GOV 391L • Statistical Analysis in Political Science II
12:30 PM-2:00 PM
This course is about "regression models," roughly and broadly defined as statistical models to explain some single dependent variable. The domain thus includes nonlinear as well as linear models and models for qualitative as well as quantitative dependent variables. I shall try to please two clienteles: aspiring methodologists and aspiring practitioners. There will be computer-based exercises to provide concrete examples, but the lectures and readings will focus on general questions of modeling, estimation, inference, and interpretation: What sorts of models imply and reflect what sorts of relationships between independent and dependent variables? What assumptions must we make, and what do they mean? How likely are the assumptions to be violated, and with what consequences? How can we tell when violations occur? What alternative assumptions might we make? What quantities should we be focusing on estimating? What estimators provide statistically desirable estimates? Where several different estimators might serve, what are their advantages and disadvantages? What do the estimates tell us, and how certainly? The lectures and readings will treat these questions practically but abstractly, referring more to x's and y's than to substantive variables. There will be much mathematical notation and mathematically phrased argument and some proof and derivation. The goal is to convey a good, relatively deep understanding of the hows and whys of constructing, estimating, and interpreting the estimates of these models. To get the taste of actual modeling and analysis, we shall examine some published examples and work through a series of mostly computer-based exercises asking you to write and analyze your own models, rooted in your own substantive interests. I shall help you find datasets, if you dont have any you are already working on or interested in. There will also be an optional term paper, to be centered on a regression model of your devising. The course requires a decent knowledge of descriptive and inferential statisticsas covered in the Government Departments Statistics I or equivalentand a reasonable facility with ordinary algebra. I encourage you to consult me if you are unsure whether this is an appropriate course for you to take. I expect there to be extra sessions, yet to be scheduled, to review the mathematical and statistical background, go over questions and assignments, and discuss concrete applications.
There will be two exams, a series of exercises, and, optionally, a term paper. The exams will be in-class and closed-book. The exercises will be a mix of pen-and-paper and computer-based, the former to help cement the math, the latter, calling for you to write and analyze models of your own, to provide a taste of actual modeling. The optional term paper should apply simultaneous equation models and procedures to a substantive problem and data of your choosing. It may build on but must go well beyond the exercises.
Damodar Gujarati. 2003. Basic Econometrics (4th ed.). New York: McGraw-Hill Jan Kmenta. 1997. Elements of Econometrics (2nd ed.). Ann Arbor: University of Michigan Press. A.H. Studenmund. 1997. Using Econometrics: A Practical Guide (3rd ed.). Reading, PA: Addisson-Wesley. Studenmund is the simplest, Kmenta the most sophisticated. You should read Gujarati and at least one of the other two, preferably Kmenta, if you only read one and you can handle it. Everyone should read Kmentas chapters 1-6, under Statistical Review.