Write-In Only Voting

[Brian Langstraat]

Are there any voting systems that only allow write-in voting?  

First-past-the-post and approval voting could have special cases of write-in only voting.

Are there any organizations that only allow write-in voting?

One example may be the Electoral College of the United States where electors may write the names of anyone for president and vice-president.

What would be the advantages and disadvantages of write-in only voting?

Only votes from "informed" voters, but more difficult to count votes.

Could reduce influence parties or could increase extreme partisanship.

[Warren D Smith]

I don't understand how write-in voting can be secured
against vote-selling and vote-coercion.
Also, a worry which might be surmountable is simply fake voting,
but overcoming this might be only at the risk of destroying
voter-anonymity / privacy.



--
Warren D. Smith
http://RangeVoting.org <-- add your endorsement (by clicking
"endorse" as 1st step)

[Brian Langstraat]

Warren,

I appreciate your response and your work on http://RangeVoting.org is the reason I joined this Google Group.

How is any voting system secured against vote-selling, vote-coercion, or fake voting?

I would assume that a private ballot would be as secure as with any other voting method, especially those that already have a write-in option.  Handwriting could be associated with a specific person, but typing or bubble-filling letters could eliminate that issue.  I assume that fake voting is defined as modifying a voter's ballot after they vote.  Voters could mark ballots in such a way as to prevent tempering with written lines (similar to checks) or filling bubbles to denote unused spaces.

Range Voting could have a special case of write-in only voting.  Voters could give high scores to their favorite candidates and low scores to despised candidates.  "No Opinion" would be assumed for all other candidates.

[William Waugh]

 "No Opinion" would be assumed for all other candidates.

That raises the racist problem, or the torturer problem, or the Hitler problem, or however you want to express it. If you use the average and you disrespect the voter's right to oppose all candidates that voter doesn't know anything about, 100 people could support a racist and get her in office. Dr. Smith will likely say that some quota rule or fake ballots resolves that, but I still say it's my right as a voter to oppose those I don't know about. I have suggested as a compromise that the ballot should allow each voter to specify a score for "all others". That way, Dr. Smith could defer to others as he apparently wants to do, and people of my ilk could set ourselves against those we don't know enough about them to support them.

William Waugh

[Warren D Smith]

On 1/10/17, Brian Langstraat <langstra...@gmail.com> wrote:
> Warren,
>
> I appreciate your response and your work on http://RangeVoting.org is the
> reason I joined this Google Group.
>
> How is any voting system secured against vote-selling, vote-coercion, or
> fake voting?

--if you vote at a polling place secret-ballot, then you
cannot sell your vote because the buyer does not KNOW
your vote because it is secret. Ditto for preventing coercion.
If you vote by mail, buyer or coercer can simply watch you
filling out ballot and mailing it.

> I would assume that a private ballot would be as secure as with any other
> voting method, especially those that already have a write-in option.

--sounds like a wrong assumption!?

> Handwriting could be associated with a specific person, but typing or
> bubble-filling letters could eliminate that issue. I assume that fake
> voting is defined as modifying a voter's ballot after they vote.

--one kind of fake voting:
how about, I send in 50000 fake ballots,
thus causing Trump to win the election?
(This may or may not be preventable.)
And it is known that scams of this kind have been organized,
e.g. in Miami FL; this is not just a hypothetical.

Also, my father did vote by mail (absentee ballot) and his
ballot was never counted, because the counters
knew my father's name and they did not like him.
Or perhaps for some other reason.
(They claimed he'd never signed it.
Which was a false claim. He was witnessed signing it.)

[Brian Langstraat]

William,

I completely agree.  The quota rule and fake ballots have seemed like an undemocratic way to deal with the extremist possibility in Range Voting.  Having the each voter specify a score for "all others" (including "No Opinion") is a compromise that I should have included in my original post.

[Brian Langstraat]

Warren,

Lets assume that their are reasonable laws and methods to prevent voter fraud.

Are there any other advantages and disadvantages of write-in only voting?

What is wrong with the assumption that a secret ballot would be as secure as any other voting method that contains a write-in option?  Why would removing partisan candidates from a ballot make it less secure?

