Putting the Polling Margin in Perspective
Here's a list of the most recent national poll numbers released by the major polling organizations (in descending order of favorability to Obama):
CBS/NYTimes = Obama 54%, McCain 41% (O +13)
ABC/WaPo = Obama 54%, McCain 43% (O +11)
Gallup Tracking = Obama 53%, McCain 42% (O +11)
USA Today/Gallup = Obama 53%, McCain 42% (O +11)
Marist = Obama 53%, McCain 44% (O +9)
NBC/WSJ = Obama 51%, McCain 43% (O +8)
Ipsos/McClatchy = Obama 50%, McCain 42% (O +8)
Reuters/Zogby = Obama 51%, McCain 44% (O +7)
CNN = Obama 53%, McCain 46% (O +7)
Democracy Corp = Obama 51%, McCain 44% (O +7)
FOX News = Obama 50%, McCain 43% (O +7)
Pew = Obama 49%, McCain 42% (O +7)
Daily Kos/R2K = Obama 51%, McCain 45% (O +6)
Rasmussen = Obama 52%, McCain 46% (O +6)
GWU = Obama 50%, McCain 44% (O +6)
Diageo Hotline = Obama 50%, McCain 45% (O +5)
IBD/TIPP = Obama 48%, McCain 43% (O +5)
What stands out here? Well, Obama is leading in every poll, and by a margin of 5 points or greater. A clear majority of the polls (12 out of 17) place the lead at 7 points or larger. And several major polling organizations have the lead in double digits.
That's certainly a good position to be in on the eve of an election. For some perspective, here were the final voting margins in other recent presidential elections (excluding the 2000 and 2004 squeakers):
1980 = Reagan 50.7%, Carter 41.0% (Reagan +9.7, electoral margin 489-49)
1984 = Reagan 58.8%, Mondale 40.6% (Reagan +18.2, electoral margin 525-13)
1988 = Bush 53.4%, Dukakis 45.6% (Bush +7.8, electoral margin 426-111)
1992 = Clinton 43.0%, Bush 37.7% (Clinton +5.3, electoral margin 370-168)
1996 = Clinton 49.2%, Dole 40.7% (Clinton +8.3, electoral margin 379-159)
The key point to underscore here is how stunning some of these poll numbers are relative to recent historical election results. If the majority of polling organizations are correct, Obama is looking at a victory the size of George H.W. Bush's victory over Dukakis in 1988. If organizations like Gallup and CBS News are correct, then Obama is looking at a victory margin matched only by Ronald Reagan in 1984.
I'm personally somewhat skeptical of the double digit projections. I think the most likely scenario is a popular vote win of 5-7 points.
I will say this, though. If McCain manages to win this election in the face of these poll numbers, we are going to face a real crisis of confidence in our democratic process.
CBS/NYTimes = Obama 54%, McCain 41% (O +13)
ABC/WaPo = Obama 54%, McCain 43% (O +11)
Gallup Tracking = Obama 53%, McCain 42% (O +11)
USA Today/Gallup = Obama 53%, McCain 42% (O +11)
Marist = Obama 53%, McCain 44% (O +9)
NBC/WSJ = Obama 51%, McCain 43% (O +8)
Ipsos/McClatchy = Obama 50%, McCain 42% (O +8)
Reuters/Zogby = Obama 51%, McCain 44% (O +7)
CNN = Obama 53%, McCain 46% (O +7)
Democracy Corp = Obama 51%, McCain 44% (O +7)
FOX News = Obama 50%, McCain 43% (O +7)
Pew = Obama 49%, McCain 42% (O +7)
Daily Kos/R2K = Obama 51%, McCain 45% (O +6)
Rasmussen = Obama 52%, McCain 46% (O +6)
GWU = Obama 50%, McCain 44% (O +6)
Diageo Hotline = Obama 50%, McCain 45% (O +5)
IBD/TIPP = Obama 48%, McCain 43% (O +5)
What stands out here? Well, Obama is leading in every poll, and by a margin of 5 points or greater. A clear majority of the polls (12 out of 17) place the lead at 7 points or larger. And several major polling organizations have the lead in double digits.
That's certainly a good position to be in on the eve of an election. For some perspective, here were the final voting margins in other recent presidential elections (excluding the 2000 and 2004 squeakers):
1980 = Reagan 50.7%, Carter 41.0% (Reagan +9.7, electoral margin 489-49)
1984 = Reagan 58.8%, Mondale 40.6% (Reagan +18.2, electoral margin 525-13)
1988 = Bush 53.4%, Dukakis 45.6% (Bush +7.8, electoral margin 426-111)
1992 = Clinton 43.0%, Bush 37.7% (Clinton +5.3, electoral margin 370-168)
1996 = Clinton 49.2%, Dole 40.7% (Clinton +8.3, electoral margin 379-159)
The key point to underscore here is how stunning some of these poll numbers are relative to recent historical election results. If the majority of polling organizations are correct, Obama is looking at a victory the size of George H.W. Bush's victory over Dukakis in 1988. If organizations like Gallup and CBS News are correct, then Obama is looking at a victory margin matched only by Ronald Reagan in 1984.
