Elixir
Made in America
I set up a scenario in which 50% say yes but only 10% were raped.
You "set up a scenario"? Well look at you putting your big boy pants on.
Did the poll do that, though? Which is the only relevant question.
No, it did not. It did not ask the women in "two families, each containing 5 women" the question.
THIS poll--the professionally conducted poll, using a representative sample of people who don't know each other and aren't members of each other's families--extrapolates to at least half of all women in America having been sexually assaulted and that this number is the lower bound, which means it's likely much higher.
So, are you done yet, or is there more pointless derail up your ass?
I used a small group to get a 100% sample so there could be no statistical error. The whole point of polling is to use a representative sample to guess at the actual value. Scale it up and sample instead of asking everyone and you should get the same results with a bit of statistical error thrown in the mix. You have flunked basic statistics here.
Might be more useful to decide who is the arbiter of "representative" when talking about the relationships between various sample sizes and the actual populations they supposed to represent. Statistical consistency doesn't necessarily correspond with empirical reality.