Talk:Instant-runoff voting/Archive 3

Archives: Oct 15, 2007 - Dec 3, 2007

Comment on POV tag removal

I'm using a Palm to write this, and if I attempt to edit the article, large chunks of text could disappear. But I can add new sections like this. The POV tag should remain. POV and political purpose are woven through the article in sometimes subtle ways. The very name IRV was a political device. By labelling a series of methods (forms of preferential voting with sometimes very substantial and significant differences in details), it can be claimed that, say, Robert's Rules describes "IRV" when, in fact, what is being proposed in many places is importantly different from what RR describes. *If* bylaws permit plurality victories -- which RR dislikes -- then, yes, what is described is "IRV," as it was described in the article.

The "Con" argument about negative campaigning has been put in this article -- with source! -- and deleted many times, by Rob Richie and by the sock puppets. We should, indeed, review the whole procedure for dealing with arguments, but the section is worded such that the needed level of proof is low, and, I would argue that the testimony of any editor that an argument is actually being made is prima facie evidence that it's being made, unless this is controverted. This, in fact, could be established as a general principle for "arguments." Yes, it is better if there are references. Similarly, the Pro argument about campaign expenses, I think I have seen. I think it's false and even silly, but ... if it is being made, it should be there. So it would reduce, in this case, to "who says so?" For simplicity, in this section, we are, I suggest, allowing arguments any of us recognize as actually being made to remain.

I'd recommend that Captain Zyrain study the history of this article and Talk on it to see what has been discussed. There has been specific discussion of the Con argument negating the "negative campaign" Pro argument. —Preceding unsigned comment added by Abd (talk • contribs) 14:32, 14 October 2007 (UTC)

Wikipedia:Verifiability says that unsourced information can be dealt with by deleting it or by adding the "fact" template. I probably am a little too impatient about giving people time to find sources for information they add. On the other hand, I think that this article is so susceptible to edit wars and challenges of the veracity of information that we should pretty much cite sources for everything at this point, preferably using the ref template, and using hidden comments if nothing else. (Hidden comments are the icon fourth from the right on that little toolbar that shows up when you're editing a page.) Plus I have nominated this article for peer review, as a prelude to a featured article nomination, and since we are entering that process I think we should raise the bar a bit, relative to how we would treat other articles. Once it gets posted to Wikipedia:Featured article candidates, reviewers will start nailing us on verifiability if that's a weak point. See Wikipedia:Featured article criteria. Captain Zyrain 16:04, 14 October 2007 (UTC)

You're welcome to dig through the history and find the sourced Con argument; I just didn't feel like doing it, partly because I'm not familiar with the history of this article and wouldn't know where to look in those ~1,500 edits. Captain Zyrain 18:10, 14 October 2007 (UTC)

The history I was referring to was Talk, and it was recent. And the word Con or the words Pro and Con is in the section title. Hard to miss. Abd 03:22, 15 October 2007 (UTC)

By the way, I understand and sympathize with your concern that IRV is a POV name. San Francisco chose to call it RCV for the very reason that the term IRV can be inaccurate, since the votes aren't necessarily retallied instantly. However, for the benefit of the reader we typically go with the more recognizable nomenclature. I would compare it to Wikipedia:Neutral point of view/BCE-CE Debate, for lack of a better example. Captain Zyrain 21:07, 14 October 2007 (UTC)

That's true for the U.S. This encyclopedia is for the world. IRV is not actually the name by which it is generally known, whatever "it" is.... Robert's Rules refers to "preferential voting." Why pick "IRV"? Certainly the article, if it is not titled "IRV," would nevertheless refer to it as a name by which it is known here, but one point I've made is that Bucklin is literally instant ... runoff .... voting. That is, it proceeds in rounds, like IRV, stops pulling in additional votes when a majority is found, like IRV, but *unlike* most implementations of IRV, it does not create a false majority by discarding exhausted ballots. It does not drop candidates. It is essentially Instant Runoff Approval.

And Bucklin saw substantial use in the U.S., and, I'm pretty sure, was dropped for the same reason that IRV was dropped from Ann Arbor and Proportional Representation from New York City. It was working. The election in Brown v. Smallwood was one where clearly the best winner was chosen. The Minnesota court reversed an election where the will of the people was manifest, and people had counted on the election method, or else they may have voted differently. It awarded the election to the plurality winner in the first round, based on totally spurious and even self-contradictory arguments. It was a corrupt decision. (And they would have tossed IRV as well, to believe otherwise you have to focus on one out of many arguments that they made, and you'd have to disregard the language of the second decision, where they reconsidered and refused to change their decision.)

My point is that the name is itself a piece of propaganda. It attempts to attach itself to precedent (runoff elections) in order to seem less strange. Further, using the same name for different variants leads to potential confusion: what exactly is it that Robert's Rules describes?

Turns out that, indeed, *some* proposed U.S. "IRV" implementations may follow the Robert's Rules procedure. The Vermont legislation that Mr. Bouricious wrote does, which may explain his difficulty understanding my objection. Why, it's obvious! Robert's Rules recommends "IRV!" His IRV. Not that of San Francisco or many other places.

There is political spin involved here, which is why having a balanced article is not as simple as making sure that the facts are correct. There is also the balance. How facts are presented can influence public impressions.

This is one reason why it was suggested here that we develop a template for election methods articles. What has happened is that proponents of a method have generally taken a large hand in structuring an article, and, in the case of this article, there was at least one paid staff member of FairVote involved, though I could only say for certain that this editor was mostly using reverts to keep criticism of IRV out. When I started editing this article less than a month ago, the Con IRV argument was pretty much "Some people are more comfortable with the status quo." There was no mention of alternate reforms. IRV was compared only with Plurality. Why? Robert's Rules, we were told in the introduction to the article, recommends IRV. Never mind that the method described in the article was not the same as what is described in Robert's Rules, that Robert's Rules recommends "preferential voting" and makes it clear that this is a whole set of methods (mostly defined by a ranked ballot), and it simply describes one way of doing it. And it also gives reasons why the method has problems. Why mention only the positive and not the negative?

Now, it's a *little* better now. Not a lot better. And some of what I objected to has come back. I'll be doing some edits tonight. I might even use a revert or two.

The introduction to the article is going to be very important to an advocacy organization. I attempted to move the Robert's Rules mention to the middle of the article, where it would be more proper to explain it in sufficient detail that the position of Robert's Rules on preferential voting was clear. It was moved back. Explanatory language that was *not* incorrect was removed as "too confusing." Yes, it is; if your goal is to convince people of something, extra detail can cause them to think too much.

I'll be putting the POV tag bag. It was premature to request peer review of the article, at a point where a consensus of the editors had not developed. We had an edit war going here, and it never did really end. Abd 03:22, 15 October 2007 (UTC)

I agree with Jeff Dahl that someone needs to create an outline from scratch and then move the text around to fit the outline, so that it will fit together as a well-organized article. Maybe sometime I will set a day aside to do that. Captain Zyrain 11:24, 15 October 2007 (UTC)

COMMENT on United States Category and edits

Someone has removed my sourced edit on IRV in North Carolina. I live in North Carolina, I watched the legislation go through 7 revisions, and I provided independent links to the the state legislature's reference for the NC IRV law, and to a news article that had not been there before.

Now someone has come in and replaced my independent links (to the legislative code and to a news article) with other wikipedia links.

I think it weakens an article to replace all independent links with either other wikipedia articles as reference, or only Fair Vote links as reference.

It looks like the article has been diluted.

Here is what I am adding back to the article and I ask that it be protected from further edits:

• North Carolina adopted a pilot program for instant runoff voting for certain judicial vacancies and municipal pilot programs starting in 2007. "The State Board of Elections shall closely monitor the pilot program established in this section and report its findings and recommendations to the 2007 General Assembly." [1] The city of Cary [2] will use IRV for mayor and city council elections in October 2007,[3] and the city of Hendersonville will use IRV for city council elections in November 2007.[4]

--Ask10questions 00:52, 15 October 2007 (UTC)

COMMENT. My bad. I apologize, please disregard above talk I posted.

I see that the link format has been grossly changed, but apparently is correct.  

Here is what I added back to the IRV in North Carolina segment under United States: "The State Board of Elections shall closely monitor the pilot program established in this section and report its findings and recommendations to the 2007 General Assembly."

--Ask10questions 01:01, 15 October 2007 (UTC)

Given that the North Carolina legislature meets in January, I would think that this report to the 2007 General Assembly has already been made? Captain Zyrain —Preceding comment was added at 01:17, 15 October 2007 (UTC)

Opposition to instant-runoff voting

I think that it will be beneficial to expand the article which I have created, opposition to instant-runoff voting. This can be a place for opponents to expound on their arguments in more detail without having to worry so much about the need to balance pro and con arguments. The major arguments, having been fleshed out and put in polished form at that page, can then be summarized at instant-runoff voting. Captain Zyrain 12:22, 15 October 2007 (UTC)

We might move the page to criticism of instant-runoff voting or arguments against instant-runoff voting. 66.208.12.125 12:31, 15 October 2007 (UTC)

The new article would be better placed in this article because it is not a separate concept but criticism of an existing concept. If you look at any other article, they will invariably include criticism sections rather than having separate articles. Green Giant 12:40, 15 October 2007 (UTC)

Please see rationale at Talk:Opposition to instant-runoff voting. Captain Zyrain 13:04, 15 October 2007 (UTC)

Rationale seen, comment added and agreement reached. :) Green Giant 13:22, 15 October 2007 (UTC)

In accordance with Abd's suggestion, it has been moved to Controversies regarding instant-runoff voting. Captain Zyrain 02:00, 16 October 2007 (UTC)

Thanks. That title is longer than it need be... I'd suggest Instant runoff voting: controversy. Presenting only one side in an article without equally presenting the other side is not balanced and will not help readers to understand the issues. Abd 14:03, 16 October 2007 (UTC)

Probably, a brief summary of arguments should remain on this page, but the exact text of this is tricky. Again, I do propose a basic standard: the article should settle in such a condition that all three classes of editors generally agree that it is fair and balanced, not just two out of the three classes. The three classes are pro-IRV editors, anti-IRV editors, and neutral editors. In this case, two out of three does not indicate a consensus, because the pro and con editors will tend to be more informed about the issue than someone who is outside the arguments. Neutral editors may easily be unaware of the political nuances of arguments and presentation, and are more likely to focus on technical issues that they can easily and quickly understand. Is a statement sourced? Is it presented in a POV way *in itself*? But questions of balance will be much more difficult for such an editor, and balance is important to NPOV as well as factual accuracy.

Given that the history of this article has largely driven away anti-IRV editors, who have found it thoroughly frustrating to try to insert even NPOV content that is disliked by pro-IRV editors, and I must say that I've had a similar experience and it is continuing, it will take some time before enough of these editors to return to allow a genuine consensus to appear. Until then, I'm acting, in a sense, as an informal proxy for them, and I can assure the other editors that when I present an argument here in Talk, there are others who agree with it. But I have *not* attempted to recruit meat puppets to come here to support my actions. Frankly, some of the people who would respond to such a call might be so highly POV against IRV in their edits that it would be distracting and could delay our process; but others will be fair-minded and will come to consensus readily, they are not fanatics.

Where we have trouble agreeing on balance, we can recruit examination by outside groups that are not controlled by either pro or con IRV activists, but which include members who are knowledgeable about the debate, and we need not limit ourselves to one single such forum.

I do believe we can do it, we can put together an article that is informative, interesting, and thoroughly neutral in both fact and balance. The more editors we have here who are willing to participate in such a process, setting aside political agenda -- which they will still use in a helpful way because of the sensitivity it creates in them over nuances -- the faster this process will be. (Note that there are very few election methods experts who are truly "neutral" on this topic, that is, they have opinions and conclusions that they have drawn from their study!)

I've started working on the Controversies regarding instant-runoff voting page, and I edited the links in this article to point to that for now. Abd 14:03, 16 October 2007 (UTC)

Robert's Rules of Order, Newly "Revisited" :-)

Last time I started a section with this title, which is a play on "Revised," the actual name of the most recent revision, someone edited the title to correct it. I guess I omitted the smiley face! This is *Talk,* folks, where it is a bit rude to edit someone else's contribution. Errors are allowed here! (It's even a tad rude to edit one's *own* contributions here, if someone has responded to them as-is ... but that's just my opinion. I'm certainly not in charge.)

The article as, as I write this, has the following in the introduction:

Robert's Rules of Order describes how to implement IRV on p. 411-414, stating that in situations such as voting by mail in which it is impractical to repeat balloting until one candidate receives a majority, "it makes possible a more representative result than under a rule that a plurality shall elect"[2].

Being placed in the introduction, it has prominence, and it is quite understandable that advocates for IRV would be pleased at this placement and attached to it. On the other hand, opponents of IRV won't like this. Now, I agree that there is a significant fact behind this, and that mention of that fact belongs in the article, but ... if it is in the introduction, it must be brief. Going into nuances is not something that belongs in a short introduction. And without nuance, the mention is not balanced.

Originally, the mention was more blatant: "Robert's Rules recommends IRV for voting by mail," or something like that. Now, when I changed that to something more accurate, the edit was reverted. When I revised the edit to reflect some of the criticism of my edit, that was reverted. When I moved the mention to the middle of the article and added, again, more detail to make the facts clear, considering what had been said, that too was reverted, and an edit war started over this. I was careful about 3RR, but one of the reverters and persistent editors was a sock puppet, originally BenB4, but BenB4 was tossed in the laundry by Wikipedia admin and was replaced by Acct4. Then an SPA appeared and immediately filed a 3RR complaint against me. When an admin looked at this, the admin was pretty clearly angry about what had been going on and banned almost everyone in sight. But not me, the administrator made it clear that I had been defending the article from socks and anonymous reverts.

