Likely voter polls usually miss mark. Here is why

[Quixote1818]

If you take a look at the two graphs listed on the link below you will see that Kerry is wining in most polls of registered voters but loosing in LIKELY voter polls:
http://www.pollingreport2.com/wh2004a.htm

The thing you have to take into consideration here is that a good chunk of voters who don't put themselves in the Likely voter column DO show up! Take a look at this Newsweek poll before the 2000 election.

Newsweek – likely voters (Oct. 31-Nov. 2)
Margin of error +/- 4%

Bush 45%
Gore 43%


Newsweek – all registered voters (Oct. 31-Nov. 2)
Margin of error +/- 4%

Bush 41%
Gore 44%
Buchanan 0%
Nader 5%


the actual tally was about half way in between both polls. So what happened? A good chunk of those who were not considered "high probability" voters showed up. Polls never factor this in. You have to basically average the registered voter poll and the likely voter poll and that will most likely give you the best results.

[janx]

1. Yup. There is no way to literally measure whether or not

people will show up to vote. That's why I've always laughed off the "likely voters" polls.

There is nothing scientific about them, and the "likely" factor is anyone's guess.

[private_ryan]

2. +/- 4%

41 could easily be 45, that's why.

[Was_Immer]

3. Its in the margin of error, so theres a need to be concerned

Dont brush this off as nothing. It could be showing a scary trend.