Darren Richardson

Oct. 8, 2012

Poll-watchers love this time of year. More accurately, poll-watchers love this time of every four years. With so many organizations conducting so many polls, poll-watchers can always find something new and exciting every news cycle.

Monday brought a new poll from the Pew Research Center, showing a drop in support for Obama and a gain for Republican nominee Mitt Romney, with Romney ahead 49-45 percent among likely voters.

That poll got a lot of attention, but what got less attention was that the latest Rasmussen poll apparently features a 101 percent sampling base. It’s either that, a widely overlooked typographical error or liberal use of "rounding." If both Obama and Romney really received 47.5 percent instead of 48 percent, is it really so much trouble to write that?

According to Rasmussen’s daily presidential tracking poll of Oct. 8, Romney and Obama are tied with 48 percent each, with “some other candidate” polling at 2 percent and 3 percent of the voters undecided. Let’s see, 48 plus 48 equals 96. Adding 2 percent for “some other candidate” gives us 98 percent, and adding in the 3 percent of undecided voters gives us 101 percent.

Talk about a poll that just doesn’t add up!

The Punditty Project has no doubt that Rasmussen always give 110 percent, so to speak, but in a world where a single digit can influence contributions to candidates by donors of all income levels, you would think Rasmussen would explain – or correct – the discrepancy.

Maybe there is some kind of perfectly plausible explanation for the 101 percent figure, but Rasmussen doesn’t give it on his site. Perhaps if someone pays for the Platinum membership, they can get the numbers that do add up.

It makes one wonder about Republicans’ pre-debate complaints against polling outlets, when they alleged a pro-Obama bias. No one who knows anything about Rasmussen would ever accuse that site of skewing the polls in favor of Obama, but this apparent ho-hum disregard for basic arithmetic does cause one to wonder about the accuracy of Rasmussen polls.

In the rapidly changing world of polls, this 101 percent figure will be mostly forgotten when the next round of polls is released. But for now, it remains to be seen if Rasmussen will clarify the Oct. 8 figures or just let them slip into the memory hole.

If you like to write about U.S. politics and Campaign 2012, enter "The American Pundit" competition. Allvoices is awarding four $250 prizes each month between now and November. These monthly winners earn eligibility for the $5,000 grand prize, to be awarded after the November election.


Rasmussen's methodology page

Additional sources and reources linked to within text.


Follow PundittyProject on Twitter

The Punditty Project on Facebook

Punditty on Allvoices.com