{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

hw1ans - Stat 500 HW 1 Solution 1 Design-based p-values and...

This preview shows pages 1–2. Sign up to view the full content.

Stat 500: HW 1, Solution 1. Design-based p-values and confidence intervals (a) Use the randomization data (lab1bigex1.txt) to compute the p-value. For the one-sided test of H o difference = 0, there are 11 values larger than 4.49, so p = (11+1)/(500+1) = 0.024 or 2.4%. For the one-sided test of H o ratio = 1, there are 11 values larger than 3.84, so p = (11+1)/(500+1) = 0.024. (b) For the two-sided test of H o difference = 0, there are 10 values less than -4.49 and 11 larger than 4.49, so p = (21+1)/(500+1) = 0.044 or 4.4%. For the two-sided test of H o ratio =1, there are 10 values less than 0.2604 and 11 values more than 3.84, so p = (21+1)/(500+1) = 0.044. (c) Use the bootstrap data (lab1bigex2.txt) to get the confidence interval. Endpoints of the 95% confidence interval are given by the 0.025*500 =12.5, rounded to 13’th sorted value, and the 0.975*500 = 487.5, rounded to 488’th sorted value. 95% confidence interval for the difference is (0.440, 10.257). 95% confidence interval for the ratio is (1.226, 9.495). Note: Notice that these 13’th and 488’th observations values are symmetrical from the ends of the sorted list. The 13’th largest is skip the first 12 and take the 13’th.

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.

{[ snackBarMessage ]}