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: Statistics 403 Problem Set 4 Due in lab on Friday, October 8th 1. Suppose we are interested in comparing the support for a certain political candidate in Michigan to the support for the candidate in Ohio. We conduct independent surveys in both states, and obtain responses X 1 ,...,X m in Michigan and Y 1 ,...,Y n in Ohio. Let μ m be the population level of support in Michigan (i.e. EX i = μ m ) and let μ o be the population level of support in Ohio ( EY i = μ o ). Our goal is to estimate μ m- μ o . (a) Construct a 95% confidence interval for μ m- μ o centered at ¯ X- ¯ Y . Solution: In order to get 95% coverage probability, we need the interval to cover two standard deviations on either side of the estimate. The standard deviation of the estimate is SD( ¯ X- ¯ Y ) = q σ 2 x /m + σ 2 y /n. Thus the interval is ¯ X- ¯ Y ± 2 q σ 2 x /m + σ 2 y /n. (b) Suppose the true levels of support are μ m = 0 . 6 and μ o = 0 . 54. When the sample sizes in the two states are equal ( m = n ), what sample size is required so that the confidence interval is 0.04 units wide? Hint: since the X i and Y i are binary, you can determine their variances using their expected values. Solution: Since the X i and Y i are binary, we need to have a general expression for the variance of a binary random variable. Suppose P ( X = 1) = p and P ( X = 0) = 1- p , so the expected value is p · 1 + (1- p ) · 0 = p, and therefore the variance is p (1- p ) 2 + (1- p )(0- p ) 2 = p (1- p ) ....
View Full Document
This note was uploaded on 02/06/2012 for the course STAT 403 taught by Professor Kerbyshedden during the Winter '12 term at University of Michigan-Dearborn.
- Winter '12