8. In SAS ADX create a new 16-run split-plot design dening three whole-plot factors A, B, and C, and four sub-plot factors P, Q, R, and S.

(a) What alternate split-plot fractional factorial designs are listed?

(b) Which one would you prefer, and why?

(c) From the design you recommend, list the eects that must be tested with whole-plot error and those that must be tested with sub-plot error.

(d) Can you create the same design using proc factex? If so, show the commands to do it.

(e) If whole-plot experimental units cost four times more than that of sub-

plot experimental units, and only information on whole-plot main eects is desired, which ADX design would you recommend?

(f) Can you create the same design using proc factex? If so, show the commands to do it.

(g) Show the complete dening relation for the design created by ADX.