somebody with more math brains than me (and that is a lot of you) if 38% were demos and 25% were pubs and kerry had a 7% lead, what would the "lead" be if the poll was 50-50 demos and pubs?
This skewing of the base is common in many polls, it is based on the assumption that there are so many more registered dems, or that dems vote with higher frequency. So a 50 50 electorate is not considered reality. The problem is that the percents that are selected are easy to "push" and may not be real either.
No one would question that there is some ratio of dems to repubs that would give a poll representative of an actual election, but you can expect the LA times to try for good news for their readers. Zogby does this too, and close to the election he seems to get the right percentage. Now, when its time to try to push undecided voters into one or the other camp, honest polls are hard to find. For that I watch FOX.
You'd have to know what percentage of each party voted with their own party's candidate. The biggest shift would occur if each party's voters voted for their guy, and independants were the only swing voters.
Bush is supposedly leading independents by 3%. I've seen where Rats and Pubbies were about equal in voter registration nationally, so if the same number of Rats and Pubbies were crossing party lines (a big presumption), Bush would be up about 1%. Bush's people complained that the poll was leaning about 10 to 12 percent too democratic.
Using just the above figures in the calculations,the result would be a 4.6% lead for the dimrats.