Overtime Politics Thread (WARNING: Possible fraud)

[Alcon]

This is coming from someone with no emotional attachment to the results of this primary:

These polls are incredibly sketchy.  The releases look like someone who's looking at polling releases and trying to duplicate their appearance, but missing nuances like age distribution on a caucus sample.  For someone who produces so many polls in so many jurisdictions, he acts like someone with very limited experience in the industry...or any political industry, for that matter.  He sounds like a guy with limited experience operating out of his basement, and that's not commensurate with the scale of operation he claims to be running.

I hate to malign someone over "bad vibes," but man there are a lot of things that just seem "off" here.  Even if there's not fraud going on, he's produced enough dubious subsample results to make me skeptical of the firm's competence.

Ockham's razor says fraud; nearly every other remotely plausible explanation says either fraud or substandard practices.

[Alcon]

naturally the black&hispanic primary voters will go down without Obama on the ballot

Why Hispanic voters?  Clinton pretty decisively won Hispanics in 2008.  It's not like there is some unified brown vote, or many Hispanic voters showed up just to vote against a black guy.

[Alcon]

"How the undecideds caucus could make or break one of the lower tier candidates, or it could add a delegate to any of the front runners."

Overtime Politics hot take: Undecided voters may decide to vote for a candidate

[Alcon]

Overtime showed a statistical tie in Iowa, and that's what actually happened. Time to stop treating them like junk for dems. For the Republicans though........

...that's really not how that should work.

[Alcon]

Overtime showed a statistical tie in Iowa, and that's what actually happened. Time to stop treating them like junk for dems. For the Republicans though........

...that's really not how that should work.

Yes it is how it should work.

oh lord I'm too tired for your silly, silly nonsense tonight, but WHATEVER:

1. You are basing this conclusion on a sample size of ONE, which is statistically ridiculous from several angles.  Think about how margin of error works, among other things.

2. You are assuming, for no apparent reason whatsoever, that this pollster used a superior methodology.  It's quite possible -- likely, even -- that the other pollsters used the sound, reasonable methodology for voter screens.  It's impossible for pollsters to be psychic about voter screens.  Using a screwy voter screen and getting lucky doesn't make a pollster good, unless they can explain the methodological rationale behind their voter screen.  It just makes them screwy.  Note the difference between Ann Selzer, who didn't nail this one but can fluently discuss her approach to voter screening, and Overtime, which talks exactly how you'd expect someone who is making things up.

3. Why would you think that this pollster had some sort of secret insight into polling the Republican Primary but not the Democratic Party?  Do you have any methodological reason whatsoever to believe that, besides that they happened to get one of the races correct?

4. Most importantly, as Adam says, any reasonable person who has followed Overtime should conclude it's more likely that this pollster is NOT REAL than that they've come up with some brilliant methodological approach based on one damn poll.

5. I really doubt you have any statistically sound argument to back up your claim that "Yes it is how it should work."  This isn't an argument against your statement, but I just wanted to reiterate how completely ridiculous you're being.

GOODNIGHT

[Alcon]

Even if their New Hampshire poll does fine, I'm going to remain completely skeptical about this firm.  Tarot readings are right sometimes too, and they're much more complex than creating a fake poll topline.