12+Analysis+of+variance

In this case the experiment has two factors brand of

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: bot II type. In this case the experiment has two factors: brand of golf ball and type of robot. Below the diagram provides an overview of the experimental process and a summary of the introduced terminology. Population of experimental units Sampling experiment ¨¨rr rr ¨¨ Sample rr ¨ ¨¨ ¨ of experimental rr ¨ rr ¨¨ r ¨ rr units ¨¨ r rr¨¨ ¨r rr ¨¨ ¨ r ¨ rr ¨¨ ¨ Apply Apply factor-level combination 1 ¨ factor-level combination p ¨ ¨ % Treatment 1 sample • • c • r rr j r Treatment p sample c Responses for Treatment 1 Responses for Treatment p 2 A. Zhensykbaev ANOVA Model of one-factor analysis. Let us given k treatments T1 , ..., Tk and corresponding samples are: k X1 = {x11 , . . . , x1n1 }, . . . , Xk = {xk1 , . . . , xknk }, n= ni , i=1 xij is j -th measurement in i-th sample, ni is the size of i-th sample, xi - its ¯ sample mean, µi - mean of treatment Ti , x - overall mean response of all ¯ sample measurements k 1 x= ¯ xi . ¯ k i=1 We present these measurements as the table Treatment 1 Treatment 2 . . . Treatment k x11 x21 ... xk 1 x12 x22 ... xk 2 . . . . . . . . . . . . x1n1 x 2n 2 ... xknk Means x1 ¯ x2 ¯ ... xk ¯ In general the model of one-factor analysis is yij = µi + εij (i = 1 : k, j = 1 : ni ), where errors εij are independent. Assumptions: 1. The probability distribution of the populations of responses associated with each treatment are normal with mean =0. 2. Population variances are supposed to be equal. 3. The samples experiment units selected for each treatment are random and independent. The null hypothesis is H0 = {µ1 = · · · = µk }. 3 A. Zhensykbaev ANOVA Alternative hypothesis is Ha = {at least two means differ}. Test statistics is M ST , M SE where M ST - mean square for treatments F= SST M ST = , k−1 k ni (¯i − x)2 , x ¯ SST = i=1 M SE - mean square for error SSE , M SE = n−k k ni (xij − xi )2 , ¯ SSE = i=1 j =1 SST - sum of squares for treatments, SSE - sum of squares for error. RR = {F > Fα }, where Fα is defined from the Tables V...
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