Team Obama: NYT/CBS poll used 'biased sample'

[Marzupialis]

Source: Politico

The Obama campaign is pushing back hard against a New York Times/CBS poll that showed Mitt Romney with advantage with women and reported that voters overwhelmingly believed President Obama backed gay marriage last week for political reasons.

"We can't put the methodology of that poll aside," deputy campaign manager Stephanie Cutter told MSNBC's Chuck Todd. "It is a biased sample."

"Aside from being outlier, CBS/NYT poll used weird methodology -- same voters as last month, reached fewer of them, more IDed as GOP," campaign spokesman Ben LaBolt said on Twitter Monday.

At the heart of the dispute is the fact that the poll used a 'panel-back survey' — polling a group of voters who have already been interviewed in previous polls. In this case, the drawback is that 852 registered voters were in interviewed last month, and just 562 agreed to be reinterviewed by the pollster.

Read more: http://www.politico.com/politico44/2012/05/team-obama-nytcbs-poll-used-biased-sample-123487.html

---

Note: Poll news should be in GD or Politics 2012 forums, but this one is special in that the Obama campaign is involved, so I ask permission to post it in LBN.

Back to the article, I wonder if it's even professional to poll the exact same sample of people polled in a previous poll, as NYT/CBS did. Isn't the randomness of a poll one of the main ingredients for accuracy?

It's possible that since these are media outlets hungry for viewers/readers, they wanted to "shock" people into reading the unlikely news that Romney leads among women, which I don't believe for a second.

[Chorophyll]

1. I heard Mara Liasson discussing this poll

just this morning on my local NPR station (WNYC.) She sounded positively gleeful. And right after that they went into "pledge drive" mode. Yeah, right. As if.

Glad to see the campaign pushing back. I think this poll is whack.

[nanabugg]

34. Maybe this will make the base get out and work harder. Especially to get people registered and get

thier IDs in order.

[PoliticAverse]

2. Note that 852 people were interviewed the first time, but only 562 agreed to be reinterviewed...

this is a potential non-response bias problem (see http://en.wikipedia.org/wiki/Non-response_bias ). Also they apparently used
registered voters and didn't screen for 'likely voters'. As to whether it was 'professional' note that this wasn't really a 'poll' as it
was a 'panel-back survey'.

I wouldn't take the results too seriously.

[Arkana]

8. Also notice that in the original sample, 24% were Republicans, 33% Dems,

and 41% Independents.

In the panel back counts, there were 30% Republicans--up 5%, 35% Dems, and only 34% indies (down 7%). Flawed methodology at best.

[madashelltoo]

3. This is why Scarborough's ass was soooooo happy this morning!

I knew this was a lie when it flew from his grinning lips! No way women simply forgave Rmoney for the stuff they're trying to do. No way immigrants have simply sighed and accepted his crap. They have to be using contaminated sampling. This is a Brooklyn Bridge sales job if ever I heard one. Nobody likes Rmoney. Nobody really believes he will be better for the country economically either. Which side of our necks do they think the screws are on?

[yellowcanine]

4. Yes, it is no longer a random poll and that is a problem.

Back to the article, I wonder if it's even professional to poll the exact same sample of people polled in a previous poll, as NYT/CBS did. Isn't the randomness of a poll one of the main ingredients for accuracy?

I don't believe that this is accepted methodology for political polling. Not only is the poll no longer random, but respondents are essentially "self selected". This is a no-no in polling.

[edcantor]

20. Exactly, there was at least a 12% skew in this poll,

That is: 5% more Republicans and 7% LESS Independents.

The sampling size was much smaller, (creating a larger margin of error), and the nature of a survey of this type cannot in any way be considered a "random scientific poll" the second time that a smaller random sampling of the same original group is preselected without accounting for non response bias.

If a truly NEW random scientific poll is conducted now with a similar sampling size as the first, and with on-one who participated in the other poll, we will know where women really are.

[pnwmom]

28. Actually, there were 25% more Republicans compared to before.

The increase was from 24% to 30% -- a difference of 6%. 6% is 25% of 24, the original percent.

So there are 25% more Republicans.

[n2doc]

5. They'll probably keep interviewing the same people until they only get 10 or 20

Who will all, amazingly, be T-party members and will show Rmoney willing by 100%!

[1StrongBlackMan]

6. I hate all this dramatic crap ...

Still, for a campaign that has openly professed dislike for process stories like this one, it's an unusually aggressive pushback against a single poll — a poll that is indeed a bit of an outlier compared to other national polls.

"So, Sir ... because, historically, you do not push back against single stories, am I to assume that didn't you stop beating your wife?"

Pure B.S.

[BlueDemKev]

7. A "biased" sample??? But....but.....

...I thought that CBS News and the New York Times were liberal-loving socialist propanganda news agencies!

My gosh, what happened?

[Andy823]

9. They can make things up all they want

Problem is the facts stay otherwise, and so the only one this poll helps is Romney's ego! Nobody likes this guy, unless of course they are multimillionaires or billionaires!

I think the right wing is doing all they can to try and make people think the right is behind Romeny, when in reality the vast majority of republicans can't stand the guy for one reason or another, but the GOP doesn't want everyone to think they have a candidate that is so disliked by his own party!

