re: Battleground polls: Will Trump defy the odds again?

quote:

---

You're assuming that these polls are attempting to be accurate as opposed to attempting to influence voters

---

In terms of who they are planning to vote for, I don't see any evidence a legitimate pollster (doesn't mean a good pollster per se), even if they have a partisan leaning, is attempting to influence voters rather than get as accurate of a snapshot of their preferences as possible.

Now sometimes the supplemental questions (e.g., how do you feel about this specific issue) might be a little more for political purposes that can be used to influence people for political reasons.

That being said, I don't really know how the result of "who are you planning to vote for," influences voters and if they do influence voters in which direction.

For example, I remember people arguing that they were manipulating the results into a larger margin for Hilary in an attempt to discourage Trump voters. But if they did manipulate it, then they probably ended up discouraging Hillary voters if anything (e.g., don't have to vote for Hillary since she's going to win).

In addition, I bought into the "unskewing" of the polls in 2012 and JoshChamberlin's (as a lurker) posts about Romney leading. Not only was that wrong, the polls actually underestimated Obama's margin. And I wonder if that actually caused them to overcompensated in 2016, which is why the overestimated Hillary's margin.

Finally, if it comes to manipulating the polls, I don't think it's the pollsters themselves that are the likely culprits. For example, OAN apparently paid for a bunch of polls with legitimate pollsters; HOWEVER, they only released the ones that were favorable to Trump and withheld the ones that weren't. So the results were not invalid, but they were lying by omission. That being said, I think OAN is an exception because even highly partisan sites like Breitbart appears to honestly release the results of the polls they pay for.