NRL News

Poll Volatility Further Evidence Presidential election likely to go down to the wire

by | Sep 24, 2012

By Dave Andrusko

The New York Post’s headline for Jay Cost’s latest analysis of presidential polls–“Inside the ‘poll-ercoaster’”–is cute, slightly whimsical, and right on the money. There is simply no rhyme or reason (we’re excluding conscious bias for the moment) that explains the wide disparity, particularly in state polls.

That gap is less the case for national polls that survey regularly, for example Gallup and Rasmussen. Gallup has President Obama up two points today, Rasmussen by one point.

Which brings us to the latest POLITICO/George Washington University Battleground Poll which finds Obama up by 3%. POLITICO headlines its story about the poll, “President Obama pulls ahead of Mitt Romney.”

What to say about this poll of 800 likely voters?  Actually the story’s next three sentences say a great deal:

“Obama leads 50 percent to 47 percent, which is within the margin of error. His 50 percent job approval rating puts him at a crucial threshold for an incumbent seeking reelection. It’s an uptick from the spring and summer, but 48 percent still disapprove.”

What jumps out?

#1. Unlike many polls, the voter sample is “actually pretty good,” according to Ed Morrissey. The number of Democrats/Republicans/Independents is defensible. Then it gets really interesting.

#2. Romney retains his lead among Independents, 46% to 44%. Obama crushed Senator McCain in 2008, winning Independents by 8 points. The “gender gap” Obama benefited from four years ago (a whopping 13 points among women, one among men) is down to 4. That is, Obama leads among women by 10, but Romney leads among men by 6.

#3. If the overall point is always look to see if numbers are way outside the norm, this is never more true than among Catholics. Last week, one poll mysteriously had Obama up among Catholics. The POLITICO/George Washington University Battleground Poll has Romney up by 8 points among Catholics.

“If Obama is losing independents and Catholics and is only down to a +4 gender gap, the road to victory looks pretty narrow indeed,” Morrissey concludes.

The POLITICO/George Washington University Battleground Poll is conducted by two firms, one Democratic, one Republican. Obviously, each side takes the same results but comes to different conclusions.

The Republican, Ed Goeas (writing for POLITICO with Brian Nienaber), takes on a major motif of critics of Gov. Romney: that he is lapsing among middle class voters. “We took a special look at middle-class voters, and middle-class families in particular, in this latest POLITICO-George Washington University Battleground Poll and found that not to be the case. In fact, on every measure it is Romney who is winning the battle for the support of middle-class families.”

To take just one specific, Goes and Nienaber write, “In our latest POLITICO-George Washington University Battleground Poll with middle-class families, which comprise about 54 percent of the total American electorate and usually split in their vote behavior between Republicans and Democrats, Romney holds a 14-point advantage (55 percent to 41 percent).”

A final consideration, borrowed from Jay Cost, writing for the New York Post (beyond the hopefully self-evident point that we ought to take all polls with a grain of salt and all those which insist that Obama is on cruise control with a pound of salt).

We are told relentlessly that only a tiny percentage of voters doesn’t already have their minds made up. That, in all likelihood, is not even close to being true.

“In fact, something like 25 percent of voters make their voting decisions after September, and anywhere from 10 percent to 20 percent will make their final choice in the last week,” Cost writes. That’s why you get elections such as the 1980 Reagan/Carter contest which was a toss-up just a few weeks out but ended up with Reagan winning easily.

“September polls are extremely volatile,” Cost adds. “And this year’s volatility is compounded by the late date of the Democratic National Convention. It was, in fact, the latest party convention in US history. And when the polls are bouncing around a lot, the chances are much greater that they will disagree with one another — which is exactly what we’re seeing right now.”

Just keep working at what you are doing. No one can predict the future, certainly not me. But I will bet you dollars to donuts that by mid-October we will suddenly be told that, whoa, this election is really close!

Your feedback is very important to improving National Right to Life News Today. Please send your comments to If you like, join those who are following me on Twitter at

Categories: Polling