in the formula above is the percentile of the standard normal distribution
associated with a choice ofa; for example,z 1 a¼ 1 :96 whena¼ 0 :05 is
chosen. In the process of sample size determination, statistical tests, such
as two-samplettest, are usually planned as two-sided. However, if a one-
sided test is planned, this step is changed slightly; for example, we use
z¼ 1 :65 whena¼ 0 :05 is chosen.
- The desired powerð 1 bÞ(or probability of committing a type II error
b). This value is also selected by the investigator; a power of 80 or 90% is
often used. - The quantity
d¼jm 2 m 1 j
which is the magnitude of the di¤erence betweenm 1 andm 2 that is deemed
to be important. The quantitydis often called the minimum clinical sig-
nificant di¤erence and its determination is a clinical decision, not a sta-
tistical decision.
- The variances^2 of the population. This variance is the only quantity that
is di‰cult to determine. The exact value ofs^2 is unknown; we may use
information from similar studies or past studies or use an upper bound.
Some investigators may even run a preliminary or pilot study to estimate
s^2 ; but estimate from a small pilot study may only be as good as any
guess.
Example 12.8 Suppose that a researcher is studying a drug which is used to
reduce the cholesterol level in adult males aged 30 or over and wants to test it
against a placebo in a balanced randomized study. Suppose also that it is
important that a reduction di¤erence of 5 be detected (d¼5). We decide to
preseta¼ 0 :05 and want to design a study such that its power to detect a dif-
ference between means of 5 is 95% (orb¼ 0 :05). Also, the variance of choles-
terol reduction (with placebo) is known to be abouts^2 F36.
a¼ 0 : 05 !z 1 a¼ 1 : 96 ðtwo-sided testÞ
b¼ 0 : 05 !z 1 b¼ 1 : 65
leading to the required total sample size:
N¼ð 4 Þð 1 : 96 þ 1 : 65 Þ^2
36
ð 5 Þ^2
F 76
Each group will have 38 subjects.
SAMPLE SIZE DETERMINATION FOR PHASE III TRIALS 463