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 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: Chapter 5 Regression with a Single Regressor: Hypothesis Tests and Confidence Intervals Testing Hypotheses About One of the Regression Coefficients Recall the problem we addressed in Ch.4. A linear relationship between class size and test score was assumed: TestScore = + ClassSize & ClassSize . Applying OLS estimation, the slope ClassSize could be estimated: \ TestScore = 698.9 2.28 & STR . Suppose an angry tax payer claims that cutting class size will not help boost test scores. To examine this claim, we are going to test whether the estimated b ClassSize is statistically significantly different from zero. TwoSided Hypotheses Concerning 1 The null and the alternative hypothesis we will test is: H : ClassSize = 0 vs. H 1 : ClassSize 6 = 0. In general, H : 1 = 1,0 vs. H 1 : 1 6 = 1,0 . Recall from Ch.4 that if the three least squares assumptions are satisfied, b 1 is approximately normally distributed in large samples. So, we can use the tstatistic to conduct the hypothesis test: t = b 1 & 1,0 SE ( b 1 ) . tstatistic is approximately distributed as a standard normal. TwoSided Hypotheses Concerning 1 The first step is to compute the standard error of b 1 , SE ( b 1 ) . SE ( b 1 ) is an estimator of b 1 : SE ( b 1 ) = q b 2 b 1 , where b 2 b 1 = 1 n & 1 n 2 n i = 1 & X i X 2 b u 2 i h 1 n n i = 1 & X i X 2 i 2 . The second step is computing the tstatistic using this SE ( b 1 ) . TwoSided Hypotheses Concerning 1 The last step is to compute the pvalue: pvalue = Pr & j Z j > t act = 2 & & t act , and reject the hypothesis if it is less than 0.05....
View Full
Document
 Fall '10
 Otusbo

Click to edit the document details