So there is history behind this. There were extensive discussions of what was actually in Robert's Rules, under descriptive section headers, just now archived.

I edited the mention to be accurate. Now, more slowly, (the socks and anonymous editor -- Rob Richie of FairVote, known by IP -- could no longer access the article to edit it, the mention has slipped back toward what it had originally been. It was placed back in the introduction. The explanation of what was *different* about "IRV" as described by Robert's Rules and what IRV is described as in the article was removed. It is now almost as POV as was the original. That is, it is a biased interpretation of the facts, prominently placed. And we have been over and over the facts. I'd suggest that anyone who wants to work on this particular problem read what came before, and then Talk about it here. I understand that this means a lot of reading, and I apologize for that; however, I certainly cannot demand this of anyone. Nevertheless, until we have an agreement on how Robert's Rules is going to be mentioned, I'm taking it out entirely, and I will defend this position against any attempt to put it back, unless what is put in is accurate *and* balanced.

I see no justification for allowing it to be in the article if we cannot agree on its language. It is not essential or necessary to the article. Not everything relevant or interesting must be in an article, and particularly not in the introduction. And I have put too much effort already into trying to find exact language that was acceptable to the pro-IRV editors. (And I am not claiming that all editors who have effectively reversed my edits on this are pro-IRV; but the view of these editors of the context hasbeen, in my opinion, as yet insufficient for these editors to recognize easily all the implications.)

What do I mean by balanced? Well, suppose there was a resume from a candidate for employment, and it quoted from a termination letter from a previous employer. The quote says that the employee did a good job making sure that certain responsibilities were fulfilled. The quote might be true, accurate, and sourced. Is that balanced? Well, it depends on the rest of what is in the report!

Robert's Rules dislikes the use of "preferential voting," in general. However, it is a manual that reflects actual practice. There are conditions where an organization may consider it impractical to hold a runoff election or other further process. (RRONR dislikes top-two runoffs, by the way, for reasons apparent to anyone who understands election methods: frequently the top two excludes the most broadly acceptable candidate, the one who would beat either of the others in a one-on-one contest). *Given* that further process is considered difficult or impossible, RRONR suggests "preferential voting" *as a possibility.* Now, there are many forms of preferential voting, with different characteristics. Some organizations, for example, use a Condorcet method, which would always select as winner a candidate like the one just mentioned, who would beat all others in pairwise elections, and it is clear that on these grounds, RRONR would prefer this. But the editors are constrained by actual practice, and the use of Condorcet methods is still relatively rare. So they do describe a "preferential voting" method, as an example, that clearly resembles IRV, so clearly that many observers, including myself, when first looking at it, have thought that it *was* "IRV." Was it? That depends on details.

Unless the bylaws specifically allow it, RRONR disallows election by a plurality. Where voters are free to vote as they choose, any ballot submitted by an eligible voter is part of the denominator when considering the fraction that have voted for a candidate, even if the ballot is spoiled; spoiled ballots can be considered No votes on all the candidates, in this sense, and a candidate must get a majority Yes to be elected. What STV (IRV is single-winner STV) allows is vote substitution, that is, you may vote for A, but if A is being eliminated by the IRV process, and you have cast a lower rank vote for B, then B may be substituted for B. In other words, if you rank a candidate, you are voting for that candidate, against all others with lower rank on your ballot, as well as against all candidates you have not ranked at all. This is, indeed, why some voters would dislike ranking all candidates, for a requirement that they do so requires that they vote *for* such a candidate, making it appear, possibly, that a majority supports that candidate when, in fact, this is not the case, the majority might actually prefer that the office remain vacant. The only way to vote, clearly, *against* a candidate if full ranking is required is to rank the candidate last, and if one wishes to so vote against more than one candidate, full ranking does not allow it.

The ballot-counting procedure described in RRONR that resembles IRV is the equivalent of sequential-elimination IRV in every respect except the final step: that procedure does not terminate unless there is a true majority, including all spoiled or exhausted ballots. Because the description was not completely explicit about this, but this understanding is clear if one understands how RRONR uses terms and considers to be valid votes, one editor claimed that my commentary on this was wrong. That was User:Tbouricius; however, when the rules in the recently proposed "IRV" legislation in Vermont were brought here, he backed up and claimed he had never argued against the basic point about "majority." Why was this an effective argument? Well, he wrote that legislation, and, I think, another commentary on FairVote saying practically the same thing. In Vermont, if there is no majority winner, the election goes to the Assembly, which votes by secret ballot to choose between the top three. How the "top three" are defined, by the way, is utterly unclear to me.... and the ballot instructions in that legislation are incorrect ... but that's another matter.

Until I started to work on this, the article seems to have assumed that there was only one form of IRV. However, batch-elimination IRV and sequential elimination IRV behave differently, and the number of ranks allowed also have major effects, and likewise this question of the majority. Lumping these all together as "IRV" is politically useful, for, then, the "IRV" movement can claim "momentum," each time we get something that is one of these forms.

A mature article will clearly delineate the various forms in use and being proposed, and, to the extent that instances of use are mentioned, the specific form used will be mentioned. Robert's Rules of order "describes" (not "recommends") a method which is IRV-sequential elimination, with no restrictions on ranking -- it can be complete, or not, as the voter chooses -- with a majority election requirement unless the bylaws otherwise specify (which it does *not* recommend, it would, in fact, discourage). What Mr. Bouricius proposed in the Vermont legislature is, I agree, what Robert's Rules describes, but most IRV implementations in the U.S. aren't like that, and IRV is being sold as "guarantees a majority winner," which is utterly untrue unless we define "majority" to exclude inconvenient votes; but this fact is easily obscure to someone who is not aware of the details.

So I can understand at least some of Mr. Bouricius' frustration with me. Since what he proposed in Vermont *was* what is described in Robert's Rules, how could I claim that it was different? But what he proposed there was *not* what was described in this article as "IRV"! Indeed, the description of sequential elimination that was here (and I think it's been taken back to that) referred only to "majority" of the ballots remaining, not of all the ballots cast by eligible voters.

(Legal voting systems may generally not count true spoiled ballots, unlike RRONR, in determining the majority, but I see no way to exclude ballots with a lawful vote from the base for "majority," unless the IRV legislation trumps a majority requirement in law that is therefore removed. Some have argued here that "IRV is like runoff elections,' but that is actually an argument of no legal force, entirely aside from being defective in another way: in a real runoff, a voter who expressed no opinion on the pair involved now has the option of expressing one, whereas "instant runoff" does not allow the voter that option.) Abd 17:30, 16 October 2007 (UTC)

In cases when you say something and then think better of it after someone has already responded, you can strike out your original comments.. On second thought, it's probably better just to avoid saying something dumb to begin with. Captain Zyrain 17:35, 16 October 2007 (UTC)

Quite possibly not. By sticking my foot in my mouth, I learn far more rapidly than if I hang back, censoring what I say because it might be wrong.... Yes, caution is good, but only to a point. Using strikeout is an excellent idea. It preserves the record and makes clear that the writer no longer supports what he or she wrote. Cool. Abd 19:34, 16 October 2007 (UTC)

True, you learn more from your failures than your successes, but the mistakes sure can be painful! Fortunately, making a regrettable comment on Wikipedia ranks pretty low on the scale compared to other ill-advised life mistakes (e.g. switching majors to philosophy). Captain Zyrain 19:41, 16 October 2007 (UTC)

Images

Here are a couple images generously contributed by Bouricius:

These are actually from an STV election, so I'm not sure whether they could find use here, but I thought I'd post the links FYI. Captain Zyrain 19:53, 16 October 2007 (UTC)

There'd be quite a few on Australian articles too as we've used a similar method at both federal and state elections for most of our post-Federation history. Most would be GFDL as they've been snapped by members of the public who've uploaded them here. Orderinchaos 22:41, 26 October 2007 (UTC)

Go ahead and post them then, please. StrengthOfNations 06:40, 15 November 2007 (UTC)

Does the American Political Science Association "use" Instant Runoff Voting?

Unless some cogent reason appears here to the contrary, I intend to remove, once again, the mention in the Introduction to this article, of APSA and its alleged use of IRV. I took this out once before, because it was unsourced (other than to FairVote, which is not generally a reliable source for Wikipedia, being an advocacy organization). However, then, a pointer was given to the APSA constitution which does describe a preferential voting procedure to be used if there are more than two candidates for the office. So I did not object when it was replaced.

However, turns out, APSA elections for President-Elect (they elect a new President-elect every year, who becomes President the next year) are not, as far as the records on-line are concerned, the last four years, contested elections, and I was informed by a correspondent that this has long been the case. There have not been *two* candidates, let alone the three required to trigger the preferential voting clause in the APSA constitution.

I have not verified that this is absolutely true, but it looks like APSA has *never* used IRV. The APSA Council is not elected by STV as one might think, but by standard multimember plurality. The Nominating Committee nominates all officers, and if there are no additional nominations by petition, the nominated officers are, in practice, elected at the annual meeting. No officer elections have been contested in the four years shown on-line, but there are eight Council seats. There have been, each year, one or two additional nominations by petition, and sometimes these win in lieu of the Committee nominees. In 2007, the additional nomination by petition won with a substantial margin over all others, and thus one of the Nominating Committee nominees was not elected. The election statement of that candidate is here: [5], and I found this from it to be remarkable: "I am running for the APSA Council in order to promote methodological pluralism and democratic representation in the association. If elected, I will work to further the following goals: .... 3. Selection of association officers through competitive elections to increase the representation and participation of all groups and interests in the profession."

Further, APSA now uses internet and email voting, so multiple balloting as is recommended by Robert's Rules could be practical.... Given the strong statement of that new Council member, we might soon see some changes in the APSA constitution. Or we might not....

Because information in the Introduction should be impeccable and not complex, the mention of APSA should be removed. Comments? --Abd 16:54, 7 November 2007 (UTC)

There being no comment in five days, I'm pulling the reference from the introduction. --Abd 03:14, 13 November 2007 (UTC)

The fact is that the APSA decided to incorporate IRV into their constitution, and did so. Whether there happen to be enough candidates to trigger its usage is not the interesting point. I have changed "uses" IRV to "has adopted" to remove any ambiguity.

Tbouricius 15:49, 14 November 2007 (UTC)

The problem is that the fact is no longer truly notable. This is in the *introduction*. Introductions should be short and quite clear, and only the clearest and most important aspects of a topic should be mentioned in them. If we are going to say that APSA "adopted" IRV, we'd need a source. Yes, it is in the Constitution of the organization, that can be cited, so at *some* time they "adopted" it. But apparently it has *never* been used, if it has, references would be of interest. When was it put in? Mr. Bouricius pretty clearly wants this in because it lends cachet to IRV. After all, if "political scientists" "use" IRV, then it must be a good method, right? But if it hasn't been used, and if APSA is not even running contested elections (let alone elections with three candidates), it could be a wheel that does not squeak because it isn't turning. Note that the huge controversy in APSA right now seems to be the idea of having contested elections *at all*!

Particularly interesting is that there *are* contested elections with APSA. The 8-member Council is elected. As with the President and other officers, the Nominating Committee names 8 nominees, but others can be nominated by petition, and, in the last years there have been, typically, one or two such nominees. So: multiwinner elections, do they use STV? If APSA is notable, is it notable that they *don't* use STV but merely use Plurality, voters can vote for up to eight candidates? In any case, the mention does not belong in the introduction, if it belongs in the article at all. The language edited in by Tbouricius is slightly better, but still biased to make it appear that the fact is a recommendation. I'm taking it out of the Introduction. I might put it somewhere else.

How does APSA *actually* choose its President? Deliberatively, in the Nominating Committee. I don't know the specific rules, but they certainly wouldn't be using IRV! --Abd 01:02, 15 November 2007 (UTC)

I just noticed that there is another article, History and use of instant-runoff voting. Detailed information about history and usage really belongs in that article, with only the briefest summary here. I would also argue that the other article should be about "preferential voting," not "instant run-off voting," but that's a topic for another day.

One more point. Tbouricius is a COI editor, clearly affiliated with FairVote, and, with this and with many other disputes which have arisen, he is attempting to keep "promotional information" in the article; "promotional information" is fact, often presented narrowly, and clearly presented because it carries some implication desired by the editor. Wikipeida articles clearly should not be promotional. I do believe that it is important that the POV of FairVote be represented in the editorial process here, which is why I suggested that the block of Tbouricius could be lifted, but he seems to have taken that lift as some kind of "exoneration" indicating that there were no problems at all with his work here. In this particular case, it is desired that the APSA mention be in the introduction, precisely because it seems to convey the idea that "political scientists" support IRV, after all, they chose it, didn't they? Actually, we don't know who chose it and when. How long has it been in the APSA constitution? APSA was founded in 1903, and preferential voting was the rage at that time. Now, as anyone familiar with Robert's Rules on voting methods would know, the kind of preferential voting that IRV represents, with sequential dropping, is *not* recommended by Robert's Rules if multiple balloting is possible, because of the well-known possible failure to select an obvious compromise, instead electing a candidate who would lose, even by a large margin, paired off with that compromise. *If* APSA were to start running contested elections, my guess is that they would drop IRV. They allow internet voting and email voting. It's no longer time-consuming or expensive to hold a real runoff, or even multiple balloting. Maybe we will find out. —Preceding unsigned comment added by Abd (talk • contribs) 01:26, 15 November 2007 (UTC)

RFC

Removed claims sourced with FairVote

I removed the following claims, because they are sourced with FairVote. Anyone is welcome to put back properly sourced information. I wish we could create "reliable sources" by doing a research paper and putting it up on an advocacy site. But we can't. See WP:RELIABLE FairVote is fine as a source for opinion, but not for fact, except as to facts about FairVote itself.

• Dozens of American colleges and universities use IRV, including as of November 2006 more than half of the 30 universities rated most highly by U. S. News and World Report.

