Sunday, October 21, 2012

Gallup pollster goes on Fox News Sunday to explain why Romney leads in his poll by 7 points, and everyone else has it tied

Posted by on Sun, Oct 21, 2012 at 1:17 PM

Frank Newport, editor-in-chief for Gallup
  • Frank Newport, editor-in-chief for Gallup
We all know that the presidency will be one by the candidate with the most electoral votes, not necessarily the popular vote. That happened in 2000, where Al Gore had 500,000 more votes than George W. Bush, but lost the electoral vote when the Supreme Court ended the recount in Florida, giving Bush the Sunshine State's 27 electoral votes (the state has 29 this time around), and getting him to 271, five more than Gore.

More pundits have been speculating lately that we could have a 2000 redux, with Mitt Romney getting more votes overall, but falling short in the Electoral College. But that's something to worry about on November 7, not October 21.

No, today we want to look at the national vote, just for the sake of it. And it looks damn close, with NBC News/Wall Street Journal showing a 47-47 tie. Real Clear Politics also shows the race even at 47 percent. Seemingly everyone has the national race tied or separated by a percentage point or two.

Everyone, that is, but the Gallup daily tracking poll. which on Sunday had Romney back up by 7 points over Obama amongst likely voters, 52-45 percent.

Needless to say, conservatives have been lovin' Gallup first reported on a seven (then six, now back to seven) point in their 7-day tracking poll last Wednesday.

What does Gallup know that the other pollsters don't?

On Fox News Sunday, Frank Newport, Gallup's editor-in-chief, attempted to explain exactly.

Newport: We do not have deep flaws in how we do things. We are highly transparent and have a team of methodologists and scientists at Gallup, best in the business, constantly looking at what we do and how we do sampling. We tweak sampling. We tweak how we interview. We just added actually more cell phones to try to take into account the growing number of people who don't even talk on land lines, with these phone samples.

So I think our methodology is extremely solid. And we are very open about how we do it.

So I would say we're doing great, and it's not unusual, of course, that people would fight back at you if they don't like what your findings.

Gallup's big poll number for Romney is based on how "likely" voters say they will vote. When it comes to registered voters, Gallup has been pretty close to the other nationally survey that show the race nationally to be neck-and-neck.

So how does Gallup determine who is a "likely voter"? Let Frank Newport explain:


We use seven questions that we have tweaked over the years.

But they include questions about knowing where people vote, or how you vote, if you vote by mail, how much attention you're paying to the campaign, how much thought you have given to it, how certain you are to voting your own self-definition and your history of voting, which we take into account for young people who couldn't have voted previously and we put all that together and isolate likely voters.

It can certainly be susceptible to events and the environment, but that's the whole idea. If events in the environment cause one group or the other to become more likely to vote based on these measures, that's what it reflects.

Newport then went on to explain that going by likely voters is why Gallup and a few others (like Rasmussen) were much more accurate than the rest of the nation's pollsters on the 1996 presidential contest.

That's when Bill Clinton went into election day that year expected to by around 16 percent points, but in fact won by just eight points. Newport says that when there a Democratic incumbent -as is the case as in 1996 (and 1980 with Jimmy Carter) "the likely voters make a big difference. "

We'll know soon enough if Gallup has been an outlier, or the one prescient voice out there when it comes to calculating the national vote. Of course as was noted earlier, that may not make the difference in this year's election, a la 2000.

Tags: , , , , , , , , , , ,

Comments (0)

Subscribe to this thread:

Add a comment

Latest in Daily Loaf

More by Mitch Perry

Search Events

© 2013 SouthComm, Inc.
Powered by Foundation

Web Analytics