This preview shows pages 1–3. 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: Section 8.5 A Nonparametric Test The twosample ttest even for the normal populations may not be the best or the most powerful size test. If the populations are not normal , we have better alternatives available. MaanWhiteney Test (Wilcoxons Rank Sum Test) Assume the two populations have same shape and spread so that they differ only in location (mean or median ). The populations need not be normal. Take samples of size m and n respectively from the populations P 1 and P 2 ; combine them and assign them ranks 1, 2, , ( n m + ) from the smallest to the largest. Assumption : Let { } m x x ,..., 1 be a random sample for continuous x ~ F with 1 and { } n y y ,..., 1 be random sample from continuous y ~ G with mean 2 . Assume both F and G have same spread and shape. 1 Consider testing 2 1 : = H . If H is true, the observations should be intermingled. Combine both the samples, order them from the smallest to the largest and assign them the ranks 1, 2, , (m+n)....
View
Full
Document
 Summer '08
 Palaniappan

Click to edit the document details