Expected Mean Squares for a One-factor between-subject ANOVA (1 of 2)
As stated in
another section, both the
numerator and the denominator of the F ratio estimate the
population
variance when the
null hypothesis is
true. Since they are both
unbiased
estimates, the
expected value of both MSB
and MSE is σ². Symbolically, E[MSB] = E[MSE] = σ². This
section covers the expected value of MSB and MSE when the null hypothesis is
false.
The MSB is based upon the sample
means: the
greater the variance of the sample means, the greater the MSB. Therefore, when
the null hypothesis is false and the population means are not equal, the
expected value of MSB is greater than when the null hypothesis is
true. It stands to reason that the more the population means differ
from each other, the greater the expected value of MSB. Mathematical
statisticians have derived the following formula for the expected
value of MSB:
where σ² is the
population variance,
µ
i is the ith population mean,
is the
mean of the population means, n is the number of subjects in each
group, and "a" is the number of population means.