--Abd 03:29, 13 November 2007 (UTC)

I restored the link to the list of 32 universities and colleges that use IRV. While the fact about the number that are on the 'U.S. News and World Report' list is also accurate, I removed it as a compromise. Many of the colleges on the list on the FairVote cite are in turn linked to articles, student government constitutions, or other sources. FairVote is a widely respected and reliable clearing house of information about preferential voting that is an acceptable source. Countless scholarly articles cite FairVote, as do countless news stories. Because of the sheer volume of material on the FairVote web site it is inevitable that there must be a few out-of-date, or inaccurate statements, but the site is exceptionally reliable.

Tbouricius 15:57, 14 November 2007 (UTC)

I've been spending the last few weeks researching Wikipedia standards, reviewing Arbcomm decisions, etc. Please reread what I wrote above, and, in particular, consider whether an advocacy web site, which is certainly not a peer-reviewed publication, nor an edited and responsible news source, and which is known for presenting one side of an issue, definitely POV, and for framing quotations of outside sources with "helpful explanations," can be used as a reference for a statement of fact in the article. The answer is pretty clear. No. I agree that FairVote is commonly cited. Indeed, I've been doing my own research to uncover primary sources and it's difficult. FairVote is often the sole source, at least the sole source that is easy to find. and the browser will come back with hundreds of hits that are exact quotes of FairVote (that is, the specific wording from the FairVote page). Now, who did the research? If it was FairVote, that's "original research," and not permitted as a source for facts. If it was somewhere else, then "somewhere else" can be cited here *if* it was published in a "reliable source." So ... I'm going to remove material that has not been properly sourced. I'm willing, personally, to compromise on material that is essentially well-known *and* important to the article; beyond that, bring the sources! --Abd 21:51, 14 November 2007 (UTC)

To continue this, I was looking at the following claim in the article that Ferndale, Michigan, had passed IRV in 2004. I tried to verify the claim, to find an original source, such as a newspaper account from Ferndale. So far, I haven't found it, but what I did find was many, many hits quoting an firv.org press release, or a FairVote press release or article. What actually happened in Ferndale? I don't know yet, but I do know what has happened with all these references: FairVote established itself as a source of information, and few, apparently, bother to look for original sources, since FairVote has conveniently placed all the information one might ever need, at our fingertips. It would seem. This information, however, has been filtered. What was actually passed? According to Ferndale IRV, "The actual text of Proposal B reads as follows":

Proposal B

Proposed Amendment to Ferndale City Charter Chapter IV, Section 17

This amendment provides for the election of mayor and council members by majority vote using an instant run-off voting procedure of counting votes as soon as the City acquires voting machine equipment, approved by the City Election Commission, to implement this amendment. Voters shall designate first preferences and subsequent preferences; if no candidate receives a majority, the candidate with fewest first preferences is eliminated and the secondary preferences for that candidate are recounted until a candidate receives a majority who shall be elected to office.

Shall the Ferndale charter, Chapter IV, Section 17 be amended as proposed?

Yes:______ No:______

(1) Ferndale may not yet be holding IRV elections. Did they get the equipment? (2) This is an inadequate description of IRV election rules. Anyone familiar with the edit warring on our article would realize that "majority" could be unclear. If a true majority is needed -- which would seem to be the intention -- then there can be election failure. I'd say this amendment is an invitation to legal action, if it is ever implemented and anyone cares, i.e., a candidate did receive a majority of votes *if not eliminated in the first round.* In small towns, legal defects can exist for a long time before anyone really cares.... The problem here is that election by "majority vote" is described, and that would make the proposal more popular, it would be important in its passing. IRV with a majority vote requirement is a substantially better method than without it (this is the default Robert's Rules example); but, of course, there might then be a (rare) election failure, a need for further election process, which would need to be in the rules.

Ferndale is a pretty small town. Is this notable? Should the article have a laundry list of every jurisdiction that decides to (perhaps only theoretically) use IRV? There is another article, History and use of instant-runoff voting. --Abd 02:12, 15 November 2007 (UTC)

It gets a bit weirder. I looked at the town charter.[6] (Charter, Chapter IV, Section 17). The language inserted into the charter is substantially more complex than what was claimed to be the actual text of Proposal B by firv.org... is something funny going on? --Abd 02:44, 15 November 2007 (UTC)

About Takoma Park and Universities using IRV

I edited the Takoma Park usage of IRV to note information taken from the source, indicating that the runoff system did not come into play. Since User:Tbouricius undid that, moving the source back to FairVote, I once again did the research to find primary sources, and this time, I got the actual election results instead of the newspaper account. Takoma Park does not appear to have contested elections, normally. The mayoral election was uncontested, so the original comment that the mayoral election was won with a majority of votes was an understatement, I think it was about 800 for the mayor to 100 write-ins. Of the six ward seats, five were uncontested. One was contested with only two candidates, and, again, a handful of write-in votes. So the runoff provisions were not only unused, they were not even *close* to being needed. Why did Takoma Park adopt IRV? This is a small town, small towns often have unopposed elections, in small towns where I've lived, I haven't seen more than two candidates, and two were actually unusual. Let's see ... does anyone on FairVote staff live in Takoma Park? Sounds familiar to me.... I'd think that getting a few more small towns to adopt IRV would help FairVote to keep their statistics up, to show one more success. Even if it is a fish bicycle there.

Yes, according to the source article for the Takoma Park adoption, Rob Richie, Executive Director of FairVote lives in Takoma Park. --Abd (talk) 18:58, 18 November 2007 (UTC)

I think we will likely move all these small town instances out of the article, to the History and Usage article, and ones like this, I'd think, shouldn't even be there. This article is not a place for FairVote to trumpet its "successes." Significant implementations are quite relevant.

Meanwhile, I've again removed the reference to FairVote for universities using IRV. I think it would be reasonable to reference a few, but, let me repeat this, the FairVote site cannot ordinarily be used for reference. That page, as User:Tbouricius noted in his revert of my previous removal, does have a few references, but many more simple claims without references, and a great deal of promotion, using language that would not be permitted here. FairVote is not a "reliable source," per WP:RELIABLE, it is an *advocacy* organization. That various authors have relied upon it (poor scholarship! -- like relying on a press release from a company for a newspaper article about the company's products) doesn't make it reliable here. Did I mention that FairVote is not a reliable source?

While someone could put back in a few universities, particularly prominent ones, all these usage details, except for very few, really should move to the main article for History and use of instant-runoff voting

By the way, try to find a source for the claim that the "inventor of IRV" was William Robert Ware that isn't FairVote or obviously copied from FairVote or Wikipedia. Not easy. Ware did apparently write a letter to Hare (of STV fame) suggesting the use of the STV technique, single-winner. However, an old source I've looked at discounted Ware's suggestion as not being new, plus as being defective, for reasons one would have a hard time finding in this article yet. I haven't found a copy of Ware's letter yet, but it's available. --Abd 04:28, 15 November 2007 (UTC)

I am restoring the link to the list of 32 colleges and Universities that use IRV. It is more useful to Wikipedia readers to have access to this entire list, rather than to just pick a one or a few to mention by name in the article.

Tbouricius 15:39, 15 November 2007 (UTC)

Depending on response to my action today, I may start the dispute resolution process. I'm sure that what I've done looks arbitrary and anti-IRV to Tbouricius, but, based on my review of *many* article conflicts and resolutions, it's not. The question is not access to the list. That access is available through the FairVote link under advocacy organizations, and it's also possible to put it in the article, properly framed. The question is Wikipedia standards for sourcing. The fact of usage of IRV by universities and colleges can certainly be in the article, but it must be properly sourced from a *neutral* source. That's easy to do for a few universities, and, it must be asked: why is it important to have a long list of universities using it? And if a university *stops* using it, would this likewise be important? If the answer to the first question is yes and the second is no, why the distinction? However, sourcing standards are *not* negotiable, it appears. While there can be exceptions, and articles can be given some time to find acceptable references before material is deleted, there appears to be no effort to provide such references. With many instances of improper referencing, over the last few days, I've done the work to find the actual acceptable references. Here, since I don't consider the mention that important, I'm not. But I would certainly not, at this point, try to take out any reference to actual university usage, properly sourced. In any case, I have edited the mention to make it, in my opinion, acceptable. And, of course, as part of the process, I found more relevant information, which I'm putting in as well. A list of usages of IRV, with the appearance that all these represent active use, is promotional, particularly if looking at the original sources reveals that it was moot that the election was IRV.

For future reference, if needed, this is the core dispute: can an advocacy organization, not a neutral party, be used as a source for statements of fact in the article? The organization in question here is FairVote, but if FairVote is allowed, what about [rangevoting.org Center for Range Voting] which contains a lot of information about Instant-runoff voting, plus advocacy material, like FairVote? --Abd 16:56, 15 November 2007 (UTC)

I've now done the editing, and have continued to add brief election details about the actual usage of IRV in election listed under "usage." I did *not* select the examples to show the irrelevance of IRV; rather, I just went down and picked some prominent examples from the FairVote site. If we are going to have usage examples, it's important to have information about *usage*! What I've been finding is that in some places, IRV has been implemented quite obviously not because there were problems with elections, but, likely, as a political statement. I have found this in in every example I've researched over the last few days. And where IRV is actually being used, and the rounds are pulling in additional votes, majority failure seems pretty common (with lots of exhausted ballots). Note that it is possible, in some cases -- and the election results I've seen give us little information -- that a majority-favored candidate existed (i.e., one who actually did receive majority support over the plurality winner -- but that candidate was eliminated because of not getting enough first-round votes. --Abd 17:36, 15 November 2007 (UTC)

Those are the same universities mentioned in History and use of instant-runoff voting in the United States. StrengthOfNations (talk) 02:02, 18 November 2007 (UTC)

That article was started by StrengthOfNations; it is essentially the old information from this article after POV and citation problems were corrected. The article should probably be AfD'd but, only so much time.... I did put a POV tag on it, the introduction has a probably straw man insertion in it. --Abd 20:58, 4 December 2007 (UTC)

Anti-IRV Selective Election Details Showing POV Being Inserted

Abd has been selectively inserting election details in the list of places that have adopted IRV in the United States. He never inserts examples where IRV saved the cost of a separate runoff, or found a majority winner after the IRV tally, but only examples that might lead one to question the value of IRV. Each fact he has inserted may be "true" and sourced, but the pattern of selection reveals an intense POV goal of his. Lists of places adopting IRV, should be simple and short with links to the source for those who want to learn more about them. Such lists are not the place to try and selectively put IRV in a poor light. Tbouricius 16:03, 16 November 2007 (UTC)

It is unfortunate that Tbouricius is choosing to take this as some sort of attack on IRV. As to his claims:

"Selective." I've been systematically removing references to FairVote as a source, and, instead of taking the easy route of deleting or flagging unsourced material, I've been replacing the original, non-qualified source, with qualified, neutral sources. In doing that, the facts of the elections become exposed to me, and I've been placing very brief details in the article, taken from the qualified sources. These details are indeed relevant to controversies over IRV; however, I have not "selected" the facts, rather I have placed them as I did the research. If we are going to have examples of usage, we should have basic *details* of usage. Not extensive election results, but very summarized ones, as I inserted. There was no selection..

"Put IRV in a poor light." I'm shining light on IRV, using sourced information, with regard to facts already mentioned in the article. I did not select these examples. Rather, I was simply beginning to clean up the sourcing, so I looked over the references and looked only at items sourced with FairVote (or other advocacy organizations). That it happens to be true that in nearly every example seen, fitting those conditions, IRV was not actually being "used," is not "putting IRV in a poor light." It's just making relevant facts clear. Relevant to what? *To the topic of the article.* No *argument* has been inserted. It's true that this, to some degree, defuses the power of the claim of FairVote that IRV is being widely adopted, and this is naturally of concern to those associated with FairVote, and *this* is a Conflict of Interest, see WP:COI. By the standards I've seen used in the past, COI is blatant here, it is not marginal.

Tbouricius was blocked indefinitely a while back for his apparent association with the cabal of anonymous editors (including the Director of FairVote), sock puppets, and SPA, COI editors, one of whom would be him. Specifically, he registered an account and immediately dove into "protecting" this article, as the cabal was losing power from the imminent blocking or actual block of one of the sock puppets. We can speculate that this was a coincidence, but

At that time, I argued that the block should be lifted, because I thought his participation would be useful for the article. My consent to the unblock was cited in the review and apparently carried weight, for the blocking administrator was adamant that the block should remain; the block was lifted.

It would *not* have been useful *to me* if my motive were promoting some anti-IRV agenda. My concern did, indeed, originally come from outside observation of FairVote, but when that led me to look at this article, I found it full of propaganda for IRV, and criticism was strangely absent. It wasn't strange when I looked at the edit history. Criticism was absent because reverts were being liberally used to keep it out, even when sourced, NPOV, and properly inserted (how about 6RR? -- by the Director). The POV was woven thoroughly through the article, and it was maintained by the cabal. Note that the article still has a POV tag. I didn't put that tag there!

Someone else has set up an RFC over the issue of using FairVote as a "reliable source." The issue here, though, is the removal of sourced facts that are seen as damaging to the subject of the article by a COI editor. I'm hoping that further process won't be necessary to resolve this, for policy is pretty clear on it. Rather, I would prefer to see the cooperation of

Tbouricius in the making of this article into one which could be Featured.

I find it odd that elaborate and extensively explained theoretical examples of the usage of IRV seem to be fine, but real-world results are to be kept out. If what I've presented is "selective," then the remedy would be to balance it out, not to remove it. Results from San Francisco, for example (overall, i.e., how many elections used additional rounds, and was a majority obtained? -- all of which would take a sentence or two), would be very interesting. It's more work, though! Instead of entering into a revert war, how about *improving* the article!

