Jump to content

Wikipedia:Arbitration Committee Elections January 2006/Proposed modifications to rules

Add links
From Wikipedia, the free encyclopedia
< Wikipedia:Arbitration Committee Elections January 2006

This is an old revision of this page, as edited by Sam Spade (talk | contribs) at 12:17, 10 October 2005 (→Comments). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

This page is to outline proposed modifications to the rules for electing Arbitrators in the December 2005 and later elections. If you have further proposals, please add them at the end.

The December 2004 elections were noted for their rancour and a large number of clearly unsuitable candidates that contributed to it. At the same time 2005 has seen the Arbitrators workload increase greatly, with a high drop-out rate of Arbitrators unwilling to dedicate an ever-greater proportion of their time to a greater number of ever-more complex disputes.

These proposals, each of which is a stand-alone proposal, are aimed at reducing these problems. It is assumed that voting this year will be from 4 to 18 December, as it was for last year.

Proposal 1

To nominate yourself for the Arbitration Committee, you must be an admin. Repeatedly nominating yourself when you are not an admin will result in your account being blocked till the election is over.

Rationale

It is unlikely that the community would be willing to accept someone as an Arbitrator who they wouldn't be willing to accept as an admin, and this would eliminate clearly inappropriate candidates at a stroke. By publicising this new rule well in advance, it would give time for good potential candidates who are not currently admins to apply for adminship before the election starts. The second sentence is to prevent trolling - it is not intended to apply to someone inadvertently breaking the rule (who would have their nomination removed and an appropriate note put on their talk page).

Comments

  • The last election was bedevilled by a number of inappropriate "candidates". This proposal would stop those candidates, whilst still allowing all genuine contenders to stand, jguk 08:16, 14 August 2005 (UTC)[reply]
  • I have mixed reactions to this proposed rule. The temperament required to be an admin and the temperament required to be an arbitrator are quite different. An admin must act quickly, such as by blocking, and so must have a sense of timing as to when a situation is urgent and when it is not. An arbitrator should act with reflection and deliberation. In fact, a good arbitrator may sometimes not be a good admin, and vice versa. If the real problem was that people nominated themselves who had not earned the trust of the Wikipedia community, I would suggest that one alternative would be to eliminate self-nomination completely, and to require all candidates to be nominated by someone else. This would in itself eliminate trolling. Robert McClenon 16:26, 14 August 2005 (UTC)[reply]
  • I support the rule change but I would delete the blocking. Simply declare that only admins may be candidates for arbitrator and that nomination of a non-admin is null and void. Requiring nomination by another user invites sock-puppetry and is somewhat arbitrary; I would prefer to keep self-nomination. David | Talk 16:31, 14 August 2005 (UTC)[reply]
  • Do we need a specific blocking provision for this? I would have thought that repeated renomination of an ineligible candidate would fall under the 'disruption' clause of the existing blocking policy. TenOfAllTrades(talk) 22:16, 14 August 2005 (UTC)[reply]
    • Is there already a policy whereby those deliberately disrupting the election process can be summarily blocked (for a reasonable period of time)? The "blocking" sentence was only added in to put it beyond doubt - it'll never apply to anyone not deliberately disrupting the election, jguk 03:46, 15 August 2005 (UTC)[reply]
  • What about requiring that a candidate's nomination be seconded by another editor with x number of votes at the time of the seconding, with y% of those votes having occurred prior to the candidate's nomination? This would allow non-admins to run, but allow for fewer frivolous candidates. Or maybe even require a thirding, as well? Zoe 22:42, August 14, 2005 (UTC)
  • I am opposed to the proposal as-is. How about instead, treating both non-admin and admin prospective candidates equally. Make a process similar to requests for adminship, but call it requests for candidacy. So that the community has a chance to reach a consensus before the actual election, which is just the same as always. This would also help mitigate some of the objections to proposal 4 expanding the number of positions, namely, the possibility of there being unsuitable candidates and more positions which must be filled than candidates: since all candidates have already been determined to be suitable (by virtue of being candidates, that issue is gone). --Mysidia (talk) 23:34, 14 August 2005 (UTC)[reply]
    • The problem with this is that it would elongate the election process (and still effectively allow anybody, regardless of their "good intentions" to run. By using a pre-existing pre-approval process, we obviate this. Maybe an alteration to allow all admins (who have, of course, already been through an earlier vetting procedure) can stand, and all non-admins who don't want to apply for adminship can apply through a similar process to be allowed to stand - but which will not make them an admin if they pass that process? Note that so far no non-admin has served or been elected as an Arbitrator, jguk 03:46, 15 August 2005 (UTC)[reply]
  • I have no problems with this, and I would also expect clearly bad-faith nominations (e.g. nominating a user that has 10 edits) to be summarily removed. Radiant_>|< 12:06, August 15, 2005 (UTC)


