12+Analysis+of+variance

The unbiased estimator of 2 being 1 nk ni k 2 xij

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: III-XI (McClave and Benson) with (k − 1) numerator degree of freedom and df = n − k for denominator. The unbiased estimator of σ 2 being 1 σ= ˆ n−k ni k 2 (xij − xi )2 . ¯ i=1 j =1 Example 2. Refer to the example 1. There are k = 4 groups of measurements for 4 level (brands of golf balls), ni = 10 (i = 1 : 4). Results are shown in the Table. Set up the test to compare the mean distances for the four brands. Use α = 0.05. 4 A. Zhensykbaev ANOVA # Brand A Brand B Brand C Brand D x1i x2i x 3i x4i 1 251.2 263.2 269.7 251.6 2 245.1 262.9 263.2 248.6 3 248.0 265.0 277.5 249.4 4 251.1 254.5 267.4 242.0 5 265.5 264.3 270.5 246.5 6 250.0 257.0 265.5 251.3 7 253.9 262.8 270.7 262.8 8 244.6 264.4 272.9 249.0 9 254.6 260.6 275.6 247.1 10 248.8 255.9 266.5 245.9 Mean x1 = 251.3 x2 = 261.1 x3 = 270 x4 = 249.4 ¯ ¯ ¯ ¯ Calculate means and than test statistic. x1 = 251.3, x2 = 261.1, x3 = 270, x4 = 249.4, x = 257.95, ¯ ¯ ¯ ¯ ¯ 4 10(¯i − x)2 = 2724.5, x ¯ SST = i=1 4 10 (xij − xi )2 = 907.1, ¯ SSE = i=1 j =1 F= 2724.5/3 SST /(k − 1) = = 36. SSE/(n − k − 1) 907.1/36 The quantity Fα = F0.05 ≈ 2.88 is defined from the Table IX (ν1 = k − 1 = 3, ν2 = n − k = 36). Since 36>2.88 we reject H0 (the brand is essential). The Bonferroni procedure. This is a procedure to make multiple comparisons of a set of treatment means µ1 , ..., µk . This is usually done by constructing the set of confidence intervals for all possible differences µi − µj of the overall confidence α. 5 A. Zhensykbaev ANOVA So each confidence interval will be of confidence α/c, where c is the total number of all possible differences µi − µj . For each difference µi − µj the confidence interval of the level α/c is (¯i − xj ) ± t 2αc · s · x ¯ 1 1 +, ni nj (i = j, i, j = 1 : k ), where tα/(2c) is defined from Table VI with df = n − k , s= S SE/(n − k ). Example 3. Refer to the example 1. There are exactly c = 6 pairs of treatment means: µA − µB , µ A...
View Full Document

This note was uploaded on 02/11/2014 for the course MATH 1390 taught by Professor Christopherstocker during the Spring '13 term at Marquette.

Ask a homework question - tutors are online