On Polls, Cell Phones, And Storms

One week before the election, and the race looks tighter than ever. Polls are all clustered right around a dead heat, with some showing Obama ahead by a point or two, some showing Romney ahead by a point or two, and some deadlocked. But are the polls reliable? This post at The Moderate Voice got me thinking about polls generally. Assuming most still rely upon telephone surveys, how do they factor in those who have abandoned landlines for cell phones? How do they factor in people who look at their caller ID and, seeing “Unknown Caller” or “Out Of Area”, assume it’s a poll calling and don’t answer? Are those people evenly split? We’ve been hearing for months about how the Republicans are more enthused about this election. Is it possible that makes it more likely that they’ll answer that call? You could then speculate that those people would cause Romney’s support to be overstated – but does that same enthusiasm make it more likely that they’ll show up to vote? The Pew Center has studied this, and has found, not surprisingly, that response rates have fallen dramatically, as shown in the graphic below:

5-15-12 #1

They note that most surveys include cell phones now, but how do they incorporate those? Do cell phone-only owners answer these calls at the same rate? And I have to wonder if, as the election draws near, response rates have fallen even further due to voter weariness, especially in the swing states.

Now, looming above and around all the polls, is this enormous storm affecting most of the northeastern US. That’s many, many households that won’t be reachable by any method in the last crucial week. With this representing large areas of Obama’s strength, will we see a spike in Romney’s number in the last polls? If we do, will that artificially add to the sense of momentum that Romney’s campaign has been claiming (with justification, to be fair). The storm could affect the election in several ways, directly and indirectly. And with Obama’s support seeming to soften, if not actually crumbling a bit at the edges, this couldn’t have happened at a worse time for him. He does get to act Presidential, being in charge during a crisis, but whether that offsets potential downsides of storm effects on actual voting will be an object of study for future analysts, I’m sure.

October 30, 2012 В· Harry В· One Comment
Posted in: Politics

One Response

  1. PW - October 30, 2012

    I\’d have to admit that this time around I\’ve slammed down the phone because of robocalls. And when it\’s a pollster, it\’s likely to be a push-poll which has a bad effect on blood pressure!

    It wouldn\’t matter if polls themselves \"push.\" That is, they can be used and often are as ways of engaging the media\’s collusion in creating or enhancing one or another candidates \"momentum.\"

Leave a Reply