Dear All
I am attempting a sample size calculation for a parallel group study, using
information from data which are highly skew using the sample size package
nQuery.
The outcome is the difference in the median percent reduction in size
between 2 treatment groups.
If I attempt a sample size calculation with the Mann Whitney Test as the
method of testing between 2 group medians, an assumption of the sample size
calculation is that the data is normally distributed. This unfortunately is
not the case for my set of data and so this method of sample size
calculation used in nQuery seems inappropriate.
Can anyone point me in the direction of a package that can cope with such
outcomes or papers that may enlighten me on this topic?
Many Thanks
Trevor Mole
Smith+Nephew
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|