Another approach would be to eliminate non-notable usages from the list of places "using" IRV. A purely theoretical adoption of IRV is good news for advocates, even if it's a fish bicycle for the jurisdiction adopting it; keeping out such non-notable examples would make some more room for notable ones. And election details (i.e., how the method is working) are certainly relevant!

A list of "adoptions" is mere promotion: part of the FairVote plan is to constantly assert "political momentum," so every adoption is treated as big news. If we look at the source on FairVote for "colleges and universities" using IRV, we will find that a certain elementary school used IRV in a mock election. Interesting, perhaps, but probably says more about the politics of the parents or staff in that particular school than about IRV as such. On the other hand, given that only a miniscule number of elections in the U.S. are currently held with IRV or IRV-like methods, it *is* notable in terms of gauging the success of the IRV movement. However, it would similarly be relevant, then, to put in examples of the failure of IRV initiatives, or the removal of IRV by initiative, as happened in Ann Arbor.

That *balancing* facts were consistently kept out of the article is why the article deserved, and continues to deserve, a POV tag. Tbouricius is, at this point, blocking the balancing of the article, and, as a COI editor, he should know that policy requires him to tread lightly, see WP:COI. ---- Abd (talk) 18:34, 16 November 2007 (UTC)

I have placed a 3RR warning on the Talk page for Tbouricius. He has made five reverts, removing sourced material, arguably relevant, not vandalism, within 24 hours, which is a violation of WP:3RR and which can result in blocking, temporary or otherwise. I do not personally plan, at this point, to file a complaint; I am not attempting to get this user blocked. But if this continues, if this COI editor continues to "defend" the IRV article from material he considers negative, I may change my mind, and, at any time, any other editor could file such a complaint based on the warning which has now been provided and failure to respond as necessary. (From the history of this user, it's possible that a warning would not be necessary, but, assuming that he has merely inadvertently violated 3RR, I consider the warning warranted.) ---- Abd (talk) 18:41, 16 November 2007 (UTC)

Removal of controversial items

I've now taken out two of the claims in the Instant-runoff_voting#History_and_current_use section.

The first is the APSA mention. The problem with this fact, as it is, is that this section is mostly about current use (and if we want to make it about history in the U.S., it will get *much* longer), whereas APSA is not, as shown by what I had placed in the article, not currently using IRV, and, quite possibly, *never* used IRV since it was founded in 1903. (Until I started fixing this, the article read that APSA "uses" IRV.) Given the context, removing the claim is probably a better edit than qualifying it to make it NPOV. "Adopted," the language inserted by another editor when the problem was pointed out is language intended to create the same impression in the reader: that political scientists favor IRV and are using it. And that is promotional and misleading.

The second is the university claim. There was nothing wrong with what I had placed in it, and I had compromised by wording the claim so that FairVote could be used as a source. (I.e., it was made into an attributed claim, not stated as a fact. That, in itself, could be questioned, but probably not by me.) However, all these claims of usage have not been verified, and the only ones I had yet researched turned out to be elections where IRV was not relevant, i.e., elections where simple Plurality would have produced the same result. The elections I looked at were major universities, and it turned out that IRV was being used, i.e., the ballots are IRV, but that the extra ballot information was moot. Therefore I don't consider these examples important to the article. The fact that IRV is being used by student associations (that's what it is, mostly) can certainly be in the article, but it should not be one-sided and misleading, overstating the importance of IRV in those elections. So I took it out for now, since my compromise language was not accepted by the reverting editor.

I'm continuing to research the actual elections being run. I'd have added a better reference for Cary, but so far I haven't found official results. A week after the election, it was still in dispute. This election *did* use the additional ranks, all three (but from preliminary results, it seems that the third rank added nothing to the votes). If I had official results, I'd put a very brief summary of how the method performed in the article, the same as I did for other elections. I'm not cherry-picking what I find for some polemic effect.

An editor has raised the question of why I have not mentioned the money saved by avoiding runoffs. So far, I haven't seen any. San Francisco may show such, but I haven't researched San Francisco yet. But the matter of saving money spent in runoffs is questionable. If, for example, we had IRV for major public offices, where party affiliation is important, it's a two-party system, and, let's say, there is a requirement for a majority, and then we have an election, and it goes into the second round, was a runoff saved? Maybe, maybe not. If IRV is the method, then we will see third party supporters voting sincerely, instead of making, before voting, the nearly inevitable compromise that is reflected by the dropping of their candidate in the first rounds and their votes's reduction to their second choice. The effect here was not that a runoff was avoided, necessarily, but that the voters were able to vote sincerely. So, what we can tell from the IRV results is not whether or not a runoff was avoided, though it is possible. What *is* relevant and important, in my opinion, is:

(1) Was the election decided in the first round, i.e., all the other choices were moot? (2) If not, by what percentage of *total* votes did the winner prevail? and, sometimes (3) Are the elections contested? If contested, are there more than two candidates? If the answer to either of these questions is No, then IRV is a complication without a justification.

It could be argued that it's still good to have the system in place, just in case it's needed. But, believe it or not, I'm not arguing about IRV here. I'm just doing research and editing for an encyclopedia, happening to have some knowledge and interest in the subject.

If the article says that the additional ranks were used (as with Cary), and if the facts support it, though, any reader could see that a runoff was avoided. Whether or not it was avoided would depend on the specific rules for runoffs in place. Additional ranks might be used but not change the result that would have come to pass anyway. My intention would be to succinctly report the election results (or a series of election results) so that what the reader would need to know about runoffs would be included, or would be readily accessible.

Whether or not a runoff was avoided is a *conclusion* that would probably require original research, which we cannot do for insertion here unless we can get it published in a reliable source. So, unless it were so blatantly obvious that it essentially is not controversial, or it is published in a reliable source, I couldn't write what the complaining editor seems to want me to write, even if I thought it true. ---- Abd (talk) 21:33, 16 November 2007 (UTC)

Information removed from article, for work

The following information shows elements of the article edited out or reverted out by User:Tbouricius. Typically this editor removed actual election information. Diffs are referenced before each block quote

[7]

• Burlington, VT held its first mayoral election using IRV in 2006 after voters approved it in 2005. Bob Kiss was elected with 48.6% of the vote, 10.6% of the valid ballots being exhausted.

[8]

• The American Political Science Association, in its constitution, provides for instant-runoff voting for the office of President-Elect, if there are three or more candidates nominated. In all years for which records are provided on the Association's web site, there was no more than one candidate nominated, so the instant runoff provision has not been used.

[9]

• Takoma Park, MD adopted instant runoff voting for city council and mayoral elections in 2006. It held its first IRV election to fill a city council vacancy in January 2007 and its second use in November, 2007. In that November election, the mayor ran unopposed, and, out of six ward seats, only one was contested. As write-in votes were less than 11% in any race, and the single contested seat was won with 66% of the vote, the runoff provisions were not exercised.

[10]

[* According to the advocacy organization FairVote, 31 American colleges and universities use IRV in some election or other. For example, the University of California at Davis used IRV in their 2007 elections for the Associated Students Presidential slate; the winning slate won in the first round with 66% of the vote, so additional preferences were moot. At the Massachusetts Institute of Technology, for the Undergraduate Association, IRV is used; in 2003 through 2006, all elections were decided by the first round. (In some elections, because there was a failure to reach a majority, additional rounds were counted, but they did not change the result, and there was still a failure to reach a majority of valid votes cast).]

For the above, the editor, User:Tbouricius wrote as his reason for the edit: "removed ridiculous amount of selective election result details inserted in university entry to advance anti-IRV POV". What might be seen about the material is that the general reference to FairVote has been left in; however, anyone checking the sources will likely find what I found: actual election information is not cited on the FairVote site, just the fact of adoption or usage. To find more takes quite a bit of digging. And what I've found, so far, is that, with all the hoopla about "adoption" of IRV, it isn't really being *used* very much. There are exceptions, I'm sure, I just haven't found any yet that I could cite; I started with examples that were sourced only on FairVote.

If any editor thinks that I have cherry-picked the examples, I'd invite the inclusion here or in the article of examples showing different results. I'm sure that San Francisco, with many contested elections with numerous candidates, would be generating some interesting statistics. Is IRV a "great success" or a "flop"? I certainly can't tell from what I've found -- though you'll find the former on the FairVote site -- but, even if I could tell, I couldn't state it in the article, and I'd probably be reluctant to state it here. That would be POV! --Abd (talk) 01:47, 17 November 2007 (UTC)

• I think he's trying to tell you something along the lines of "facts have a well-known liberal bias". - McCart42 (talk) 04:12, 17 November 2007 (UTC)

Perhaps. He is correct that selective presentation of facts can be POV; however what he seems to miss is that selecting a whole series of facts to make it appear that IRV is having ongoing success, with no mention of either problems or simple irrelevancy of the decision to implement IRV in some cases, is itself POV. This article has a POV tag largely because it was, and still is to some degree, such as with this list of adoptions, a compilation of facts and claims about IRV without balance. I have *not* selected elections that happen to make IRV look useless. I was quite surprised by what I've found. All I haven't done is to do the research for San Francisco. If my facts are out of balance, the remedy would be to balance them, not to delete them. The section is titled "History and Usage." Not "Places which have decided to adopt IRV." What I'm doing is adding usage information, highly condensed, with sources.... In any case, if User:McCart42 thinks the information should go back, I'd appreciate some help with that. I don't want to push the 3RR limit! Part of the whole purpose behind 3RR is to prevent some editor with strong opinions from controlling an article, it's a very rough form of "vote," pending consensus or other resolution. I may do a couple of reverts. --Abd (talk) 05:24, 17 November 2007 (UTC)

The material is back in the article. I reverted it back, then Tbouricius made an edit to the Burlington entry.[11]The edit summary was "Removed POV bias in presentation of facts, though I still think election results are inappropriate at all for this list." If any editor thinks the language of the text is biased in some way, then suggesting alternate language that is without the bias but which conveys the same information is an obvious way to proceed. I find little or no problem with the replacement language. I find it fascinating, however, that one of these statements was considered biased and the other not. This echoes earlier bouncing edits over describing exhausted ballots as "discarded" or "disregarded." Apparently, any direct mention of exhausted ballots -- the technical term one finds in many election reports -- is "biased." This was my edit:

• Burlington, VT held its first mayoral election using IRV in 2006 after voters approved it in 2005. Bob Kiss was elected with 48.6% of the vote, 10.6% of the valid ballots being exhausted.

and it was replaced with this:

• Burlington, VT held its first mayoral election using IRV in 2006 after voters approved it in 2005. Bob Kiss was elected with 48.6% of the vote, to Hinda Miller's 40.7%, with the remainder of ballots expressing no preference between the two finalists.

And I am consterned to find something: the governmental elections page that I had sourced the election results claim from gives actual results; it has been replaced with a page that I had assumed was advocacy, from its appearance and topic. However, looking closer, it appears to be, in fact, a Burlington city site. However, it contains information about Instant Runoff Voting which is essentially standard FairVote propaganda, not merely information for citizens about how IRV works (or, in this case, election results). I'm going to guess that this web site was prepared for the City of Burlington Clerk/Treasurer's Office by IRV advocates. Helpful. Alternatively, the City Clerk's office researched and prepared it. Where would they look? Well, probably at the Wikipedia article as it existed when it was under the control of the cabal. The same information, Robert's Rules, etc. Essentially, what is on that site is not permitted here, and that makes the site undesirable as a reliable source. I'm going to take the reference back to the actual election results page, which does not contain biased information or analysis. Now, interesting question: Why did User:Tbouricius, editing the text, bother to change the source back to the old one, the burlingtonvotes.org source, when that page doesn't have the actual election results on it? This editor is a Vermonter, active promoting IRV there. Did he write the information on that site? I suppose we could find out....

--Abd (talk) 03:30, 18 November 2007 (UTC)

COMMENTS ABOUT GOVERNMENTAL ELECTIONS, SPECIFICALLY BURLINGTON INFO

Its possible that the burlingtonvotes.org site was prepared by IRV advocates. It happened to the site about Cary NC's elections. If you key in caryvotes123 to your browser, it re-directs you to the Wake County North Carolina Board of Elections website, Wake County Gov The problem is that the caryvotes123 domain name is registered to - Robert Richie

--Ask10questions (talk) 07:35, 18 November 2007 (UTC)

Response on cost savings

Proponents are right that if you have IRV, you dont have a runoff election, not usually. But it is arguable as to whether IRV decreases costs to elections departments, and with some consistency.

But look at San Francisco, the largest real life example of IRV in the United States:

San Francisco forecasts doubling their budget in 2007-2008. San Francisco’s higher expenses include special voting software, special poll worker training, more laborious and costly recounts, and IRV related voter education costing about $1.87 per registered voter. Fact: The Department of Election’s proposed $19,809,917 budget for FY 2007-2008 is $10,683,599 or 117.1 percent more than the original FY 2006-2007 budget of $9,126,318. [12]

San Francisco had 418,285 registered voters in Nov 2006. Their current budget is almost 10 times that of Wake County North Carolina's budget with 460,821 reg voters in 2004.

Compare that to Wake County North Carolina, which has been in at under $2 million a year subtracting income from expenditures) from 1999 to 2004. See actual Wake NC Budget [13]

San Francisco may not be paying for traditional runoff elections, but they have new costs. San Francisco’s new expenses include special voting software, special poll worker training, more laborious and costly recounts, and IRV related voter education costing about $1.87 per registered voter. San Francisco is also seeking a replacement voting system, and recently considered spending about $12 Million on Sequoia Voting machines.

Maybe IRV saves money, but there isn't a solid cost savings analysis using actual election department's net annual expenditures.

