My admittedly limited understanding of this process is that some of the
older technology like punch cards are unreliable but not predictable. In
other words, if you do a recount using the same set of ballots and the same
machine to count them, you will likely get a different count but it is not
necessarily more accurate. Unless the machine has more trouble reading some
patterns than others, you would expect multiple runs to cluster around a
mean that would approximate the true count. I did notice on the Palm Beach
ballot, the punch holes were right next to each other for the various
candidates (i.e., 4,5,6,7,etc.). In my precinct we also use punch cards but
the choices are spread further apart (e.g. Bush might be 5 and Gore 10). Is
it possible that having them so close could have led to some choices
"bleeding" over into the next either through a punch being too large or a
slight misalignment when it is run through the counting mechanism? This
might explain the high number of ballots being read as having two punches.
What do you think?
Michael B. Quanty, Ph.D.
Psychology Professor
Senior Institutional Researcher
Thomas Nelson Community College
PO Box 9407
Hampton, VA 23670
Phone: 757.825.3500
Fax: 757.825.3807
-----Original Message-----
From: Paul C. Smith [mailto:[EMAIL PROTECTED]]
Sent: Friday, November 10, 2000 8:56 AM
To: 'James D.Dougan'; [EMAIL PROTECTED]
Subject: RE: Recounts and Statistics
To add to that question, it should be noted that the "swing" resulting from
the recount was not uniform across counties. Here are the numbers (change in
votes from count1 --> count2, with one county yet to report):
Towards Gore: Bush:
Alachua + 65 + 62
Broward + 43 + 0
Dade + 62 + 0
Duval +184 + 0
Gadsden +170 + 0
Hillsborough + 0 + 47
Lake + 0 + 47
Martin + 0 +106
Nassau - 73 -124
Okaloosa + 0 + 50
Orange +105 + 0
Palm Beach +751 +108
Pinellas +417 - 61
Polk +216 + 79
St Johns + 0 + 49
These look to me to be consistent with the notion that there were
regular
processes causing the incorrect counts, and that those processes tended to
operate against Gore's voters for some reason _other than_ the fact that
they were Gore's voters (in other words, that they acted directly on some
kind of demographic or geographic basis and only indirectly on the basis of
voting decision). At the same time, unless you believe that there was fraud
on both sides, the fraud hypothesis looks pretty weak (look at Nassau
County, for example, or the infamous Palm Beach, or Polk).
Paul Smith
Alverno College
Milwaukee
