a. If x successes result from a binomial experiment with n = 1000 and p = P(success) and the 95% confidence interval for the true probability of success is determined, what is the maximum value possible for the “maximum error of estimate”?
b. Compare the numerical value of the “maximum error of estimate” found in part a with “the margin of error” discussed in Applied Example
9.9.
c. Under what conditions are they the same? Not the same?
d. Explain how the results of national polls, like those of Harris and Gallup, are related (similarities and differences) to the confidence interval technique studied in this section.
e. The theoretical sampling error with a level of confidence can be calculated, but the polls typically report only a “margin of error” with no probability (level of confidence). Why is that?