Hello there:

I have a couple of questions about comparing proportions.

1) why do we use a z-test to compare proportions but t-test to compare means? What is the theory behind that?

2) what would be an appropriate test to compare proportions where there may be some respondent overlap? Currently I am using McNamara.

Very grateful for any insights. I work in a field where very often statistics are applied simply by memorizing a list of rules of thumb. I would like to understand some basic theory as well.