What the hell is with these pollsters?

[asaratis]

Why is that poll bogus?

Because it is not represenative of the political landscape of our country. Not even close.

It is only the November election results that will constitute a valid polling of the American public (and whatever alien/dead votes the Democrats are able to garner) and this will still be a poll confined to a sampling of ACTUAL VOTERS. There will be many people that will not vote (for whatever reason) and will not be represented in the poll.

All these polls in the run up to the election have great potential for fraudulent interpretation or accidental skewing of the true feelings of those sampled. The way the question is asked, who is asked, when they are asked...all have bearing on the poll results. Politicians love to publish favorable polls in hopes that they will convince fence walkers to jump on their "winning" bandwagon.

83.5% of all statistics are made up.

 

[Liability]

So they had 9% more Dems? Who the fuck cares? That doesn't mean you toss out the whole fucking poll. It was a random selection of people. Jesus Christ you guys are thick.

Anyone who cares about honesty and integrity.

Clearly not you

So pollsters are supposed to go out of their way in a blind sampling to make sure they get a 50/50 split? What if people lie? What if in this poll, people lied about how they're registered?

Seriously. You're making a mountain out of a molehill. Because your guy is getting beaten like a red-headed step child. All that money spent on a fucking TURD of a candidate.

Hey, how come Mittens hasn't spent any of that $43 million of his own money he spent last time? Think it's because it made him look like an out of touch rich bitch trying to buy the Oval Office?

Your guy sucks, and the polls are showing it. Sorry, Bagtards.

No no, you motherfucking stupid, ignorant and/or dishonest twat.

But sampling that OVERsamples for one group in a way that is so disproportionate as to not even come close to that which it is trying to capture makes the entire exercise useless.

As a purely illustrative hypothetical: If 53 Dims are going to vote for every 47 Republicans (on average), then having 6% more Dims the Republicans in the random sample makes the poll, arguably, more accurate as a barometer.

But if the random sample leads to 6% more Republicans, its predictive power is going to go out the fucking window. And even if it has 10% more Dims than Republicans, although closer, that still won't validly capture the likely outcome of the actual vote.

This error is compounded when they sample REGISTERED voters rather than LIKELY voters.

But I'm sure ALL of this is sailing HIGH over your stupid pinhead at mach speed.

 

[Dick Tuck]

Utter nonsense. The numbers are fluid and change regularly. Party Affiliation: 36% Democrats, 33% Republicans - Rasmussen Reports™

How'd Rasmussen arrive at those figures?

(hint: random sample survey)

How'd Fox arrive at their figures?

(hint: random sample survey)

A "random sample" survey doesn't act like the Mr. Clean magic eraser.

If the random sample over samples for Dims over Republicans, it still over samples. This still leads to skewed results.

Why do you think they weigh their results when they oversample?

 

[8537]

I'm still waiting for someone to explain how the pollsters is supposed to know exactly the mix of Dems, Reps and independents to use in order to NOT oversample anyone.

 

[onecut39]

What is the logic for Fox to publish a bogus poll

Maybe it is remorse for their previous twisted ways.

 

[Liability]

How'd Rasmussen arrive at those figures?

(hint: random sample survey)

How'd Fox arrive at their figures?

(hint: random sample survey)

A "random sample" survey doesn't act like the Mr. Clean magic eraser.

If the random sample over samples for Dims over Republicans, it still over samples. This still leads to skewed results.

Why do you think they weigh their results when they oversample?

(A) that would assume that they do.

(B) it would also assume that their methodology of weighting the results properly corrects for the over-sampling, and

(C) that the weighting simultaneously compensates for the skewing caused by randomly sampling Registered voters rather than Likely voters.

 

[Remodeling Maidiac]

I'm still waiting for someone to explain how the pollsters is supposed to know exactly the mix of Dems, Reps and independents to use in order to NOT oversample anyone.

Im still waiting for one of you to stop excusing bogus polls just because they favor your team.

 

[PLYMCO_PILGRIM]

I'm still waiting for someone to explain how the pollsters is supposed to know exactly the mix of Dems, Reps and independents to use in order to NOT oversample anyone.

Your curiousity for an explaination is legit as the pollsters typically do not know, in advance, the political affiliation of those polled.

Hence why I take no issue with this poll considering the pollster (fox news) told us the breakdown of political affiliations within the sample.

Now if they just put up the poll but didn't clarify I might have an issue IF I later found out there was a discrepency that was large enough to impact the results.

 

[8537]

I'm still waiting for someone to explain how the pollsters is supposed to know exactly the mix of Dems, Reps and independents to use in order to NOT oversample anyone.

Im still waiting for one of you to stop excusing bogus polls just because they favor your team.

So I'll take this as an indication that YOU aren't going to be the one to explain how pollsters are supposed to know the exact mix to sample.

I'll wait - maybe someone else can explain it.

 

[onecut39]

What is the logic for Fox to publish a bogus poll

Maybe they are doing penance for their past wicked ways.

 

[Liability]

I'm still waiting for someone to explain how the pollsters is supposed to know exactly the mix of Dems, Reps and independents to use in order to NOT oversample anyone.

Im still waiting for one of you to stop excusing bogus polls just because they favor your team.

As part of the polling questions, would it be impossible to ask, as ONE of the questions, do you consider yourself to be a Democrat, a Republican an Independent or "other?"

And maybe, since they are randomly calling REGISTERED voters, they could ALSO ask, "how likely is it that you will be voting on Election day?"

