October 03, 2006
THE LATEST POLL shows a Ford-Corker dead heat.
Hmm. Just yesterday we had one with Ford up by 5; not long before that there was one with Corker up by 5. Is it just me, or is this more variation than we usually see? Are voter sentiments that volatile (or superficial)? Or is there something about this race that makes minor differences in polling methodology more important? Or is this normal?
UPDATE: Michael Barone emails:
I don't think the poll numbers in TN are unusually volatile. When you look at all the public poll numbers since the primary (excluding the Zogby Interactive), link, you see that Ford is getting between 42% and 48% and that Corker is getting between 42% and 48% (except for one Democratic poll that shows him at 39%). With an error margin of plus or minus 4%, all those numbers could be the same.
Interestingly, Ford does best (48%) on the two robopolls, Rasmussen and SurveyUSA.
This is about the same level of variation as in the MO Senate race, which seems the closest thing we have this cycle to a tie. Link.
Interesting. And that Ford is doing better in the robopolls casts doubt on a theory I've heard around here, that some people are telling pollsters they'll support Ford over Corker in order not to sound racist, but that they'll vote for Corker in the privacy of the ballot box. Seems to me that a robopoll would vitiate this phenomenon if it existed.
ANOTHER UPDATE: Some interesting thoughts on robopolls from Mickey Kaus. For this to be true, though, I'd expect us to see the same pattern elsewhere. Do we?
YET ANOTHER UDPATE: Mark Blumenthal emails:
Barone beat me to the first half of the answer (though I expanded on that a bit), but a very interesting thing happens when we do the Pollster.com thing and draw a picture: The phone surveys track well with each other, but Zogby/WSJ Internet trends are completely different: Link