Reply #52: A variety of errors almost never add up in the same direction [View All]

Also, from their methodology webpage they say "An interviewer gives every nth voter exiting the polling place a questionnaire to complete...Participation is voluntary and anonymous."

So, the only real possibility of bias is that Bush voters were far more likely to turn down the exit poll questionnaire than Kerry voters. There has been a claim that fundamentalists did just this, and they turned out in numbers greater than previous elections. But, if they turned down the exit polls (and thus weren't counted) how could anyone know they voted in greater numbers than usual, since it is the exit poll questionnaire itself that gives us information on the religious affiliation of the voters. The two claims can't be true at the same time.

The methodology also states that "interviewing starts when the polls open in the morning, continues throughout the day until an hour before they close at night." Democratic voters have historically been more likely to vote late (the don't get time off from work), so if anything this should bias the polls against Kerry.

The whole thing just doesn't add up.