OMG, Sanders campaign cites "online polls" to explain away real polls.

[sufrommich]

Last edited Mon Nov 16, 2015, 01:18 PM - Edit history (1)

Briggs was asked, "Are you concerned about the scientific accuracy of the PPP poll?"

He replied, "I'm just suggesting that your readers might be in a better position to consider the source if they knew that 'poll' was paid for by a PAC supporting one of the candidates."

Briggs also pointed out that Sanders won the debate according to several unscientific polls, including C-Span’s Facebook survey and Time’s Web survey.



Tom Jenson of PPP's reply:

Public Policy Polling conducted the survey on behalf Correct The Record, which is described as “a strategic research and rapid response team designed to defend Hillary Clinton from baseless attacks.”

Tom Jensen, director of Public Policy Polling, defended the survey’s accuracy.

“Most polls are conducted for paying clients, it doesn’t impact how they’re conducted,” Jensen wrote in an email Sunday night.

Jensen noted that his organization’s poll found similar results to the one conducted by CBS.




http://www.usatoday.com/story/news/politics/elections/2015/11/16/sanders-camp-responds-poll-debate-clinton-won/75865814/

[MaggieD]

1. Oh good lord

How unserious can a campaign get? Not to mention desperate.

[sufrommich]

4. I'm looking forward to reading what went on inside

the Sanders campaign once this is over.

[Agnosticsherbet]

2. His suppoerters do it all the time, and he is their leader.

And if he admitted that scientific polls show his revolution is falling a bit short, it would not play well.

[MoonRiver]

3. Dream on!

[Jarqui]

5. Nate Silver on online polls

http://fivethirtyeight.blogs.nytimes.com/2012/11/10/which-polls-fared-best-and-worst-in-the-2012-presidential-race/?_r=1

Among the nine polling firms that conducted their polls wholly or partially online, the average error in calling the election result was 2.1 percentage points. That compares with a 3.5-point error for polling firms that used live telephone interviewers, and 5.0 points for “robopolls” that conducted their surveys by automated script. The traditional telephone polls had a slight Republican bias on the whole, while the robopolls often had a significant Republican bias. (Even the automated polling firm Public Policy Polling, which often polls for liberal and Democratic clients, projected results that were slightly more favorable for Mr. Romney than what he actually achieved.) The online polls had little overall bias, however.

I'm not so sure I'd be so quick to dismiss so many online polls that line up. Sure there are potential arguments against them but there are also with the methods PPP used for example - only allowing 19% of those under 45 years old to participate in their poll when 51% is what that number was for 2012 Democrat votes.

[Jester Messiah]

6. Shush, you're harshing the narrative!!

Everyone knows that only land-line polls like Rasmussen and Gallup are "real."

[sufrommich]

8. There's a big difference between what Nate Silver is describing

and self selected clickbait polls that are used to generate ad revenue and can be voted in as many times as the "voter" would like.

[hack89]

9. That is not the kind of online poll we are discussing

we are talking about online polls where no sample is defined and where people can vote multiple times.

Nate is talking about scientifically constructed polls that are conducted online.

[Jarqui]

19. I didn't check every one but I did see on a number of them that you couldn't vote twice

because I gathered those polls results originally after the debate and then went back after midnight to update them - which is when I saw a number of them wouldn't let one vote again. I suspect if I went over them, there are probably some there where one could vote twice - but I'm certain it wasn't all of them.

And there is a difference with many of them as you point out - they don't look at demographics and "scientifically" construct them.

But collectively - looking at places like FOX (the media enemy of Democrats), it is striking that they all had Bernie winning handily.

We could say the straw poll wasn't scientific either but many don't seem to have as much of a problem them with them - dismissing them as quickly as online polls. The straw poll had Sanders winning.

Now for the two "scientific" polls that favored Clinton - their results are not that close - so the science being applied is rather suspect - a much bigger margin of error than they claim because they fall outside of each others margin of error. As I've stated before, the demographic of 19% under 45 years old voting in the PPP poll is a joke - that's not science. It's indefensible. CBS didn't publish all the "science" behind their numbers - we can't check.

