by JS O’Brien
Conventional wisdom in presidential elections is that they almost always tighten near the end. Today’s Rasmussen Reports tracking poll, which shows Obama’s lead shrinking to +3, 50% to 47%, not only reflects a tightening, but may reflect deeper trouble for Obama, as well.
Or maybe not.
The Rasmussen Reports poll has been the steadiest of all the polls during this election, primarily because of its methodology. Unlike most polls, Rasmussen’s is automated, meaning that the questionnaire items are always delivered in exactly the same way. In addition, Rasmussen’s sample size is larger than most, reducing its margin of error, and this large sampling allows its weighting by Republican vs. Democratic responses to be more statistically meaningful. Since my last post on the polls, Rasmussen has shown very little movement from the +5 to +7 or so for Obama that it’s been showing for over a month. Other polls have been coming in with not-terribly-credible, outlying numbers (Obama +15 yesterday in a Pew poll and Obama +1 in an IBD/TIPP poll from last Thursday), but Rasmussen has been steady as a rock, and has tracked within a point or two of both the Real Clear Politics (RCP) average and the FiveThirtyEight trend line for at least a month.
This is why it’s so surprising that Rasmussen should move from +7 on Monday to +5 yesterday (a movement I attributed to normal margin-of-error fluctuation), to +3 today. That’s four points in three days from what has been a rock-steady poll.
Is Rasmussen reflecting reality? Well, there has been some movement toward McCain since last week. Obama’s RCP average soared as high as +8 and was down to +6.3 yesterday (it’s down to +5.9 so far today, largely because of the Rasmussen results, so it will change during the day as new polls come in). I think the +8 number was too high, but +7 certainly wasn’t, so it’s fair to say that McCain has picked up around a polling consensus point, or perhaps a bit more, from his low-water mark. That would be normal tightening, but a move like the one Rasmussen is reporting — four points in two days — would be something else, entirely, especially since tracking polls reflect multi-day, rolling results, indicating that Rasmussen’s recent results show an even stronger-than-reflected trend for McCain.
Adding further to the confusion is that polls on the presidential race at the state level aren’t showing this sort of trend, at all. I checked the dates on the most recent state polls to see if they could be lagging the data from the tracking polls and found that, if they are, it’s not by more than a day on average. I don’t notice any strong movement toward McCain in the battleground states, but perhaps today’s new state polls will demonstrate some of that.
So, what have we got? No one knows. Statistical noise? A strong movement toward McCain in states so red or blue that no one is bothering to track them anymore? A real movement toward McCain nationwide that will soon be reflected in the vital battleground states?
Stay tuned. As polling results accumulate today, we should get a clearer picture of what’s really going on.