The Pew poll is a joke.

[ClusterFreak]

Romney 50 Obama 45 or whatever it is....?

About 1100 people surveyed....4 in 10 of those come from the South...about 500 are self-described Conservatives, around 200 are self-described Liberals (yes...200. 200, that's right. 200!) and the remainder are self-described moderates.

The poll is a fucking joke. It is soooo heavily skewed Repuke...it has no legitimacy whatsoever. None!

Polls which oversample Repukes, or oversample Dems, are bogus and need to be outed as such.

Now if we could only get the MSM to see things the same way. Of course that's wishful thinking cuz by gawd they have a new narrative to push for a while. The old one, the one with Obama pulling away and well on his way to an electoral rout, was getting a little tired I guess.

Gonna go have a brewski and relax now.

[budkin]

1. Do you mind posting the breakdown?

I'm too lazy to go googling. Thanks.

[SidDithers]

2. Here you go...

Hopefully, it's not too small to read.

Sid

[regnaD kciN]

4. I normally don't feel much sympathy for those looking to "unskew" polls...

...especially not from respected pollsters like Pew, but those numbers seem a little...unusual.

One thing's for sure, though: if this poll is accurate in having almost 10% Republicans than Democrats show up at the polls, not to mention 2.5 times as many conservatives as liberals, we will lose -- and no one should be surprised about that.

[budkin]

6. Wow that really is skewed, Wtf?

Makes it a little more easy to swallow though.

[tarheelsunc]

7. So...

I've been a bit put off by all the poll criticizers, but a few of these demographics are just wrong.

5.1% Hispanic
7.6% 18-29 years old
30.4% Below 50

I did a quick search and found the same groups from 2008:
9% Hispanic
18% 18-29
47% 18-44 (unsure about 45-49)

I would imagine the Hispanic electorate has grown in the past 4 years, and I don't think the youth vote will be more than halved. It's also quite a stretch to say 70% of voters will be above 50.

[Shanti Mama]

12. Thanks for ths info.

I can't take the time right now to dig this deep, but you did it for all of us.
Much appreciated.

[Patiod]

18. Are these demographics representative of likely voters

based on the previous election?

[tarheelsunc]

27. I just took them straight from the exit polling in 2008.

I wouldn't expect those same numbers again, but I imagine the real 2012 numbers will be closer to those than what Pew had in these categories.

[Bjorn Against]

11. The regional numbers prove the national number is skewed

Of the four regions he is only losing one of them and that is the south while the midwest is dead even. In the west however Obama has a solid lead and the northeast he is winning by an even larger margin than Romney is winning the south. How could Obama be ahead by large margins in the two most heavily populated regions of the country and tied in the midwest yet still be losing to Romney who is ahead only in a single region of the nation and by a margin that is less than Obama's margin in a more heavily populated region? The numbers don't make any logical sense.

[Honeycombe8]

19. The Pew poll from 9/12-9/16 (Obama +8 over Romney) also had more conservatives & southerners.

IDEOLOGY
Conservative 27 67 6 979 25 70 5 911
Moderate 55 37 7 832 58 36 6 741
Liberal 86 8 5 518 90 6 4 470

That poll also had more respondents from the south:
REGION
Northeast 56 34 10 436 56 36 8 399
Midwest 51 42 7 589 53 41 5 534
South 48 45 7 898 47 47 6 804
West 52 43 6 501 52 42 5 455

Results of that poll had Obama 8 pts over Romney.

http://www.people-press.org/files/legacy-pdf/09-19-12%20Political%20release.pdf

[hrmjustin]

3. $ out of ten are southerners. I do not think so. What a bunch of bull.

[UTUSN]

5. R#4 & K for, thanks!1 n/t

[CakeGrrl]

8. Translation: POTUS is doing pretty darn well with that skewed demographic.

This is why I watch my premium movie channels rather than cable news.

[Hutzpa]

10. This Pew poll is extremely skewered

Regions

Northwest 201

Midwest 271

South 417

West 223

These numbers are for the month of October

Now lets take a look at September

Northwest 219

Midwest 294

South 449

West 239

So basically the Pew poll is over sampling the South more than the other regions.

There are no Hispanic and less African American, according to this poll, the only people that votes are Whites from the South.

In conclusion, whenever we see or hear about Pew research we should assume it is a skewered poll just for White Southerners only.

We get it.

[Drunken Irishman]

13. 37% of the U.S. population lives in the South...

So, I don't think their numbers are far off. If they're three-points off, it does change the results, but it's all within the MOE.

[Hutzpa]

21. Within the Margin of Error?

this is what I am looking at;

[ProSense]

15. That shows LV and RV for Oct. Here are

the numbers for Sept. likely voters

Northwest 399

Midwest 534

South 804

West 455



Conservative 911

Moderate 741

Liberal 470





Conservative Republican 501

Mod/Lib Republican 159

Mod/Cons Democrat 435

Liberal Democrat 338



http://www.people-press.org/files/legacy-pdf/09-19-12%20Political%20release.pdf


With the conservative skew, it appears undersampling Democrats will produce the kind of results in the current poll.

[Honeycombe8]

22. To judge the accuracy of THAT poll, you need to compare it w/the prior Pew poll (Obama +8 over R)

See my post above.

The Pew poll from 9/12-16 was also skewered in the same way. The results of that poll had Obama 8 pts over Romney.

