When do you use a z-value rejection region when testing a claim about a population mean?

Answer 1

If you know the population variance or if you have a very large sample then you could reliably use a Z test. Otherwise you should use a t test and use s^2 as an estimator for the population variance.