likely voters v.s. registered voters

[ginnyinWI]

Today there is a disparity between to two--more registered voters for Obama than likely voters. Which one is a more realistic way to predict the vote? Is a likely voter only someone who has voted in the recent past? What about newly registered voters--how should we factor them in?

If the likely voter count for Obama is smaller than the registered voter count, does that mean a lot of new voters for Obama, or what?

[mohc]

1. Democrats almost always do better among RV than LV

Likely voters are just a subset of registered voter that pass the particular pollster's likely voter screen. It can be as simple as asking the voter "How likely are you to vote?" to other things like asking them how many times have they voted in the past.

[last1standing]

2. No one can answer those questions definitively.

However, Likely Voters are usually a better indicator of actual results - with a big caveat - they are a better indicator when turnout is low than when it is high. We have every expectation that there will be a larger than usual turnout this year - it could even break records, but we can't really know until election day.

Oh, and likely voters are generally counted as people who have voted in the last election so it would not include the millions of newly registered voters who are likely to vote. Make sense?

[sandnsea]

3. They always do this

They go from "all people" to "voters" every 4 years. Then slice it up even more to "likely voters". That's how come the polls never match what people think. If the half the country that doesn't vote would, things would be very different in this country.