|
Early Development
of a Data Analytic Plan
The analysis aspect
of randomized control trials is simpler than non-randomized trials,
since the groups should be comparable on all known and unknown
factors, and relatively straightforward comparisons are required
to test the hypotheses at hand.
As emphasized before,
it is critical to have a biostatistician or epidemiologist involved
from the beginning in planning the trial, so that there is an
early development of the data analytic plan. The availability
of easy to use statistical software programs should never substitute
for experience and statistical expertise (Peduzzi
et al., 2002).
|