New Fox News Poll- Obama 46, Romney 39

[Liability]

* * * *

By snipping all of FrenchiTubs' generally irrelevant babbling bullshit, we leave the kernel of value.

In this latest case, of course, once again there was no kernel of value.

 

[The T]

* * * *

By snipping all of FrenchiTubs' generally irrelevant babbling bullshit, we leave the kernel of value.

In this latest case, of course, once again there was no kernel of value.

So in effect AND AFFECT...you have quoted nothing of value that matters anyway...

 

[Vast LWC]

Rassmussen - Oct. 13, 2010 Inouye 53% Cavasso 40%
Actual Nov. 2, 2010 Inouye 74.81% Cavasso 21.57%

Yeah "dead on the nose"

In the 105 polls released in Senate and gubernatorial races by Rassmussen and Pulse Opinion Research they missed the final margin by 5.8 points (way too high for "professional" pollsters) 13 polls missed by 10 points or more, in 55 polls they overestimate Republican support by 3 points or more (as opposed to overestimating Democratic support by 3 points or more in 12 races)

Their average house bias is + 3.9 in favor of Republicans. An honest "margin for error" would swing both ways Einstein.

You ever met anyone with a high school math class under their belt? (You wouldn't get along)

Woah, those are excellent points.

Is there a link I can find this information on all in one place, or was this a compilation?

Because I'd love to peruse your source material.

 

[The Infidel]

* * * *

By snipping all of FrenchiTubs' generally irrelevant babbling bullshit, we leave the kernel of value.

In this latest case, of course, once again there was no kernel of value.

So in effect AND AFFECT...you have quoted nothing of value that matters anyway...

 

[The T]

By snipping all of FrenchiTubs' generally irrelevant babbling bullshit, we leave the kernel of value.

In this latest case, of course, once again there was no kernel of value.

So in effect AND AFFECT...you have quoted nothing of value that matters anyway...

I hope that Liability understands that I was bolstering...

 

[kiwiman127]

How's "None of the Above" doing in the polls?

 

[NYcarbineer]

How's "None of the Above" doing in the polls?

Apparently about 15%. I think the tea party should run someone.

 

[Liability]

* * * *

By snipping all of FrenchiTubs' generally irrelevant babbling bullshit, we leave the kernel of value.

In this latest case, of course, once again there was no kernel of value.

So in effect AND AFFECT...you have quoted nothing of value that matters anyway...

Sort of.

Well, other than the fact that I snipped.

In his case, that actually ADDS value!

 

[NYcarbineer]

Rassmussen conducts Fox polls via Pulse Opinion Research

Which makes you wonder why they diverge from Rasmussen.

Foxnews and CNN hit the 2008 election number on the nose btw,

Obama over McCain by 7

 

[NYcarbineer]

This Foxnews poll said that 43% of those supporting Romney did so because he 'wasn't Obama'.

Dude... Make up your mind... Foxnews or Opinion Dynamics...

You're all over the place...

The poll that Opinion Dynamics conducts FOR Foxnews is called the Foxnews poll.

Now stop. I'm not gay, I told you that over to hannity. enough with the flirting.

 

[Liability]

This Foxnews poll said that 43% of those supporting Romney did so because he 'wasn't Obama'.

Dude... Make up your mind... Foxnews or Opinion Dynamics...

You're all over the place...

The poll that Opinion Dynamics conducts FOR Foxnews is called the Foxnews poll.

Now stop. I'm not gay, I told you that over to hannity. enough with the flirting.

Stale material is -- redundant in carbuncle's case.

 

[The T]

By snipping all of FrenchiTubs' generally irrelevant babbling bullshit, we leave the kernel of value.

In this latest case, of course, once again there was no kernel of value.

So in effect AND AFFECT...you have quoted nothing of value that matters anyway...

Sort of.

Well, other than the fact that I snipped.

In his case, that actually ADDS value!

Dead on...YOUR quoting it gave him more value than desreved.

 

[Liability]

So in effect AND AFFECT...you have quoted nothing of value that matters anyway...

Sort of.

Well, other than the fact that I snipped.

In his case, that actually ADDS value!

Dead on...YOUR quoting it gave him more value than desreved.

I would agree except that I don't think "* * * *" constitutes all THAT much of a quote.

 

[The T]

Sort of.

Well, other than the fact that I snipped.

In his case, that actually ADDS value!

Dead on...YOUR quoting it gave him more value than desreved.

I would agree except that I don't think "* * * *" constitutes all THAT much of a quote.

But if one backtracks through the fog of "***"? YOU did him a huge favour...

 

[Liability]

Dead on...YOUR quoting it gave him more value than desreved.

I would agree except that I don't think "* * * *" constitutes all THAT much of a quote.

But if one backtracks through the fog of "***"? YOU did him a huge favour...

Ok. I will give you that one.

 

[The T]

I would agree except that I don't think "* * * *" constitutes all THAT much of a quote.

But if one backtracks through the fog of "***"? YOU did him a huge favour...

Ok. I will give you that one.

Following the logic brother.

 

[BluePhantom]

I have been hesitating to say much about this particular Fox poll but I am bored so I will. The big problem here is that most people look at poll results on RCP or wherever, see the spread, and start cheering or crying. In reality it's way more complicated than that. RV polls will give a natural lean toward this party, and you have to factor in demographics, and the laws of probabilities...it's just not as simple as looking at the bottom line of a poll and taking it as gospel because different polls measure different things. LV polls measure one thing, RV polls measure something completely different, A polls measure something completely different than the other two.

