STATISTICS 402 - Assignment 7
Due April 15, 2011
1. An experiment was done with patients suffering from agoraphobia (the fear of going
out in public places). There are three factors of interest; drug (a selective serotonin
reuptake inhibitor (SSRI) e.g. Prozac, a tricyclic antidepressant e.g. Elavil, an
antianxiety e.g. Xanax and Placebo – offered as a new drug for agoraphobia), type of
psychotherapy (psychodynamic, cognitive/behavioral, and group), and depression
(yes/no). The experimenters were able to assign the 12 combinations of the drug and
type of psychotherapy to 3 depressed and 3 non-depressed agoraphobics. After 6
months the severity of agoraphobia was measured for each subject (maximum score
30, minimum score 0). The data are given below and are also available in a JMP data
table available on the course web page.
Not Depressed
Psychodynamic
Depressed
Not Depressed
Cognitive
Behavioral
Depressed
Not Depressed
Group
Depressed
Prozac
11
12
7
10
8
8
8
9
11
9
12
10
14
11
10
12
9
8
Elavil
13
10
11
11
7
8
10
11
12
13
13
16
15
13
11
11
11
14
Xanax
18
19
20
16
14
13
12
13
11
8
10
12
20
19
19
16
17
18
Placebo
21
24
21
18
18
21
19
16
16
17
16
15
11
13
15
10
10
13
a) What are the response, conditions and experimental material?
b) How do you know that this experiment was run as a completely randomized
design?
c) The severity of agoraphobia for each participant at the beginning of the study will
vary. Where will this variation end up in the analysis of the data? Explain
briefly.
d) Use JMP to calculate means for each of the factors and display these means in
appropriate plots. What factors appear to have effects on the response? Explain
briefly.
e) Are there statistically significant differences among the 24 treatment sample
means? Report the appropriate F ratio, P-value, decision and reason for the
decision.
1
f) What factors are statistically significant? For each significant factor report the F
ratio and P-value.
g) What interactions are statistically significant? For each significant interaction
report the F ratio and P-value.
h) Use JMP to calculate means for the combinations of factors taken two-at-a-time
and display these in interaction plots. You should have three interaction plots.
Describe each interaction plot and indicate how your description matches the
results of the tests conducted in g).
i) Based on the analysis above, what statistically valid recommendations for
treatment of agoraphobia can be made?
j) You should turn in the following JMP output. If additional output is turned in, I
will deduct points.
 Factor level means and plots.
 A complete analysis of variance table showing sources of variability: Model,
Error and C. Total.
 Effect tests for all factors and interactions.
 Means for the combinations of factors taken two-at-a-time and 2-way
interaction plots.
2. The maker of Prozac makes the claim that for depressed patients receiving
Cognitive/Behavioral, Group or Psychodynamic therapy taking Prozac will result in a
lower mean severity score than Placebo and the difference in mean severity scores is
statistically significant. The FDA makes the counter claim that for some depressed
patients receiving therapy taking Placebo is just as effective, on average, as taking
Prozac. Analyze a subset of the data, just depressed patients taking either Prozac or
Placebo, to see whose claim is supported by the data. For each claim provide, and
explain, statistical evidence either for or against the claim.
2