I'm personally somewhat skeptical of the double digit projections. I think the most likely scenario is a popular vote win of 5-7 points.
I will say this, though. If McCain manages to win this election in the face of these poll numbers, we are going to face a real crisis of confidence in our democratic process.
posted by A.L. at 2:15 PM
10 Comments:
I will say this, though. If McCain manages to win this election in the face of these poll numbers, we are going to face a real crisis of confidence in our democratic process.
You do remember a little thing called the Electoral College? The Founding Fathers put that in to keep from having too much democracy in what is supposed to be a republican nation of united republican states (small "r" republican, not the Republican Party); the idea is that lesser populated states wouldn't be bowled over by more heavily populated states. If Obama wins the popular vote, but McCain wins the electoral vote, the one that counts, it won't be the first time and the country will get through it just fine.
It was the same with how members of the Senate were decided upon before "progressives" screwed that up with the 17th Amendment; each state's general assemblies selected the U.S. Senators, not the people through an election.
Shouldn't a McCain win cause you to face a real crisis of confidence in the faith you put in polls?
Interesting post on the poll phenomenon! With the 2008 presidential campaign finally wrapping up tomorrow, I thought you might be interested in some light-hearted political fun. Here at Public Agenda, we think that political movies have the potential to encourage public engagement, as many films explore intriguing social and political themes. That’s why Public Agenda is asking what your favorite political movies are!
We’ve posted a list of our favorites at http://www.publicagenda.org/pages/our-favorite-political-movies and we’re asking you to post yours at http://www.publicagenda.org/forum/election-2008/your-favorite-political-movies.
Also, be sure to check out our non-partisan Voter’s Survival Kit at http://publicagenda.org/citizen/electionguides.
Shouldn't a McCain win cause you to face a real crisis of confidence in the faith you put in polls?
Of course. What I mean, though, is that many people will feel (rightly or wrongly) that the game was somehow fixed. If the real numbers depart dramatically from the polls, they'll think that the election was stolen somehow. This sentiment was widespread after 2000 and 2004 and in those elections the polls were close.
I will say this, though. If McCain manages to win this election in the face of these poll numbers, we are going to face a real crisis of confidence in our democratic process.
To say the least.
As to questioning the polls: that's why it's helpful to have so many polling organizations using different methods. Statistics are meaningful and science matters. It's time we started getting used to dealing with reality again.
It is well known that random sampling is generally more accurate than trying to count everything.
There used to be an exercise in trying to count the number of e's on a sheet of text. Picking ten lines at random, counting their e's and multiplying up was more accurate than laboriously counting the e's in every line, where human error crept in.
Modern polling is by definition not an exact science, but is an accurate a science as any. Statisticians think the US census would be more accurate done by sampling than by the inaccurate 100% count. The errors from sampling would be less than the errors from the 100% counting process.
That is not an argument for replacing the democratic vote with a sample. In a democracy, everyone's vote should get an equal weighting. But it does point up the importance of ethical management of the voting process, and the conduct of the count.
I am a bit shocked at the long voting lines shown on TV. Surely, America voting in 2008 should not look like the Zimbabwe election 2008. The delays tend to suggest bad planning and mismanagement, and mismanagement can easily turn into fraud.
After Ohio 2004, and Florida 2000, this result hopefully will be above suspicion. I mean, how lucky was George W. Bush to win two of the closest elections in history? And then, suppose it gets followed by a shock turnaround, even narrower victory for a Republican, against ALL the polling predictions (McCain has been behind in over 100 successive tracking polls.)
Wouldn't you smell a rat there, somewhere?
Wouldn't you smell a rat there, somewhere?
Yeah, in the polling.
Hypothetically, steveil, what temporally plausible occurrence would convince you otherwise?
Here's the problem, eludication. National polls, and A.L. knows this, are immaterial. Obama could win the popular vote by carrying 12 states that contain the nation's largest cities. But if McCain carries the other 38, it's over and McCain becomes the President. That is due to the Founding Fathers far-sighted use of having an Electoral College vote for the President.
It's the state polls that carry more significance than the national ones. And it doesn't matter until all the votes are counted anyway since a poll response is not a vote.
That's a dodge; the state polls project Obama to win between 300 and 400 electoral votes. So, again, what would convince you otherwise? Or is this belief of yours simply unfalsifiable by any plausible empirical evidence whatsoever?
Post a Comment
Links to this post:
Create a Link
<< Home