Two-proportion z-test

Also written: z-test

An unpaired test for an A/B experiment that asks how likely a gap between two rates from different users would be if both versions had the same true rate.