Mail-in ballots can be convenient, but have inherit risks.  Voters need to decide whether the convenience is worth the risk.  As an Iowan, I could try to mail-in one ballot AND vote in person with another ballot (http://www.desmoinesregister.com/story/news/crime-and-courts/2016/10/28/voter-fraud-suspect-arrested-des-moines/92892042/).

[William Waugh]

Let's split this discussion into two dimensions:

- Dimension of ballot with candidate's names printed on it vs. requiring the voter to fill in the names of the candidates they wish to score (but allowing a blanket "all others" in addition to the names);

- Dimension of permitting ballots to be mailed in vs. only filled out in the voting booth.

Unless I am missing something, the risk of coercion comes from the second dimension only.

[Brian Langstraat]

William,

Lets focus on the first dimension (Write-In Only Voting), since the second dimension (Mail-In/Fraud) are sidetracks.  Looking back, I think Warren was confusing Write-In with Mail-In.  Write-In is where there is an empty space on the ballot where each voter may write a candidate's name.  Personally, I have never mailed-in a ballot, but I have been tempted to write-in a candidate.

Range Write-In Only Voting with each ballot containing many range selections for blank spaces and a range selection for "All Other Candidates" may be a surprisingly decent way to vote in single-member elections.  I am not sure whether this voting system could be simulated, or has already been used somewhere.

Ranked Write-In Only Voting (similar to Single Transferable Voting) where each voter can list candidates on many blank spaces from top to bottom may be a decent way to vote in multi-member elections.  What are your thoughts on this multi-member voting system?

[Warren D Smith]

if there were no security problems, then I would be perfectly happy
with everybody voting from their home by mail or telephone
or internet or whatever. In fact I'd be happier.
More convenience, less expense,
same (by assumption) security.

But it astounds me that hard historical lessons about
election frauds just seem to be utterly ignored by the
vote-by-mail crowd. Give up your election security and
it may be the last mistake you ever need to make.

[Brian Olson]

My software is capable of it.

http://betterpolls.com/do/lIrBkrORm5Nf

--
You received this message because you are subscribed to the Google Groups "The Center for Election Science" group.
To unsubscribe from this group and stop receiving emails from it, send an email to electionscience+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Brian Langstraat]

Brian Olson,

That digital ballot is almost exactly what I was picturing for the Range Write-In Only Voting for single-winner elections.

Each ballot contains many range selections for blank spaces, but it does not include a range selection for "All Other Candidates".

Did you make this program before or after reading this post?  Either way, I am impressed.

Are you able to make a multi-winner version?

Who runs BetterPolls.com?

Looking back, I think Warren was confusing Write-In with Mail-In.  Write-In is where there is an empty space on the ballot where each voter may write a candidate's name.  Within voting booths, your digital ballot could be printed as a paper ballot to be hand written-in by a voter or digitally filled-in by the voter on a computer and printed as a completed paper ballot.  Then the paper ballot could be inserted into a counting machine and saved for a potential recount.

To unsubscribe from this group and stop receiving emails from it, send an email to electionscien...@googlegroups.com.

[Brian Langstraat]

Warren,

I think that you are confusing Write-In with Mail-In.  My definition of Write-In is an empty space on the ballot where each voter may write a candidate's name.  Personally, I have never mailed-in a ballot, but I have been tempted to write-in a candidate on my paper ballot while in a voting booth.

[Brian Langstraat]

Warren,

I think the Quorum Rule and the comparison of how various voting systems deal with Write-In candidates (http://www.rangevoting.org/RuleD.html) would be worth considering for this discussion.

With the Quorum Rule, there would be no need for a ballot to include a range selection for "All Other Candidates" for single-winner elections.

The Quorum Rule may be preferable, but how was the threshold of 50% the score-sum arrived at?

Would a lower percentage help unknown, favorable candidates and a higher percentage help known, unfavorable candidates?

A range selection for "All Other Candidates" was not considered in the comparison of how various voting systems deal with Write-In candidates.

I think it would be:

Range (averaging, blanks allowed and "All Other Candidates", no quorum rule)  ---  Unknown write-in candidates difficult to elect unless voters give unfavorable ratings to known, unfavorable candidates (e.g. Trump/Clinton). No bias against lesser-known (rather than lesser quality) candidates, provided they are well-enough known and favorable enough to outweigh voter's fear of the unknown. Great behavior.

[Brian Olson]

I made betterpolls.com, I run it.

Multi-winner version is coming. I wrote it in Go and it turns out that reimplementing STV with precise accounting has a lot of messy edge case logic. More likely I'll deploy my pet proportional representation algorithm first and fix STV later.

To unsubscribe from this group and stop receiving emails from it, send an email to electionscience+unsubscribe@googlegroups.com.

[William Waugh]

- Dimension of ballot with candidate's names printed on it vs. requiring the voter to fill in the names of the candidates they wish to score (but allowing a blanket "all others" in addition to the names);

A problem I see with this comes from breaking down the possibilities with regard to the logistics of the tally.

- Case 1: the people doing the tally interpret the names written (or typed) by the voters.

- Case 2: the tally is done with software.

Each of these cases raises a concern.

In case 1, if candidates' names are spelled similarly, I'm concerned about how far off the tally could get as a result of misreadings by the people doing the tally.

In case 2, I'm concerned that there is no way to convince a significant class of critics that the tallying software works correctly and is free from fraudulent provisions.

If it were possible to choose from the above alternatives and address the concern I associated with whichever one is chosen, I would see some benefits to voting by entry of names by the voters.

- It would get rid of the current advantage to parties that receive big donations from the 1%, in that by the laws of the several States, these parties get candidates being on the blank ballots. No candidates would be listed on the blank ballots. This seems to me as a huge advantage if there is a safe way to buy it.

- Maybe a little bias in favor of voters who are smart enough to know whom they are thinking about well enough to be able to spell their names at least.

I repeat my stance in favor of an ability to down-score "all others".

On top of that, it might be good to include NOTA. A win by NOTA would keep the current officeholder, or maybe a legally determined next selection, like Vice President, as a temporary caretaker officeholder, and there would have to be a new campaign with different candidates or additional candidates.

Waugh

[Brian Langstraat]

Brian Olson,

I appreciate your work.  When I posted about "Write-In Only Voting", I thought that it was just a fun topic to discuss with this Google Group, but it may actually be a worthwhile voting system to consider.

Before I login with Google to http://betterpolls.com/ , how will my e-mail be used?  I do not want any spam. :-)

When I mention STV like "(similar to Single Transferable Voting)", I am aware that there is a plethora of voting systems to choose from.  STV is just one of the most common and well understood, even though it has flaws.  As a mathematician it is frustrating that such a simple thing as voting does not have a single perfect system. I have been convinced by http://rangevoting.org/ that some form of Range Voting is probably the best (effective, fair, simple) voting system for single-winner elections, but I have not been convinced of a best voting system for multi-winner elections.  Could you point me to more information about your proportional representation algorithm?

[William Waugh]

If the design of an election is going to exclude all candidate names from the pre-printed ballot, that's a decision that is independent of the voting system itself, as I think the participants in this forum by and large understand "voting system". The voting system is a mathematical pair. The first element of the pair is the abstract grammar of what the voter can express on the ballot. Another way to say that is that it is the voter's freedom of movement relative to the election. The second element is an algorithm that given the valid cast ballots produces the electoral outcome. The distinction of whether the voter marks a pre-printed ballot or writes the candidates' names on the ballot is a distinction of concrete syntax, not abstract syntax, and so isn't part of the voting system.

William Waugh

[Brian Langstraat]

William,

So far, I see the main weakness of Write-In Only Voting to be coordinating voters written or typed preferences with the names of candidates.  This is already being done most current election ballots, but usually with a trivial number of write-in votes.

I am sure that there are many methods used to count write-in votes, one example is from http://minnesota.cbslocal.com/2016/10/17/good-question-write-in-votes/ :

By Minnesota state law, write-in candidates for local or city elections must be tallied. (In some cities, including Minneapolis, there are additional requirements for charter city offices.)  For county, state and federal elections, the write-in votes are only counted if a candidate is registered as a write-in.  Candidates must fill out paperwork at the Secretary of State’s Office to register.

Write-in ballots are counted at the city level.  First, electronic ballot counters take digital images of the ballots and separate out which ones have candidates written in.  Then, those digital ballots with write-in votes are sent to the city election judges to be counted. The entire process can take a couple of days.

Below, I discuss a possible method to address both of your Cases, but this is just a rough idea.

