Statistical analysis on redistricting and women in office; sample size, ideas??

Okay. I have this huge paper that I've been working on for umpteen thousand years. I really need to graduate from college.

The question: how does redistricting affect the number of women elected to office??

It's a pretty straight forward answer: protecting incumbents hurts women because majority of incumbents are men.

I now want to tease out certain states from my equation, and need to know what sample size would be best

If I have 50 states, 2 political parties, roughly 330 million people/200K per district.

I want to tease certain states out of the equation because they are multi member districts. How many can I tease out before I ruin my sample size.

I'm thinking two tailed t test, is there any other test that might be more doable??

Thanks in advance and sorry if nothing makes sense, I just woke up.