More election nonsense

Democratic Rep. John Conyers is trying to get the media to turn over its raw exit poll data from the November 2004 presidential election. His rationale? According to this report, "so that any discrepancies between the data and the certified election results can be investigated." Conyers' letter claims that "[w]ithout the raw data, the committee will be severely handicapped in its efforts to show the need for serious election reform in the United States."

It is difficult to comprehend how anyone, especially the ranking Democrat on the House Judiciary Committee, can believe that exit poll results are more accurate than the actual vote count! While the exit poll data may be slightly different than tha pre-election polls we all watched anxiously, I recall that those pre-election polls all included a "margin of error." The margin of error, presumably, is to account for discrepancies in the sample, or inaccurate answers given to the poll-takers. There is no reason to believe that a poll of voters leaving the pools would not be subject to the same influences (e.g., over-sampling of one candidate's voters, false answers given by the voters, etc.) that the pre-election polls were, and would not also have a margin of error. What we do know, though, is the actual vote count leaves no room for error. It gives hard numbers on which the electors are chosen.

Why would hard numbers be less accurate than polls that include a margin of error? The election is over. To borrow a phrase, MoveOn.