Tests whether two groups differ in proportion. The two samples are pooled to form the standard error, and the difference between the proportions is divided by it to give z. This is the test behind A/B testing.