Reply #12: Great God A-mighty. The highlight of this fantastic paper comes at the

[bleever]

end where they declare their intention to create and analyze a database of precinct-level data for the 2004 election in order to do a mathematical analysis of it.

Other highlights for me:

HIGHER response rates to exit polls in Bush-leaning precincts, which Edison-Mitofsky fails to point out is demonstrated by their own data, and which blows the whole responder bias issue away;

E-M claims that there is no skew related to type of machine used, but fails to back it up, while the data suggests just the opposite;

E-M go out of their way to say it must be their fault, and their unsubstantiated hypotheses about why they got it wrong make it appear that their own analysis has interests other than the purely scientific and mathematical.

Also, it feels REALLY good to read someone saying it straight out:

Final polls showed Kerry won by 3%

without talking about "adjusted polls" or how they "initially" showed him winning.


Thank you TIA for everything you've done to get us to this place.


:thumbsup: