Calling all pollsters
Wednesday, October 15, 2008 at 09:05AM Ok, if there are any semi-qualified pollsters or otherwise knowledgeable people out there, someone please explain to the wide spread between the following three polls (paging Dr David Rockland ...)
The most recent CBS/New York Times poll has it 53-39 -- a whopping 14 point spread. That's as big as any lead I've seen since June 1st.
Zogby's tracking poll on Tuesday has it Obama 48, McCain 44 -- a mere 4 point spread. He's consistently had a narrower spread (than most other polls) throughout the fall. Zogby's claim to fame is that he was closest (among all the major polls) to the final tally in 2000 and 2004.
And finally, Rasmussen has it Obama 50, McCain 45. Like Zogby, Rasmussen has had a fairly narrow spread throughout the fall, usually between 4-8 points.
Gallup daily tracking has it at 9, and some other polls have it at high single digits, too. But I don't get the gulf between CBS and Zogby ... any answers out there? Beuller? Anyone?
Reader Comments (5)
Another good question: how in fifteen days did West Virginia go from McCain up by eight points to Obama up by eight points? Your question about the discrepancies gets to the heart of the problems with polling in general, and with polling this year in particular: what are their samples, are they of adequate size, and are they representative of the folk who will actually cast ballots? This year, add one question (which is virtually unanswerable until after November 4 when the analysts hit the fan): to what extent does latent racism, or its opposite--fear of exposure that one is actually voting for a part-African American--NOT show up in the polls and, more importantly of course, in the election results?
Are part of the undecideds--or part of the alleged Obama votes--actually folk who don't want others to know that they can't vote for a black solely on racial grounds? or are some of them folk who, in the sanctity of the voting booth, will vote for Obama but claim to vote for McCain, because in their social circle they would be ostracized for doing otherwise?
Aside from these considerations, and based on the movement in the polls, Obama could be expected to garner more than 300 electoral votes and probably 52 to 53% of the popular vote. But this year, even more than in the last two elections, the polls just may not be very meaningful. And by the way, Ferris has left the building, Nick. Maybe to vote early.
-- triton --
I, too, find these polls to be confusing. At least with the erratic DOW, the numbers are real and the losing candidate is always me (SIGH).
The problem is in error in the individual polls. Some favor dems, others repubs. I've said it before, I'll say it again, look for analysis that allows for pollster-induced error. Not just margin of error.
www.fivethirtyeight.com (the newest and most accurate source for pollling analysis)
www.pollster.com (all the polls, weighted)
www.realclearpolitics.com (although they've been found guilty lately of adding more GOP-leaning polls into their data)
Major media over-sample dems; and get a higher Obama favorable reading, why is this surprising?
This one could be a shocker, hold on to your seats...
<h1>replica designer bags</h1> Affordable bags
<h1>replica designer handbags</h1> Come ah
<h1>air max 2012</h1> Favorites
<h1>nike shox turbo</h1> Discounts and more
<h1>men puma shoes</h1> Online Sales
<h1>air max 90</h1> Wholesale sales
<h1>wholesale puma shoes</h1> Wholesale sales
<h1>puma shoes sale</h1> Look comfortable
<h1>timberland mens boots</h1> Buy comfortable
<h1>gucci women shoes</h1> You are wrong?
<h1>louis vuitton outlet</h1> There must you like