Quote:
Originally Posted by rubecube
|
Head to head polls where they pit a party nomination candidate against and incumbent are generally fairly useless in terms of predicting the general. I still think they have value in an electability argument as between the party candidates (for example, it would be meaningful to observe that "Klobuchar vs Trump is at 46% while Bernie is at 52%", if those were real poll numbers). But I don't think they have any value at all in telling you who will win a particular state, or the whole thing, if the candidate becomes the party's nominee.
Unless you have some really striking polls, anyway. I guess if you had a half dozen 1000-person-sample polls that showed Trump losing to someone 60-40, or a bunch of polls of people who voted Trump in 2016 in Ohio, Pennsylvania, Michigan and Florida that said that a big chunk of those polled would vote for the Democrat, that might be of value. But these "52-47, 50-46" ones, I don't put much stock in.