Polls are beginning to turn Hillary's way.

[DCBob]

This is from Huffington's Pollster range 8/1 to current.. with less smoothing option.



Looks like Bernie is flatlining while Hillary is starting to move back up slightly. I suspect this is the result of many voters coming to realize the email "scandal" is not actually a scandal.

[jfern]

1. Wow, you managed to get it to show a slight short term improvement for Hillary

Here are the polls from 2007-2008. You can see a substantial improvement for Hillary in September and the first half of October 2007.

http://www.realclearpolitics.com/epolls/2008/president/us/democratic_presidential_nomination-191.html

[DCBob]

3. 2007-2008 ??

That's ancient history dude and as far as I know there is no candidate Obama anywhere in sight.

[dsc]

84. call me crazy but I see a down trendline for Hillary in the latter half of sept

and early october

[Fawke Em]

101. That's what I see, too.

The OP's chart looks like it's leaving out a lot of polls.

Here's what I get July-Sept including all the polls:

[DCBob]

110. I didn't leave out any polls.

I simply did a range from 8/1 to current and selected less smoothing option as I indicated in the OP.

[DCBob]

109. Not crazy but confused because the latter half of Sept and early Oct haven't occured yet.

[Persondem]

2. Thank you for posting this. I am still not sure why Biden is included as a matter of course.

Seems like including him clouds the true picture .... though I could see if every so often a pollster included him as a possibility.

[DCBob]

4. I really wish they would all drop Biden until he decides to run or not.

No doubt Hillary's numbers would go up.

[jfern]

5. Found the Pollster from 2007

No tweaking was done. And you can see that most of Q3 as well as early October 2007 were great for her and bad for Obama.

[img][/img]

[DCBob]

7. Do you not understand what happened in 2007-2008 is not relevant to the current situation??

[jfern]

8. Why not?

[DCBob]

12. Must I explain it??

Barack Obama was probably the most inspiring candidate to run for President in 5 decades and his campaign was probably most brilliant and effective in history.

Bernie does not compare in any way shape or form.

[jfern]

14. Almost anyone is inspiring compared to Hillary

[DCBob]

16. She is inspiring to many..

and Bernie is quite uninspiring to many.

[jfern]

20. Hillary lost almost half her support in NH

That doesn't happen to inspiring people.

[DCBob]

23. The latest poll shows she is starting to gain that back.

http://elections.huffingtonpost.com/pollster/polls/massinc-wbur-npr-22747

[jfern]

24. She's only 31% there, compared to 68% in an earlier poll

Also the MOE is 4.9%, so Sanders' lead is probably closer to what the other polls have shown.

[DemocratSinceBirth]

27. The poll that preceded it was Sanders +7

BTW, if you look at the internals she is already leading among Dems in the last two polls...

I know it's a open primary but if Trump is still viable Indys will vote for him and to stop him.

[jfern]

30. The pollster should figure out who is a likely Democratic primary voter

And a drop of 3 points between 2 polls that have MOE of 5 is completely statistically insignificant.

[DCBob]

28. She is only 4 points back on that poll.

The 5 previous polls she was down 7,22,9,7,7. Granted they are from different polling organizations but seems significant to me.

I think its a fairly good bet she is gaining back on Bernie.

[jfern]

31. Sounds like Sanders has about a 7 point lead

Ignoring that obvious outlier of 22 points, none of the differences there are remotely close to being statistically significant.

[DCBob]

32. We shall see.

I suspect the next poll out of NH has Bernie and Hillary statistically tied.

[jfern]

33. Well, the last one did, but it's also has no statisticall difference from that 7 point lead

that the other polls show. Large MOEs of 5 points tend to do that for you.

[DemocratSinceBirth]

49. Can we all agree that CBS should have been embarrassed to release that dreck.

Here's the methodology

The CBS News 2016 Battleground Tracker is based on interviews conducted on the internet between September 3 and 10 with 4,860 registered voters in Iowa, New Hampshire, and South Carolina. The poll was conducted by YouGov, an online polling organization.

Respondents were selected from YouGov’s and two other online panels. These are “opt-in” panels which are open for anyone to join. However, YouGov also randomly selected persons from voter registration lists who had previously voted in primary elections and contacted them by phone. (Note: though some respondents were initially contacted and recruited by phone, all interviews were conducted online.) A total of 19,047 registered voters were contacted by phone and the YouGov sample includes 1,163 phone recruits.

https://today.yougov.com/news/2015/09/13/methodology-2016-cbs-news-battleground-tracker/

What possibly could go wrong?

