Question: what is the margin of error? I would guess + -3% or 4%.
Since the assertion is that this 40% figure is of "likely" voters, what is the margin of error in identifying likely voters? Is it also + -3% or 4%? If so, should these two sums be added for a total margin of error of 3% or 8%?
The report also says:
"Obama leads 60% to 37% among these voters."
Question: is there a margin of error here? Could this be 57% to 40 %, for example? Again, is this a margin of error cumulative? If it is the fact that Rasmussen has gotten corrupt data in that nothing like 40% have already voted, rather 28% according to official statistics have voted, is all of this corrupt? Is it possible that this poll is correct and that the official statistics are lagging and will eventually come up to his poll? How much credibility does this deserve?
The report goes on:
"Ninety-three percent (93%) have made up their minds whom they will vote for, and it’s Obama 50%, Romney 49% in this group."
Since Rasmussen obviously did not question everyone who has made up his mind among likely voters, is it correct to assume that there is a margin of error in identifying the group that he says is 93%? Is it also fair to say that the margin of error here should be an added to the original margin of error? Is it fair to say that the cumulative margin of error could be as high as 8%?
If Rasmussen has one point separating Obama and Romney (50%-49%) "in this group" it seems to me the margin of error is cumulative in identifying the group and identifying the result within the group. Is the margin of error 3% +3% or only 3% of 3%?
We are talking about subgroups within subgroups and a difference of 1%. How can we draw any conclusions from these data in such a close race?
