Applied Statistics and Probability for Engineers

(Chris Devlin) #1
12-3

The data will be coded by defining a new set of scaled regressor variables as

where is the corrected sum of squares of the levels of wj. The coded data
are shown in Table S12-2. This transformation makes the column of ones in Xorthogonal to the

Sjjg
n
i 11 wijwj^2
2

xij

wijwj
1 Sjj

i1, 2,p, 15 j1, 2, 3, 4

Table S12-1 Cement Data
Observation
Number yw 1 w 2 w 3 w 4
1 28.25 10 31 5 45
2 24.80 12 35 5 52
3 11.86 5 15 3 24
4 36.60 17 42 9 65
5 15.80 8 6 5 19
6 16.23 6 17 3 25
7 29.50 12 36 6 55
8 28.75 10 34 5 50
9 43.20 18 40 10 70
10 38.47 23 50 10 80
11 10.14 16 37 5 61
12 38.92 20 40 11 70
13 36.70 15 45 8 68
14 15.31 7 22 2 30
15 8.40 9 12 3 24

Table S12-2 Coded Cement Data
Observation
Number yx 1 x 2 x 3 x 4
1 28.25 .12515 .00405 .09206 .05538
2 24.80 .02635 .08495 .09206 .03692
3 11.86 .37217 .31957 .27614 .03692
4 36.60 .22066 .22653 .27617 .33226
5 15.80 .22396 .50161 .09206 .39819
6 16.23 .32276 .27912 .27617 .31907
7 29.50 .02635 .10518 .00000 .07647
8 28.75 .12515 .06472 .09206 .01055
9 43.20 .27007 .18608 .36823 .27425
10 38.47 .51709 .38834 .36823 .40609
11 10.14 .17126 .12500 .09206 .15558
12 38.92 .36887 .18608 .46029 .27425
13 36.70 .12186 .28721 .18411 .24788
14 15.31 .27336 .17799 .36823 .25315
15 8.40 .17456 .38025 .27617 .33226

PQ220 6234F.CD(12) 5/20/02 10:52 M Page 3 RK UL 6 RK UL 6:Desktop Folder:TEMP WORK:MONTGOMERY:REVISES UPLO D CH114 FIN L:Quark

Free download pdf