I've gone for a new look for my comparison of the Gallup and Rasmussen daily tracking polls on the Dem primary, as I found
the graph in which I mapped both polls parallel to become confused.
Also, I didnt like that it didnt include a representation of the share of voters who said they didnt know, or preferred a third candidate. Whereas that can be an interesting indicator.
Between March 15 and March 18, for example - the days after the Wright thing first erupted - Obama lost an average of 3 points in the two polls, but Clinton stayed stagnant. It was the "dont know" camp that went up 3. (Thats more reassuring for Obama than if they'd simply crossed over. Although from 14 to 15 March, he did lose an average of 2.5% directly to her.)
So here's the new look: the Gallup and Rasmussen numbers the mere thinnest of lines, and fat lines for the averages of the two:
Most striking element: the total lack of overall movement since March 17. On the 18th Obama still dropped a point in both polls (and that would be in interviews up till the evening of the 17th), and then in the last two days nothing. So that suggests that Obama's speech did succeed in stopping the bleeding, but hasnt created a rebound, or not yet.
It's hard to tell for sure though, because Gallup and Rasmussen disagree strikingly about what's been going on the past couple of days. You can see this in these separate graphs for how they have been polling Obama and Hillary:
Both polls agreed that Obama slipped significantly around the March 4 primaries - from 50% or close to 43-44%. They both agreed that he recovered afterwards, culminating a few days after the Mississippi primary when he hit 50% in both polls. And both agreed that he slipped right back when right then, the Wright affair erupted.
But then what? Rasmussen had him drop 4 points in a day, but then holding roughly steady at 45-47%. Gallup saw him dropping further, to 42% - his worst score in the Gallup poll since February 8.
And the speech? Rasmussen had him immediately picking up 2 points, then dropping again 1; Gallup only now sees him picking up just 1 point.
There's some cautiously hopeful news in the accompanying description to today's Gallup numbers though:
Quote:Obama made a major speech on Tuesday addressing the race issue in large part to help move past the controversy. While Tuesday night polling showed no immediate benefit for Obama, the Wednesday results were more favorable to him, as reflected in the slight drop in Clinton's three-day average lead. [..] The tracking data over the next several days will be a crucial indicator of the lasting impact, if any, of the Wright controversy.
Following up on
yesterday's description, that means that the day-to-day interviewing has gone as follows:
Interviewing up through Sunday - Hillary leads Obama narrowly
Monday night interviews - Obama drops, Hillary takes a significant lead
Tuesday night interviews - Same numbers as on Monday
Wednesday night interviews - An upturn in Obama's numbers
That means that even if tonight's interviews just got the same results as last night's, we'll see Obama climb back up in the tracking poll a little further again.
The disagreement between the polls is actually larger when it comes to Hillary's numbers:
Not for the first time, Gallup has Hillary's numbers higher than Rasmussen - yesterday, Gallup pegged her at 49%, Rasmussen at 42% (the largest 'disagreement' between the polls since Feb 14). The reason: Gallup saw Hillary picking up not just on the 15th, but also in the several days since; Rasmussen did not agree.
The contrast between the two polls is visible most starkly in the familiar graph showing the lead (or deficit) Hillary has on Obama:
In short: in the aftermath of the Wright revelations, Gallup saw Hillary rebounding to a 7-point lead: the largest she'd had since Feb. 8. Today that's chipped down a little again, hopefully tomorrow some more. But Rasmussen never believed it in the first place: it kept polling a narrow lead for Obama.
The volatility and variance in these numbers, I think, primarily underscore the ambivalence of the Democratic voters, many of whom just still find it hard to choose when asked, even now that most of them have already cast their ballot.