So what we're trying to find out when we look at a poll is really "who is going to win the election"...because "who do the majority of American citizens support" and "who will win the election" are two completely different questions because almost half of the people who are eligible to vote don't actually go vote.

Well there are numerous formulas and calculations for all this but the standard breakdown of people who actually show up on election day is generally 32% Democrat, 32% Republican, and 36% Independent. That will fluctuate a bit depending on how energized the base is, and all that crap, but usually speaking that's roughly how it turns out.

So people shouldn't be looking at the final results of any poll and taking them as gospel because as I pointed out earlier, polling is a game of probabilities just like anything else. If you flip a coin four times in a row on average you will get heads twice and tails twice, but every now and then you might get heads four times in a row or tails four times in a row or any mixture in between. The more times you run the exercise of flipping the coin four times in a row the greater your variation of results will be.

This is what I was getting at earlier when talking to nobraininafight about margins of error, standard deviation, and how they shrink when you expand the sample of a given set but expand when you increase the number of sets in an average.

So basically everyone is looking at the wrong thing. Instead of looking at this FoxNews poll and saying "Obama +7" and reacting according to who they support, it's important to look deeper. Fox has provided the crosstabs and the total sample size and with that and the generic 32/32/36 model all it takes is some very basic math to convert it into a generic model of who would win according to this poll if the election were held today. So let's do that.

Now we know that the FoxNews poll in question was 913 registered voters. According to the demographics question at about the middle of the link I provided, the sample was made up of 42% Democrats, 34% Republicans, and 20% Independents. Wow...that's a long way from the 32/32/36 that historically shows up on election day. Democrats are oversampled by 10 points, Republicans by 2, and Independents are undersampled by a whopping 16 points.

Thankfully they provided crosstabs just slightly further down; question #2. Democrats supported Obama 88% to 5% for Romney. Republicans were 6% Obama and 84% Romney. Independents were 29% Obama 34% Romney and the rest haven't made up their mind yet.

So let's take that total sample of 913 and weight it according to historical norms on election day:

Republicans: 913 x 32% = 292.16
Democrats: 913 x 32% = 292.16
Independents: 913 x 36% = 328.68

So we have split the sample of 913 (could be any number really) according to a generic average of how many Republicans, Democrats, and Independents will show up on election day. For now we will ignore undecided voters because we can move that overall sample of 913 to any number and the same thing will happen as we are expressing things as a percentage, but lets just go with the 913 they gave us.

Well according to the crosstabs 88% of Democrats would vote for Obama and 5% would vote for Romney and we know there are 292.16 Democrats in our generic LV sample. Pretty easy math:

292.16 x 88% = 257.1 Obama votes among Democrats
292.16 x 5% = 14.61 Romney votes among Democrats

Republicans were 6% Obama and 84% Romney

292.16 x 6% = 17.53 Obama votes among Republicans
292.16 x 84% = 245.41 Romney votes among Republicans

Independents were 29% Obama and 34% Romney (lots of undecided voters there)

328.68 x 29% = 92.03 Obama votes among Independents
328.68 x 34% = 111.75 Romney votes among Independents

So let's add them up

Obama: 257.1+17.53+92.03 = 366.66 votes
Romney: 14.61+245.41+111.75 = 371.77 votes

Obama: 366.66 / 913 = 40.16%
Romney: 371.77 / 913 = 40.72%

Difference: Romney +0.56

Now let's calculate margin of error against a standard confidence interval of 95%. In a single sample the formula is:

(1.96)sqrt[d(r)/(n)]

where
d = Obama percentage
r = Romney percentage
n = sample size

so we get
(1.96)sqrt[(.4016)(.4072)/913]
(1.96)sqrt[.163532/913]
(1.96)sqrt[.000179]
(1.96)(.013383)
.026231
or
2.62%

So we have is a 95% chance that it's Romney +0.56% with a MOE of 2.62%.

The result is less than the MOE and so according to this poll we are 95% certain that we have a statistical tie which is exactly what Rasmussen, Gallup, Quinnipiac, and everyone else is showing.

So this FoxNews poll of Obama +7 when weighted according to a generic likely voter model is right on the nose with everyone else...dead even.

 

[JoeB131]

obama has lost North Carolina and Colorado, Wisconsin has become a toss up state.

.

Ehhh, not so much. Both are pretty much still tossups...

 

[BluePhantom]

obama has lost North Carolina and Colorado, Wisconsin has become a toss up state.

.

Ehhh, not so much. Both are pretty much still tossups...

Obama's announced support of gay marriage just handed North Carolina to Romney on a silver platter. Wisconsin I have leaning to Obama but it's holding on to that status by its fingernails. One more poll within the MOE in Wisconsin and it's time to move Wisconsin to a toss up....Colorado I still have leaning to Obama

 

[JoeB131]

obama has lost North Carolina and Colorado, Wisconsin has become a toss up state.

.

Ehhh, not so much. Both are pretty much still tossups...

Obama's announced support of gay marriage just handed North Carolina to Romney on a silver platter. Wisconsin I have leaning to Obama but it's holding on to that status by its fingernails. One more poll within the MOE in Wisconsin and it's time to move Wisconsin to a toss up....Colorado I still have leaning to Obama

You can believe whatever you want there, guy.

But your boy is so desperate he's sucking up to Liberty University, which teaches his religions is a cult and how to avoid it.

It would be like Obama going to Bob Jones University when they still had the no-race-mixing policy.

Romney's biggest problem- when people get to know him, they don't like him. He's been doing well lately because he's not been in the spotlight. He's "Generic Republican" right now. He's back in the "MIttness Protection Program."