Skip to main content

View Diary: Revote in Ohio -- More Votes Than Voters Recorded on Diebold machines (199 comments)

Comment Preferences

  •  actually, his name is Freeman (none)
    (I know, you were quoting the previous poster.)

    There's not much wrong with citing Freeman on the point that the 2004 exit polls were off by large margins. Yeah, he goosed the p values, but anyone who takes p values very literally is already in deep, deep trouble; the number of zeroes should be a side issue. (Citing Freeman will not work so well if one tosses in 2000 and 2002. The 2002 U.S. exit polls were basically dead on arrival -- the data system melted down -- but I don't think we know whether they would have been biased if the system had worked. The 2000 exit polls -- at least the presidential ones -- were mostly much more accurate than the 2004 exit polls, despite Florida. The 1992 exit polls had problems.)

    But Freeman jumps the shark IMHO when he depicts  exit polls as normally bias-free and accurate within sampling error. It just ain't so, but a lot of people think it is, and arguably Freeman deserves a lot of the credit for that. I'm not sure.

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site