Multiple Regression Analysis: Estimation, Diagnostics and Modelling

Short Bio Bernhard Kittel is professor of Economic Sociology at the University of Vienna. Previous engagements include the University of Oldenburg, the University of Amsterdam, the University of Bremen, and the Max Planck Institute for the Study of Societies in Cologne. His research interests cover collective decision-making, political economy, and comparative research methodology. Prerequisite knowledge a) The course makes use of the freeware statistical package R, which can be downloaded from http://www.r-project.org, and assumes basic skills in using the package. For students unfamiliar with R, a preparatory course will be offered prior to the course. A good introduction into using R for statistics is Field et al (2012). b) Students are expected to understand the logic of inferential statistics. Prior knowledge of multivariate statistics is useful but not necessary. Remedial reading: any good statistics textbook, for example: Alan Agresti & Barbara Finlay (2008). Students familiar with R but in need of a refresher in basic statistics are encouraged to take part in the preparatory course on statistics. c) Furthermore, although a brief session on the use of matrix algebra in regression analysis is scheduled at the beginning of the course, it is assumed that the logic of matrix algebra is understood at the level presented in Gujarati. Short outline The course starts with a discussion of the logic of the multivariate regression model and the central assumptions underlying the ordinary least squares approach. This part includes an extensive discussion of the derivation of the ordinary least squares estimator, its standard error and summary statistics. Then it proceeds with testing for the adequacy of the assumptions and suitable corrections and extensions to the estimation techniques in the context of cross-sectional data. Particular emphasis will be laid on influential cases, multicollinearity, heteroskedasticity, and autocorrelation. Finally, categorical predictors, interactions, and nonlinearities will be considered.