Consider the data from Exercise 9.2.14. Suppose a new study of the effectiveness of eSET is being considered for women ages 35-39, and it is desired to have a standard error no greater than 0.04. How large a sample would be needed
(a) if the data from Exercise 9.2.14 are used to get a preliminary estimate of  ?
(b) if we don't trust that data from women ages 40-44 are useful and thus no estimate of  is available (i.e.. if we wish to guard against the worst case)?