
    Alternate reading:
  • Wonnacott & Wonnacott, pp. 120-150
    Recommended reading:
  • Aldrich & Nelson, entire
Apr. 7 Causal Analysis Inference in Regression
    Alternate reading:
  • Wonnacott & Wonnacott, 194-205
  • Hilton, Chapter 10-12.
Problem Set 7
Apr. 13 Multiple Regression: Selecting Predictors and Assessing Models
    Alternate reading:
  • Hilton, Ch. 8
  • Wonnacott & Wonnacott, Ch. 12,14
Apr. 20

Diagnostic Procedures in Multiple Regression (example)

    Required reading:
  • Kutner, Chapt. 10
Problem Set 8
Apr. 27

Remedial Procedures in Multiple Regression (example1 example2)

May 4 Special Problems of Time Series Data: Autocorrelation (example1, example2, example3)
    Required reading:
  • Kutner, Chapt. 12
  • [example reading to be determined]
Problem Set 9

Extra Topics (examples)

  • Measurement and Reliability
    • Suggested reading:
      Spector, Summated Rating Scale Construction (Sage)

  • Factor Analysis and Principal Components Analysis
    • Suggested reading:
      Kim & Mueller, Introduction to Factor Analysis (Sage)
      Kim & Mueller, Factor Analysis: Statistical Methods and Practical Issues (Sage)
      Paul Wahlbeck, James F. Spriggs, II, and Lee Sigelman, "Ghostwriters on the Court: A Stylistic Analysis of U.S. Supreme Court Opinion Drafts," American Political Research 30 (March 2002), 166-192.

Bert Kritzer, 608-263-2277, Kritzer@PoliSci.Wisc.Edu
Last modified, May4, 2004