Tri State Bank wants to determine the proportion of its customers who would take advantage of its proposed new interest bearing checking account, and plans to select a simple random sample to make an appropriate estimate. It wants a sample size that will ensure a 95% probability that the actual population proportion will not differ from the sample proportion by more than 3 percentage points (.03).
a. Assume you have no prior information allowing you to estimate what the sample result might be, and you are not interested in doing any sort of pilot study. How large a sample would you recommend?
b. Assume a preliminary study suggests that the population proportion is approximately 20%. How large a sample would you recommend for the survey?