6. Consider an experimental situation where the investigator was interested in detecting a maximum dierence in treatment means that is twice the stan-

dard deviation of the response measured on replicate experimental units assigned to the same level of the treatment factor, i/e.
= 2:0. If there are 4 levels of the treatment factor:

(a) Modify the SAS data step commands in Section 2.7 to calculate the power for various numbers of replicates r per treatment level.

(b) Use the SAS analyst tool and proc glmpower to reproduce your results.

(c) Calculate the number of replicates necessary to have 0.90 power of de-

tecting a dierence as large as
= 2:0.

(d) How would the result you got in

(c) change if the number of levels of the treatment factor increased to 8, or decreased to 2?