Introduction to Regression:
Brian Schaffner, University of Massachusetts
7 - 18 July (two week course / 35 hrs)
Brian Schaffner is Professor of Political Science at the University of Massachusetts, Amherst. He was an Assistant Professor at Western Michigan for two years
before spending four years as an Assistant Professor at American University in Washington, DC. He has also served as a Program Director for Political Science at the U.S. National Science Foundation.
His research focuses on public opinion, campaigns and elections, political parties, and legislative politics. His research has appeared in more than two-dozen refereed journal articles
- The primary focus of the course is the development of the linear regression model and its utility in helping researchers make causal inferences. We begin by discussing the challenges researchers face in making causal inferences from observational data. The course then turns to introducing the regression framework as a way to address those challenges. Students will gain practical experience in estimating regression coefficients using Ordinary Least Squares (OLS) and they will learn the assumptions underlying this model. This includes a discussion of hypothesis testing and measures of goodness of fit. We also examine the limitations of regression analysis, including violations of the assumptions underlying the OLS model and how to correct for them. This will teach you how to address the most common problems encountered when using regression analysis.
- Use STATA to generate and present regression output.
- Interpret regression output, including the coefficient estimates, their statistical significance, and summary statistics designed to capture the quality of the regression as a whole.
- Identify (and where possible correct for) weaknesses associated with the regression model.
- Participants should have some background knowledge of basic descriptive statistics and be comfortable with linear algebra.
- Barakso, Maryann, Daniel M. Sabet, and Brian F. Schaffner. 2013. Understanding Political Science Research: The Challenge of Inference. Routledge Press.
- Pollock III, Philip H. 2006. A Stata Companion to the Essentials of Political Analysis. Washington, DC: Congressional Quarterly Press.