What is the difference between a z-test and a t-test? When estimating population mean, how would you determine the optimal sample size when the maximum error limit is known; however, population variance is unknown and your initial sample size is less than 30? Provide a numerical example.