Example 12.9 Suppose that in Example 12.8, the researcher wanted to design
a study such that its power to detect a di¤erence between means of 3 is 90% (or
b¼ 0 :10). In addition, the variance of cholesterol reduction (with placebo) is
not known precisely, but it is reasonable to assume that it does not exceed 50.
As in Example 12.8, let’s seta¼ 0 :05, leading to
a¼ 0 : 05 !z 1 a¼ 1 : 96 ðtwo-sided testÞ
b¼ 0 : 10 !z 1 b¼ 1 : 28
Then using the upper bound for variance (i.e., 50), the required total sample
size is
N¼ð 4 Þð 1 : 96 þ 1 : 28 Þ^2
50
ð 3 Þ^2
F 234
Each group will have 117 subjects.
Suppose, however, that the study was actually conducted with only 180
subjects, 90 randomized to each group (it is a common situation that studies
are underenrolled). From the formula for sample size, we can solve and ob-
tained
z 1 b¼
ffiffiffiffiffiffiffiffiffiffiffi
N
4
d^2
s^2
r
z 1 a
¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffi
180
4
32
50
r
1 : 96
¼ 0 : 886
corresponding to a power of approximately 81%.
12.9.2 Comparison of Two Proportions
In many other studies, the endpoint may be on a binary scale; so let us consider
a similar problem where we want to design a study to compare two propor-
tions. For example, a new vaccine will be tested in which subjects are to be
randomized into two groups of equal size: a control (nonimmunized) group
(group 1), and an experimental (immunized) group (group 2). Subjects in both
control and experimental groups will be challenged by a certain type of bacteria
and we wish to compare the infection rates. The null hypothesis to be tested is
H 0 :p 1 ¼p 2
versus
464 STUDY DESIGNS