Microsoft Office Tutorials and References
In Depth Information
Table 6.12 Data for four software products experiment
Obs.
(Analysts)
average
Block
assignment
Software
treatment
1
12
A
d
23
2
13
A
a
14
3
13
A
c
12
4
13
A
b
21
5
16
B
a
16
6
17
B
d
25
7
17
B
b
20
8
18
B
c
15
9
21
C
c
18
10
22
C
d
29
11
23
C
a
17
12
23
C
b
28
13
28
D
c
19
14
28
D
a
23
15
29
D
b
36
16
31
D
d
38
17
35
E
d
45
18
37
E
b
41
19
39
E
c
24
20
40
E
a
26
is exhausted, beginning with the top 4, and so on. Then analysts will be randomly
assigned to one of 4 software products, within each block. Finally, a score will be
recorded on their task time and the Excel analysis ANOVA: Two-Factor without
Replication will be performed. This experimental design and results is shown in
Table 6.13.
Although we are using the Two-Factor procedure, we are interested only in a
single factor—the four software product treatments. Our blocking procedure is more
an attempt to focus our experiment by eliminating unintended inﬂuences (the skill
of the analyst prior to the experiment), than it is to explicitly study the effect of more
capable analysts on task times. Table 6.12 shows the 20 analysts, their previous 6
month average task scores, the 5 blocks the analysts are assigned to, the software
product they are tested on, and the task time scores they record in the experiment.
Exhibit 6.8 shows the data that Excel will use to perform the ANOVA. Note that
analyst no. 1 in Block A (see Table 6.12) was randomly assigned product d .In
Exhibit 6.8 the cell (C8) associated with the cell comment represents the score of
analyst no. 20 on product a .
We are now prepared to perform the ANOVA on the data and we will use the
Excel tool ANOVA: Two-Factor without Replication to test the null hypothesis
that the task completion times for the various software products are no different.

Search JabSto ::

Custom Search