After the way the media is latching onto Kerry after the debates I was prepared for Bush to slip some, but something's screwy with the results:
The poll says they interviewed 391 Republicans, 300 Democrats and 270 independents.
They then say that Bush has the support of 93% of the Republicans (364 respondants), 7% of the Democrats (21 respondants), and 39% of the independents (105 respondants). Add them together and Bush ends up with 490 of the possible 961 votes, or 51% of the total respondants.
Kerry, meanwhile, ends up with 87% of his party's vote, or 261 respondants, he gets 4% of the Republican vote, or 16 respondants, and 45% of the independent vote, or 122 votes for that group. Add them all up and Kerry ends up with the support of 399 of those polled or just under 42% of the total number of respondants.
So if you apply the total percentages of the party vote that each candidate receives in this poll and multiple it against the number of respondants by party affiliation then Bush is actually up 50%-42%.
I don't think these numbers are right
"So if you apply the total percentages of the party vote that each candidate receives in this poll and multiple it against the number of respondants by party affiliation then Bush is actually up 50%-42%."
Hhhmmmm...this is odd...I'm not good with numbers- but this looks like a worthwhile question to send Newsweek, pronto.
Looks like they "corrected" for party affiliation. They weighted the published results according to some prediction about party turnout. Rerun your numbers weighted for the party affiliation reported to exit polls in the 2000 election and I bet you get the numbers published by Newsweek.