how can a confidence interval estimate for the population mean provide conclusions for the corresponding two-tail hypothesis test for the population mean?