63
Adaptive Sampling Plans
At each analysis of the data, the sampling plan
can be modified to account for changed perceptions of
possible results
•E.g.,
Proschan & Hunsberger (1995)
–Use
conditional power considerations to modify ultimate sample size
•E.g.,
Self-designing Trial (Fisher, 1998)
–Prespecify
weighting of groups “just in time”
•Weighting for each group only need be specified at immediately
preceding analysis