This course serves as an introduction to statistical inference and econometrics at the graduate level. First, basic concepts in probability theory (i.e., random variables, their distributions, and their moments) are reviewed. Next, we move to statistical inference. Methods for point and interval estimation (least-squares, maximum likelihood, and method of moments) are introduced and criteria for evaluating their performance (bias, efficiency, consistency, and various risk measures) are discussed. Procedures to perform hypothesis testing (t-test, F-test, Wald, Likelihood ratio, and Lagrange multiplier) are also developed. Finally, estimation and testing in the linear regression model is studied in detail.
Lectures complemented with problem sets.
27, rue Saint-Guillaume 75007 Paris France T/ +33 (0)1 45 49 50 51 - F/ +33 (0)1 42 22 39 64 www.sciences-po.fr
The main reference text for the course is Statistical Inference by G.C. Casella and R.L. Berger (2008, Brooks/Cole).
Slides, class notes, and additional material will be made available.
It is useful to complement the main reference text with a book that devotes particular attention to estimation and inference in the linear regression model. All of the following econometric texts contain chapters on linear regression at the appropriate level: Estimation and Inference in Econometrics by R. Davidson and J.G. MacKinnon (1994, Oxford UP), A Course in Econometrics by A.S. Goldberger (1991, Harvard UP), Econometrics by F. Hayashi (2000, Princeton UP), and An Introduction to Classical Econometric Theory by P.A. Ruud (2000, Oxford UP).
A technical compendium on the asymptotic theory behind least-squares estimation is Asymptotic Theory for Econometricians by H. White (2000, Academic Press).
Other statistical texts that provide treatments of probability theory and statistical inference at the level of the course are Basic Concepts of Probability and Statistics by J.L. Hodges Jr. and E. Lehmann (2004, SIAM) and Testing Statistical Hypotheses by E. Lehmann and J.P. Romano (2005, Springer).
Niveau requis : Familiarity with calculus, linear algebra, and elementary statistics and regression analysis is useful. Some experience with statistical software such as Stata is also helpful but not required.
Modalités d'évaluation : Grading is based on (i) a final exam at the end of term and (ii) problem sets during term.
Dernière mise à jour : lundi 12 mars 2012