2016 Postmortem
Related: About this forumFirst Primary And Senator Sanders Neighboring State Poll (New Hampshire)-HRC 46% SBS 43%
2016 New Hampshire Democratic Presidential Primary
Asked of 600 likely voters - democrat
Hillary Clinton (D) 46%
Martin O'Malley (D) 3%
Bernie Sanders (D) 43%
Other 0%
Undecided 7%
http://elections.huffingtonpost.com/pollster/polls/arg-23409
still_one
(92,325 posts)onehandle
(51,122 posts)Feel The Hil.
DemocratSinceBirth
(99,710 posts)onehandle
(51,122 posts)draa
(975 posts)American Research Group ranked 21st out of 24 pollsters used during the 2012 election cycle. Even worse than Rasmussen and Washington Post/ABC.
ARG had an error rate of 4.5 in 2012. Their error rate was 50 times greater than the most accurate (IDB/TIPP with a .09).
ARG skewed Republican in 2012 by 4.5 points. Higher than any other pollster but Gallup (7.2 Repub).
I understand the need to post a new poll daily showing how each on our candidates are doing but could we at least stick to ones with a little more respectability. It would be like trusting Fox Polling. Not very accurate and historically bad.
http://fivethirtyeight.blogs.nytimes.com/2012/11/10/which-polls-fared-best-and-worst-in-the-2012-presidential-race/
edit: corrected misspelling
CoffeeCat
(24,411 posts)Every other NH poll released within the last month has Bernie in the lead.
It's possible that he could have lost ground, but if this pollster isn't reputable, according to fivethirtyeight, that means something.
There are a lot of junk polls out there.
draa
(975 posts)The thing about blogger activists and polling is we need accurate polling to know what's happening with our candidates. With polls like this you don't get a feel for gains or losses. Or what might be a polling trend up or down.
And that goes for supporters of both candidates. Even more so considering Democrats are the reality based community in politics.
Anyway, we're not Karl Rove. We need accurate polls when possible.
CoffeeCat
(24,411 posts)...and it appears that some polling outfits are purposely using polling methods that will achieve a certain result for a campaign.
The Monmouth polls are laughable. It's too bad that organizations engage in this sloppy science.
A poll should be accurate and scientifically sound. If you can't do it right, then get out of the business.
Whenever I see a poll released, I immediately read the methodology before I make any judgments about the poll being sound or unsound.
Some polling outfits are fine; others are terrible. I'll wait for the next Ann Selzer Iowa poll, as her methods are sterling and her organization received an A+ rating from Nate Silver.
Samantha
(9,314 posts)Dretownblues
(253 posts)Is known to be a poor poller.