So none of these quick polls I've seen are "perfect". Nate may well fail them all if he looked into it. But I will not look at PPP's demographics while realizing Clinton paid for the PPP poll and give it any more weight than the ones online.

[JaneyVee]

21. You can vote twice. Clear cookies. Vote again. Clear cookies. Vote again...

[DemocratSinceBirth]

24. Ummm, they are closer than they appear

PPP included Democrats in their poll. CBS included Democrats and independent, the latter group where the Vermont independent is stronger.

[Jarqui]

35. The PPP poll said

"A Public Policy Polling survey of Democratic primary voters nationally"
http://www.publicpolicypolling.com/pdf/2015/DemocraticPostDebatePoll111415.pdf

If you look at exit polls, "Democratic primary voters" includes independents and even some Republicans

[DemocratSinceBirth]

36. You are conflating exit polls with polls that measure other things.

PPP specifically asked respondents if they are Democrats and if they would be watching the primary and then contacted them to get their reaction. CBS/GFK did essentially the same thing but included independents.

What does that have to do with how voters responded in exit polls?

[Jarqui]

39. Show me where they

"specifically asked respondents if they are Democrats"
http://www.publicpolicypolling.com/pdf/2015/DemocraticPostDebatePoll111415.pdf

This was question 1:
"1 Did you watch the Democratic Presidential candidate debate tonight?"
(nothing in there about being a Democrat)

They refer to them as "Democratic voters" on the basis of voting in the past primary - which Independents and Republicans can do.

"PPP interviewed 510 Democratic primary voters nationally" Independents do vote in primaries.

[Recursion]

57. That's a screening question; it's not on the poll (nt)

[Jarqui]

59. Where is the reference to that type of screening taking place?

Where does it state that in the poll?

You made this claim above:
"PPP specifically asked respondents if they are Democrats and if they would be watching the primary and then contacted them to get their reaction."

I've asked before and I'm asking again, where does it state in the poll "asked respondents if they are Democrats"? Where? Please quote it.

Here's what they said about screening:
http://www.publicpolicypolling.com/pdf/2015/DemocraticPostDebatePoll111415.pdf

"PPP interviewed 510 Democratic primary voters nationally by telephone after the debate
who had been pre-screened on Thursday and Friday as planning to watch the debate and
willing to give their opinions about it afterward. "

That does not say they screened only for Democrats. It says "PPP interviewed 510 Democratic primary voters nationally ... who had been pre-screened on Thursday and Friday as planning to watch the debate and willing to give their opinions about it afterward."

That's all. "Democratic primary voters" includes Independents and they do not claim to have screened out Independents. The screen was to see if they were going to watch the debate and if they were willing to answer the poll afterwards.

So again, there is no evidence that demographics for this poll in terms of party affiliation was dramatically different than CBS. Yet the two polls did not fall within their claimed margins of error - which means the "science" in at least one of those polls is wrong. And we know the age demographics PPP used was absolutely wrong.

[Recursion]

61. I didn't claim anything upthread

It's also perfectly possible (expected even) for two polls to have a non-overlapping MoE. The MoE just means that if you asked the entire population it would be 95% likely to be within the MoE of that poll.

That said, I agree with you that the screening question was for "democratic primary voters", not simply "democrats".

[hack89]

26. If HRC and Bernie had identical demographics supporting them

then I might give credence to the online polls.

But it is clear that there is a generational divide with Bernie attracting younger voters. I think that fact alone would significantly impact the self selected online polls - older voters are less likely to be immersed in social media and participate in such polls. I do a lot of voter education in RI - older voters simply do think to go to the internet to voice support for politicians. They expect old fashion retail politics.

Every poll is paid by someone - that is how polling companies make money. That does not mean that polling companies will skew the results just to please a customer. Such companies live and die by their reputations - if no one trusts them then they will go out of business.

[tammywammy]

10. Nate Silver is talking about online polling like Google Consumer Survey

Not click-bait widget "polls".

