This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Multivariate Regression and the Omitted Variable Bias Ch1 Appendix pp.2024 All Else Equal • Each of the predictions of economic theory presented earlier included the phrase “all else equal.” • Sometimes the same phrase is written “holding all other factors constant” or “ceteris paribus” (from Latin). • In any case, the phrase emphasizes the point that there are always multiple independent variables that affect the same economic dependent variable. • Because the predictions of economic theory assume that the other factors are held constant, a test of the theory should account for those factors as well. Multivariate Regression Analysis • The univariate regression analysis presented earlier implicitly assumed that no other factors influenced quit rates in firms. • To account for other factors, one can simply add an additional independent variable into the regression equation. • In this case, the average age of the workers in a firm might affect the quit rate. • Older workers are less likely to quit their jobs than young workers. Data • Firms 13 have employees whose average age is less than 40 while firms 46 have employees whose average age is greater than 40. 1 1 1 A i > 40 > 40 > 40 < 40 < 40 < 40 Average Age 10 12 6 15 10 5 20 8 4 30 8 3 35 6 2 40% 4 1 Quit Rate Average Wage Firm Dummy Variables • A i is called a dummy variable because it only takes on the values of 0 and 1. • In this case, A i takes on a value of 1 when average age is greater than 40 and 0 when average age is less than 40. • Dummy variables can be thought of as true/false indicators where 1 is true and 0 is false. • A i might be referred to as the “older workers dummy” because it takes on a value of 1 when “older workers” is true. Multivariate Regression Analysis • Adding the older workers dummy variable gives a new equation describing the data: Q i = α ’...
View
Full
Document
This note was uploaded on 10/22/2008 for the course ECON 470 taught by Professor Yeonsookim during the Fall '08 term at Maryland.
 Fall '08
 YeonsooKim

Click to edit the document details