At 10:36 PM 11/8/2004 -0800 Doug Pensinger wrote:
>> We're looking at that issue right now, actually. Two
>> complementary initial guesses, but they're both just
>> that. Exit polls tend to be taken in
>> higher-population districts.
>
>Well how stupid is that? Why in the hell would they bias their polls in
>such a way as to make them useless? Do they only take regular polls in
>higher-population districts? That makes no sense at all to me.
You have to understand that polls are a sample, and that sample is based
upon reasonable guesses of what the population looks like. In this case,
precincts are selected for exit polling based upon a weighted sample....
i.e. each precinct is given a weighted chance of selection for exit polling
based on the likelihood of a voter voting in that precinct. If those
weights were based upon turnout in the last election ( a reasonable
assumption ), and if the Republican turnout machine greatly altered turnout
patterns in rural districts, then those weights would be flawed and produce
a systematic error in the results.
That is one guess.
JDG
________________________________________________________
John D. Giorgis - [EMAIL PROTECTED]
"We have one country, one Constitution, and one future that binds us."
-George W. Bush, 11/3/2004
________________________________________________________
_______________________________________________
http://www.mccmedia.com/mailman/listinfo/brin-l
