Lecture 13 Hypothesis testing in multiple regression model, Part 1

# Lecture 13 Hypothesis testing in multiple regression model, Part 1

This preview shows pages 1–6. Sign up to view the full content.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Economics 326 Methods of Empirical Research in Economics Lecture 13: Hypothesis testing in the multiple regression model, Part 1 Vadim Marmer University of British Columbia May 4, 2009 The model I We consider the classical normal linear regression model: 1. Y i = + 1 X 1 , i + . . . + k X k , i + U i . 2. Conditional on X &s, E ( U i ) = 0 for all i &s. 3. Conditional on X &s, E & U 2 i = 2 for all i &s. 4. Conditional on X &s, E & U i U j = 0 for all i 6 = j . 5. Conditional on X &s, U i &s are jointly normally distributed. I We also continue to assume no perfect multicolinearity: The k regressors and constant do not form a perfect linear combination, i.e. we cannot nd constants c 1 , . . . , c k , c k + 1 (not all equal to zero) such that for all i &s: c 1 X 1 , i + . . . + c k X k , i + c k + 1 = . 1/15 Testing a hypothesis about a single coe cient I Take the j-th coe cient j , j 2 f , 1 , . . . , k g . I Under our assumptions, its OLS estimator j satis&es that conditional on X s: j & N & j , Var & j , where Var & j = 2 / n i = 1 X 2 j , i (see Lecture 12). I Therefore, & j j / r Var & j & N ( , 1 ) . I The conditional variance Var & j is unknown because 2 is unknown.The estimator for Var & j is d Var & j = s 2 n i = 1 X 2 j , i , where s 2 = n i = 1 U 2 i / ( n k 1 ) (see Lecture 11). 2/15 Testing a hypothesis about a single coe cient I We have that conditional on X &s, j & j r d Var & j t n & k & 1 . I Standard error: SE & j = r d Var & j = q s 2 / n i = 1 X 2 j , i . 3/15 Testing a hypothesis about a single coe cient: Two-sided alternatives I Consider testing H : j = j , against H 1 : j 6 = j , ....
View Full Document

## This note was uploaded on 03/24/2012 for the course ECON 326 taught by Professor Whisler during the Spring '10 term at The University of British Columbia.

### Page1 / 16

Lecture 13 Hypothesis testing in multiple regression model, Part 1

This preview shows document pages 1 - 6. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online