But a poll in CA shows Bush leading Edwards by 4% and trailing Kerry by 4%. So which poll is accurate?
The one with the best result, of course... 
But, remember, 4% is alot in CA, because of all the people that live there.
Not really...if the polls has a similar margin, the MoE is still the same...I'll have to think about whether it's easier or not to swing a certain percentage if the whole is bigger...it should depend on resources per voter.
And where the vote is coming from.
By knowing where the vote is coming from you will know how much support each of the canidate will get.
You mean if the poll was done the wrong way, sampling an unrepresentative portion of the electorate?
I second that.
Wich poll is most acutte? It's a problem of 3 var:
1. The dispertion of vote in each sample, the most the dispertion the most the MoE
2. The size of the sample itself, the most the sample the minus the MoE
3. And finally, as Gustaf remember us, the technical quality of the poll itself, especially the questions, sample selection and display.
End of the Resumé, so if someone has the features (and dates) of both polls we can answer the question of wich's most accute.
