ABSTRACT
As we detailed in the third edition of DOE Simplified (2015) in a new chapter (11) devoted to this topic, experimental factors often cannot be randomized easily. Split plot designs then become attractive by grouping HTC factors, albeit with a loss in power (Anderson and Whitcomb, 2014). While split plots for factorial experiments go back to nearly a century (Fisher, 1925), application of these restricted-randomization designs to RSM is a relatively recent development, finding use, for example, in wind tunnel testing,
where some factors, such as wing-tip height, cannot be changed easily (English, 2007).
