Comparing Means with a Control (2 of 4)
The procedure for comparing each experimental mean with the
control mean is called "Dunnett's test" after the statistician who
developed it. Dunnett's test controls the
EER and is more
powerful than tests
designed to compare each mean with each other mean. Dunnett's test is
conducted by computing a
t-test between
each experimental group and the control group using the formula:
where M
i is the mean of the ith experimental group,
M
c is the mean of the control group,
MSE is the mean square error as computed
from the
analysis of variance, and
n
h is the
harmonic mean of the
sample sizes of the experimental group
and the control group. The
degrees of
freedom (df) for the test are equal to N-a where N is the total
number of subjects in all groups and "a" is the number of groups
(including the control).