Hell even the powers that be don't like him, and they know he is going to lose, they just don't have the guts to admit they are running a sacrificial lamb like they did with McCain back in 2008. The GOP does not want to win because they have no idea how to fix the things they messed up during the Bush years. They will wait, Obama will fix the mess, and then they will try and take credit for it and hope the american people are dumb enough to vote them back into power in 2016.

[cindyperry2010]

10. Delegate counts folks electoral college delagate counts nothing else

[slackmaster]

11. It strikes me as bizarre that the President's campaign would bother with such a non-issue

[SalviaBlue]

13. Since this poll is being touted on all M$M as an indicator,

I think the Pres. should push back and call them on their bullshit. Maybe make people think twice about what they are hearing.

[slackmaster]

14. I didn't hear about it on the M$M

I read about it here; and I still don't give a rat's ass about it.

[Dawson Leery]

18. When Dead Intern Joe and the MSM are touting this BS poll.

The Whitehouse is right to hit back.

[slackmaster]

21. It's unfortunate that so many people are brainwashed by TV

[CreekDog]

29. why shouldn't they?

it's flawed methodology helps portray a move the president made as unpopular with women.

[slackmaster]

30. I didn't say they shouldn't. I said I think it's bizarre.

Not the same thing.

it's flawed methodology helps portray a move the president made as unpopular with women.

It's just a stupid poll done by a media corporation to generate buzz so they can sell advertising slots. All polls have imperfect methodology. The only poll that really matters is the election.

[CreekDog]

31. why is it bizarre? bizarre has a negative connotation.

Republicans respond to poll results too.

[truebrit71]

12. Any poll that shows Rmoney, or any other rethuglican with an 'advantage with women' is pure bullshit

...surprised that CBS is going into full whore-mode this early...

[maxrandb]

15. It's stupid

You'd think they would learn. No matter how much Repub dick CBS, CNN, ABC, NBC, the New York Times, or any other media outlet suck, they WIL ALWAYS...ALWAYS...ALWAYS! be the "Liberal Media".

You would think they'd quit trying to win the hearts and minds of the Tea-baggers. Hell, they even accuse Faux News of "liberal bias", if they stray one milimeter off the reservation.

I've never understood it. CNN or CBS could give a show to Limpballs and have Wolf Blitzer lick his taint on live TV, and the TEA-BAGGERS STILL WOULD HATE THEM!

[HotRodTuna]

17. I haven't heard the 'taint' referenced since college

thanks for the laugh.

[Alcibiades]

16. Much of whether this is a good poll depends on our assumptions

In order for this poll to be valid, we have to assume that there is no systemmatic difference between the people who agreed to be reinterviewed and those who did not. If there is, there is a problem: it's quite likely that this systemmatic self-selection played a role, and the best evidence of it is the difference in the numbers. If people who don't like the president were more likely to agree to another round, that's a problem for these results. In fact, it's such a big problem that it means these results are not results, but "results."

Also, the margin of error increases from 3% to 4%, but this understates the problem of self selection.

[AzDar]

19. After the debacles of 2000/2004, I am always, ALWAYS concerned about theft. False/ biased

polling is a necessary component of election theft, and I'm glad they're not letting this bullshit stand.

[LynneSin]

22. Well you have to do whatever it takes to keep the race close

pretty sad isn't it. Shame on the NYTimes and CBS - I expect better from these organizations. Someone should lose their job over bad sampling just to create a news story.

[LiberalLovinLug]

25. Exactly

The MSM have to get it closer to a 50/50 horse race before the election. They'll do whatever they have to. Inflammatory magazine covers, skewed polls....

[NRaleighLiberal]

23. zzz... polls this early. zzz. wake me in the fall! n/t

[liberal N proud]

24. They are shameless in the crimes they are perpetrating against the republic

This is leading to the election fraud they are planning. Polls must make plausible that Romney wins in November.

[AndyTiedye]

36. And the Teabaggers Control Several States We Have to Win

If they can steal Ohio, Florida, Michigan, Wisconsin, and Pennsylvania, we're fukued.

[goclark]

37. I totally agree ~ we can either wake up

now or nap until November and wake up saying ~ "What happened, we got out the vote????"

[AngryAmish]

26. Well, that settles that.

[pnwmom]

27. It's not the same sample -- it's a self-selected portion of the original sample,

which I think would be even worse.

[DallasNE]

32. MSM Always Attempt To Create

A horserace atmosphere because it drives up their revenue and that means some creative things with poll numbers so this kind of thing is fully expected. This poll looks to be even more creative than some of the things they have done in the past, possibly reflecting on the enormity of the task. Frank Luntz would be proud.

[fasttense]

35. You have that right DallasNE

Fake polls are used to discourage the opposition.

"See all those polls, it means that I must be in the minority. I guess I should just accept what everyone else wants." Or some such rot.

[elzenmahn]

33. I've never taken polls seriously...

...as even if they are accurate/reflective/whatever of the public's opinion, that they are used purely for propaganda purposes.

This poll, however, reeks of slant - especially if it's the outlier that it is being made out to be. I would just consider who is funding and sponsoring these polls, and what their agenda might be. The business community has made it's dislike of Obama known for a long time, and they're just itching to get "their guy" in the White House. If the same kinds of people aren't above stealing votes (2000, 2004), then an opinion poll like this should be nothing for them...