Here’s our release today on a special edition of our monthly polling.
One thing I do want to emphasize is this poll differs from our usual monthly polling in that it samples all registered voters.
Our monthly polls only samples voters who voted in either the 2002 General, 2004 General, 2006 General or registered to vote since 2006.
This poll is probably more accurate of the larger electorate, but not necessarily those most likely to actually show up and vote.
You seem to be saying here that this is a poll of registered and not likely voters, yet this poll is labeled as “likely voters” on RealClearPolitics. I am assuming you are right about your own poll and they are wrong, but this is leading to some confusion with people following RCP.
Chad
Yes, this special edition poll is of all registered voters. Our normal monthly polls are likely voters, that could be causing some of the confusion.
So, a likely voter is one that voted in all these elections (’02, ’04, ’06, or at least registered to vote in ’06), and a registered voter is simply anyone that is registered, right?
Is there any difference?
I’m taking this to mean that turnout numbers for these likely voters will be higher than turnout for registered voters in the general election? One last question, how often do they purge voter rolls, if they even do this?
Most county boards of elections will purge a voter off the rolls if they do not vote in 2 consecutive Presidential elections elections. But it is up to the elections office on how aggressive they want to pursue it.
Each pollster will use a different model for determining a “likely voter.”
It has been shown that the best predictor of likeliness to vote is previous voting behavior.
Thus, for our monthly polls, we consider a likely voter one who has voted in one of the three previous general elections. We added newly registered voters since 2006 to attempt to capture the increased number of registrants. We then further screen by the first question we ask is “How likely are you to vote in the upcoming election”? If they say they aren’t likely to vote, we terminate the call.
The poll we released last week called all registered voters, thus someone who registered in 2001, but has never voted has equal chance of being surveyed than any other. Is that person as likely to vote as someone who has participated in elections? History tell us no.
The other thing we do that is different than some/most of the IVR firms, is we call from the voter file, so we know certain things about you before we dial the phone number, like you party registration.
Most IVR firms use Random Digit Dialing (RDD) which, as it is called, randomly dials phone numbers. Thus, they rely on the individual to tell them how they are registered. Some people, though, may be registered as a Unaffiliated, but at that moment may identify more with Democrats, and say they are a Democrat.