Re: "Unbelievers are allowed to depart" Now wholesale/retail fraud

From: Richard C. Johnson <dick_at_iwwco_dot_com>
Date: Mon May 23 2005 - 16:28:27 CDT

There is at least one other compromise option in the publishing of ballots. The Bureau of the Census puts out what used to be called Public Use Samples at fairly high sample fractions of the actual census forms being tabulated every 10 years. The catch is that the Bureau takes care NOT to publish anything which would serve to identify any individual. No identifiers are published in these samples. Writing in something weird would simply mean that such entries would be gathered into an "other" category. But the Bureau actually uses statistical algorithms to spot cells with insufficient numbers to protect confidentiality, and even "other" would lead to geographical aggregation if it were not numerous enough.
With ballots in a given election, one could publish a very large fraction of the total, scanning for unusual or sparse cell populations, and aggregate anything that wasn't sufficiently populated to hide individuals ballot choices.
Such a large sample of the total voted ballots would do far better than exit interviews to allow public supervision of the voting process. Large precincts would almost always be auditable and smaller precincts would be auditable in the aggregate.
This compromise can preserve the secret ballot, which, like the secrets in the census form, are part of our political heritage and tradition. A Public Use Sample of the ballots will do far more than exit polls to provide public oversight of the voting process. But...the secret ballot must be preserved.
-- Dick
Richard C. Johnson, Ph.D.

Ron Crane <> wrote:
I think it's very difficult to formulate a quantitative measure for any
of this, which is why there's so much debate on it. But at least we can
state the tradeoffs. As I see them, prompt publication of all ballots
would permit the public to audit each precinct's totals by retabulating
all the published ballots. However, it would also open elections to
vote selling or coercion.

The ability to retabulate would deter frauds that mis-state the totals,
but, without more (e.g., access to sign-in books) it would not deter
box stuffing. It would not deter casting frauds (e.g., the casting
station makes it easy for the voter to vote as it wishes, and hard to
vote otherwise), since these frauds leave no traces for retabulation to
detect. Realistically it would deter pre-tabulation ballot substitution
only to the extent that each voter individually examines her own ballot
during retabulation.

The likely extent of vote selling and coercion is difficult to
quantify. Some of the risks are that voters will mark their ballots in
various ways to allow buyers or coercers to verify the desired votes
(e.g., by writing in a coded name for a "don't care" race), or that
they'll use (or be forced to use) ballot numbers or ids for the same
purpose. The coded name risk applies to existing mail-in ballot schemes
that allow write-ins; the ballot number or id risk does not unless the
mail-in ballots also are available for public inspection. Mail-in
ballots are susceptible to buying or coercion at the point of casting,
but this fraud requires a buyer or coercer to be present as each ballot
is sealed into its envelope and mailed. Public disclosure of ballots
extends this risk by permitting anyone to peruse the ballots at her
leisure; thus, a single buyer or coercer could verify whether her
sellers or victims have complied with their sides of the deal.

Buying and coercion risks can affect the general public, via crooked
bosses, strongarm political machines, and the like. There are also
conditions under which they could affect voters who favor unpopular
parties or candidates more than voters who favor popular parties or
candidates, particularly if write-ins are not allowed and ballot ids
are not provided; it's easier to correlate the votes of a small number
of voters than of a large number.

Got more risks? Please add them. Then maybe we can try putting numbers
on them.


On May 23, 2005, at 9:40 AM, Edmund R. Kennedy wrote:

> Hello Clint,
> I was just looking over my letter and I realize that
> I'm not being particularly clear. Everyone has been
> arguing this issue on a qualatative basis. I think
> you could be persuasive if you can muster up a
> quantatative argument. HTH.
> Thanks, Ed Kennedy
> --- "Edmund R. Kennedy" wrote:
>> Date: Mon, 23 May 2005 09:27:44 -0700 (PDT)
>> From: "Edmund R. Kennedy"
>> Subject: RE: [OVC-discuss] "Unbelievers are allowed
>> to depart" Now
>> wholesale/retail fraud
>> To: Open Voting Consortium discussion list
>> Hello Clint:
>> I'm not completely up to speed on this discussion
>> but
>> it sounds like your are stuck on the issue of what
>> we've come to call wholesale versus retail fraud.
>> Retail fraud has to do with a voter by voter
>> subversion along the lines of voter coercion and
>> vote
>> buying, while wholesale fraud has more to do with
>> ballot box tampering kind of along the lines you are
>> suggesting. Is this correct? I know that a number
>> of
>> us went multiple rounds on which one was more
>> important. After a great deal of weeping, wailing,
>> tearing of hair and gnashing of teeth; the light
>> dawned on us that both issues are equally important.
>> Perhaps we need to do a little more formal risk
>> analysis on this issue? I know I'd like to see you
>> put up some sort of cost/benefit analysis of voter
>> receipts. Intuitively and experientially we
>> generally
>> believe voter receipts are a problem. However,
>> we're
>> all reasonable people here and I consider the issue
>> open, -- to a very persuasive analysis. So far, I
>> haven't seen the actual benefits to the voter or
>> even
>> how a voter would be able to audit their vote
>> through
>> the entire process. The paper ballot generated
>> seems
>> like an adequate system without the downside. So, I
>> look forward to a risk or cost/benefit analysis from
>> you for voter receipts.
>> As always, I'm only speaking for myself here but I
>> understand this an issue that fate has thrust upon
>> you
>> so I'm personally willing to give you the type of
>> respectful hearing that I'd like to receive.
>> Thanks, Ed Kennedy
>> --- wrote:
>>> Kind of testy. Have you considered that possibly
>>> others may have
>>> opinions that you will not be able to unilaterally
>>> dictate. If you are
>>> the final word on this matter then I will depart.
>> If
>>> you are not then
>>> expect my resistance to anything which gives vote
>>> fraud comfort.
>>> Clint
>>>> -------- Original Message --------
>>>> Subject: [OVC-discuss] "Unbelievers are allowed
>> to
>>> depart"
>>>> From: Edward Cherlin
>>>> Date: Mon, May 23, 2005 5:04 am
>>>> To:
>>>> There have been a number of people on this list
>>> (and several
>>>> others that I have received) who are remarkably
>>> resistant to
>>>> facts. For example, it does no good to refuse to
>>> recognize that
>>>> the law is as it is, however much we may wish it
>>> to be
>>>> different. We can argue that the law *should* be
>>> different if we
>>>> are also willing to discuss whether we are
>> likely
>>> to be able to
>>>> convince people to change it. However, insisting
>>> that settled
>>>> principles of law and human rights are invalid
>>> gets us worse
>>>> than nowhere.
>>>> One of the leading principles of democracy is
>> the
>>> secret ballot.
>>>> If you don't believe in the secret ballot, you
>>> have no
>>>> credibility in voting matters. This is not our
>>> private opinion
>>>> on this list. It is established fact nationwide
>>> and beyond. If
>>>> you would rather argue for a non-secret ballot
>>> than work on
>>>> securing the secret ballot, we are not the
>> people
>>> you want to
>>>> work with.
>>>> We can argue for laws permitting, encouraging,
>> and
>>> even requiring
>>>> Free/Open Source Software and paper ballots with
>>> an electronic
>>>> audit trail. We cannot argue for publishing
>>> ballots in a way
>>>> that allows anybody to determine routinely whose
>>> ballots they
>>>> are. This is a thorny issue, because publication
>>> of all ballots
>>>> does allow some voters to be identified, while
>> not
>>> publishing
>>>> them prevents independent analysis and detection
>>> of fraud.
>>>> Denying either of these plain facts is not
>>> helpful.
>>>> What we need is more facts. Perhaps someday we
>> can
>>> secure
>>>> ballots, using open software and a robust audit
>>> capability, to
>>>> the point where the results are not routinely
>>> called into
>>>> question, without having to provide for routine
>>> outside
>>>> analysis. Perhaps this is not that day, and we
>>> have to take some
>>>> risks with privacy in order to secure
>>> accountability. And then,
>>>> perhaps, we will have to choose where we come
>> down
>>> along that
>>>> axis, and explain to others why that is the best
>>> choice in the
>>>> circumstances. But we certainly cannot just
>>> dismiss either
>>>> privacy or accountability.
>>>> I have faced this issue many times in my work on
>>> spam. There are
>>>> people who believe that spam should be legal,
>> and
>>> should be
>>>> fought by technical rather than legal means.
>> They
>>> can believe
>>>> what they like, but we didn't let them come in
>> and
>>> interfere in
>>>> the workings of a group devoted to creating laws
>>> against spam
>>>> and getting them passed.
>>>> If any of you think the rest of us at OVC (or
>> any
>>> of us in
>>>> particular) are blockheads for not seeing the
>>> brilliance and
>>>> inevitability of your ideas, you have your
>> reward.
>>> Please don't
>>>> try to foist it on the rest of us.
>>>> --
>>>> Edward Cherlin
>>>> Generalist & activist--Linux, languages,
>> literacy
>>> and more
>>>> "A knot! Oh, do let me help to undo it!"
>>>> --Alice in Wonderland
>>>> _______________________________________________
>>>> OVC discuss mailing lists
>>>> Send requests to subscribe or unsubscribe to
>>> _______________________________________________
>>> OVC discuss mailing lists
>>> Send requests to subscribe or unsubscribe to
>> --
>> 10777 Bendigo Cove
>> San Diego, CA 92126-2510
>> 858-578-8842
>> Work for the common good.
>> My profile:
>> I blog now and then at:
>> _______________________________________________
>> OVC discuss mailing lists
>> Send requests to subscribe or unsubscribe to
> --
> 10777 Bendigo Cove
> San Diego, CA 92126-2510
> 858-578-8842
> Work for the common good.
> My profile:
> I blog now and then at:
> _______________________________________________
> OVC discuss mailing lists
> Send requests to subscribe or unsubscribe to

OVC discuss mailing lists
Send requests to subscribe or unsubscribe to

OVC discuss mailing lists
Send requests to subscribe or unsubscribe to
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
Received on Tue May 31 23:17:46 2005

This archive was generated by hypermail 2.1.8 : Tue May 31 2005 - 23:17:53 CDT