[kenn3d]

93. Well +22 Sanders did drop a few jaws I think... even some of his supporters questioned that.

... and yet, when I chose to eliminate that pollster from my customized chart... you said:

"when you start cherry picking polls you are introducing bias into your averages."

FWIW, I considered that yougov poll to be an outlier... (if not an out and out liar).
And that is why I was interested to see what eliminating the internet polls and low-rated pollster data from the composite would show. YouGov is a C rated pollster according to the 538.com rating list and was therefore not included in the chart I posted in #34 below (which calculates only data from A and B rated pollsters):
http://www.democraticunderground.com/?com=view_post&forum=1251&pid=605336

Do you really think the accuracy or value of a composite polling chart is improved by including data from poorly conducted polls? How about the ones Nate rates as D or F or bans altogether?

I still think averages which include outlier biases from poorly conducted polls will only be more biased and less accurate.
Not science (social or otherwise), just my opinion.

[DemocratSinceBirth]

100. If I was accumulating polling data I would include it as to not poison my sample.

However I used to make small wagers on intrade. I certainly would have ignored that poll in planning my wager.

You always get outliers. Zogby's final CA poll in 00 had Gore down by 1. I would have bet the farm he would win CA and he won it by 12.

[kenn3d]

106. I understand the scientific (statistical) concept of your argument.

But fwiw, http://fivethirtyeight.com/interactives/pollster-ratings/ includes two Zogby polls... one is C rated, the other is rated F.
So if you placed a wager on Gore in CA, I hope you excluded the Zogby poll from your calculations.
just sayin'...

I'm not a betting man, but what few donation dollars I can afford are going to Sanders.

Peace.

[ReallyIAmAnOptimist]

58. Bernie is "uninspiring" record crowds...

...well beyond what Obama had done at this point in the cycle.
Bernie has cross party appeal, and the young vote.
Finally, people are fed up with establishment candidates, so everything the pundits and professional insiders think they know is pretty much out the window.

[DCBob]

71. Yes, Trump and Sanders have both tapped in the anti-establishment movement but..

it doesn't mean either one can win.

[frylock]

61. Arenas are filled with the people who are uninspired by Bernie..

did you see the pictures from New Hampshire University of Hillary talking to a half-empty room of baby boomers? About 350 showed up. Meanwhile, in fantasy unicorn rainbow land, Sanders already has 1300 RSVPs for his appearance at the same venue on Sunday. The overflow will be at least double what Hillary drew.

[DCBob]

68. Filling arenas is not necessarily indicative of political success.

Don't forget the Ron Paul phenomenon in 2008 and 2012. His supporters packed arenas and were the most aggressive and dedicated out there but in the end it was just a footnote in the political history books.

I see Bernie ending up much the same way.

[Bluenorthwest]

73. That's some hype for Ron Paul there. His largest crowd was on campus at UCLA, 7,000.

Where you get that he was packing arenas all over the place I do not know. Obama drew thousands more than that in March of 08 at the far smaller University of Oregon, located in not Los Angeles.
Obama's top Oregon crowd was the week of the election, 70,000+ in Portland.

Sanders in Los Angeles, not on a campus, 27,500 estimated crowd. Almost 4X what Paul got on campus. Where is the comparison? There isn't one.

[frylock]

96. Yes, yes. Ron Paul..

Been hearing the comparison for months now. This isn't Ron Paul's coalition. Bernie Sanders has a Posse.

[Joe Turner]

108. Hillary is inspiring to

Lobbyists, DC insiders, the DNC leadership, 3rd Wayers, Free trade Democrats, wall street bankers. Not so much anywhere else.

[Scootaloo]

64. Clinton is the hare in this race

She's campaigning as if she doesn't think she should have to. She's cancelled public events to make room for closed fundraisers. She demurs on issue after issue, and when cornered on one - say, black lives mattering - she throws it back in the questioner's face, along with her finger. When she does rise to actually making a move on the campaign, it's to use surrogates against sanders in what, I swear to god, looks more like a parody of attack efforts than actual effort. Trotting out Carlos Danger? This SuperPac bullshit? Naw, she's not taking this seriously at all.

