Prerequisites: Univariate mathematical statistics (Statistics 36-325 and 36-326 provide adequate background) or 90-905, or consent of instructor. An introduction to classical analysis of multiple linear regression models. Basic concepts in multivariate statistics (such as joint, conditional, and marginal distributions) are briefly reviewed. The emphasis is on presentation and derivation of theoretical results for the standard linear regression model. Topics include the algebra of least squares, sampling theory for least-squares and maximum-likelihood estimators, specification analysis, confidence intervals, hypothesis tests, pretesting, and prediction. The generalized linear regression model is introduced late in the course.
