Ar. Feinstein et J. Concato, THE QUEST FOR POWER - CONTRADICTORY HYPOTHESES AND INFLATED SAMPLE SIZES, Journal of clinical epidemiology, 51(7), 1998, pp. 537-545
To have the ''power'' of avoiding undersized clinical trials, the cust
omary statistical strategy used in the past few decades is aimed at re
jecting both a null stochastic hypothesis and a contradictory alternat
ive hypothesis. This approach gives a trial the ''power'' to confirm t
he ''insignificance'' of differences much smaller than the large value
of delta desired in trials done to show efficacy. In many instances,
however, a prime problem is that the current ''double-significance'' a
pproach produces sample sizes 2-3 times larger than needed for stochas
tic confirmation of large differences (greater than or equal to delta)
. The inflated sample sizes and consequent problems can be avoided if
a realistic value for delta is chosen and maintained thereafter, and i
f an adequate ''capacity'' is calculated for ''single significance.''
(C) 1998 Elsevier Science Inc.