You're right; Bernie Sanders isn't Barack Obama. But he's putting in the work. He's taking his campaign seriously. clinton is as poor a campaigner as she was in 2008, and really seems to be taking this one for granted.

[DCBob]

67. After what happened in 2008, I can assure you Hillary is definitely not taking anything for granted.

I suspect her campaign HQ is like a war room with all sorts of strategies, risk assessments, what if scenarios, etc. What you are seeing is the result of all that. The current strategy is to play it cool. That might change if Bernie closes the gap more.

[treestar]

72. Different people. Different year.

8 years have passed since then. Stuff happened in those 8 years, making the world different than it was in 2007.

The candidates here are different. Hillary may be the same person, but Hillary has also lived through the past 8 years and experienced and done other things and is thus a different candidate from what she was then. The race she is in is different. Then she was running against Obama, Edwards, Kucinich, Richardson, Biden and others. Now she is running only against Bernie basically. Bernie isn to the same person as Obama, etc.

[pinebox]

79. You're right

A lot has happened. A lot has happened involving controversies and her. A crapload.
When it comes down to it, indy's won't touch Hillary at the voting booths and her favorability shows that.

[Cali_Democrat]

6. It also coincides with Bernie having Cornel West campaign for him

Having Cornel West campaign for him wasn't the smartest move.

[DCBob]

9. Yep, that could be part of it.

[jfern]

11. Yes, it totally doomed him by having him drop 0.2 points in some tweaked average

[artislife]

13. How can it be bad if AAs weren't into Bernie anyway? nt

[Scootaloo]

65. It's the same logic we saw from this type in 2014

Pre-election: "We don't need the fucking left! They only want ponies, and we need the conservative vote!"
Post-election: "We lost because the fucking leftists didn't vote! I hope you're happy, hippies!"

Democratic conservatives live inside Schrodinger's Box.

[artislife]

81. You know...it did seem familiar. nt

[morningfog]

10. Bernie's peaked again. Until his next peak.

[DCBob]

15. if you look closely at the polling trend line it indicates this started a couple of weeks ago.

[jfern]

17. Sanders is up since 2 weeks ago

The gains for Hillary are at Biden's expense.

[DCBob]

25. Not according to the Pollster graph..

On Sept 7 Bernie was at 25.1%
On Sept 16 he was at 25.4%

That's flatlining to me.

[jfern]

37. If Sanders is gaining only 0.3 points in a totally cherry-picked 9 days, he's doing fine

[DCBob]

40. Right.. good luck with that.

His opponent has gained 2 points in that same time frame.

BTW, these are not "cherry picked".. they are all polls.

[jfern]

41. Mostly from Biden

[DCBob]

43. So what?

If Bernie is stalled at ~25% then he cant possibly catch Hillary.

[jfern]

44. Obama was stalled at 25% or less for all of 2007

[DCBob]

46. Here we go with Obama again.

You will never learn.

[jfern]

47. When Hillary loses again, we'll see who never learns

[DCBob]

48. That is unlikely.

[Pope Sweet Jesus]

59. You may think it's unlikely, but it is feasible...

since Bernie is still on a upward trend, and Clinton is about to take a nosedive after the middle of October

[HereSince1628]

74. Interesting how that graph doesn't look at all like the one at RCP

How do suppose they get to be so different when they say they are doing the same thing?

[Fawke Em]

102. Because the poster created a custom chart.

Here's what July-September looks like without cherry-picking polls:

[HereSince1628]

107. Interesting...well

DCB gets to go to ignore... what I do with name-callers, cheats, devievers, dissemblers and well, liars.

I don't really object to haters for Hillary if they don't screw with the data

[DemocratSinceBirth]

18. I have seen "smoothing" mentioned several times.

What does it mean?

I do think it's instructive to compare polls to polls performed by the same pollster since the methodology is the same. There are multiple polls from yougov, ipsos-reed, and morning consult.

[DCBob]

21. The trend line smoothing option hides some of detail.

This might be useful when the data are very inconsistent. I like the less smoothing option. It provides more insight into the data trends.

[morningfog]

19. We'll see.

The national polling is still not most important. The current state of things, Hillary isn't due a win til march 1. That's a long time, and a lot can change between now and then.

[DemocratSinceBirth]

26. Wow, you are calling NV and SC for Senator Sanders.

Hillary isn't due a win til march 1.

Wow, you are calling NV and SC for Senator Sanders.