[DemocratSinceBirth]

11. This obscurantism won't stand as long as DemocratSinceBirth has a pulse.

This obscurantism won't stand as long as DemocratSinceBirth has a pulse. You are either intentionally or inadvertently conflating internet polls with rigorous controls that try to approximate a random sample by weighting their results, casting out as wide as net as possible to capture a large number of respondents, and apply rigorous controls with simple online polls where anybody can vote as many times as he or she wants and there are no attempts to make sure every demographic is represented,

[mythology]

13. Those are polls from polling firms

The polls showing Sanders winning the debate are not. They are Facebook or CNN etc without any attempt to check the demographics of responses to ensure a correct estimated sample.

[Jarqui]

23. That's not accurate

The Wall Street Journal Poll that said Sanders won the debate was done by Google Consumer Surveys - a polling firm - and somewhere on this site, I linked those results and the demographics behind their poll (which favored Sanders some but not nearly as badly at PPP's demographics favored Clinton)

[Gman]

33. That is not the kind of online poll they're talking about.

They're talking about a poll where random people who have agreed to be polled are scientifically selected and sent an email asking them to participate. Zogby pioneered the method starting years ago.

This is not the same thing as clicking on a Facebook page post anyone can do.

[Metric System]

53. He's talking about scientific online polls, not click polls where you can vote multiple times, the

kind you see on Time.com, People.com and CNN.com

[Recursion]

56. Delivering a scientific poll online isn't "an online poll" in the sense he's talking about

There's nothing in particular about using the Internet that makes a poll useless; what makes it useless is simply putting up a button and letting anybody who wants to click it. There's no actual "sampling" going on there in any real sense.

[NCTraveler]

7. A poster here is literally touting an instant FOX post-debate online poll...

as proof that Sanders is the candidate of the people. It's flat out gut busting funny.

[NurseJackie]

12. +1 (nt)

[seaglass]

15. I saw that. When I first read the headline I though Bernie was up for a People's Choice Award.

very entertaining.

[fredamae]

14. Fact is-

No poll, no person "knows" whats gonna happen until the votes are tallied After the primary.

We All have hope for our candidate and try to find "evidence" with which to anchor our hope and belief.

No one knows who will win until after the votes are counted.

[Attorney in Texas]

16. In the 36 hours after the debate, the only line that got any media was Clinton's 9/11-Wall St gaffe

Seemed to me that O'Malley did well in the debate, but you wouldn't know that unless you watched it. I thought Sanders did well, too, but that will be lost on those hundreds of millions who did not watch the debate. I also thought Clinton did a very good job aside from her unfortunate 9/11-Wall Street slip up, but almost all of America won't know that either.

The only widely repeated story from the debate was the Clinton gaffe and so -- whoever won the debate -- it wasn't Clinton.

If you have a poll that makes you sleep better, rest soundly on that bit of comfort.

[AgingAmerican]

17. Your the one trying to explain away

No link?

[sufrommich]

18. Sorry,link added. nt

[AgingAmerican]

28. Well, I can see why you didn't initially add the link

Because that David Brock poll was debunked immediately after the debate as coordinated with the Clinton campaign.

[sufrommich]

29. Did you read the OP? It covers the PPP poll and did before

I added the link.

[Iliyah]

20. 2012 unskewing polls that are not in favor of one's candidate, sounds familiar?

The presidential candidate had prepared an acceptance speech and fake news was in shock. I'll respect PPP because they are fair even if its in favor of the candidate I'm not supporting.

[Betty Karlson]

22. "Real" Polls?

Would that be the polls which consistently call elections wrong, because landline-based polling seriously underpolls young people?

Whereas online polls have a smaller margin of error?

Which is why camp Weathervane wants those online polls to be unreal?

Except when CNN does it, because that one had Clinton as a winner? Except she wasn't but they called it for her anyway? And then removed the online poll because why would anyone disagree with their preconceived opinion?

