Читать книгу Applied Regression - Colin Lewis-Beck - Страница 7

In this second edition of Applied Regression: An Introduction, we maintain our firm commitment to the method of ordinary least squares (OLS). We are not alone in our defense of OLS. Peter Kennedy (2008), author of a leading econometrics text, observed the following: “The central role of the OLS estimator in the econometrician’s catalog is that of a standard against which all other estimators are compared. The reason for this is that the OLS estimator is extraordinarily popular” (p. 43). This popularity was recently affirmed in a methodological content analysis of the articles in the three leading general political science journals, with the finding that “OLS is by far the most popular method” (Krueger & Lewis-Beck, 2008, p. 3). This is not surprising, since OLS is the analytic tool of the classical linear regression model.

As Jan Kmenta (1997), author of our favorite econometrics book, reminds us, “The need for familiarity with the basic principles of statistical inference and with the fundamentals of econometrics has not diminished. . . . Most econometric problems can be characterized as situations in which some of the basic assumptions of the classical regression model are violated” (pp. v–vi). In our monograph, we pay special attention to these basic regression assumptions. Also, in the writing, we are inspired again by Kmenta (1997) and his “philosophy of making everything as simple and clear as possible” (p. vii). It is our hope that readers agree that we have realized this goal. Indeed, if readers are interested in further analyzing or replicating the results presented in this monograph, the datasets are available for download through the SAGE website.