Reconsider the cake mix experiment described in exercise 9 of Chapter 8.

The control factors that can be selected by the manufacturer are F: the amount of our; S: the amount of shortening; and E: the amount of egg powder in the mix. The noise factors are T: baking temperature, and t:

baking time.

(a) Did you detect any control-by-noise interactions in the split-plot analysis conducted in Chapter 8? If so, can these interactions be exploited to reduce the eects of variation in baking time or baking temperature?

(b) Reanalyze the data using the location-dispersion modeling approach. Do you reach the same conclusion about the optimal choice of levels of F, S, and E?

(c) Which method of analysis gives you more information?