Opinionated Polls and Who to Believe - Poll Week


By Anna Swann-Pye

As Election Day approaches, America is looking toward the polls (or rather, squinting with shielded eyes) in hopes of being enlightened – what is the likelihood of President Obama being elected again? What are the chances that Governor Romney will be chosen to take his place? According to the most recent Gallup poll, Romney holds a slight lead over Obama in the wake of the post debate (49 percent to 47 percent). But Gallup is not the only pollster American voters are following.

Image by Sean MacEntee.

Different polls have different answers, perhaps because polling is a faulty art. Jerry Remmers writes in his article for The Moderate Voice, “at best [polls] are a broad indicator of public sentiment on an issue at a specific point. At worse, they are a manipulative weapon to calculate the efficacy of a political point of view.”

Los Angeles Times writers Phil Trounstine and Jerry Roberts suggest that a poll’s methodology, from the size of the pool to the process of selecting question phrasing, has a drastic effect on any given poll’s results. They suggest that this is why in 2009 the left-wing daily weblog The Daily Kos could come up with drastically false numbers while predicting the primary race for governor of California.

Nate Silver, New York Times poller and author of the blog FiveThirtyEight also thinks that polls can often be biased. In one post, he mentions what he calls “House Effects,” or “systematic differences in the way that a particular pollster’s surveys tend to lean toward one or the other party’s candidates.” House Effects are a result of sponsored polls – ones that benefit from showing a specific party’s strength and popularity.

Like Phil Trounstine and Jerry Roberts, Silver attributes this to polling methodology – e.g. what time pollsters collect information and how often they target the same group. Unlike Trounstine and Roberts, Silver doesn’t blame any specific poll. “When I publish something like this,” Silver writes, “my intent is not to chide the various pollsters into some kind of orthodoxy — I frankly like the difference of opinion!” What Silver is suggesting is not that sponsored polls are wrong or even that wrong polling is wrong. What he is suggesting is that we, as Americans, shouldn’t blindly trust the numbers. We should be weary and we should be informed.

In a recent article for Rolling Stone, Matt Taibbi writes, “Banning poll numbers would force the media to actually cover the issues. As it stands now, the horse race is the entire story – I can think of a couple of cable networks that would have to go completely dark tomorrow, as in Dan-Rather-Dead-Fucking-Air dark, if they had to come up with even 10 seconds of news content that wasn’t centered on who was winning. That’s the dirtiest secret we in the media have kept from you over the years: Most of us suck so badly at our jobs, and are so uninterested in delving into any polysyllabic subject, that we would literally have to put down our shovels and go home if we didn’t have poll numbers we can use to terrify our audiences.”

We shouldn’t be angry at the opinionated polls, nor should we look to them as a be-all-end-all in these next few weeks. Instead, we should take the numbers with a grain of salt and find some substance in the pre-election news.