[EM] Approval fraud prevention (was Re: A conversation with an English woman about IRV)

[Jameson Quinn]

2011/5/4 ⸘Ŭalabio‽ <Walabio at macosx.com>

>        2011-05-0416:39:26Z, “Jameson Quinn” <Jameson.Quinn at Gmail.Com>:
>
> >       Unfortunately, there is no task that you can manually ask the
> voters to do, which won't lead to unacceptably high levels of spoiled
> ballots. My ballot doesn't count because I didn't vote against Wingnut
> Moonbat? Or because I didn't count up my approvals correctly? Once I failed
> to win a competition because I incorrectly counted and self-reported my
> score. Since it was a math competition, perhaps that was just. But voting is
> not a math competition; spoiled votes should be avoided.
>
>         Since the ballots are both human/machine-readable, the voter should
> run them through a ballotvalidator (an optical scanner).  The
> ballotvalidator would find errors such as:
>
>        [+]      [-]     Candidate       P
>
>        And
>
>        [*]      [*]     Candidate       W
>
>        And tell the voter.  The voter would then fix the errors and
> revalidate the ballot.  If a voter spoils a ballot, it goes into a
> privacyenvelope and goes into a locked transparent spoiled ballotbox.  When
> the ballot is perfect, it goes into an opaque privacyenvelope.  The voter
> puts the ballot into a transparent locked cast ballotbox:
>
>        The transparent locked cast ballotbox, as well as the spoiled locked
> stransparent ballotbox, start off empty and is always keep viewable to the
> public and on a live webcam.  After the polls close, pollworkers unlock the
> cast locked transparent ballotbox and count, at the precinct, count the
> votes in front of anyone who wants to watch and on live webcam.
>
>        The pollworkers count the votes both manually and using optical
> scan.  If the totals differ, different humans and different optical scanners
> recount the votes.  The process iterates until both the humans and the
> optical scanners arrive at the same total.
>
>        Because of the ballotvalidators, the locked transparent cast
> ballotbox should have no spoiled ballots in it.  A spoiled ballot in the
> locked cast transparent ballotbox should lead to an immediate investigation
> into fraud.
>
>        The government should archive the cast ballots, the blank ballots,
> the spoiled ballots, and the video from the webcams.  That is how we should
> do it.


You have come up with one of many possible systems which would be
sufficiently secure. The security of the system you propose is not
significantly affected by the fact that a voter is required to explicitly
approve or non-approve each candidate. Yes, an automated check for spoiled
ballots is a good thing. But, given sufficient other security measures such
as you suggest, ballot security is not significantly impacted by the choice
of system, so should be discussed separately.

In fact, under approval, with a much lower intrinsic level of accidentally
spoiled ballots, there would be less chaff in the air when it came to
looking for evidence of fraud. Thus you could just as easily — and just as
irrelevantly — argue that it is more fraud-secure than plurality. In my
opinion, though, even seriously entertaining these arguments is a rhetorical
mistake.

JQ
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20110505/063287e5/attachment-0003.htm>