GOV 391L • Statistical Analysis in Political Science II
12:30 PM-2:00 PM
This course is a continuation of Statistical Analysis in Political Science I (391J). The course assumes that students are familiar with the absic concepts of statistics and probability theory including: descriptive statistics, probability, distributions, sampling, point estimation, confidenc eintervals, hypothesis testing, ANOVA and basic OLS. This course will build on 391J and we will discuss the "classical" multiple regression model, and lean how to deal with problems that arise when the basic assumptions of OLS are violated. The course will also provide an introduction to dichotemous dependent variable models (probit and logit). Topics to be covered are: - The bivariate OLS regression model (a review) - R-squared and correlation (a review) - The multivariate OLS regression model (a review) - The Gauss-Markov theorem - Evaluating OLS resoluts: goodness of fit - Evaluating OLS results: hypothesis testing - Predictions: Extrapolation and interpolation - Multi-collinearity - Heteroscedasticity - Serial correlation (autocorrelation) - Endogeneity - Specification errors - Qualitative response models (Probit and Logit) The focus of the course will be to train graduate students to begin to use statistics in analyzing problems in political science. While we will go through a rigorous mathematical proofs of some of the key concepts the emphasis on the course will be on the application (using hands on exercises in STATA) rather than the theoretical.
Grading: Homework Assignments (5-8): 30% Midterm: 30% Final: 30% Participation: 10%
Gujarati, Basic Economietrics, McGraw-Hill, 1995.