In searching for biomarkers, one mightmaintain that it is prudent to havepreliminary studies based on smallsamples. Given a satisfactory result, alarge-sample follow-on study can thencheck reproducibility. However, is thistwo-fold approach, namely the decisionto conduct the small study and the subsequent decision to spend signi?cantresources on a large study, actually prudent? There is the risk that in the ?rststudy nothing at all might be discoveredwhile in the second study the ?rstresults are not reproducible.
展开▼