[ride the whirlwind]

Thanks. I've been hearing, leading up to this election, that polls of likely voters are more reliable than those of registered voters. Last night, I read something by the host of another site (democRat site) attempting to explain why he feels that registered is better. Maybe he just didn't like that 8 point lead by GWB!
It's too bad that gallup used more Rs than Ds in their sample though. Can you think of any good reason for doing this?

[dvwjr]

Gallup does not re-weight their sample for partisan political party affiliation. Therefore the R/D/I percentages 'float'; other polling organization such as the American Research Group and Zogby use some form or re-weighting based on the year 2000 VSN presidential exit polls to 'fix' their raw data. That is why I created the far right column in the master data table of the Gallup partisan preference internals to show what would happen if Gallup went from being a polling organization that 'fixes' rather than to 'floats' its polling sample...

We will see by the year 2004 exit polls conducted by the successor organization to the now defunct VNS show which polling philosophy was correct.


dvwjr