[morningfog]

56. Doh! No I am not (yet!). I meant Feb 20.

Thank you for the correction.

[Renew Deal]

22. Take Biden out and it's more dramatic

[DCBob]

29. It would be very interesting to see data without Biden.

I suspect Hillary gains big.

[DemocratSinceBirth]

35. Morning Consult and PPP is sans Biden

In WAPO and CBS she goes from 20-30 without Biden.

yougov has a subsample where Biden voters break 2-1 to 3-1 for Clinton.

I thought the new Hillary and apology was silly but it seems to be working.

[DCBob]

36. Yeah, that's about what I would expect.

It's ridiculous that all the polls dont just leave him out.

[kenn3d]

34. Yeah... customizing the composite charts is cool fun.

Here's the same HuffPost Pollster 2016 National Dem Primary chart as in the OP but from Jan 1 2016 to current ... with normal smoothing, AND all the internet polls removed, AND all the poorly rated pollster data removed. Only the six top rated pollsters with A and B ratings from FiveThirtyEight.com are included in this graph.



Probably more accurate than either the default graph at the HuffPo site or the unsmoothed 45 day chart in the OP...
But no national chart will likely predict much beyond name recognition until more nationwide campaigning gears up amongst all the contenders. imo

GOTV!

[DCBob]

39. Data from more than 4 or 5 months ago is meaningless.

[DemocratSinceBirth]

42. It's also sloppy social science to willy nilly ignore polls you made the subjective determination...

It's also sloppy social science to willy nilly ignore polls you made the subjective determination theyaren't up to snuff...The Law Of Large Numbers equals out the 'good' and 'bad' polls.

[DCBob]

45. I think you meant this for somebody else.

[DemocratSinceBirth]

50. Yes

Even Nate doesn't throw out polls he doesn't like. He just weights them differently.

Sam Wang at Princeton Consortium doesn't weight polls. He just averages them all. Nate gets marginally better results but Wang is truer to the math.

[DCBob]

52. Yep. That's a better way to handle it.

[DemocratSinceBirth]

54. Wang got up to 97% , 98% accuracy... Nate nailed everything in 012.

eom

[kenn3d]

57. Ok... I'm sure no social science expert.

But I didn't make the determinations as to the quality (and historical accuracy) of the pollsters... fivethirtyeight.com did. All the major poll charting sites (like RealClearPolitics and HuffPollster) use various and different polls, including some and excluding others for who knows what reasons. I just thought is was interesting to see what effect using only top-rated pollsters would have. I really can't see how including lesser quality polling data can improve accuracy once sufficient sample sizes have been met.
Anyway graphing 9 months of data makes trends a bit easier to see than graphing 45 days of data. But I still say the national polls (no matter how they are graphed or customized) are simply not very predictive at this stage. And just getting a graph to show a slight short term improvement for any given candidate is simply fun and folly.

Peace

[DemocratSinceBirth]

62. The problem is when you start cherry picking polls you are introducing bias into your averages.

eom

[progressoid]

51. So just look at the last 5 weeks.

Still not pretty for Hillary.

[DCBob]

53. Yes, but the last two weeks though are looking better.

It might be just a blip but it might be significant and perhaps a turning point.

[Ed Suspicious]

80. When you condense the date range to two weeks out of a 6 month campaign . . . you are pretty

much showing the essence of manipulative and meaningless.

[OilemFirchen]

88. Which pollsters were including Biden in March?

Just curious.

[kenn3d]

92. Nearly all of the pollsters found in the HuffPollster charts

have included Biden since late 2012 or before. You can see this by holding your mouse over the individual plot points.
I suppose it's reasonable for them to assume an incumbent VP is a likely contender whether announced or not. I think the question is at what point in time should he no longer be considered in the polling if he has NOT announced, but speculation is still rampant on that.

[OilemFirchen]

94. Thanks.

Looks like they included Warren as well.

I wouldn't attempt a trend analysis until all the polls reflect only declared candidates.

[Thinkingabout]

38. As time passes and more information comes out on the candidates stand on the issues it will become

Clearer where voters will go. There will need to be a day of facts and figures, programs are good if there is a way of paying for them.

[vadermike]

55. Well

Let's see what happens. Are her GE polls improving as well? That's the real questioner must win next year with the strongest candidate. Let's hope this is for real not just in the primary but the GE as well about this while email stuff cause if voters don't trust you they won't vote for ya

[frylock]

60. More national polls..

Do you know how the primaries work?