For both Cases (People or Machine):

Within voting booths, a digital ballot could be digitally filled-in by the voter on a computer (e.g. drop-down menu with a randomly arranged list of registered write-in candidates with clearly unique full names and no corresponding party), double-checked on a full digital ballot confirmation screen, and printed as a completed paper ballot as a triple-check.  Then the paper ballot could be privately inserted into a counting machine and saved for a potential recount.

Determining requirements for registering as a write-in could be controversial. Voters could select the wrong candidate(s), but their fault after the opportunity to triple-check their ballot.  A voter could be allowed to bring in (and leave with) a single example ballot with their favorite candidates secretly listed.

If Case 1 (People / Recount):

Misspelled and similarly spelled names would be corrected by using the digital drop-down menu.  Identical names would be an issue for most voting systems with individual candidates, even with party identifiers.

The chances of misreading names would be less likely than hand-written write-ins and about the same as reading filled-in bubbles.

If Case 2 (Machine):

Counting machines could use an Optical Character Recognition (OCR) program to read clearly printed names on each paper ballot and automatically count them.  If there is a problem with a paper ballot, then it would be hand-counted.  For redundancy, the results of vote totals between digital ballots and counted paper ballots could be compared with any significant discrepancy triggering a recount.

Does this method address your concerns?

What would be fair requirements for registering as a write-in?  Perhaps, petition signatures from 0.1% of the voting population.

How could it be improved?

Replying to your further comments:

"It would get rid of the current advantage to parties that receive big donations from the 1%"

Maybe not, since being popular (known and favorable) would be very important. Candidates would still need lots of ads and strong ground games to have a chance of winning.  The 1% might spread out their money to non-party candidates.

"No candidates would be listed on the blank ballots."

My suggestion of a drop-down menu does list the choices of write-in candidates without noting parties.  There are plenty of ways to avoid the drop-down menu, but they could affect counting accuracy.

"This seems to me as a huge advantage if there is a safe way to buy it."

Could you clarify this statement?

"Maybe a little bias in favor of voters who are smart enough to know whom they are thinking about well enough to be able to spell their names at least."

A drop-down menu would resolve this issue.  Then, even poor spellers could vote and politically disinterested people could vote randomly.  Of course, most politically disinterested people usually do not choose to vote.

I am starting to think that this is why US Founding Fathers decided to vote for the President using the Electoral College with Write-In Only Voting to avoid the "ignorant masses".

[Brian Langstraat]

William,

An option of None Of The Above (NOTA) seems like a poor idea, but I am not very familiar with it.

By definition, it would heavily favor incumbents who may not even need to run with other candidates.

NOTA would improve political consistency at a great cost to political diversity.

NOTA would not work well with term limits and other restrictions on elected positions.

NOTA would often add rounds of voting until a non-NOTA solution.

It could cause long-term minority party rule.

What is the advantage gained by using NOTA?

On Thursday, January 12, 2017 at 1:01:14 AM UTC-6, William Waugh wrote:

[William Waugh]

Randomly ordered drop-down:

This isn't in any sense write-in. It's a choice from candidates already known to the system. This is also essentially what a printed ballot does, other than the psychological difference between random order and fixed order. Printed ballots could be printed in random order at some finite extra expense. Maybe not all orders would be equally likely. There could be a few variants, say 12, chosen at random from among the possible orderings, with thousands of copies of each variant then being run off the presses. Anyway, if the voter can't bring a name not already known to the system, I don't see that the proposal adds much social value beyond that of the conventional ballots printed with the qualified candidates' names.

Warren D. Smith has made a point in this forum, which I agree with, that if the State (Virginia for example, where I live, does this) classifies candidates as ballot-qualified or write-in qualified or unqualified, and does not count votes for unqualified candidates, that this creates two classes of qualified candidate, which seems to Warren and me morally indefensible. If voters can vote for candidates and their vote is to count, we don't see why the State should be able to bias the election against some of those supposedly qualified candidates, by making it harder or mechanically different to vote for them.

Of course, none of your proposals creates two classes of qualified candidate. When you first wrote "write-in", I assumed that by that you meant that the only qualification required would be for the voter to write in the name. But now in your pull-down suggestion, you back away from that, and say that the voters can only refer to qualified candidates, which pretty much restores the meaning of an election with the names of the candidates printed on the ballot as everyone is used to.