 

[8537]

I'm still waiting for someone to explain how the pollsters is supposed to know exactly the mix of Dems, Reps and independents to use in order to NOT oversample anyone.

Im still waiting for one of you to stop excusing bogus polls just because they favor your team.

As part of the polling questions, would it be impossible to ask, as ONE of the questions, do you consider yourself to be a Democrat, a Republican an Independent or "other?"

And maybe, since they are randomly calling REGISTERED voters, they could ALSO ask, "how likely is it that you will be voting on Election day?"

They do ask which party you consider yourself. That's where they get the sampling figure.

 

[ConservaDerrps]

I'm still waiting for someone to explain how the pollsters is supposed to know exactly the mix of Dems, Reps and independents to use in order to NOT oversample anyone.

Im still waiting for one of you to stop excusing bogus polls just because they favor your team.

As part of the polling questions, would it be impossible to ask, as ONE of the questions, do you consider yourself to be a Democrat, a Republican an Independent or "other?"

And maybe, since they are randomly calling REGISTERED voters, they could ALSO ask, "how likely is it that you will be voting on Election day?"

They obviously did ask that question, or at least somehow gleaned it. Otherwise the poll wouldn't say that it had 9% more Democrats than Republicans.

It's a legit poll. Those who disagree with the findings are the ones talking shit.

 

[Remodeling Maidiac]

I'm still waiting for someone to explain how the pollsters is supposed to know exactly the mix of Dems, Reps and independents to use in order to NOT oversample anyone.

Im still waiting for one of you to stop excusing bogus polls just because they favor your team.

So I'll take this as an indication that YOU aren't going to be the one to explain how pollsters are supposed to know the exact mix to sample.

I'll wait - maybe someone else can explain it.

Im not in the polling business but it would seem to me that asking prior to administering the poll would be a good starting point for fairness in the results. Why not do a 40/40/20 split or whatever number is represenative of the documented numbers? When you reach your % marker you stop polling that group. Seems pretty fucking simple to me but what the fuck do I know.

 

[8537]

I'm still waiting for someone to explain how the pollsters is supposed to know exactly the mix of Dems, Reps and independents to use in order to NOT oversample anyone.

Your curiousity for an explaination is legit as the pollsters typically do not know, in advance, the political affiliation of those polled.

Hence why I take no issue with this poll considering the pollster (fox news) told us the breakdown of political affiliations within the sample.

Now if they just put up the poll but didn't clarify I might have an issue IF I later found out there was a discrepency that was large enough to impact the results.

My question was partly rhetorical. The only way to know the proper sample would be to do a census of the entire country. Short of that, any attempt to survey and arrive at the figure would involve a random sample.

 

[Caroljo]

I have NEVER had anyone contact me. And i always vote......

That is proof right there. All polls must be bullshit.

Did i say that?

I just made the point that i've nevered been polled on an election.

 

[8537]

Im still waiting for one of you to stop excusing bogus polls just because they favor your team.

So I'll take this as an indication that YOU aren't going to be the one to explain how pollsters are supposed to know the exact mix to sample.

I'll wait - maybe someone else can explain it.

Im not in the polling business but it would seem to me that asking prior to administering the poll would be a good starting point for fairness in the results. Why not do a 40/40/20 split or whatever number is represenative of the documented numbers? When you reach your % marker you stop polling that group. Seems pretty fucking simple to me but what the fuck do I know.

What are the documented numbers and how did the documenters arrive at them?

hint: The documented numbers are from a random sample survey.

Are you beginning to see the problem here?

 

[Remodeling Maidiac]

Im still waiting for one of you to stop excusing bogus polls just because they favor your team.

As part of the polling questions, would it be impossible to ask, as ONE of the questions, do you consider yourself to be a Democrat, a Republican an Independent or "other?"

And maybe, since they are randomly calling REGISTERED voters, they could ALSO ask, "how likely is it that you will be voting on Election day?"

They obviously did ask that question, or at least somehow gleaned it. Otherwise the poll wouldn't say that it had 9% more Democrats than Republicans.

It's a legit poll. Those who disagree with the findings are the ones talking shit.

The findings are accurate. They arent however represenative of our country. Sorry that point is too complicated for you to grasp.

 

[AmyNation]

As part of the polling questions, would it be impossible to ask, as ONE of the questions, do you consider yourself to be a Democrat, a Republican an Independent or "other?"

And maybe, since they are randomly calling REGISTERED voters, they could ALSO ask, "how likely is it that you will be voting on Election day?"

They obviously did ask that question, or at least somehow gleaned it. Otherwise the poll wouldn't say that it had 9% more Democrats than Republicans.

It's a legit poll. Those who disagree with the findings are the ones talking shit.

The findings are accurate. They arent however represenative of our country. Sorry that point is too complicated for you to grasp.

How would you make them more accurate?

 

[Liability]

Im still waiting for one of you to stop excusing bogus polls just because they favor your team.

As part of the polling questions, would it be impossible to ask, as ONE of the questions, do you consider yourself to be a Democrat, a Republican an Independent or "other?"

And maybe, since they are randomly calling REGISTERED voters, they could ALSO ask, "how likely is it that you will be voting on Election day?"

They do ask which party you consider yourself. That's where they get the sampling figure.

Right. So that answers YOUR question, then. Randomly SELECT the roughly correct percentage of respondents from that group of the entire sample of respondents to correspond with the actual percentages of Dim vs. GOP voters.

That's one thing they could do.