So......it does look like, as far as THAT poll goes, Obama has dropped quite a bit. Maybe neither of them is right, but the difference between teh 2 results is still significant.

[ProSense]

26. It was not

"The Pew poll from 9/12-16 was also skewered in the same way."

...skewed in the same way. With the conservative skew (http://www.democraticunderground.com/?com=view_post&forum=1002&pid=1497715), it appears undersampling Democrats will produce the kind of results in the current poll.

Current LV

Republican 36 percent
Democrat 31 percent
Independent 30 percent

September LV

Republican 29 percent
Democrat 39 percent
Independent 30 percent


Current RV

Republican 33 percent
Democrat 32 percent
Independent 31 percent


September RV

Republican 28 percent
Democrat 37 percent
Independent 31 percent

http://www.people-press.org/files/legacy-pdf/10-8-12%20Political%20Release.pdf

[Honeycombe8]

28. It's close to the same. Except a few more Dems before. Still, way more "conservatives."

You can think what you want. I can see the numbers. There has been a large swing in that Pew polling, and the Pew polling has more conservatives in it, in both of them. And a small % of more Repubs in the latest, but nothing at all to explain a 12 pt swing.

I take all polls into consideration. It's a loser's game deciding the poll that you don't like isn't fair, but the ones that are good for you, are the fair ones.

No one complained about the Pew poll when it showed Obama 8 pts over Romney, even though it had more conservatives and over sampled Dems.

[ProSense]

30. If you have a conservative sample, poll more Republicans and fewer Democrats

you get a shift among women that the current poll shows.

Women didn't suddenly shift to Romney. It shows the increase in Republicans.

"No one complained about the Pew poll when it showed Obama 8 pts over Romney, even though it had more conservatives and over sampled Dems."

Well, if they're going to go exact opposite, they should oversample Republicans and skew liberal/moderate.

I mean, how the hell does anyone expect a conservative sample packed with more Republicans to produce anything but a Romney lead?

The poll that showed Obama ahead was still skewed conservative.

[Patiod]

31. I am a little wary of criticizing polls as well

Since I work in market research, early in my career I spent hours involved in lengthy discussions about sampling, weighting, methodology. I do not have the quant chops any more to evaluate their sampling, but there are usually so many people who have input into a poll like this - one that's so focused-on - that someone internally would pipe up if it were truly intentionally skewed one way or another.

And it's skewed, not skewered, although I know people are trying to be funny. I just don't want to sound like one of those dopey tea baggers/freepers "it's hugh!"

Also, not that anyone here came out and said it (although it may have been implied), but the size was good. About 1100 is right for a large national poll. After that, you get only the smallest of increases in accuracy.

[reformist2]

29. When you reduce Dems by 7% and increase Repugs by 7%, you can expect a 14 point swing.

And the Pew poll wet from +8 Obama to -4 Obama, a 12 point swing. So basically it showed what you'd expect.

The only news here is how bad the Pew Polling methodology is.

[Mass]

14. This just means that Democrats did not answer the poll in the days after debate. They are coming

back.

A data that pollsters should publish is what is their answer rate. At this point, it is only about 10 % and of course, those are only the most motivated that answer. In the previous weeks, Democrats were energized and Republicans were not. In the few days after the debate, the opposite happened (this is what the Pew Poll captures). If this is only a fluke, number should come back soon, but we have to learn to stop react at every single poll, good or not.

[reformist2]

16. The most glaring problem is the swing in party identification between the two polls: 39% -->31% dem?

Really? As I've said before, one can't conclude one poll is "right" and the other "wrong," but I think it's safe to say the population doesn't shift party identification that much that quickly, and so the shift toward Romney between the two polls is equally suspect.

Furthermore, if the two polls were reconciled by party identification, the results of the two polls would be practically identical.

[LisaL]

17. If they polled a 100 % mormon republicans, Romney might have gotten a 100 % lead.

I don't know what their deal is, but clearly they changed their sampling to include more republicans in Oct vs Sept. Giving Romney a bump.

[Patiod]

20. Maybe with all the voter suppression, they figure more Republican will be voting

A little OT, but have you seen this?
http://en.wikipedia.org/wiki/Voter_suppression

I'm extremely jaundiced about Republican voter shenanigans, but even I was stunned by the wide, wide range of things they are doing to suppress the vote.

And even though I'm against disenfranchising felons who have "paid their debt", I never considered that the laws disenfranchising for life anyone who has committed a felony is, indeed, a form of "civic death penalty"

[alcibiades_mystery]

23. More like the Pew Suppression: the GOP attack on poll skewing had its intended effect

Scare the pollsters into adjusting their demographics.

[reformist2]

24. I didn't know about felony disenfranchisement. Take a look at this map (with figures):

It's a real eye opener...

[MrMickeysMom]

25. Obama will serve an additional 4 yrs...

You can't say where in the south they polled. There's a lot a cherry picking for this poll. 55% of registered voters thought RMoney did better.... Hmmm.... I wonder why.

There are a lot of deranged, disconnected out there in media who'd like you to RESPOND to this poll. Their fucking jobs depend on it. What else can they do, but make this into a horse race, given the tiny brain and lies from the R/R camp.

Romney's improved performance has not sustained his ability to "over-take" the whole voting population.

I'm gonna join Lady Gaga if I hear any more of this puke storm.