Friday, October 26, 2012

When 'probable' voters are getting filtered out of a likely voter screen ...

A woman walks into a polling station for early absentee voting at Courthouse Plaza in Arlington, Virginia October 25, 2012. Two weeks before what could be one of the closest presidential elections in U.S. history, efforts to mislead, intimidate or pressure voters are an increasingly prominent part of the political landscape.    REUTERS/Gary Cameron A real voter in Virginia who may or may not be counted as a pollster "likely" voter. I'm going to pick on ABC News' poll because they provide information on their likely voter screen that most other pollsters do not. (So kudos on that.)

Their latest poll is here.

Obviously, they first screen for registered voters. Then they ask those registered voters whether they are:

certain to vote
probably vote
chances 50/50
less than that
don't think I will vote

The likely voter questions are then prefaced: "(IF CERTAIN TO VOTE) ..." In other words, if you say you'll probably vote, you're excluded. If you say you're 50/50 to vote, you're excluded.

Now I could see excluding 50/50 being defensible, but someone who says they'll probably vote? Seems ludicrous.

How does this affect the toplines?

In their latest poll, 79 percent registered voters are "certain to vote" and Romney leads 50-47 among them.

On 10/13, 85 percent of their sample was "certain to vote," and Obama led 49-46.

That's a six-point shift in voter certainty, which is coincidentally (or not) the same as the six-point Romney gain. I don't think it's controversial to say that it hurts Democrats when registered voters are filtered out in a "likely voter" screen. And if you're going to filter out people who say they'll probably vote, you're being particularly restrictive.

Remember, registered voter screens are historically more accurate than likely voter screens. The fact that more Democrats are voting at this point in early voting states than in 2008 also belies any claim that they won't turn out. If nothing else, the Obama GOTV operation is dragging them to the polls.

The national polling is fairly split on who has the actual lead, but there's no controversy that Obama leads with registered voters in polling that provide those numbers. This suggests two things, one, that the polling is underestimating the level of Obama support (even though he's winning where it matters with LV's anyway), and two, that GOTV is critical to making sure Obama's supporters'certain to vote or just probable voters'are counted at the polls.

You can play a key role in getting Democratic voters to the polls in swing states with Workers' Voice, the largest independent Democratic voter turnout operation in the country.


No comments:

Post a Comment