If you've followed the polls over the last couple of days, you might be tempted to sing a chorus of "Me and My Shadow." In real time, you see, PPP has released polls within days (or, in today's case, hours) of a new Rasmussen poll. It happened in North Carolina, and it happened in Colorado.
And, not surprisingly, the two firms came up with very contrary results. As in, high single digits of variance between the two pollsters. This will undoubtedly leave some on the Right to conclude that PPP is trying to jump in and change perceptions before Rasmussen's decent numbers for Mitt Romney can take hold.
However, I think this is one of those cases where the appearance is counter to the reality. I think either (a) the timing is coincidental, or (b) it is the House of Ras that is chasing PPP and trying to muddy perceptions about the state of play. My reasoning for this is actually pretty simple, and I will share it after the jump.
But, first, on to the numbers:
PRESIDENTIAL GENERAL ELECTION TRIAL HEATS:
NATIONAL (Gallup Tracking): Obama tied with Romney (46-46)
NATIONAL (PPP): Obama d. Romney (48-46)
NATIONAL (Rasmussen Tracking): Obama d. Romney (46-45)
COLORADO (PPP): Obama d. Romney (49-43)
COLORADO (Rasmussen): Obama tied with Romney (47-47)
NORTH CAROLINA (PPP): Obama d. Romney (49-46)
DOWNBALLOT POLLING:
HI-SEN—D (Benenson Strategies for Hirono): Mazie Hirono 50, Ed Case 33
A few thoughts, as always, await you just past the jump...
Now, if I have any readers on the Right, they are probably chortling right now. "Ha! Singiser is saying Rasmussen is shadowing PPP, when PPP's polls came after Rasmussen's. Typical delusional lefty! Bwahahaha!"
(Sorry, pals on the right, but this is how I imagine your internal monologues...)
Well, here's the thing: my guess is that it probably is a coincidence that the two pollsters came out with numbers in the same swing states, with contradictory data, within days of each other.
But, if it is not a coincidence, I would say there is a 90-plus percent chance it was Rasmussen following PPP's lead.
The reason for this is actually really quite simple: PPP has a well-earned reputation for being very solicitous of input for where and how they poll. Therefore, astute poll junkies have known for nearly a week that PPP was going to poll in both North Carolina and Colorado. They asked for input about what to ask in those states last Wednesday afternoon.
Therefore, North Carolina is almost certainly a coincidence. Rasmussen went into the field on the same day that PPP revealed they were polling the Tar Heel State. However, Colorado went in the field several days later.
The one rather instructive thing about these two very disparate pollsters offering nearly contemporaneous data is that we can see exactly WHAT drives the differences between the two polls. Take North Carolina: PPP's poll there had a chasm of a gender gap—they had the president up nineteen with women, but down sixteen with men. Rasmussen had a much tighter spread among women—their take was that the gap between Obama and Romney among female voters was a mere five points. Among men, their margin was almost identical (15 percent) to PPP's. The 2008 exit polls essentially split the difference between those two ends of the spectrum, as it related to women voters.
Furthermore, the two camps see independent voters dramatically differently. Rasmussen gave Romney a 30-point lead with "voters unaffiliated with either of the two major political parties." PPP had it at four points.
In 2008, John McCain carried Indies by 21. However, Barack Obama's numbers in the PPP poll showed a much greater defection from Democrats than the 2008 exit polls (or, presumably, Rasmussen, given Romney's comically large lead with Independents).
Herein lies the danger in making a lot of conclusions based on self-identified party ID. My guess here is that when PPP asked about party, some conservative Democrats identified their registration, as opposed to their common preference. That would explain Obama losing 18 percent of Democrats in the PPP poll, but only 9 percent in the 2008 exits.
In other polling news...
- Well, the streak of days where both Gallup and Rasmussen gave the president the lead in their daily trackers has ended. At one day, for those having trouble keeping track. However, professor and polling expert Alan Abramowitz actually was on Daily Kos this morning with some analysis of precisely why that might be the case. Worth a read.
- Another poll out of the great state of Hawaii graces us this Tuesday, and it is of the Senate race. Democrat Mazie Hirono's team drops a poll, courtesy of Benenson Strategies, which shows her with a comfortable lead over former Rep. Ed Case. Two items worth noting: first—while little early polling in this race showed such a lofty margin, the most recent one (courtesy of the Honolulu Star-Advertiser a couple of weeks ago) had a similar lead for Hirono. Second—it cannot go unmentioned that Ed Case's crew has been pretty quiet on the polling front as we head to the wire this weekend (yup...Saturday primary) in Hawaii.
- And...y'all will have to let me know how that one turns out. It is that time of the summer, which means that the Polling Wrap and its trusty curator will be taking a brief hiatus while the family truckster heads eastbound. You can look for the next edition of the Wrap on Monday evening, August 20th. In the meantime, continue to check out Daily Kos and Daily Kos Elections for all your numbers needs. And behave, or you are in for it when I get home, kids!