Re:What I learned trying to classify abortion access across ...
by lesuorac on Monday July 18, 2022 @05:53PM
(#32144416)
> Polls look pretty extremely variable. Some polls show half of Americans are against abortion.
> Some polls show that of the Americans that support abortion, support drops after the first trimester.
Those are two separate claims that can easily both be true. A big problem with polls about abortion is that its not actually a well defined term (how many weeks? is mothers life in danger? rape? etc).
> This seems like the shy republican polling problem in 2016 all over again.
Not quite on topic, but polls did not give hilary a 100% chance to win the 2016 election. Polls are extremely good at predicting events (compared to say a literal coin flip) but lower probability events do occur. Go flip a coin a bunch of times and you'll get 2 heads in a row (25% chance! thats even worse than what 538 gave trump [1]).
Those are two separate claims that can easily both be true. A big problem with polls about abortion is that its not actually a well defined term (how many weeks? is mothers life in danger? rape? etc).
> This seems like the shy republican polling problem in 2016 all over again.
Not quite on topic, but polls did not give hilary a 100% chance to win the 2016 election. Polls are extremely good at predicting events (compared to say a literal coin flip) but lower probability events do occur. Go flip a coin a bunch of times and you'll get 2 heads in a row (25% chance! thats even worse than what 538 gave trump [1]).
[1]: https://projects.fivethirtyeight.com/2016-election-forecast/