[DCBob]

70. The latest primary poll I am aware of also shows Hillary gaining back.

Previous five polls in NH had Bernie up by quite a bit more than that. Granted they were from other polling orgs but still seems significant.

[MisterP]

63. anything to salvage the political equivalent of a "participation medal"

[liberal_at_heart]

66. ah, so the undemocratic practice of having no debates, and having a super PAC launch

attack ads is temporarily working. Congratulations! Enjoy it while it lasts. Once Bernie's message can no longer be suppressed the numbers will change again.

[DCBob]

69. I think it's more that the fake email scandal has run its course.

[Skwmom]

75. The AHA moment is fast approaching

An Aha moment is a moment of sudden insight or discovery.

She used an unsecured private system and wiped the server. Why, oh why, why would she do this. It just doesn't make any sense.

SOS Transactions

Speaking fees

Clinton Foundation

Unsecured wiped server

AHA....

[DCBob]

76. Is this the DrudgeReport headlines??

[OilemFirchen]

90. Well...

"unsecured wiped servers" would definitely qualify, as it's completely untrue.

[99Forever]

77. Hogwash.

Partisans read into "polls" exactly whatever suits their own agendas. Pure, unadulterated propaganda.

[DCBob]

78. BoarBathing.

I guess you don't get the concept of polling trends.

[99Forever]

82. I "get" the concept...

... of poll trolling, just fine.

[DCBob]

83. trolling??

[99Forever]

86. It is what it is.

Hope you get to feeling better.

[DCBob]

89. you have no idea.

Move along.

[99Forever]

91. Of course I don't.

Only "experts" such as you do.



Oh...


... btw...


I'm not "moving along" anywhere. I'm not quite sure what ever gave you the fantasy that I would EVER take orders from the likes of you.

[Tom Rinaldo]

85. Too little, too late

Hillary needed to create a public perception that Sanders was essentially a very minor protest candidate before he gained any real traction in order to avoid a real contest against him. Bernie has the traction he needed already now. He is already highly competition in the first two states up on the calendar. His grassroots fundraising has him well funded already. He just hit the cover of TIME, he just got a hero reception on Colbert, and he has already been validated as a serious candidate because the larger "Team Clinton" has already gone negative on him. Main stream media coverage of Sanders has increased over the last 10 days. And now, finally, a live Democratic candidate debate is looming. Sanders can't be minimized now before then, and that debate is the start of the a new phase in the contest. I'm not Anti-Hillary, I'm Pro-Bernie, and most polling seems to show that Sanders gets support from those who like his message rather than from those who are worried about and so called Clinton scandal. Biden may have gotten some of the latter.

[magical thyme]

98. the one-day "trend" (pretend-trend?) has already reversed since Huffpost updated their chart from .

this morning. It's back to its normal course, with Hillary down to 45 and Bernie up to 26

[AgingAmerican]

87. Until she takes a hard stance

And starts doing public events, she will keep slipping further and further.

[workinclasszero]

95. Awesome!

Go Hillary!

[magical thyme]

97. odd....I see them at 45:26 today. also, if you're going to use the term "flatlining" you may want to

check what it means first. Bernie is hardly at "zero" with no heartbeat.

seriously, you can't look at results from one day or week to the next to call a trend. Otherwise, I could claim that Bernie is trending upwards because 2 days ago showed them at 46:25 and now they're at 45:26. Or maybe they have resumed their respective trends.

[dorkzilla]

103. Just wait till after the first debate

HRC is going to seriously tank. She is an abysmal debater.

[magical thyme]

104. also another round of Beghazi hearings.

More Clinton drama.

[Android3.14]

105. Here's to hoping this is more propaganda bullshit

The GOP won't win unless they face Hillary in the primary, so compared to anyone on the progressive side, Hillary would be the worst choice, to lead the USA, out of a group of Democratic candidates that also includes my occasionally incontinent cat.

Just say no to bitter rich people.

[jfern]

111. LOL, now it's not so good for Hillary

According to the OP's chart from Friday
Clinton 46.6
Sanders 25.4
Biden 17.8


Today using the less smoothing
Clinton 42.2
Sanders 25.9
Biden 17.9

[DCBob]

112. Looking good to me..

in Iowa!