skewed description of polls

[Literate Tar Heel]

from an article by Tim Funk on 1A of today's Charlotte Observer (headlined "Edwards has little impact on polls"):

"A poll conducted Thursday and Friday by Newsweek found Kerry is still locked in a dead heat with President Bush; nearly 70 percent of those surveyed said the choice of Edwards won't make a difference in the race for the White House.
An earlier poll for The Associated Press, released two days after Kerry chose Edwards on Tuesday, even gave Bush a slight lead over Kerry as voters expressed increasing confidence about the economy. The numbers showed Bush at 49 percent, Kerry at 45 percent and independent Ralph Nader at 3 percent."

a couple points -- it doesn't mention that the "dead heat" is Kerry at 47 percent, Bush at 44 percent, Nader at 3 percent, and 6 percent undecided ... so a 3 point lead with Bush at 44 percent is a "dead heat" which doesn't warrant mention of the numbers, but a 4 point lead with Kerry at 45 percent is a "slight lead" with voters expressing increased confidence about the economy ... I understand margin of error and all that, but I can't imagine this would be worded this way if the polls were reversed

the other thing, "nearly 70 percent of those surveyed said the choice of Edwards won't make a difference in the race for the White House"? I also understand subtraction, so that means that over 30 percent of those surveyed, almost a third, say that he will or may make a difference ... with only 6 percent undecided, over 30 percent sounds like a big number for that question

[blm]

1. They are only using LIKELY VOTERS numbers now.

LIKELY VOTERS means those who have a track record of voting. That leaves out newly registered voters and registered voters who have not recently voted but plan to now. That list always skews to Republicans, but the media fails to report that info.