**** Elections thread *****

[wufwugy]

Of course you think that.

Fwiw, I don't really support Clinton so much as find her less objectionable than Trump. If there was a viable third option of 'anyone but those two' I would almost certainly support them.


And while I certainly think there's something funny going on when the range of poll results is 14% between the most pro-Clinton and the most pro-Trump, I still haven't seen any actual evidence that any of those polls are 'rigged' in the sense of deliberately overstating one candidate's support.

The only poll I've really looked at in detail is the USC/LAtimes one where they used methods that can most generously be described as 'experimental'. There's other polls that are probably doing something suboptimal as well, but I can't be bothered to investigate every single polls' methods in detail.

If someone can actually point out a valid issue with a poll (or link to someone else who does), I'd be interested. Hasn't happened yet. The only complaint I've come across is 'zomgrigged D+ sample' which the poll in question corrected for and which doesn't constitute any evidence at all of 'zomgrigging'.

That said, any pollster who deliberately tried to rig their result wouldn't be stupid enough to not hide the evidence of rigging. So it could be going on.


As for vote rigging, the only evidence I've seen is that it's being done to improve Trump's chances, not Clintons.

Over a week of IBD getting a +R sample yet weighting it to +D at a greater tick than in 2008. Considering that in 2008, the sample that voted in the general election closely mirrored the sample that voted in the primaries, this weighting by IBD is hilarious since 2016 primaries sample was IIRC R+1. This is an example where the person (loosely) considered the best pollster of 2012 (Ann Selzer) would say IBD is making a big mistake. Selzer would say this because the poll that got her fame in the polling community was one where she did the opposite of what IBD is now doing.

If you're looking for proof of poll foul play, you're unlikely to find it. Although I have come across some pieces of supposedly hard evidence, they're small and can be argued away by anybody who wishes to do so. What we're dealing with instead is our political brains and political logic. As somebody who has followed politics and polling very closely for years, the polls do not come close to passing my smell test. My attempts to explain why tend to fall flat, apparently. I do believe I have given you enough to claim that they don't pass your smell test either. Things like the D+34 sample point to some serious issues with response bias.


I've seen a bunch of people claim they've been polled and hung up on after they said they're voting Trump or that they're Republican. Are they all lying? Maybe.

A statistically weird percentage of polls have had R+ sample data. Does this suggest something is fucky? Yes.

The polls are wildly inconsistent. Does this suggest something's wrong with them? Yep.

The pollster are weighting based on their own opinions, not on data. Evidence of an agenda? You betcha.

[Poopadoop]

This is an example where the person (loosely) considered the best pollster of 2012 (Ann Selzer) would say IBD is making a big mistake. Selzer would say this because the poll that got her fame in the polling community was one where she did the opposite of what IBD is now doing.

The last Bloomberg poll (which was overseen by Selzer) on Oct. 19 gave Clinton the biggest lead (9 points) of any poll done around that time. Maybe she forgot how to be the best pollster.

http://www.bloomberg.com/politics/ar.../national-poll

[wufwugy]

The last Bloomberg poll (which was overseen by Selzer) on Oct. 19 gave Clinton the biggest lead (9 points) of any poll done around that time. Maybe she forgot how to be the best pollster.

http://www.bloomberg.com/politics/ar.../national-poll

Iowa Republican Caucus.

I didn't bring her up to say that she was the best or infallible. She did not perform spectacularly outside of the Iowa Caucus. But what got her the best result in an election that many others got very wrong is by using a specific tactic that would be very relevant now yet pollsters are not using.

[Poopadoop]

Iowa Republican Caucus.

I didn't bring her up to say that she was the best or infallible. She did not perform spectacularly outside of the Iowa Caucus. But what got her the best result in an election that many others got very wrong is by using a specific tactic that would be very relevant now yet pollsters are not using.

You don't think she used that technique in the Bloomberg poll?

[wufwugy]

You don't think she used that technique in the Bloomberg poll?

She probably did, and something I covered earlier can help explain why it was not a telling choice in that situation. The Iowa Republican Caucus was a particular time when the race was so sporadic and hard to pin down that it turned out the best polling strategy was to (her words) "listen to what the voters were saying" instead of using standard assumptions. In the general, that was unlikely the case. However, in this cycle we have evidence to believe Selzer's strategy would work better than the orthodox ones.

[Poopadoop]

She probably did, and something I covered earlier can help explain why it was not a telling choice in that situation. The Iowa Republican Caucus was a particular time when the race was so sporadic and hard to pin down that it turned out the best polling strategy was to (her words) "listen to what the voters were saying" instead of using standard assumptions. In the general, that was unlikely the case. However, in this cycle we have evidence to believe Selzer's strategy would work better than the orthodox ones.

So if you give her credit for being a competent pollster (which I think you do), how is it the Bloomberg poll she oversaw gave Clinton a bigger lead than the ones you're claiming are fucked up, like the IBD poll?

[wufwugy]

So if you give her credit for being a competent pollster (which I think you do), how is it the Bloomberg poll she oversaw gave Clinton a bigger lead than the ones you're claiming are fucked up, like the IBD poll?

It could be all sorts of reasons. It's also irrelevant. I merely pointed out one example where her Iowa Caucus logic would apply. IBD continues to get samples that are telling it turnout is not going to look like the weights they're using suggest. They should take that into account, but they are not taking that into account.

BTW I'm not sure if I've said the polls are rigged. I think it is mostly momentary incompetence derived in part from a few different agendas (not being outliers from the pack since doing that and being wrong would destroy their business, and preferring a Clinton presidency).