> In a pilot experiment, from 20 patients, only one had returned for the
> second and last dose. The question is... how to test this ?
I'm not convinced that this is a statistical problem. The 20 patients
given the first dose comprise your "population". Out of this, 1
returned. This is an absolute number, it is not an estimate and there is
no sampling error. One out of 20 failed to return - that is the end of
the story. There is no "statistical significance" involved.
In the same way, if you want the mean performance of a class of
students, the class comprises the population. You are not purporting to
make inferences or estimations about other populations and so the mean
is mu (not x-bar) and the standard deviation is sigma (not s). Since you
have sampled the whole population, standard errors and confidence
intervals are meaningless.
Jim Fowler
___________________________
>From the Royal Maalie Court
60.38N, 01.10W
|