[sufrommich]

25. "Whereas online polls have a smaller margin of error? "

Show your proof of this.


"And then(CNN) removed the online poll because why would anyone disagree with their preconceived opinion"


Debunked:

http://www.politifact.com/punditfact/statements/2015/oct/19/nowthis/no-internet-cnn-did-not-delete-its-poll-showing-be/

[NurseJackie]

31. Oh, thank god! (No, what I mean is... thank YOU!)

That claim is so tiresome. I assume it persists because it helps to add "credibility" to their claim that online polls are accurate.

[DemocratSinceBirth]

27. Self appointed spokesperson for Camp Weathervane

Would that be the polls which consistently call elections wrong, because landline-based polling seriously underpolls young people?

Most credible polls call landline and cell phone users. The landline only argument is so 2004.

Whereas online polls have a smaller margin of error?

Opt in online polls have no margin of error because no sample is defined and consequently a margin of error can't be established.

[Recursion]

58. ^ That

Mass opt-in polls are so far from a scientific sampling that you can't even meaningfully talk about "error".

[bobbobbins01]

30. What seems like the more likely scenario?

All the online polls, which show similar results of Bernie Sanders winning by a very large margin, are wrong....or...

The one poll, paid for by a candidate, which flies in the face of all the other polls, and just happens to show the candidate who paid for it as winning, is wrong...

[sufrommich]

32. What about the CBS poll with similar results?

other than self selecting online polls where one can vote multiple times, where is Sanders "winning by a very large margin"?

[NurseJackie]

34. Easy! "All the ONLINE polls, which show similar results ... ARE WRONG" Yes! Exactly!

That's it!
That's the one!
Your search is over!
We have found the correct answer!


The one poll, paid for by a candidate, which flies in the face of all the other polls...
HA! I see what you did there! You should probably apply for a job writing online poll questions.

[bobbobbins01]

49. I'll have some of whatever you're having.

[NurseJackie]

51. Aww! Hey there you, come here! Looks like you could use some! Let me pour you a strong glass of …

reality.

[Starry Messenger]

37. Well, this confirms a suspicion of mine.

There isn't much daylight between the campaign and several of the supporters, and know I now why. I don't really care, since sloppiness like that is going to tank their efforts, but it is sure funny as fuck to watch.

[Tierra_y_Libertad]

38. Are "Real" polls like "Real" men? "Real" Democrats? "Real" Liberals? "Real" Coke?

There are no facts, only interpretations. Friedrich Nietzsche

[Metric System]

41. Scientific polls are real polls. Simple.

[Tierra_y_Libertad]

42. Oh, well...if they're SCIENTIFIC..

Kinda like saying "It's the Bible!"

[tammywammy]

43. Take a statistics class. n/t

[Tierra_y_Libertad]

44. I took a poll. Very scientific.

Of myself. I guarantee the results are scientifically sound.

Results: Bernie 1, Hillary 0.

[tammywammy]

45. I'll just repeat my advice that you should take a statistics course. n/t

[Metric System]

52. +1000. I mean, seriously.

[brooklynite]

46. What the margin of error?

[Recursion]

60. By definition a poll of 1 has a 100% MoE (nt)

[NuclearDem]

62. Kind of ironic you're mockingly bringing the Bible into it.

Considering you sound like a raving creationist lunatic attempting to dismiss an entire field of science because he didn't like the results it gave him.

[Metric System]

40. They know his supporters well, then.

[zappaman]

47. It's like some websites have their own version of Baghdad Bob!!!!

OMG!!!!
Sanders won an online poll!!!!

[SunSeeker]

48. K & R

[Cha]

54. Let them continue living in denial. Better for the country and the Planet.

[stevenleser]

55. Sanders and his campaign really seem to have a race to the bottom following your nuttiest supporters

kind of non-leadership style.

His goal seems to be to demonstrate he is weaker at leadership and ability to stand up to bad behavior by his followers than John McCain was.

Say what you want about McCain but he at least once publicly stood up to the nuttiest of his followers. Sanders is too weak to do that.