We were spoiled by the New Hampshire and South Carolina polling. Those states weren’t stagnant in voter opinion, but they at least moved at reasonable speeds, and allowed for a clear understanding of what was going on.
Florida is different. After swinging 20 points to Newt Gingrich, has now gone 10-15 points right back to Mitt Romney.
I hate inconsistencies. The less consistent the polling is, either across pollsters or between readings, the less confident we can be in making predictions based on that polling. We can never be sure whether the changes are the result of real movement in the electorate, or the result of some form of measurement error.
In the case of Florida right now, the polling is looking to be inconsistent over time. Rasmussen Reports (750 LVs, MoE 4) shows Romney 39, Gingrich 31, a 17 point swing from the Romney 32, Gingrich 41 reading of just three days before. Inside Advantage (530 LVs, MoE 4) is very close to that: Romney 40, Gignrich 32, a 16 point swing from three days before.
CNN/Time (369 LVs, MoE 5) didn’t poll a few days ago, and in fact didn’t poll in a week. So CNN apparently missed the Gingrich bump and kept Romney ahead all the time. Be careful though, while Romney 36, Gingrich 34 though is a worse result for Romney than the other two, don’t be misled by the fact that CNN shows a 22 point pro-Gingrich swing since its last poll. This is why only looking at swings within a single pollster’s results isn’t a technique I use.
Estimated win probabilities based on each poll: 84% Romney, 84% Romney, 57% Romney. So it’s very near even if CNN is right, otherwise Gingrich needs pull another surge from his sleeve to get Florida on his side.
Some will say Gingrich clearly can do it, since Florida has been volatile. Others say Gingrich’s lead was a mirage, and was a bump he was never going to sustain, just basking in great media coverage after South Carolina and a debate. My suspicion is that voters just aren’t sure. In the YouTube era, every candidate is now visibly flawed.