Abstract
ABSTRACT Besseris (2009) recommended a rank sum test for eight-run, fractional-factorial designs. Our article makes two points regarding such tests. First, the backward elimination procedure mentioned by Besseris leads to excessive overfitting of the model; that procedure is virtually guaranteed to include inactive terms when in fact there are no active effects. Second, though one could correct for this overfitting problem, there is an even more fundamental flaw to using rank sum tests for analyzing factorial designs. It is for this second reason that we suspect that no rank sum tests will be useful for analysis of two-level unreplicated factorial designs.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