--Ask10questions (talk) 07:25, 18 November 2007 (UTC) (moved by Abd (talk) 15:23, 18 November 2007 (UTC))

currently, there is some argument in the article, but ultimately serious discussion of arguments for or against IRV should be in the article referenced from this one, Controversies surrounding instant-runoff voting. It's relevant here, however, in Talk, because there are implications to a possible finding that most IRV implementations are not actually avoiding runoffs, which would explain why one editor seems to think including election information in this article is "POV pushing." If most elections aren't using the additional information on the ballots, then the argument that IRV is worth the extra election cost because it saves money for runoffs is ... misleading. But actual argument over this does not belong in this article, and, I expect, we will be moving almost all argument to the Controversies page where it can be addressed in proper detail. Arguments can be *complicated*, what can be presented here is only a few bullet points, and that is difficult to keep balanced. Bullet points on political issues are often *designed* to be imbalanced and misleading.

--Abd (talk) 16:03, 18 November 2007 (UTC)

== there is no proof that IRV saves money. Its like saying that you save money on gas heat by using electric plug in heaters. In fact, if you look at San Francisco, the largest jurisdiction in the US that uses IRV, and that has actually used IRV in elections, its clear that this city spends 5 times what the average same size area in North Carolina, with the same brand and type of voting machines, the ES&S optech eagles.

Now San Francisco plans to double the budget, which will make their costs 10 times that of the average same size area in NC. This claim that IRV saves money must be challenged or removed from this article because it can be disproven. Also, the SOS of Vermont issued a report that shows IRV will increase costs. See http://vermont-elections.org/elections1/2007IRVReport3.8.07.doc

--Ask10questions (talk) 18:14, 22 November 2007 (UTC)

About the exit poll allegedly showing that voters prefer IRV over runoffs

I've pulled this claim from the article. For convenience, here is the diff:[14].

I was going to leave it, just doing a little work on the language preceding it (which was a bit POV) but then I looked at the source, [15] and found that the poll was "managed by Bob Hall, director of Democracy North Carolina." For Mr. Hall and that advocacy organization, see, for example, [16]. In any case, if the source given supports the claim made *in the text of it*, I certainly missed it. It was in the headline, and no claim was made like that in the text. I'm beginning to wonder, how come, so many times, when I follow up on the claims being made about IRV, I find stuff like this? We could chalk it up to editorial ineptitude -- certainly I'm guilty of that myself at times -- but so often?

It's possible that there is a publication elsewhere from this study that actually gives evidence for the claim, but it still would be unusable except *maybe* as an attributed statement. Not as a fact.

If we can keep this up, however, it may become possible to remove the POV tag from the article. --Abd (talk) 19:34, 18 November 2007 (UTC)

Ah, yes, about the STV mention. IRV is not less complicated than STV when STV is used single-winner; indeed, it *is* STV single-winner. Looks to me like someone was trying to make excuses for IRV's complexity, so compared it with *multiwinner* STV, which can be quite complicated because of the fractional vote reallocations used to create proportional representation (and, in my opinion, may be worth the complication). IRV is less complicated than Condorcet methods, maybe. Depends. --Abd (talk) 19:39, 18 November 2007 (UTC)

As I understand it the exit poll in North Carolina was designed and run by Prof. Cobb of NC University, and Hall is a resource for reporters to contact for additional information. As to whether the exit poll supports the statement that voters prefer IRV, the study report summary states that "Of those with a preference, 72 percent of Cary voters said they preferred IRV while just 28 percent said they preferred voting for a single candidate." Abd, just seems to need to find fault with every source that doesn't agree with his world view.

Tbouricius (talk) 17:17, 19 November 2007 (UTC)

First of all, my "need" is not relevant to content. It's a given that, as a strong supporter of IRV, paid consultant to FairVote, and active political proponent of IRV in Vermont, Tbouricius is going to have a different POV from me, and things that may seem perfectly normal and straightforward to him, since he's had them in mind for years, may look skewed to me. However, if we pursue this process, we will find, I expect, that we can agree on what is NPOV. The question here is whether or not the mention was accurate, balanced, NPOV. In reporting a poll, it's crucial what the question was! Now, let's assume that the question was as implied: "Do you prefer IRV, or do you prefer voting for a single candidate?" Right away, I might suspect the numbers reported. They add up to 100%, and I'd vote No on the IRV question (maybe, depends on context) and No on the single candidate question. Now, it's true. People like me, in some respects, are about 0.1% of the population, a fact I've had to live with for much more than fifty years. But that question did not ask about, and therefore the reported results did not show, the preference claimed in the article. It was Mr. Bouricius' assumption. Or maybe it did ask that. We have to look at the actual poll! And then we might also ask whether such a poll is a reliable source. Published? Original Research? But we don't need to go there on this one. We don't have the study.

Now, the claim has been replaced with something better sourced. This is a copy from the source:.

In a separate question, respondents were asked whether they preferred "this system with no December runoff election," whether they preferred "the former system with a December runoff election," or whether it made no difference to them. A majority of respondents to our exit poll (61%) preferred the new system; 13% said they preferred the runoff system, and 27% said it made no difference.

Now, does this mean that we can state, as the article currently states, "IRV may also be less likely to induce voter fatigue, and exit polls indicate voters prefer IRV over two-round runoffs"? No. First of all, there is no source for the statement that IRV may be less likely to induce voter fatigue. The whole concept of voter fatigue is a debatable one; it could be argued, for example, that runoffs are a kind of rough Range Voting: the votes of voters who really don't care are approximately excluded because of differential turnout. Is this an improvement? As I'm saying, it's debatable. I was going to leave it in until I wrote this.... on reflection, it is argument and, as such, must be sourced. So I'm taking it out. As to the exit poll, the report is overstated. This is one study, one set of exit polls, one election, one particular jurisdiction with one particular set of election problems. However, it *is* notable, and so I'm leaving it in with proper qualification sourcing it. It is not sourced such that we can state it as a fact (about voters), rather it is an assertion or conclusion from one study, not confirmed in a reliable source. So I'll fix that so that what is stated is inarguably a fact. As to voter fatigue, we can talk about it. But I don't think we can make up statements like that and then put them in, even if we *do* agree on it.

Am I nitpicking? Perhaps. But one person's nit is another person's serious lice infestation, or the beginning of it. It can take nitpicking to negotiate consensus language between opposing POVs. If we *don't* do this, someone, later, will, or, worse, will think that the article is POV even if they can't put their finger on why. --Abd (talk) 20:35, 19 November 2007 (UTC)

North Carolina

House Bill 1024 defines "instant runoff" as follows:

"Under 'instant runoff voting,' voters rank up to three of the candidates by order of preference, first, second, or third. If the candidate with the greatest number of first‑choice votes receives more than fifty percent (50%) of the first‑choice votes, that candidate wins. If no candidate receives that minimum number, the two candidates with the greatest number of first‑choice votes advance to a second round of counting. In this round, each ballot counts as a vote for whichever of the two final candidates is ranked highest by the voter. The candidate with the most votes in the second round wins the election."

Isn't this Sri Lankan contingent voting, rather than instant-runoff voting? Markus Schulze 16:09, 19 November 2007 (UTC)

People who actually know about election methods can be remarkably touchy about the names of things, and Markus is indeed an expert. One of the problems with the Instant Runoff Voting article is that "IRV" is a coined term, *recently* coined at that, and it is used to cover all kinds of systems varying in details. In U.S. history, the method has been called various names, what I've seen is "Preference voting," as in Ann Arbor, Michigan, in 1974. When what eventually became FairVote set out in the early 1990s, I think it was, to promote proportional representation in the U.S., a tactical decision was made by a small number of activists to first promote the single-winner form of STV, and, as a promotional device, they apparently decided that people here were familiar with runoff elections, at least in many places, so naming it "Instant Runoff" made it sound less strange, glossing over the differences (some good, some not so good) between IRV and top-two runoff. But then the full ranking used, and even compulsory in some places, for STV was considered impractical here. And so we see truncated ranking. The version above is what is called Batch Elimination in this article.

(STV is quite a good method for proportional representation; single-winner, well, not so good. Most election methods advocates support the goal of proportional representation, using one method or other, but object to collecting the data needed for a decent Condorcet discovery and then passing over an ignoring it, in order to use the sequential dropping which was originally invented for proportional representation, by Hare, and which has problems with single-winner, and I have found references from about a hundred years ago pointing out that Ware's method, they called it was problematic, for the same reasons that election methods experts know today.)

There are, from my own examination of the literature and characteristics of these methods, two basic classes of "IRV" elections: those which fail when a majority of voters don't vote for the plurality winner (after transfers), and those which go ahead and declare this candidate a winner even though a majority of voters voted against that candidate. Some implementations or proposed implementations in the U.S. are the former kind and some are the latter. Robert's Rules of Order, as has been stated in the article, suggests that if repeated balloting is not practical, with a mail election, that preferential voting may be used, and then it gives, as an example, a method which is sequential dropping, like IRV, and which is so close to what we think of as IRV that I thought it *was* IRV until I realized the difference. The Robert's Rules version does not pretend that there is a majority winner from the majority artificially created by dropping all the other candidates but the top two. And so the Robert's Rules method, unless there is a bylaw specifically permitting election by a plurality, will fail unless a majority is *actually* created by vote transfers.

This detail, in fact, makes IRV much less objectionable as a method, since it won't violate majority rule. I'd say that if a majority of voters have voted *against* a candidate, no matter who they voted for, democracy and majority rule would require that this candidate not be elected, period, unless enough voters change their minds to make it happen, or new voters vote in a new election. But then IRV can't be sold as avoiding runoffs (though, in fact, it would still reduce the need for them.)

So, basic problem: there are actually a series of methods of preferential voting, not just one called "IRV," such that with a list of places which have implemented "IRV" we have a series of successes for one method. What we have is different methods in different places, and I haven't seen the information gathered in one place yet detailing the exact details in each place and how they match and differ.

And then there is the fact that there are other methods which could also be called "instant" and "runoff," such as Bucklin voting. Bucklin proceeds in rounds, like IRV. In the version found unconstitutional in Minnesota in what really should come to be seen as a scandalous decision (it's worth reading, Brown v. Smallwood)[17], there were three rounds. Voters were only allowed to vote for one candidate in the first and second rank, and for as many as desired in the third rank. If there was a majority in the first round for a candidate, that candidate won, with no attention paid to subsequent rounds. Thus Bucklin satisfies the Majority Criterion, no matter how we slice it. However, if there is no majority in a round, the next round votes are added in. Thus the election becomes Approval voting. But it was actually called, as I recall, "preferential voting," and it appears it was very popular when it was killed by the Court, which was also apparently contrary to the bulk of legal opinion at the time. Politics.

This wanders, I'm sorry, but I find it interesting that Bucklin was called "preferential voting," and the court specifically ruled against allowing additional preferences to be expressed. FairVote has claimed that Brown v. Smallwood does not apply to "instant runoff voting," but that argument is clearly not based on a careful examination of the case, and lawyers have issued opinions that IRV is contrary to the precedent of Brown v. Smallwood, and I agree. I also think it is a corrupt precedent, ripe for being overturned, and, in this case, I'm encouraging Approval supporters to aid in that, when, as seems likely, this is tested soon. --Abd (talk) 21:10, 19 November 2007 (UTC)

"doesn't guarantee a true majority-supported winner" con

Can someone explain what voting system does guarantee a true majority supported winner? The current U.S. presidential election system certainly doesn't. Having trouble seeing why this is listed as a "con"... --WayneMokane (talk) 00:33, 22 November 2007 (UTC)

We have a little wild Elf named Abd who is upset by a claim that IRV guarantees a majority-supported winner on the grounds that some voters may not rank enough candidates to support one of two candidates in the final instant runoff round and lose their vote, and so a winner may still have less votes than half the ballots. So it is a con because it is a claim judged false - IRV can't promise a true majority. You see this Elf likes Approval voting which offers no guarantees at all that a singular winner can be found with majority support, so in order to compensate, he has to make sure everyone knows that IRV might also suffer this fate. SERIOUSLY, there's a reference link if you want to read an external source. It all comes down to the fact that majority supported winner is the only real selling-poing of IRV, and so naturally it is attacked by those who don't want IRV.

IN REPLY, a two round system and blanket primary ALSO both promise a majority-supported winner, IF everyone are willing to vote twice, since with two choices you're guaranteed one will have more than 50% unless there's a tie. Tom Ruen (talk) 01:00, 22 November 2007 (UTC)

Looks like all the linked page says:

• IRV DOES NOT REQUIRE A MAJORITY. The 2000 London mayor won with a 45% plurality.
• Rebuttal: ALL VOTERS CAN PARTICIPATE IN THE RUNOFF, SO MORE PEOPLE ELECT THE WINNERS. If there's no immediate majority winner, second and further choices are counted. The winning candidate usually gets more than 50% of the votes in the instant runoff.

Tom Ruen (talk) 01:05, 22 November 2007 (UTC)

Please treat other Wikipedia editors with respect. Name-calling is out of line; stick to discussing the content of the article. Thank you. Ka-Ping Yee (talk) 08:53, 22 November 2007 (UTC)

In the scientific literature, the desideratum, that the winner should always be a candidate who can win a majority of the continuing ballots in a runoff with some other candidate, is usually called "Condorcet loser criterion". Instead of saying that the task of a runoff is to guarantee that the winner is "a majority winner", it makes more sense to say that the task of a runoff is to guarantee that the winner is "not a Condorcet loser". Instead of saying that IRV always finds a majority winner, it makes more sense to say that IRV satisfies the Condorcet loser criterion. However, because of well-known reasons, IRV supporters usually consider the term "Condorcet" to be pejorative.

This is also the reason why some people say that the contingent vote and the two-round system always choose a "majority winner". What they mean is that these methods never choose a Condorcet loser.

