Weighing the polls: Which pollsters are most reliable?
Fordham University political scientist Costas Panagopoulos, analyzing the same data, found that Rasmussen, SurveyUSA and Public Policy Polling produced results close to the final outcome. Mason-Dixon Polling & Research and YouGov were among the further off-course.
But during the 2010 elections, Rasmussen missed the mark by a far wider median margin than did SurveyUSA, YouGov and PPP, according to Journal analysis of data compiled by NCPP from Senate polls. The Quinnipiac University Polling Institute did better than all of them.
Gallup has received a lot of scrutiny this year as its tracking poll over the past month gave a large lead to Mr. Romney. Its final poll, released Monday afternoon, shows Mr. Romney leading, 49% to 48%.
Lee Miringoff, director of the Marist Poll, which is conducting polls of battleground states this year for the Journal and NBC News, notes that pollsters who complete their last poll closer to Election Day have an advantage in predicting the final result. Assessments of accuracy also can differ depending on whether they measure how close a pollster got to the candidates’ final vote totals or their margins of victory.
Blowback
Note from Hot Air management: This section is for comments from Hot Air's community of registered readers. Please don't assume that Hot Air management agrees with or otherwise endorses any particular comment just because we let it stand. A reminder: Anyone who fails to comply with our terms of use may lose their posting privilege.
Trackbacks/Pings
Trackback URL
Comments
Anyone who uses a sample similar to 2008 or skewed even more towards Dems is not reliable.
Doughboy on November 6, 2012 at 8:53 AM
About all of those polls that we’ve done for the past several months showing a completely different outcome…
NEVERMIND.
aunursa on November 6, 2012 at 8:53 AM
..wonder what Gallup and Rasmussen will show in their daily tracking polls on November 7th.
(It’s supposed to be a joke, people.)
The War Planner on November 6, 2012 at 8:53 AM
Whattaya want to bet there will be polls tomorrow:
Was this a fair election?
Are you happy with the results?
Do you think Obama cheated? (yes)
Should the police have arrested the New Black Panthers for voter intimidation?
Do you think Obama trying to declare victory at noon on Tuesday was intended to suppress voting? (duh)
If you never see Stephanie Cutter again, will you be sad? (snort)
BettyRuth on November 6, 2012 at 8:57 AM
This doesn’t look like an article worth signing up with the WSJ over.
rhombus on November 6, 2012 at 9:07 AM
oh great…
Can.I.be.in.the.middle on November 6, 2012 at 9:12 AM
The polls on who has the most reliable polling are in!
Quis custodiet ipsos custodes?
MassVictim on November 6, 2012 at 9:17 AM
Here is the Art Machado – final poll – in CT suburbs and exurbs.
117 voters in my suburban polling place with a full parking lot before 6:15am I have the sunrise pictures and in another exurb town parking lot so full police directing traffic outside.
So Linda McMahon wins for Senate and Obama wins CT by a small margin but is toast everywhere else.
I can’t figure out how to download pictures here but then end of this 4 year ordeal is here.
ArthurMachado on November 6, 2012 at 9:25 AM
Did you poop your pants again?
wargamer6 on November 6, 2012 at 9:33 AM
What’s the point of “predicting” an election result two days before it happens? Campaign contributions are in; media buys are set; and it’s way too late to start scheduling new rallies. The only possible purpose of provide a CYA to the pollster.
I saw a study a short while back, which looked at poll results during the two months before the elections. At that time, liberal media polls weren’t just off; they were WILDLY off in giving an advantage to the Democrat candidate.
logis on November 6, 2012 at 9:43 AM
It seems to me that all the polling companies are trying to make the poll results a tie by any means necessary.
Polls with D+5, D+8, D+11. D+20 all end up with a tie.
Polls with R+25 independent advantage, end in a tie.
They are all scared to show Romney winning.
Why?
faraway on November 6, 2012 at 10:07 AM
None of them are reliable. They’re in bed with the lamestream media.
trigon on November 6, 2012 at 10:30 AM
Great point.
My theory is that the pollsters, who are paid by the news media after all, provide the results their customers desire. The news media (of all stripes) has a vested interest in claiming everything is a tie in order to boost interest and viewership. It’s a simple business model.
MassVictim on November 6, 2012 at 10:42 AM
True, and it has the added advantage of allowing the MSM to fantasize that their best boyfriend ever will win.
drunyan8315 on November 6, 2012 at 10:57 AM