This course will consider topics in regression analysis aimed at people wanting to use regression for empirical work in a fairly sophisticated way. The focus is on regression with cross-sectional data. Topics include a brief review of the assumptions used to conclude that ordinary least squares (OLS) has desirable statistical properties; how one confronts problems of multicollinearity; estimating and interpreting models with flexible functional forms (such as interaction terms); selection of regressors and proxy variables; using and interpreting linear regression with discrete outcomes; making inference robust to heteroskedasticity and cluster correlation; the pros and cons of weighted least squares versus OLS; linear regression and the estimation of average treatment effects.

Examples will be shown using the software package Stata. The course assumes an understanding of multiple regression at the level of Chapters1 through 5 in Introductory Econometrics:A Modern Approach, 4e, by Jeffrey M. Wooldridge.

Location:	106 Farrall Agricultural Engineering Hall [map]
Price:	Free to Graduate Students, Faculty and Research Staff
Sponsor:	public
Contact:	Dr. Brian Silver bsilver@msu.edu 517-353-9288
Share: Facebook Twitter Download event (.ics) Print