Link: Stats
Use computer to randomly take samples from a distrbution curve.
If we do lots of these tests, we can get lots of t-test result which generates P-value. From how frequently the t-test gives a small p-value (means t-test worked), we know if we need to increase the sample size or not.