Tom Ruen wrote: "Looks like all the linked page says: IRV does not require a majority. The 2000 London mayor won with a 45% plurality." Actually, the London mayor is elected by the Supplementary Vote and not by IRV. Markus Schulze 17:04, 22 November 2007 (UTC)

Hi Markus - thanks for pointing this out, the Condorcet loser criterion and that the Supplementary Vote is used - which only allows 2 rankings, making it more likely to have exhausted ballots. Runoffs can guarantee a majority winner in the final round from remaining ballots, but not from original ballots. That's the meaning of the con argument. Tom Ruen (talk) 18:38, 22 November 2007 (UTC)

The question has been asked if any voting system guarantees a majority winner. We have to ask, first, what a majority winner is. In my view, a majority winner is one who will be elected if the electorate is presented with the question, in isolation, "Shall this candidate be elected," and a majority answer affirmatively. So election by ordinary motion, as is allowed by Robert's Rules as one option, guarantees a majority winner or there is *no* winner. Again, many election methods guarantee that *if there is a winner*, that winner will be a majority winner, and this is common, actually, in the U.S. Plurality satisfies "majority winner" (considering this informally as a criterion) if there is a requirement that the winner gain a majority. By the same line of thinking, top-two runoff satisfies "majority winner." However, the problem is that the method ordinarily does not provide an opportunity to expressly vote *No* on the question of "Shall any of these candidates be elected?" For this reason, contrary to what Mr. Ruen has written, my favorite voting method is not Approval Voting but full-on deliberative process. My support for Approval Voting is based on its simplicity, the trivial cost of implementing it, and its reasonable performance (in my opinion, better than IRV). Like IRV, it cannot guarantee a "majority winner." No method can force a majority to approve a candidate and have that really mean approval. --Abd (talk) 00:32, 23 November 2007 (UTC)

Instant Runoff Voting, if a true majority is required (i.e., after vote transfers, the winner has received a vote from a majority of all valid ballots cast, not merely of ballots remaining in the last round), also satisfies "majority winner." Essentially, the only problem with IRV in this respect is when a winner is declared based merely upon a majority in the last round, that is, there remain only two candidates, and one of them has more votes than the other, but there are other ballots which are exhausted. Attempting to maintain the claim that IRV "guarantees a majority winner," some IRV advocates seem to blame the voters who cast the exhausted ballots for having "failed" to rank those remaining candidates, but the only thing we know for sure about them is that they did not sufficiently support those candidates to rank them. Essentially, we have to think of these votes as being an answer, "No," to the question of electing them. And thus it can come to pass that, absent a majority election requirement, IRV can declare a victory for a candidate ("by a majority in the last round") when a majority of voters have actually expressed a vote in opposition to that candidate. Essentially, in this, IRV is quite like Plurality or Approval or any other method. No *method* can guarantee that a majority of voters will *actually* approve the election of a candidate.

Because the default under Robert's Rules is that a majority is required, the method of preferential voting that is described in Roberts Rules is one which does require, for victory, a true majority. Or it will fail to produce a winner. Absolutely, if all voters rank all candidates, IRV does guarantee a true majority winner. In some jurisdictions, election rules require voters to rank all candidates. To my mind, this is fundamentally undemocratic, it is requiring voters to vote *for* a candidate whom they may actually despise, to quite possibly contribute to the victory of that candidate. If there is a ballot with, say, options to vote for Jimmy Carter, Adolf Hitler, and Genghis Khan, I am *not*, if not forced, going to vote for one of the last two, I will bottom-rank them both. And if there is an actual runoff election between the last two, I'm not going to vote at all, I'm going to do what I can to leave the place.

So details matter. In any case, most IRV implementations in the U.S. use quite limited ranking, so ballot exhaustion is practically guaranteed if there are enough candidates. --Abd (talk) 02:01, 23 November 2007 (UTC)

Abd's rephrasing of the Con about IRV not guaranteeing the election of a true majority winner is confusing. By using the phrases "pass-over" a candidate with "majority support" he leaves the false impression that a candidate might have over 50% of the vote (first-preferences?), but be passed over for some other candidate with less support. The brief headings for these pro and con arguments should summarize the claim and not embellish it with implied arguments that are not actually being made. the previous "does not guarantee..." wording was concise and fairly summarizes the argument.

Tbouricius 20:00, 2 December 2007 (UTC)

Part of the confusion here may be that a more powerful argument is being substituted for a less powerful one. As others have pointed out, no voting system can "guarantee" a majority winner -- though I think I have seen exactly that wording of an argument for IRV. However, there are systems where, if a majority prefer a candidate over another, expressing that on the ballot, the second candidate cannot win. Any Condorcet method does this. IRV does not. Now, it may take some time to find an argument out there that is exactly on point. But this issue is one of the most common criticisms of IRV among election methods experts. Tbouricius is concerned that the argument is "confusing." But arguments are *often* confusing, sometimes, even, in real political contexts, that is their very purpose. The article itself, in stating facts about IRV, should make clear what IRV does and does not do. Further, the difficulties here are a fairly good argument for dumping the Controversies entirely from this article and putting all that effort into the Controversies regarding instant-runoff voting article, where arguments and facts can be juxtaposed, and where there is the space to attribute every opinion. The constraint in the Controversies section of this article is that it should be brief.

It's not simply an *argument*, and we could state it as a *fact*, that IRV can pass over a candidate who would beat all others in pairwise contests, a "compromise" winner, and this happens to be explicitly stated in Robert's Rules, talking *specifically* about IRV, not about preferential voting in general.

Certainly the wording might be better. But the fact stated is true, the argument is made. I'm working on the Con arguments at this point, sourcing them better (I just added a sourced Pro argument). I'm likely to change the wording yet again.

This particular section might be confusing to a new editor. Normally, arguments regarding a controversy must be attributed. In order to make this section read more clearly, that is not being done; this is why there is a disclaimer at the top. These are *arguments*, not *facts.* All this can be avoided if we eliminate the Pro and Con sections entirely and refer to the Controversies regarding instant-runoff voting article, where every opinion is attributed and sourced (or at least that is the goal; certain very well-known arguments are not yet sourced there, I think. As are many Wikipedia articles, it is a work in progress. Editors will, by consensus, allow unsourced facts that are broadly accepted; when there is a reasonable challenge to the verifiability of these facts, then it becomes necessary to source them. Writing an article can be done much more efficiently if the writer can simply write what the writer knows, particularly if the writer is an expert. Adding references is a big task; but anyone can do it here, or can challenge suspicious "facts."

However, challenging facts that are common knowledge on the basis that one does not like the implications can be disruptive. Rather, if something needs a source, adding a citation-needed tag is reasonable.

--Abd 02:58, 3 December 2007 (UTC)

I restored the simple clear version of the argument without the miss-leading phrases like "pass-over" and the miss-leading statement that implies some ballots are simply disregarded (rather than some rankings on some ballots, where the voter's first choice is still in the running, are not used, etc.) The phrasing proposed by Abd creates a false impression to most Wikipedia readers about the nature of this CON argument. Nobody is arguing that a candidate with a majority of first preferences is passed-over in favor of a weaker candidate, nor that some valid ballots are simply ignored...yet that is the impression that Abd's choice of words leaves with most readers. His summary of the issue does not capture the thrust of the main claim that exhausted ballots can lead to sub-50% winners, nor even the much rarer claim that candidates with much less core support, but broader secondary support, may get eliminated.

Tbouricius 17:57, 3 December 2007 (UTC)

Negative campaigning Con argument

A newly registered user rewrote a Con argument ("Does not reduce negative campaigning") to dilute it, allegedly on the basis of what was "actually said" in the referenced article. However, that article is only referenced as one example of this argument being presented in the public debate over IRV. This section on arguments, like the article, Controversies regarding instant-runoff voting, is about *arguments*, and arguments are presented without weasel words, generally. The source itself, at least as I read it, is quite a bit stronger than this editor seems to claim. Here is a relevant quote from the source:

2) Negative campaigning. It was a vote winning argument for Proposition A but it's a myth. [...] But there is no evidence IRV is stemming the flow of hit pieces. In the 2004 supervisor contests Districts 1, 3, 5, 7, 9 and 11 all witnessed negative campaigning.

Some negative campaigns used robo calls, others used the mail while others used both. The District 1 race broke new ground with negative street signs. In this year's contests, there were hit pieces.

Negative campaigning is basic to our politics and no voting method is going to eliminate it.

I think a fair summary of this is: "did not reduce negative campaigning," and, given that there is no evidence I have ever seen that it actually reduces negative campaigning, but this appears to be only a theoretical idea that sounds reasonable -- typical of many political arguments -- it seems quite reasonable that there is a Con argument to match the Pro one.

Remember, arguments don't have to be true to be reported here. It's enough that they are actually being made, that they are relevant and reasonably notable. We would not allow straw man arguments to be presented here, but we can include, for example, even blatantly false arguments if they are actually being made in the public arena. This is why we have the warning that some of the arguments may be false or misleading. The inclusion of that warning could be controversial: however, including it made the text much simpler; without it, every statement would have had to have been accompanied by attribution or similar framing.

I reverted the change. --Abd (talk) 05:09, 26 November 2007 (UTC)

Why is "does not reduce" a fair summary? Are you asserting that if "there is no evidence" that x is true, then it must be that x is not true? Is there anything in the article that compares the amount of negative campaigning to previous elections, which would allow us to say that the amount changed? As it stands, the source does not support the statement. MilesAgain (talk) 12:31, 26 November 2007 (UTC)

(unindent)I appreciate that User:MilesAgain has explained the action here. However, these are arguments, not "facts." First of all, the article cited clearly is making the point that IRV has failed to reduce negative campaigning; it also states that it expects no election method would do that. The claim about reducing negative campaigning is substantially promotional hype, the kind common in political campaigns, where both sides will sometimes dredge up any reasonable-sounding argument. I personally find it likely that IRV (like just about any form of preferential voting, even the additive forms like Approval or Bucklin) would reduce negative campaigning, but only the relatively minor negative campaigning that takes place between relatively aligned candidates. Not the most offensive and serious negative campaigning that takes place between deeper rivals.

That IRV will not substantially reduce negative campaigning does not make it a bad method. It simply means that *this particular argument* is misleading.

Secondly, this is, as this editor later noticed, a section about arguments. When arguments are made in the public media, rarely do we see them qualified with weasel words that are necessary to make them, arguably, 100% true. For example, the Pro argument is stated with the weasel word "may." Absolutely, the Pro claim is true, with "may" in place. And still misleading, since it is unlikely to reduce serious negative campaigning. In real political campaigns, the weasel word will be eliminated. Imagine a bumper sticker or a headline on a promo piece. Will it say "IRV May Reduce Negative Campaigning?" I don't think so. Much more likely, it would say something like: "IRV: Get Rid of Negative Campaigning." Ideally, we should find some actual references, and if, indeed, all of them have the weasel words, then it would be absolutely appropriate to have them in place. Otherwise, what should really be there is something like: "IRV reduces negative campaigning." But I'll wait to find some specific sources.

It's very important in the Controversies section to present actual arguments, as they are actually made, not diluted with weasel words by editors seeking to make them more palatable to nitpickers. The question of balance (what if some arguments are accurate and some are hyperbole, presented by different proponents and critics?) is one we will need to work on or negotiate. We have an active proponent of IRV, officially connected with FairVote, as an editor of this article; actually we have two of them. I don't think they are going to allow the article to swing to anti-IRV NPOV, and, indeed, I did what I could to make sure that one of them was able to return to editing this article when he was banned. My goal is that the article be interesting, informative, and truly balanced, which isn't easy in the presence of strong feelings about the issue, people politically committed to a cause. But I think it can be done, and, in fact, it's happening.

Third, the sources cited for arguments are simply examples that have been found. Theoretically, we *might* be able to include arguments made here simply on the basis that, clearly, at least one editor espouses them. But that gets to or crosses the boundary of notability; in any case, if there is any doubt, facts should be sourced, and that an argument is being made is a fact. The distinction between facts and opinion is quite important: an opinion is still a fact, but is only a fact if attributed to the holder or the one who expressed it.

Lastly, the editor here is newly registered but shows marks of being quite familiar with Wikipedia, or of being an SPA. There are lots of reasons why new editors may still be legitimate, but this article also seems to attract sock puppets and meat puppets with regularity. As is clear from Wikipedia policy, edits by sock puppets are not rejectable merely because they come from a sock. But if I'm convinced that an editor is a sock puppet -- or if the suspicion is strong and reasonable -- I may be much freer with reverts than I would with a known and established editor, I may (and have on one occasion) crossed the 3RR boundary, and I've been sustained in that action. In theory, all editors have the privileges of administrators, just not the buttons. And just like an administrator, I might myself be blocked if I go too far. --Abd (talk) 16:14, 26 November 2007 (UTC)

You say, "the article cited clearly is making the point that IRV has failed to reduce negative campaigning," but where do you get that? All it says is that there isn't any evidence that it did. In your extremely lengthy reply, you don't actually support your point, you just "clearly" handwave. Are you accusing me of being a sock-puppet because you have no valid argument? And yes, I am a legitimate sockpuppet under the rules and I have a good reason for being one which I do not intend to share. I will not revert you again: I ask that others look at the source and decide for themselves. Look what the article also says: "Negative campaigning is basic to our politics and no voting method is going to eliminate it." It's an OP-ED, for goodness' sake! Op-eds are explicitly not allowed as sources; you know that, right? MilesAgain (talk) 16:37, 26 November 2007 (UTC)

(1) I quoted from the article, and apparently, I must do so again. "Negative campaigning. It was a vote winning argument for Proposition A but it's a myth." I find this crystal clear, particularly considering what else was written -- but even without it -- that the author is saying that IRV did not reduce negative campaigning, or at least not significantly.

