ECON 3740 Lecture Notes - Lecture 1: Standard Deviation, Conditional Variance, Squared Deviations From The Mean
3 Dec 2017
School
Department
Course
Professor
Document Summary
The statistical properties of least squares: assumptions. The statistical properties of the least squares estimator are based on a number of assumptions. These assumptions are also important for establishing valid methods of testing hypotheses and the construction of confidence intervals. Since so much rests on the assumptions, it is extremely important in practice to check for signs of violations. Econometricians have developed a long list of diagnostic statistics that warn the researcher against all sorts of errors in the assumptions. First, we focus on a simple model in which a single variable x explains the dependent variable y. In the population, x and y are jointly dependent random variables. Assumption a1: the population regression function of y on x (the conditional mean function of y given x) is linear and is written e(y|x) = + x. The parameters and are population parameters and can never be known exactly by the researcher.
