POTUS Debate II: The Empire Strikes Back at the Wrath of Electric Mittensaloo

[DGuller]

Not sure how Gallup is a crap poll.  They have the largest sample size of all the polls (2700).

Sample size is one of the many considerations when considering the quality of the poll, and by far the easiest to quantify.  Sample size won't help you one bit if your sampling methodology is flawed.  You can sample 100,000 people and still be off if all your respondents are in Utah.

[MadImmortalMan]

Hasn't Gallup been the most accurate one?

[garbon]

I think D has been on edge since that quiz told him to vote for Romney.

[DGuller]

And in any case, the cherry-picking point applies much more strongly.  Any poll is going to have some margin of error, even the most well-designed, consistent, and unbiased ones.  If you're going to go with the most jaw-dropping one each time, citing Pew Research one week, Gallup the next, and then whichever one looks the most ominous for Obama the week after, you're going to be off by a lot. 

Your strategy of picking a poll to cite is going to always find a poll whose house effect combined with random error swings the most from the true number, and that combined effect may be much higher than the stated margin of error.  Margin of error, apart from assuming perfect sampling, also assumes that the poll result is picked at random.  By picking the most favorable poll for Romney, no matter how reputable it is, you're blowing that assumption out of the water.

[DGuller]

Hasn't Gallup been the most accurate one?

Actually, it's the one most tilted towards Republicans, at least according to 538.

[Eddie Teach]

It's a conspiracy. Romney's winning because of internet quizzes telling susceptible Democrat leaners they like Romney. 

[garbon]

It's a conspiracy. Romney's winning because of internet quizzes telling susceptible Democrat leaners they like Romney. 

Damn Christian Science Monitor and their dastardly plot!

[OttoVonBismarck]

Hasn't Gallup been the most accurate one?

Actually, it's the one most tilted towards Republicans, at least according to 538.

There is no evidence of an ideological bent in Gallup's organization, unlike with say, Rasmussen. Polling isn't exactly as simple as just calling a few thousand random people and asking them questions. You have to try to create a "representative sample." Every pollster has a different technique for determining what is a "representative sample" some polls might do this to intentionally make one candidate look stronger, but most of the "big polls" like Reuters, WaPo/ABC, WSJ/NBC, Gallup, NyTimes/CBS I don't think have any intentional biases along party lines. They have different ways of constructing their sample which, in different elections will be more or less favorable to different parties and candidates.

I'm not a Karl Rove fan per se, but he made a good point in the Wall Street Journal the other day, that people ascribe a certain aura of scientific accuracy to polls that just aren't real. When individuals look at a poll they often ignore the margin of error, assuming it's just a "hedge" and the poll result is all that matters. In fact the margin of error is extremely important, as are the questions asked, how they were asked etc.

That is why things like RCP's Poll of Polls and Nate Silver's 538 are a good way to look at the big picture, because they look incorporate polls from many agencies throughout the election.

Something I find interesting, we all know Nate Silver is a lefty, but some people actually think his "percent" means something other than it does. I've met people who think his current 65% chance to win for Obama is akin to Obama "polling 65%." Nothing could be further from the truth, what it actually means is that under Silver's model on election day he's saying there is a 65% chance that the election goes to Obama. That's not the same as Obama polling 65%--if he was polling that there would be a 100% chance of him winning election.

Within the context of Silver's own model, he considers anything less than 60/40 for one candidate to be "up in the air." So he's projecting an Obama victory, but under his model if Obama's chance to win hit 60% or lower he'd consider the election "up in the air."

[HVC]

I have officially entered the race to get a rise out of meri

There. You win!

woo hoo

[Josephus]

Here's the breakdown of yesterday's debate viewers:

NBC: 13.8 million total viewers
ABC: 12.5 million total viewers
Fox News: 11.1 million total viewers/3.460m adults 25-54
CBS: 8.9 million total viewers
CNN: 5.8 million total viewers/2.581m adults 25-54
MSNBC: 4.9 million total viewers/1.920m adults 25-54
FOX: 4.6 million total viewers

So Fox News viewers kicked MSNBC viewers in the chin. What to make of that?

[Ed Anger]

Msnbc sucks, and continues to suck in the ratings.

[derspiess]

It's a conspiracy. Romney's winning because of internet quizzes telling susceptible Democrat leaners they like Romney. 

He needs to chill out a bit and stop with the stalking. 

But as to the polls, I said several months ago that the two I mainly follow are Rasmussen and Gallup.  I cited the Pew poll the other day with a bit of skepticism, but apparently it still offended him that I dared mention it.

I know it makes him uncomfortable to even think of the possibility of Obama losing, but it might be wise for him to grow up a bit and confront that fear.

FWIW, I'm still sticking with my prediction from a while back of Romney winning the popular vote and Obama winning EC.

[derspiess]

Here's the breakdown of yesterday's debate viewers:

NBC: 13.8 million total viewers
ABC: 12.5 million total viewers
Fox News: 11.1 million total viewers/3.460m adults 25-54
CBS: 8.9 million total viewers
CNN: 5.8 million total viewers/2.581m adults 25-54
MSNBC: 4.9 million total viewers/1.920m adults 25-54
FOX: 4.6 million total viewers

So Fox News viewers kicked MSNBC viewers in the chin. What to make of that?

That's pretty representative of cable news network viewership.

[DGuller]

He needs to chill out a bit and stop with the stalking. 

You seriously need to chill out with the accusations of stalking.  It was amusing at first, but now it's getting personal.  It's a very dirty tactic to discredit someone and de-legitimize their posts, so I would ask you kindly to fuck off and take it like a man if someone disagrees with you.

[Eddie Teach]

So Fox News viewers kicked MSNBC viewers in the chin. What to make of that?

That most lefties consider CNN and the networks to be "fair" and so don't need a channel to explicitly reaffirm their own biases.