This preview shows pages 1–5. Sign up to view the full content.
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 Document
Unformatted text preview: Chapter 7 Hypothesis Tests and Confidence Intervals in Multiple Regression (Part 1) Hypothesis Tests and Confidence Intervals for a Single Coefficient Recall the model we examined in Ch.6. We had a multiple regression model: TestScore = + 1 & STR + 2 & PctEL + u . Applying OLS estimation, the slope , 1 and 2 could be estimated: \ TestScore = 686.0 1.10 & STR 0.650 & PctEL . You may want to test whether the estimated b 1 is statistically significantly different from zero. Hypothesis Tests and Confidence Intervals for a Single Coefficient Testing the hypothesis H : j = j ,0 vs. H 1 : j 6 = j ,0 1. Compute the standard error of b j , SE ( b j ) . 2. Compute the tstatistic, t = b j & j ,0 SE ( b j ) . 3. Compute the pvalue = 2 & & t act ,where t act is the value of the tstatistic actually computed. Reject the hypothesis at the 5% significance level if the pvalue is less than 0.05 or, equivalently, if t act > 1.96. 95% confidence interval for j : j = [ b j & 1.96 SE ( b j ) , b j + 1.96 SE ( b j )] Confidence Intervals for a Single Coefficient For our example, suppose SE ( b ) = 8.7, SE ( b 1 ) = 0.43 and SE ( b 2 ) = 0.031, \ TestScore = 686.0 & 1.10 STR & 0.650 PctEL ....
View Full
Document
 Fall '10
 Otusbo

Click to edit the document details