Information Technology Reference
In-Depth Information
Ta b l e 8 . 2 .
Notation for a 3
×
2factorialdesign
Design
Factor
A
a
1
a
2
a
3
Factor
B
b
1
b
2
Data Set (
Y
ij k
)
a
1
b
1
a
2
b
1
a
3
b
1
a
1
b
2
a
2
b
2
a
3
b
2
Y
111
Y
121
Y
131
Y
112
Y
122
Y
132
Y
211
Y
221
Y
231
Y
212
Y
222
Y
232
Y
311
Y
321
Y
331
Y
312
Y
322
Y
332
Y
411
Y
421
Y
431
Y
412
Y
422
Y
432
Y
511
Y
521
Y
531
Y
512
Y
522
Y
532
AB
11
AB
21
AB
31
AB
12
AB
22
AB
32
AB
Matrix of Sums
Factor
A
a
1
a
2
a
3
Sum
Factor
Bb
1
AB
11
AB
21
AB
31
B
1
b
2
AB
12
AB
22
AB
32
B
2
Sum
A
1
A
2
A
3
T
in the second matrix of Table 8.2, labeled
Data Set (
Y
ijk
)
.Werefertothe
series or observations in this matrix collectively as
Y
ijk
to denote three
pieces of information. The subscript
i
designates the ordinal position of
each participant or case within a treatment combination. The subscript
j
denotes the level of the
A
treatment and the subscript
k
represents the
level of the
B
treatment. This notational device reminds us of the exact
treatment combination each participant receives.
The six columns of raw data in the
Y
ijk
matrix can each be summed
to produce treatment combination totals referred to collectively as
AB
jk
(again, the
j
subscript refers to the level of the
A
treatments and the
k
subscript signifies the level of the
B
treatments). These six
AB
jk
treat-
ment combination sums can be configured into their own matrix, which
we depict in Table 8.2 as the
AB
Matrix of Sums
.These
AB
treatment
combination sums can be summed vertically (to obtain the
A
treatment
sums) or horizontally (to obtain the
B
treatment sums) and will pro-
vide the ingredients for our future main effects analyses. The
AB
matrix
of sums can be converted into an
AB
matrix of means by dividing each
AB
jk
treatment combination sum by
n
.Wehavedonesoatthebottomof
Ta b l e 8 . 3 .
Before proceeding with the computation of sums of squares, degrees
of freedom, mean squares, and
F
ratios,weneedtocomputesomepre-
liminary calculations (e.g., treatment combination sums, squared sums
of the treatment combinations, and treatment combination means, and
standard deviations), which are depicted at the bottom of the
Y
ijk
matrix