Home | Entries [Previous |Next]
0 Comments
Analysis for October 04, 2006
Commentary: Polling on the Senate Race
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), 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.
- Michael Barone
Instapundit is leading an interesting discussion on Harold Ford, Jr's showing in the US Senate race here in Tennessee, with particular discussion on what robocall polling may show differently than live-call polling. Reynolds speculates that the live-call polls inflate Ford's numbers because live-call respondents are less willing to buck Ford for racist reasons. The trends as Barone notes, however - show that Ford is doing best among robocalls, where callers should be more likely to buck Ford for racist reasons, if that were the case.
Confer, Mickey Kaus speculates that "voters in Tennessee don't want to admit in front of their conservative, patriotic fellow citizens that they've lost confidence in Bush and the GOPs in the middle of a war on terror and that they're going to vote for the black Democrat." Of the two speculations, this one seems more plausible, but in my opinion is still unlikely, because it would assume Tennesseans are voting with the ebb and flow of Bush's approval ratings, which does not seem to be the case. Against this point, one might cite pollster.com: "We did a similar 'pollster compare' graphic on the Virginia Senate race over the weekend. The pattern of automated surveys showing a slightly more favorable result for the Democrats was similar from July to early September, but the pattern has disappeared over the last few weeks as the surveys have converged."
With that said, Ford's trending is undeniable - he has been ascendant since beginning his extensive television campaign - and all of his ads have been excellent (with the possible exception of the recent 'Lifestyles' ad). His message has been on point in appealing to the average moderate Tennessee voter. The trending is explained in as simple terms as that. Earlier, Tom Ingram commented on the polling saying that Corker came out of the GOP primary with inflated polling numbers because the television advertisements were not competitive in terms of total dollar commitments at that point between Corker and Ford. What Ford's ads since have done is bring Ford's positive votes up by remaining relatively tame in terms of negativity, while still knocking Corker down on issues important for this cycle like security and illegal immigration. All of the ads thus far produced by the Ford campaign have had the dramatic effect of boosting Ford astronomically by appealing to traditionally conservative issues.
The one thing working to Corker's benefit is that for the most part, I do not believe Tennessee voters have made the Senate election a referendum on Bush. It is clear that despite the slightly higher positive showing for Bush in Tennessee than national averages, if this were the case Corker should be much lower. The fact that Corker, despite not having a good answer to Ford's TV spots is still tied with Ford in the polls shows that voters are not following the Kaus thesis and making the question of who represents them in the Senate for the next six years a decision about whether or not they like Bush. For the most part, Ford has realized this lack of connection and played up attacks on Corker's personal record rather than the wider record of Republicans in general. This is probably a wise strategy in a state where more folks might be put off by a wide-net criticism of political identity as opposed to specific criticisms where the Democrat is doing everything possible to push himself to the center-right.
The conclusion here is that the slight disparity in the numbers between robo-calling and live-calling has no significance because there is no difference, as Barone has stated. This disparity is "interesting" only if we can conclude that robo-polling is a less-valid methodology than traditional polling. But up to this point, we have failed consider other methodological factors in each of these polls, including sampling issues (what quotas were used and how were they determined) and the definitions of likely voters (what was the methodology for determining propensity for inclusion in the sample universe). The relatively closeness of all of these polls makes these questions somewhat unless, however. The simple reality is the race is a statistical tie - and that is illustrated in all of the recent polls, robo, live or otherwise.
