Punch ballots not so inaccurate?

[Frank Cross]

Mark Scarberry notes that the error rate for punch card ballots checked
was only 0.1%, rather than the 0.3% estimate, but higher than 0.03%
estimate for scantron ballots.  This does suggest that the earlier statements
about error rate may be wrong.

But it simultaneously says something much more important!  It pretty
clearly indicates that the hand recount is not a partisan game to dig up
Gore votes.
The counters are finding fewer new Gore votes than would be anticipated,
which would seem to add to the validity of the hand recount process.


At 10:53 PM 11/14/2000 -0800, you wrote:
> The NY Times is reporting that Dade County officials decided not go through
>with a full manual recount because a manual recount of 5,870 ballots
>resulted in a change of only 6 votes.
>http://nytimes.com/2000/11/15/politics/15DADE.html. That is about a one
>tenth of one percent difference, far less than some have said the inaccuracy
>is with machine counting of punch ballots (even assuming all of the manual
>ballot decisions were correct). It seems to be in the same range that others
>have said result from optical scanning of "fill in the bubble" ballots.
>There seemed to be a similar result in Broward County when they did a sample
>hand recount.
>
>Should this perhaps (1) lower the temperature of the argument about the need
>for manual recounting, and (2) cast some doubt on the whether large changes
>from hand recounting in Palm Beach County reflect accurate, unbiased
>counting? The NY Times story also reports that the 3 Dade County precincts
>hand recounted contain predominantly African-American, Hispanic, and elderly
>Jewish voters. Maybe the arguments that lower socioeconomic groups are
>disadvantaged by punch ballot systems are thus also less persuasive than we
>thought.
>
>Mark S. Scarberry
>Pepperdine Univ. School of Law
>
Frank Cross
Herbert D. Kelleher Centennial Professor of Business Law
CBA 5.202
University of Texas at Austin
Austin, TX 78712