If you look at the limited internals of the polls by clicking on each link, you can see the breakdown for D/R/I and M/F.
They seemed to somewhat closely follow the 2000 exit poll data for the M/F breakdown (with a point or two variance).
So the gender weight is not much different than 2000.
I also think they weighted the poll to account for political party affiliation, but haven't had a chance to compare the two numbers.
What SUSA, Mason-Dixon, and Gallup have been doing is not weighing and letting the chips fall where they may.
From The Kerry Spot:
UPDATE: Kerry Spot reader Matthew points out, "If you click on each state of this poll individually, you will see how distorted the polls are. For example, look at Illinois. The sampling is 29% Republicans, and 42% Democrats. And Bush only trails by 6? Same with PA. 42% are Republicans and 48% are Democrats. The poll also is weighted more towards women than men."
Yup, as I look over this, I see Florida is 55 percent women, 45 percent men. Ohio, 54 percent women, 46 percent men.
Also Kerry Spot reader Michael notes that this ARG poll has essentially the same results as the ARG poll completed 9/1, a 47-47 tie.
__________________
As you can see, 2000 data is not adequate for weighting since 9/11, too much has changed.