(2) I did not accuse User:MilesAgain of being a sock puppet. Read what I wrote. It's right above! I noted marks of sock puppetry and that I'm less restrained in reverting sock puppets, but I also indicated that sock puppet-provided content must be judged on the content, not on the identity of the user. If I engage in an edit war with a sock puppet, I can still be sanctioned under some conditions, particularly if the sock puppet is following Wikipedia guidelines in the edits. And the editor confirmed that he or she is a sock. And, yes, this *can* be legitimate.

(3) Op-eds are not allowed as sources for fact. However, an op-ed, in this case, *is* the fact. Thus referencing the op-ed allows the reader to personally verify the *fact*, and that ability is the core of reliable sourcing. I.e., the argument or claim is actually being made, it's not a straw man argument or something simply made up by an editor here. This is a section about *arguments*, and arguments can be completely false, what matters is whether or not they are being made. Now, if this were just an isolated argument, we would simply attribute it in the text, i.e., "So-and-so argued that...." But that would make this section cumbersome, which is why, instead, the whole section is preceded by a warning that the arguments aren't facts, they are claims, and may be totally false or merely misleading.

(4) This section on Controversies is a shadow of the full article on them, and it is intended that it be brief, and not an attempt to explore them in depth. That happens in the main article on Controversies. Only the most common arguments are presented here, and it's important that there be balance. Since the claim that IRV reduces negative campaigning is in the Pro section, the counterclaim is quite appropriate for balance. If that particular source did not exist, we would still need to deal with the imbalance to avoid a POV effect. It should be understood that this has been, to some degree, a negotiated settlement between pro-IRV editors and critics. Quite possibly, this entire section, except perhaps for the most significant arguments, will be eliminated in consideration of the existence of the main article on Controversies. I have *not* attempted to establish as a fact in the article that IRV does not reduce negative campaigning, nor has any source been cited to show that it *does*, only that the claim and counterclaim exist. Balanced. NPOV.

(5) Ah, yes. The editor here is repeating arguments used in the past by a sock puppet, who was ultimately blocked indefinitely for the activities. That may or may not be a coincidence.

--Abd (talk) 02:13, 27 November 2007 (UTC)

I removed this Con bullet. If someone claims in an op ed that IRV does not significantly reduce negative campaigning, that is an argument that it is neutral on that issue. It is a NEUTRAL rather than a CON. An argument that IRV actually INCREASED negative campaigning would be a CON appropriate for this list. A failure to IMPROVE is not a CON. IRV also does not assure peace and prosperity, but that is not a CON argument. Tbouricius (talk) 15:44, 28 November 2007 (UTC)

Were there no Pro argument "IRV reduces negative campaigning," yes, there would be no Con argument as described. But there is one. The Con argument is the negation of a Pro. Now, I could remove the Pro argument, but it is, in fact, an argument made. We don't have to have all arguments made, full detail could be in the Controversies article, where the arguments are presented and considered. In that article, the counterargument to the Negative claim would be in the discussion of it, most likely. But here, there is just a list of arguments, and having the Pro argument without the corresponding negation of it -- which is also being asserted in public discussion of IRV -- is imbalanced and thus POV in effect.

--Abd (talk) 06:04, 29 November 2007 (UTC)

So, by Abd's logic we could look at the CON arguments and add a PRO argument that simply is the negative of each of those, and do the same for the Cons. In other words we could add these PRO arguments:

• takes no more effort for voters compared to Plurality or some other systems;
• avoids forcing voters to spend additional time and effort to re-evaluate candidates who may change their campaign message in an actual runoff;
• guarantees a true majority-supported winner more fully than would a separate runoff with reduced voter turnout;
• is not significantly more expensive to count than Plurality voting or Approval Voting, requiring only minor changes to vote counting procedures or voting equipment

That seems silly. But I could add those and find sources of "argument" that they are true.

Tbouricius 15:26, 30 November 2007 (UTC)

The source says, "there is no evidence IRV is stemming the flow of hit pieces." But ABD says that source supports that "opponents of IRV argue [that it] does not reduce negative campaigning." Well, there's no evidence for that either. This is the classic argument from ignorance: "absence of evidence is not evidence of absence". MilesAgain 15:56, 30 November 2007 (UTC)

As to Tbouricius's comment, what may seem silly to one, may seem necessary to another. Consensus is found when we are considerate of each other's views and perspectives. We have addressed the problem in the Controversies regarding instant-runoff voting article. Arguments are classified as Pro and Con based on a "side" which most frequently asserts it. So the negative campaigning argument is a "Pro" argument. Then each argument is discussed in a neutral manner, so the counter argument would be presented there. Since, in my opinion, the negative campaigning argument is really a silly one,if we want to get rid of "silly," I'd be quite comfortable with simply removing all of it, both the Pro and the Con. What should be in *this* article is only a few major arguments, balanced. The balance is necessary with any true controversy.

as to MilesAgain's argument, it seems to be difficult for him to understand the issue. "Reduces negative campaigning" and the mirror "Does not reduce negative campaigning as claimed" are not facts to be established by evidence. They are arguments or opinions or conclusions; another way of putting this is that the "fact" is that the argument exists. Sure, the *wording* might be tweaked. MilesAgain doesn't like the statement "does not reduce negative campaigning." He's right, that's stated as a fact and it should be changed, which I now intend to do.

The *fact* is that a named source has argued and claimed that IRV reducing negative campaigning is a "myth." And more detail is given. Further, when we look at it, there is utterly no reason that I can think of, or that I have ever seen asserted by IRV proponents, as to why IRV would stop truly negative campaigning, the kind that takes place in campaigns between truly opposing candidates, which is also the kind that people truly dislike. The Negative Campaigning argument is a pretty good example of political rhetoric, that is, an argument which can sound reasonable as long as the target audience doesn't think too much about it or actually investigate it. I have *never* seen any evidence asserted that it is actually true, which has nothing to do with whether or not it's a good idea to use IRV. It's not actually a major argument for those who understand the issues. --Abd 20:19, 30 November 2007 (UTC)

I would tend toward eliminating the entire Pro and Con argument here. Against this idea would be the fact that IRV advocates *do* stress this argument, and without the weasel words that were once a part of this articles mention of this.

"IRV decreases "negative campaigning" because candidates need to address the majority this increases the attention to the "real" issues at hand." irvwa.org

"IRV opens politics to new candidates and their ideas, increases political debate, and even discourages negative campaigning as candidates try to win rankings from the supporters of their opponents." "Instant Runoff Voting Is Catching On" By Steven Hill see fairvote.org/about_us/stevenhill

"IRV results in a more positive campaign and better information for voters because candidates have an incentive to appeal to their rivals' supporters to be their 2nd choices. Candidates will be less likely to distort the differences with a rival and will be more likely to say how they are similar to a rival." fairvotemn.org

"One-on-one races in traditional runoffs usually get nasty. Instant Runoff Voting reduces negative campaigning. Because candidates want the second choices of their opponents, they are less likely to attack their opponents with negative campaigning. Instead, IRV promotes more focus on platforms and real issues." [http://www.sarasotairv.com/whysupport.php sarasotairv.com

I get 1400 Google hits for "instant runoff voting negative campaigning." I've only quoted from the first few, above. I have before seen official city publications that made the same claim about IRV. Apparently, the promotional claim has been widely accepted as true. But I have *never* seen evidence that it actually happens in reality, nor would I expect IRV to have *much* effect on negative campaigning in major elections. Maybe in minor, nonpartisan elections, though: there, I would expect a reduction in negative campaigning, if any, to come from rivals who are really quite similar, attracting the same set of voters. What I've written above about this issue is what I continue to think to be likely true: this is a piece of political polemic, invented to make IRV sound better, not based on any actual experience. It's notable, obviously, but quite POV if not balanced, and see how much trouble it has been to keep the counterclaim in the article! I'd be happy to relegate it all to the Controversies regarding instant-runoff voting article, where the presentation can be much more thorough and clear, and where the Pro argument is presented and the denial of its cogency -- as well as any evidence on the matter -- is immediately discussed, not presented in some different Con section. --Abd 22:03, 1 December 2007 (UTC)

Reversion of my edit re "voters have one vote."

I edited the introduction to change this:

Instant-runoff voting (IRV) is a voting system used for single-winner elections in which voters have one vote, but can rank candidates in order of preference. In an IRV election, if no candidate receives a majority of first choices, the candidate with the fewest number of votes is eliminated, and ballots cast for that candidate are redistributed to the surviving candidates according to the voters' indicated preference.

to this:

Instant-runoff voting (IRV) is a voting system used for single-winner elections in which voters may rank candidates in order of preference. In an IRV election, if no candidate receives a majority of first choices, the candidate with the fewest number of votes is eliminated, and ballots cast for that candidate are redistributed to the surviving candidates according to the voters' indicated preference.

[User:Tbouricius|Tbouricius] reverted it. I have reverted back, and will here explain why. Meanwhile, this editor has been previously warned about contentious editing, due to his Conflict of Interest. I request that he refrain from it. Reverting a change, not vandalism, without discussion is not appropriate for such editors, who are generally advised to *discuss* changes likely to be controversial before making them.

The comment about the voter having one vote is an interpretation. Yes, it is parallel to the "single" in Single transferable vote," which is identical to IRV, the latter being the application of STV to single-winner elections. However, it is utterly unnecessary in the introduction, the "one vote" adding nothing to the explanation that is not already clear. Why is this important? Several reasons:

(1) This is the introduction, which should be concise.

(2) The Minnesota decision that outlawed Bucklin voting, Brown v. Smallwood, hinged on the analysis of the court that "alternative votes" were a violation of one-person, one-vote. Legal opinions have been issued that this applies not only to the Approval voting-like characteristics of Bucklin, but also to the alternative votes in IRV. (I don't agree that the legal reasoning in Brown v. Smallwood was correct, neither Bucklin nor IRV, in my view and apparently according to the majority legal opinion of the time, violated the Minnesota Constitution. Nevertheless it *was* true that some voters cast "more than one vote," and this gives them some possible advantage over voters who only cast a single, first preference vote, and this was the reasoning of the court majority in the case mentioned.

(3) The voter, with IRV, may literally cast more than one vote on a single ballot for a single office. That only one vote is considered at a time does not contradict that. Thus if the statement is going to be made that the voter "has only one vote," it would need to be more accurately specified, since, clearly, the voter may cast more than one. I considered doing that revision, in response to the reversion, but since the phrase is entirely unnecessary except as an attempt to intercept a conclusion by the reader that IRV violates one-person, one-vote, I decided against it. That would be argument disguised as fact, which should be avoided and which I intend to avoid even though I agree that IRV does not violate the 1P1V principle. The Court in Brown v. Smallwood, quite clearly, would have disagreed with my view (as it did with similar views expressed in the dissent filed in that case and by petitioning lawyers). --Abd 06:14, 1 December 2007 (UTC)

This is wildly senseless to me. Obviously supporters of IRV care to make it clear this runoff system is a one vote system. It is perhaps the only "preference ballot" single winner method that only counts one vote per ballot. It is the first question people will ask when they are told they can rank their preferences - what does it mean. Saying "One vote" says quickly what the following text explains in further detail.

I'm adding the phrase back. I believe it is wrong to remove something that is correct because you don't like it and expect it to remain removed simply on your thoughts. Let it stay UNTIL there's a consensus it is not helpful. Tom Ruen 07:01, 1 December 2007 (UTC)

Sigh. First of all, Tom Ruen hit it on the head why this change, so minor in appearance, is considered so important that one clear COI SPA editor, Tbouricius, fights for the status quo over this detail even though he's been warned (not only by me, but by others), then is joined by another editor whose status is more marginal: Tom Ruen, while not an SPA or sock, is "affiliated" with FairVote. FairVote supporters, up to and including the Executive Director, Rob Richie, largely created and maintained this article, secure from criticism, for a long time. As I've written many times, subtle POV bias -- and much that was not so subtle -- was woven through the article in many places. We have eliminated *most* of the obvious stuff, and it's now possible, even, to include criticism of IRV in the article; previously, sock puppets and Rob Richie made sure that critical information was kept out, until they were finally blocked indefinitely. We are still seeing likely sock activity here and in related articles, AfD's, etc., but whoever the puppet master is, it's no longer so serious.

This is background. The core point here is that the introduction is a critical part of the article, to advocates, critics, and to encyclopedists as well. The latter would want it to be a concise summary of the most important points a reader needs to know, so that the reader is situated to read the rest of the article, and, at the same time, won't go away with any false or misleading impressions. We often don't require references for claims in the introduction; the trade-off for this, though, is that claims in the introduction should be thoroughly non-controversial.

Is the claim uncontestable that IRV is a system in which "voters have one vote"? I presented evidence, above, that it is not. Certainly *there is a way of looking at IRV* that can be expressed in that way. But there is also another perspective from which voters have more than one vote, and that perspective happens to be enshrined in an unfortunate Minnesota Supreme Court decision. Ruen, I'm sure, must know that IRV is almost certain to face legal challenge with the Minneapolis implementation because of the precedent. FairVote and its allies have issued opinions that Brown v. Smallwood does not apply to IRV, based essentially on the one-vote-at-a-time argument, but B v. S was fundamentally based on the very idea of expressing more than one vote on a single ballot. It was about "preferential voting." If we look at an IRV ballot for a single-winner race, we will see, quite simply and clearly, that the voter may express more than one vote on it. Absolutely, the counting process ensures that only one vote from each voter -- at most -- is considered in each round. As what turns out to be a debatable and complex point, "one vote" should not be mentioned in the introduction. It is a *debate* point. Hence I am once again taking it out.

