Consider planning an experiment to compare the taste of chocolate brownie recipes. The treatments will consist of three brands of boxed brownie mixes and one from-scratch home recipe. A panel of tasters will be used to eval-

uate the brownies. Each taster will taste each brownie and give it a rating from 1 to 10 (10 being the best possible).

(a) Why do you think it might be appropriate to use a crossover design for this study? What is the alternative?

(b) Use the williams function in the R package crossdes (shown in Section 9.4.1) to create Williams's crossover design for four treatments to use for this experiment.

(c) Decide how many subjects to include in your study and create a ran-
domized list assigning each subject to one of the four sequence groups.

(d) Write the model you would use to analyze this data.

(e) Make a data input form and show how you would code the indicators for the subject, period, treatment, and carryover eects.