OCR:

This has been talked over probably better by others than I can recall or come up with off the top of my head. Some factors in favor of convincing at least _some_ people (maybe even me) that the technique is accurate and contains defenses against fraud (insider or outsider), would be that of course the machines are off-the-shelf general-purpose OCR machines marketed and sold for general purposes and not specifically marketed in connection with election tallying. A point against using published OCR interpretations of the ballots has been brought up (by Smith, I think, in this forum) is that coerced voters could signal their coercers by marking an unusual combination of down-ballot candidates.

There are at least a couple of groups on Facebook concerned with black-box voting. I found it very hard to convince the frequenters of those groups that there is any way to trust any machine with a CPU in it to be free from insider fraud, or any machine that reads and interprets the ballot papers. Some of them will accept a machine that counts a pile of papers already sorted by humans. Someone claiming to have had experience with counting said that counting the piles with machines like that is less error-prone than counting them by hand. She said she has seen that technique used in practice. In (my) theory, a machine like that simply lacks the sensors to look at the ballots and do anything based on what is marked on them; it is just something to count up papers. Much harder to hack than a computer, if the machine can be opened to a simple inspection of the mechanism by a few parties, and can be tested a few times on dummy piles of paper.

If you want to make a point that social benefits could be gained by some combination of arrangements that depends on using OCR, I think you would have an interesting time convincing the Facebook denizens. Are they right? Do they have reason for their level of skepticism? I don't know. Are they so numerous and/or vocal that there is no way to convince enough of the public, without convincing these people, to get changes through? I don't know. One can get an impression by trying to argue with people. But I may be interacting with people in a bubble. Or they may represent a large fraction of those who are willing to think about voting considerations at all at any level of mentality more sophisticated than those who say voter ID will solve all possible fraud.

On Thursday, January 12, 2017 at 2:30:39 PM UTC-5, Brian Langstraat wrote:

https://groups.google.com/d/msg/electionscience/nOS7FHstz_g/tCnHpFZBAQAJ

[William Waugh]

NOTA:

Come to think of it, even though I mentioned NOTA in a context of talking about a write-in (not mail-in) only system, it's probably least useful there. It could mean at most that the voter finds that there is no one good to vote for and they want better people to campaign.

In a system without write-ins, i. e. a system in which votes only for qualified registered candidates count, NOTA could signify the voters' decision that none of the qualified candidates is good enough and they want somebody else. It's a vote of no confidence in the mechanism that qualified the candidates.

When people propose NOTA, I tend to say, "Fine", but if it is not proposed, I might not have a strong argument in favor of including it. Especially not if people get to vote in the general for anyone remotely resembling a candidate (e. g. Bernie), untrammeled by a party or primary process, and if the voting system conforms to the Frohnmayer balance constraint. In that case, I could say to a NOTA voter, "if you want someone else, why don't you vote for them?".

All of my discussions are about single-winner systems if I don't specify otherwise. I don't have an opinion on whether the balance constraint is important in a PR system. That's for smarter people than me to argue. Also, I do not oppose proposals to change the system so there would be no single-winner office anymore. However, I think that is a very hard sell in the US and in each State within the US.

[William Waugh]

"It would get rid of the current advantage to parties that receive big donations from the 1%" -- I didn't mean to say or imply that it would get rid of all the advantages these parties receive from donations. Just the particular advantage that they can get their nominees named on the ballots without having to collect oodles of signatures (in some States) but third parties and independents face a higher cost in getting the same prominence on the ballot. And when I said this, it was of course based on the assumption that "write-in" meant no prequalification would be required. And what I meant by "buy this" was the event that a way would be found to bypass one of the concerns I had listed.

[Brian Langstraat]

William,

Perhaps, this is a solution to the "randomly ordered drop-down" method, which requires computers and candidates to register.

Every citizen is assigned a unique (random ten-digit) ID number similar to a Social Security Number.  Each candidate would publish their ID number as part of their campaign.  Voters would fill-out their ballots using the ID numbers (with filled-in bubbles).  Voters would be allowed to bring a list of ID numbers with them to the voting booth.

This method would be more difficult for voters (with more spoiled ballots), but it would actually be easier than filling in my name for standardized tests in grade school (100 bubbles vs. 520 bubbles).  Hand counting may be difficult.