RE: Contract: need Hart support for random sampling

[Paul Walmsley <paul@xxxxxxxxxxx>]

On Tue, 27 Apr 2004, alkolwicz wrote:

> Paul,
> 
> Will you take the lead to ensure that the sampling process is not limited to
> "checking the addition" of the wrong set of votes?  
> 
> Equally important to be sampled are the qualification of ballots and the
> interpretation of votes.
> 
> 1.  Eligible ballots go uncounted, and ineligible ballots get counted. 
> 
> 2.  The interpretation of what constitutes a vote is not reliable.  As I
> understood the Hart system, only those ballots that WERE initially rejected
> by the equipment will be reevaluated for a recount.
> 
> I am available to discuss further.

Hi Al,

there are many open issues with the stat sample hand count at this point
that will need to be resolved in the upcoming weeks.  in general, I would
encourage everyone who is interested in this issue to pay attention to it
by reading the memos, watching the recent County Commissioners meeting
recordings, etc.  

I would hope that any formal proposals for stat sampling would be posted
to the cvv-discuss list.  most of the discussion so far has been informal 
and mostly focused on figuring out ways to get the state to approve the 
general process.

I can say, regarding your point #2 above, that merely hand-counting a 
sample of the residual ballots would be unacceptable.  If the point 
of doing the statistically-significant sample hand count is to verify that 
the counting machines are counting ballots correctly, we'd need to count 
plenty of non-residual ballots also.  Perhaps you are referring to the dup 
board process for the residual ballots?


- Paul