Tom Ruen wrote "I believe it is wrong to remove something that is correct because you don't like it and expect it to remain removed simply on your thoughts. Let it stay UNTIL there's a consensus it is not helpful." I think he has it backwards, on principle. First of all, I see no sign that the inclusion was ever considered as an issue before. Even people who know election methods would not be particularly likely to notice this, especially if they have Single transferable vote in mind: the word "single" makes the statement. Few experts, let alone regular editors, are familiar with the actual text of Brown v. Smallwood. The point here is that there has not been, to my knowledge, an informed consensus on this point. Further, it's the introduction; in my view, the standards for what are in the introduction are much higher than elsewhere in the article. Does IRV only allow one vote per voter? If that were true, then there would be no legal issue in Minnesota, and people could be confident that any suit filed based on the precedent there would be quickly dismissed. I think that these arguments did come up in the Minneapolis campaign for IRV, and FairVote was quick to provide legal advice that there was no problem. Yet an actual city attorney with no apparent axe to grind came up with a different opinion, plus there are the opinions of others opposed to IRV. In my view, Brown v. Smallwood is likely to be overturned, but not on the basis that IRV is different from Bucklin voting; rather on the basis that alternative votes *should* be permitted. The plaintiff in B v. S convinced a court majority in Minnesota (and, more likely in my opinion, there was a political conspiracy involved, for Bucklin voting, contrary to FairVote claims, was *working*) but that opinion was not replicated elsewhere; the Oklahoma decision eliminating a Bucklin variant was on a completely different basis (and, it happens, I agree with that court).

An opinion piece in the St. Cloud Times argued:

The state's highest court ruled in 1915 that preferential voting systems were "contrary to the intent of our Constitution." In this decision, the court emphasized that the Constitution, by implication, forbids any elector to cast more than a "single expression of opinion or choice."

In an attempt to circumvent this ruling, Schultz argues that the 1915 court ruled preferential voting unconstitutional merely because it involved the Bucklin method of counting, which "had the effect of giving some voters more than one vote."

He argues that IRV "does not share this fatal flaw." The claim is that even though each voter may cast multiple votes, only one vote is actually counted — the one applied to the highest preferred candidate eligible to receive it.

This is pure rhetorical trickery. Where did the other votes go, cyberspace? It only stands to reason that if you rank three choices on the ballot, you've cast three votes!

.

Consistently, even when I have provided sources to show that my claims and arguments presented here in Talk are not merely my own, these FairVote editors have dismissed me as merely pushing my own crazy, senseless opinion; Ruen, above, expresses his view of the situation:

We have a little wild Elf named Abd who is upset by a claim that IRV guarantees a majority-supported winner on the grounds that some voters may not rank enough candidates to support one of two candidates in the final instant runoff round and lose their vote, and so a winner may still have less votes than half the ballots. So it is a con because it is a claim judged false - IRV can't promise a true majority. You see this Elf likes Approval voting which offers no guarantees at all that a singular winner can be found with majority support, so in order to compensate, he has to make sure everyone knows that IRV might also suffer this fate. SERIOUSLY, there's a reference link if you want to read an external source. It all comes down to the fact that majority supported winner is the only real selling-poing of IRV, and so naturally it is attacked by those who don't want IRV.

IN REPLY, a two round system and blanket primary ALSO both promise a majority-supported winner, IF everyone are willing to vote twice, since with two choices you're guaranteed one will have more than 50% unless there's a tie. [quoted from Tom Ruen on this Talk page, above]

It's off topic for this narrow point about "one vote," but we will eventually return to the Robert's Rules issues, which revolve around "majority": FairVote editors really want it to be mentioned in the article that Robert's Rules "recommends" -- the original wording -- "IRV." And these POV editors see edits that are attempting to clean up the article, so the POV tag can be removed, as hostile, and think that the IRV claims are being "attacked." I find it fascinating that Ruen appears to acknowledge that "majority supported winner is the only real selling-point" of IRV," when, in fact, there are much cheaper, simpler, and more effective ways of accomplishing that goal. (I don't agree that it is the only selling point, dealing with the spoiler effect is another important goal; there, too, there is a cheaper, simpler way. Bucklin voting did it beautifully! The case Brown v. Smallwood was one where the court reversed a clearly just election result, given the votes; IRV would have probably produced the same result as Bucklin. What these one-person, one-vote arguments don't consider is that Approval voting -- and thus Bucklin -- *also* is one-person, one vote, i.e., in the end at most one vote per voter counts toward the winner. The only way in which later votes hurt the first preference with Bucklin occurs when the first preference did not gain a majority of first preference votes; in that case, it could happen, theoretically, that the second preference vote causes the second preference to win over the first preference; and, with serious rarity, I expect, the first preference might have a majority in that later round, but the second preference has a *larger* majority. Otherwise there was no "harm" at all. For too long, FairVote controlled the debate on these issues, controlled the relevant Wikipedia articles, and was being used as if it were a reliable source. No longer. FairVote is still a legitimate source for *arguments*, for the positions of the IRV movement, but not for fact, and, until I began cleaning up this article, and attracted the ire of the socks and thus the attention of an administrator -- a sock was created just to attempt to get me blocked! -- nobody, including Tom Ruen said "Boo!"

--Abd 19:28, 1 December 2007 (UTC)

I give up, don't care enough to fight. This debate is insane to me. Tom Ruen 00:02, 2 December 2007 (UTC)

Indeed. However, attention to detail and nuance is necessary for this article to see its POV tag removed and possibly to become a featured article. We are getting there, slowly. --Abd 05:32, 2 December 2007 (UTC)

I disagree. I would object to the removal of the tag because you have biased the article against IRV by including misleading half-truths and advocacy, against the consensus of all of the other editors. MilesAgain 09:02, 2 December 2007 (UTC)

MilesAgain is clearly a sock of an experienced editor, see Special:Contributions/MilesAgain, and he acknowledged being a "legitimate" sock.(diff) Whether or not this is the same banned puppet master who has so afflicted so many voting methods articles, I don't know. I agree with an edit of MilesAgain to User Talk:Jimbo Wales that devil's advocates are necessary, but we don't put devil's advocates at the steering wheel, we merely listen to them, To a degree. And we don't allow them to waste more time than necessary. The problem with even "legitimate" uses of socks -- which is never tendentious editing, as displayed by this sock, is that they can be throwaway accounts, there is no personal responsibility, unless the master gets pinned and banned by IP, which sometimes fails.

Now, MilesAgain has made two serious charges. The first is that I have allegedly biased the article against "IRV." That would be technically an error, because "IRV" is not a POV. He would presumably mean that I've incorporated, he claims, a POV that is supposedly opposed to the use of Instant Runoff Voting. If so, I have failed, that is not my goal. However, MilesAgain has shown no example of such bias. Each argument he raised here was corrupt; however, where there has been opportunity for improvement based on his edits, it has been followed. There is no charge of POV bias in the article, introduced by me, from the FairVote editors (one with clear WP:COI the other with a stated "affiliation"), only from, now, MilesAgain. In the present example, the mention of "voters have one vote" in the Introduction, no POV bias against IRV is introduced by my edit removing it. I happen to agree with the claim, *but* it's controversial, and some of the pro-IRV campaign, in Minnesota, has depended upon an assertion that there is no legal problem for IRV in that state, based on this claim. I just happen to be familiar with Brown v. Smallwood, the relevant precedent, and believe that the FairVote argument on this is corrupt, aimed toward facilitating implementation. But my opinion or knowledge on this is not relevant: what is relevant is that there is apparently unbiased and official legal opinion to the contrary of the claim, *plus* argument being made in public against that claim. It is, without doubt, controversial, and thus should not be stated as a fact in the Introduction. It could be placed in the Controversies section or the Controversies article.

The other claim of this editor is that I'm editing contrary to consensus. It could appear that way, and the reason is an example of why Wikipedia is not based on voting. There is participation bias introduced by the patterns of interest and the availability of editors. We have to look at the identity of the active editors. Basically, there is me, two editors affiliated with FairVote, the sock MilesAgain who easily could be COI or even banned, plus occasional contributions by others. What the three solid editors (COI and myself with my POV) agree upon is a pretty good start toward consensus. Where we do not yet agree, there is not yet consensus. It's that simple. And what is not agreed by consensus *of fully legitimate editors* does not belong in the article, unless an editor can prove the case beyond doubt (which would presumably be judged through an RFC, mediation, or arbitration). It's essential to take each edit, case by case.

Anyone reviewing this record later, please note how we have generally managed to settle on mutually acceptable language, which may or may not be fully pleasing to each POV, but which is recognized to be NPOV, true, supportable, and at least arguably relevant.

--Abd 17:21, 2 December 2007 (UTC)

There is a very important reason to explain that IRV uses a one-vote, multiple ranking methodology in the introduction. This is one of the key features that distinguishes IRV from other ranked (and some unranked) voting methods. The voted ballot might look identical for Borda, Condorcet, Bucklin, but these other methods can utilize voters' rankings simultaneously, and in effect exercise multiple votes for more than one candidate. This is not a question of whether the fact that IRV uses a single transferable vote is pro-IRV (as Abd seems to believe), since the advocates for these other systems trumpet this distinction as an advantage of these other methods over IRV (considering all of voter-supplied information). This is simply an important (KEY) fact about what IRV is, and it needs to be in the introduction. Abd's mention of the ancient Smallwood case is irrelevant, since that case did not deal with IRV (but rather Bucklin). More recent cases ACTUALLY ABOUT IRV, such as Ann Arbor case, clearly confirm legally that IRV uses a single transferable vote.

Tbouricius 19:30, 2 December 2007 (UTC)

(unindent)The alleged "fact" is controversial. I've given sources showing contrary legal opinion, plus contrary argument in media. Apparently the FairVote propaganda has been quite effective: Tbouricius believes it. I've acknowledged that IRV only counts one vote at a time; however, the same could be claimed about other methods, there being a way to describe, for example, Condorcet methods that only considers one vote at a time. Further, there are two aspects to "one vote," which is the *maximum* vote *and* the *minimum* vote. IRV only counts one vote at a time, but it discards expressed votes under some conditions. It is possible that this could be asserted as a violation of one person, one vote, for not all votes are created equal with IRV. It's a complex issue, however, and I'm not trying to establish one particular interpretation, only noting that there *is* controversy over this. The "ancient" Smallwood case is still legal precedent in Minnesota, and if the city of Minneapolis believed the FairVote claims that it did not apply, not to worry, they are going to be spending real dollars fighting a challenge based on Smallwood, it is a near certainty. As I've written, I'm quite familiar with Smallwood, having written an extensive article on it and having read it many times. I recommend it. FairVote propaganda can be quite effective; many times I've read something having read the FairVote argument about it first, and, big suprise, I came to the same conclusion as FairVote had, a good example would be the claim that "Robert's Rules of Order recommends IRV." I looked at the source, conveniently quoted on the FairVote site, prefaced, of course, with their "explanation," and believed it to be true. Until I came back at least a year later and looked again, because something didn't seem right.

Here is a newspaper opinion piece on IRV and Brown v. Smallwood: [18]

Brown v. Smallwood itself:[19]

The point is that the statement that was in the article is controversial, it is not a simple fact, and thus it really does not belong in the introduction. It is not essential to understanding IRV, it is "important" as part of an argument that IRV is legal. A manifested desire to lead the reader to conclude that IRV is legal is, by definition, POV. This editor is simply showing his POV, which is normal and not surprising.

One more point. It is often argued that in IRV, your vote for a second preference cannot harm your first preference's chance to be elected, and this is then used to support a claim that Brown v. Smallwood does not apply to IRV. If there is a majority election requirement, this isn't necessarily true. As an example, we could look at the legislation that I think was originally proposed in Vermont by none other than our own Tbourcius. It had a ballot instruction that I think was the same in a later version of the bill:[20]:

"Instructions on the ballot shall include the following statement: “In addition to your first‑choice candidate, you may rank alternate choice candidates if you wish. Marking a second or other choice cannot help defeat your first choice.”"

There is a majority election requirement for Governor in the Vermont constitution. IRV, like some other methods such as Approval, does indeed make it easier to get a true majority winner; however, if there is no majority winner, the Vermont constitution provides that the top three candidates (apparently only first place votes are considered here, which is itself a problem, but I'm not sure about that) are then presented to the General Assembly for choice by secret ballot.

Now, suppose that the first place choice of a voter was eliminated, but would be the third candidate in the top three based on first place votes. If the voter adds no second place choice, the election, under some conditions, would go to the General Assembly, which *could* choose the favorite of this voter. However, if the voter adds a second place choice, and that gives this choice a majority, the election is finished with the second place choice winning. The voter helped defeat the first choice by marking a second choice, contrary to the proposed ballot instruction.

Robert's Rules, in discussing the implications of preferential voting, notes that voters may choose not to rank all candidates, believing that this may help their favorite to win. They consider this an error; however, they also point out that the balloting may have to be repeated if the voter does this (Robert's Rules, by default, assumes a majority election requirement). If the voter considers this a better outcome than completing the election with his second choice (Robert's Rules recommends *against* top-two runoff), the voter quite rationally does not add a second choice. It is only if the voter prefers to avoid voting in a runoff election that the voter has a stronger incentive to add additional choices, or if the voter knows that the voter's favorite isn't going to win, period, and wants to help a frontrunner. —Preceding unsigned comment added by Abd (talk • contribs) 04:38, 3 December 2007 (UTC)

Abd states above that the simple fact that IRV uses a SINGLE transferable vote is "controversial." That is nonsense. It is the core of the definition of what IRV is. Abd states that he has "given sources showing contrary legal opinion..." That is simply untrue...He has cited a discredited ancient case (Smallwood) that is about a DIFFERENT voting method called Bucklin Voting, which did allow voters to have more than one vote for more than one candidate at the same time. Abd appears to have a POV agenda with this Wikipedia edit, which is to try and link IRV to this irrelevant old legal case about Bucklin in hopes that it may hinder the adoption of IRV.

In short, the "one vote" explanation belongs in the introduction because it is the defining characteristic that distinguishes IRV from most other ranked voting methods.

Tbouricius 18:26, 3 December 2007 (UTC)