I'd like to thank the House of Rasmussen for giving me the catalyst for today's Wrap analysis. They're not the first to exhibit this odd phenomenon, as of late. But theirs might be one of the most clear examples of it.
You see, this morning, Rasmussen's daily tracker put Mitt Romney ahead of Barack Obama by a single point. Later in the day, however, they dropped a new poll in Ohio showing the president leading Romney in the Buckeye State by four points.
In other words, the Democratic nominee is running better in Ohio, relative to the Republican nominee, than he is nationally. At least, according to Rasmussen. The last time that has happened in Ohio? 1972. So, yeah, there's that.
Here are all the numbers on a (almost ridiculously) light Friday:
PRESIDENTIAL GENERAL ELECTION TRIAL HEATS:
NATIONAL (Gallup Tracking): Romney d. Obama (47-44)
NATIONAL (Rasmussen Tracking): Romney d. Obama (46-45)
OHIO (Rasmussen): Obama d. Romney (46-42)
DOWNBALLOT POLLING:
OH-SEN (Rasmussen): Sen. Sherrod Brown (D) 44, Josh Mandel (R) 41
A quick word about that whole Ohio thing, and what to make of it, after the jump.
In fairness to the House of Ras, they are far from the first to find that odd pairing of results. Last week, Fox News said that Romney was ahead by two points nationally. But they came out with a poll this week showing Barack Obama up by two points in Florida, and six points in Ohio.
For those who are fans of arithmetic, that means that FNC's new polling crew have Barack Obama outperforming his national margins by eight points in Ohio, and four points in Florida. It has, as I said in the intro, forty years since a Democrat has run better in Ohio than he has nationally, relative to his GOP rival.
For Florida, it goes back almost as far—1976, when the native son of a neighboring state (Jimmy Carter) graced the Democratic line on the ballot.
I am sure that the math majors in the room can drop all kinds of knowledge explaining this, but this PoliSci major has a very simple theory: national polls are simply more prone to wild swings.
I keep a separate "national" projection of the presidential race, based on the most recent state polls. I use an absurdly (stats guys may scoff now) simple way of doing it: I multiply the percentage in each state by the number of House members in each state. Then, I get the percentages from the net totals for each candidate.
By that admittedly simple metric, Barack Obama has never led by more than 4.6 points since last November. He has also never trailed. The total range was right around 4 percent, give or take a few tenths. By comparison, Rasmussen has swung by eight points during just this month alone, and their swing has been as high as 15 points over the life of their tracking survey. Gallup has also made some wild swings, including one of nine points in the last few weeks.
The volatility of state polls does exist, but taken as a whole, they don't seem nearly as prone to wild swings as these national polls (particularly those daily tracking polls). Something to consider the next time you want to exult when reading a sudden surge for the president in the latest Gallup bowl, or the next time you are despondent when the president fades in the latest Rasmussen tracker.
The flaw in doing such a state analysis, however, must be conceded: some of the polls you are using to make that assessment are really, REALLY old. Until PPP chimed in a couple of weeks ago in Colorado, the "battleground" state in the intermountain West had not been polled in 2012. You had to go back to late 2011 to find data. While that is also something to factor in, it is also something that is highly unlikely to continue to be a problem as we head closer to November. After all, the volume of polling by the end of the cycle is expected by just about everyone to be positively absurd.
Which, for folks who like reading this Wrap (and the dude who writes it), is exceedingly good news.