Skip to main content

View Diary: IN-09: Dems demand a recount (131 comments)

Comment Preferences

  •  exit polls (4.00)
    are compleatly different then the polls leading up to the election. They are generally VERY reliable
    •  Exactly: exit polls have always been MUCH better. (4.00)
      Pre-election polling relies on calls to people's houses.   There are plenty of sources of error:

      a) some people don't like being bothered at dinner

      b) what is the relation between people who are called versus those that actually vote?

      c) what is the relation between random phone numbers and actual distribution of precincts across demographic groups?

      Exit polls eliminate the primary sources of systematic error:  sampling bias in choosing who to ask (as it is a scientifically generated sample across precincts), and bias in who actually shows up and votes.

      The only remaining bias is in who may choose to answer the pollsters.  But even this is different from pre-election polling.  In exit polls, people just voted, and had voting on their mind and not the Broncos game or whatever.  And usually there's personal bias in favor of who they voted for and people are proud of it.   In other countries they have proven to be very accurate as well.  

      The issue with exit polls is (was!) that they were so accurate even with incomplete samples that by publicizing their results they may influence who shows up to vote later while voting time is still left.

      But for some odd reason, despite great attention to these issues and the strictest scrutiny (from 2004 where they called Florida early, and incorrectly because of the tiny margin of difference---maybe) somehow the exit polls were consistently wrong and in chimpoleon's favor.


      Also, the exit polls in 2000 weren't really wrong in Florida.  The exit polls measured who people thought they voted for.  Some ended up marking votes for Buchanan based on a poor ballot design, and in other precincts, which just happened to be frequently minority, the rate of "spoiled" ballots was much higher than the norm.   There's no question that if one were to consider the true will of all those people who were eligible to vote in Florida and actually did show up and vote, Gore won.

      •  Another reason (none)
        exit polls are accurate is that there is no "likely voter" uncertainty thrown into the mix. Exit polls are often used in emergent democracies by outside election monitors to detect fraud.

        As for FL, all one need to do is look at the double marked ballots (because some idiot ballot designer  didn't know the difference between "write in candidate here" and "write-in candidate here") and one will see a change in the outcome. Exit polls were on the money, and what was the response? They dropped exit polling from the 2002 election and Max Cleland loses Georgia, a state served by Diebold. In case you can't tell, I'm still pissed.

        The greatest blessing bestowed on a people is the absence of ignorance in public office. - Confucius

        by cavanaghjam on Mon Nov 15, 2004 at 11:52:41 PM PST

        [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site