Y romerillo!

Can you post a link to the story you quote instead (or in addition to) the PDF of the study

Henry:

The AP apparently did not find the tightening of the race to be worthy of mention. They do have a nice long article, however, detailing McCain's falling favorability numbers. Otherwise, no mention is made in the AP article about the head-to-head numbers.

I just posted about that HERE

Where do you get this conclusion:

"which sampled 7% more democrats than republicans on the premise that more democrats will turn out to vote this year than republicans"

According to what is being reported about the poll,
the data comes from 873 Democrats and 650 Republicans.

"Interview dates: October 3, 2008 – October 13, 2008 Interviews: 1,769 adults; 1,528 registered voters: 873 Democrats; 650 Republicans."

That's the number of respondents. There may weighting of those numbers to correspond with known voter registration trends etc.

For example forget it's a political poll let's say you want to poll adults and have the right mix of men and women. Women currently outnumber men (of the top of my head it's like 51% to 49%. But when you survey 1000 people you might get 550 women and 450 men because the women tended to answer the questions more. Then the data has to be totaled and weighted to the known proportions of the universe being measured.

The pdf does not describe any weighting but you just can't assume that the raw numbers were used just as is.

Another thing to note. No independents. That makes this poll useless as a predictive poll.

That poll is not a conventional poll like Gallup, Rasmussen or Zogby (the most reputable firms). It's a trend poll meant to measure the trends favorability over several waves. I wouldn't put to much stock into it.