This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Stat 665 (Spring 2011) Kaizar Solutions 5 Exercises 20 points total 1. Problem 3.11, b and c (b) Using R, I found the following output: Estimate Std. Error z value Pr(>z) (Intercept) 1.6094 0.1414 11.380 < 2e16 *** x 0.5878 0.1764 3.332 0.000861 *** The prediction equation is: log( ) = 1 . 6094 + 0 . 5878 x I estimate that the log expected number of imperfections on wafers treated with B is 0.5878 higher than the log expected number of imperfections on wafers treated with A. That is, I estimate that wafers treated with B are expected to have e . 5878 = 1 . 8 times as many imperfections than wafers treated with A. (c) From the above output, the pvalue associated with the Wald test of the the null hy pothesis that = 0 is 0.000861. The Likelihood Ratio Test statistic is 27.85716.268 = 11.589. If the null hypothesis that = 0 is true, then the LRT has a 2 1 distribution. The pvalue is 0.00066. Regardless of which test I use, I reject the null hypothesis and conclude at significance 0.05 that wafers treated with B have more imperfections than wafers treated with A. 2. Modified Problem 3.12 (a) Using R, I found the following output: Estimate Std. Error z value Pr(>z) (Intercept) 2.0541 0.1132 18.142 <2e16 *** z0.2296 0.17011.349 0.177 The prediction equation is: log( ) = 2 . 0541 . 2296 z I estimate that the log expected number of imperfections on wafers treated with high coats is 0.2296 lower than the log expected number of imperfections on wafers treatedcoats is 0....
View Full
Document
 Spring '10
 Staff

Click to edit the document details