[Date Prev][Date Next] [Chronological] [Thread] [Top]

Fw: Logging Scanner Jams (was: Audit trail improvement suggestions)



Title:
 
----- Original Message -----
Sent: Friday, December 10, 1999 9:48 AM
Subject: Re: Logging Scanner Jams (was: Audit trail improvement suggestions)

Count is off by 4 votes.
Administrator calls Inspector and asks:
    What happened yesterday?  How many Jams did you have?
Inspector says I remember 1 jam around 10:00am but I went to lunch between 2-and 3 in the afternoon.
Admin checks log and notes that there were 3 jams between 2 and 3 while inspector was gone.
Now calls other judges and asks them about jams, and who handled, it.
Finds person who handled jams and asks them what procedure was.  Determines they just refed them when they were already counted (or vice versa).
 
Again, this is just a "reconstructive" tool.  The statistical approach is ok.  Right now we simply say, "well maybe there were some jams" and we are never able to focus on what the problem was and don't know if its a training issue or the AccuVote passing ballots.  Now if we don't want any log because we're actually "passing" ballots, then I'd understand, but the statistical approach to logging jams would be helpful (especially if we have real estate issues with the logging approach).
----- Original Message -----
From: Ken Clark
Sent: Friday, December 10, 1999 9:21 AM
Subject: RE: Logging Scanner Jams (was: Audit trail improvement suggestions)

> > If the ballot jam had happened at 2:23:52PM instead of
> 1:13:24PM, what are
> > you going to conclude as the election administrator?
> >
> >
> Sometimes they actually call the inspector and pollworkers to ask what
> happened.  The more information they have the better they can discuss with
> the pollworkers.  As I said, the statistical approach would help them
> explain to others, the log approach would help them reconstruct the
> explanation.


I don't mean to be pedantic here, and certainly don't want to pick on Steve, but this relates to do with other log requests we have had, so I will take one more crack at this. 

What, exactly, are you going to "reconstruct"?  Give me a realistic scenario where an election administrator, armed with the times that the ballots jammed, could conclude anything interesting about the election that they couldn't conclude armed with the stat.

Ken