Lecture 9Fitting Data vs. Estimating a Model课件.ppt

Lecture 9:Fitting Data vs. Estimating a Model Least Squares Derivation of OLS (Chapter 5.7) R2(Chapter 5.5) Superfluous Variables (Chapter 8.1) Omitted Variables (Chapter 8.2) Review To test a null hypothesis, we: Assume the null hypothesis is true; Calculate a test statistic, assuming the null hypothesis is true; Reject the null hypothesis if we would be very unlikely to observe the test statistic under the null hypothesis. Six Steps to Hypothesis Testing State the null and alternative hypotheses. Choose a test statistic (so far, we have learned the t-test). Choose a significance level, the probability of a Type I Error (typically 5%). Six Steps to Hypothesis Tests (cont.) Find the critical region for the test (for a 2-sided t-test at the 5% level in large samples, the critical value is t* = 1.96). Calculate the test statistic. Reject the null hypothesis if the test statistic falls within the critical region. Today’s Agenda The Least Squares Derivation of OLS (Chapter 5.7) Goodness of Fit Measure: R2 (Chapter 5.5) Superfluous Variables (Chapter 8.1) Omitted Variables (Chapter 8.2) Fitting a Line to Data (Chapter 5.7) So far, we have interpreted our estimators as “best guesses” of the true underlying parameters of the Data Generating Process. Our economic model (the DGP) has been central to our derivation of our estimators. Fitting a Line to Data (cont.) The Gauss–Markov Theorem states that, under the Gauss–Markov DGP, our Ordinary Least Squares estimator is the Best Linear Unbiased Estimator. Fitting a Line to Data (cont.) We could also interpret Ordinary Least Squares as the best way to summarize the particular dataset in front of us, without consideration of an economic model. Suppose we simply want to find the “best” line to fit a scatterplot of (X,Y ) points. What makes a line “best”? Fitting a Line to Data (cont.) A good line should be as close as possible to all the points. The vertical distance between the line and each point is called the