  • Oppose, as per Theresa knott. Inappropriate/troll candidacies should be removed. --Merovingian (t) (c) 16:58, August 15, 2005 (UTC)
    • How would you decide whether a candidate is inappropriate or a troll? The last election was bedevilled by them, and all they did was make the whole process a lot more acrimonious. This is an attempt to define candidates that will clearly fail and to eliminate them immediately from the process. If you have a better way of doing this, what is it?
  • I can't imagain someone who could get elected to arbcom not being able to make it as an admin. If they had some reason that they didn't want to be an admin then they could ask for thier admin powers to be removed after the election.Geni 00:03, 16 August 2005 (UTC)[reply]
  • I think instead of making only admins eligible, why not set the eligibility based on certain number of edits, as it is already done for a number of elections. For the Arbcom elections we can set the bar fairly high like 1000 or more article space edits. This should prevent most of the problems of inappropriate candidates. The people who disregard the rule would simply have their nomination voided. No need to ban anybody for this. I think it is not really neccessary to make people go through the admin process before nominating themselves for arbcom. people should be free to nominate themselves if they wish and meet the criteria. kaal 05:54, 16 August 2005 (UTC)[reply]
  • I also oppose this modification to the rules, for one thing it's overreaching and it stops many potentially good nominations as well as bad ones, second it seems to be to just be more uneccessary rule creep. Jtkiefer T | @ | C ----- 18:04, August 16, 2005 (UTC)
    • Are you saying there is a realistic possibility that WP would accept someone as an Arbitrator whom it wouldn't also accept as an admin? I certainly agree with Raul that I can't see this ever happening (it certainly hasn't happened or come close to happening so far). There are many good editors who for various reasons have not wished to stand for adminship to date. My guess is that these are not people who are likely to want to become an arbitrator anyway - but if they did, they'd have little problem becoming admins first. Without some restrictions to candidacy, the 2005 will be as plagued by disruptive candidates with no chance of success, in much the same way as the 2004 election was, jguk 07:45, 19 August 2005 (UTC)[reply]
  • I was under the impression that adminship was not supposed be a big deal, and that it consisted entirely of a handful of tools intended for maintenance. This proposal would seem to fly in the face of that. Aquillion 08:29, 16 September 2005 (UTC)[reply]
  • I don't think restricting it to administrators is a good idea. There is a problem looking for a solution, but I don't think this is the solution needed. Perhaps requiring all candidates to have been wikipedians for x amount of time (6 months?) or to be an administrator would be better, as I personally think anyone who has only joined recently would be very unlikely to be elected. However if they are already trusted to be an administrator then imho they are extremely unlikely to be an inappropriate candidate. Thryduulf 20:39, 20 September 2005 (UTC)[reply]
  • Oppose - there are several people who could be admins if they wanted, but who repeatedly turn down nominations because they have no interest in administrative tasks. They ought not be prohibited from being arbitrators. Snowspinner 15:36, 5 October 2005 (UTC)[reply]

Proposal 2

The vote should be conducted on First Past the Post (technically bloc voting in this case). If there are X vacancies, each Wikipedian entitled to vote for up to X candidates. The X candidates with the most votes win.

Rationale This voting method is the easiest of all to understand, and the least open to manipulation - the only thing you can do to help your candidate win is to vote for him/her! It also encourages positive voting, which may get rid of some of the rancour of earlier elections. Although the current method is called "approval voting", it is in effect a negative vote - you select which candidates you cannot stomach and vote for everyone else.

Comments

  • I support this change. I'm a great believer in 'the right electoral system for the right election'. Here, the result should be the candidates the most people are content with. It is not the aim of the ArbCom elections to produce a committee with a balance of opinions, which is the explicit aim of proportional voting systems, so I would rule out STV. We also do not want cliques, so all arbitrators should run in one pool. I would however suggest that voters be compelled to use all their votes rather than be allowed to plump for one or two of their 'friends'. David | Talk 16:37, 14 August 2005 (UTC)[reply]
    • Suppose a voter only knows enough about a few of the candidates to rationally vote for them, and is indifferent to the others. should such a voter randomly vote for unknown candidates for hir votes to count? That doesn't make sense to me. DES (talk) 16:41, 14 August 2005 (UTC)[reply]
  • Ummm, what is this? "[ first past the post ] is the easiest of all to understand, and the least open to manipulation." In what universe? Clearly one with different mathematical properties than ours. Please see Plurality electoral system#Tactical voting for a thorough treatment of this subject. -- RobLa 21:45, August 14, 2005 (UTC)
    • Surely that does not apply here. No-one can be sure who will win, and the candidates are running as individuals rather than as party nominees. In any case tactical voting isn't "manipulation" - it's part of the system. It's not a bug, it's a feature. David | Talk 22:00, 14 August 2005 (UTC)[reply]
      • Two questions: 1. Can you point me to the research that supports this claim: "first-past-the-post is the least open to manipulation"? If not, I'd ask that it be retracted from the proposal. 2. How is "manipulation" different from "tactical voting"? Please provide an example of each. Thanks. -- RobLa 22:28, August 14, 2005 (UTC)
        • I don't have to justify a claim I didn't make. As far as 'manipulation' goes, giving a voting preference to your preferred candidate in order to make their election more likely is a manipulation of the electoral system. It just happens to be an expected one. Surely the key question in all of this is: Who do we want elected? Do we want a spread of opinion so that every small clique can get their own pet arbitrator, or do we want a group of qualified and trusted arbitrators in whom the Wikipedia community has expressed confidence? If it's the former, then use a list system or STV. If it's the latter, we should use FPTP. David | Talk 22:35, 14 August 2005 (UTC)[reply]
          • FPTP is most difficult to manipulate (in a non-party system) as it is so widely understood and follows the principle that everyone has an equal say in the matter. The only thing you can do to elect the candidates you want elected is to vote for them - so you have to vote positively. And every vote is equal. A positive voting system (rather than the negative one we currently use) may help get rid of some of the rancour of the 2004 election. Remember that tactical voting happens with every voting system, it's just less manipulative with FPTP - other systems are biased towards (1) those that understand them; (2) those that are particularly organised, jguk 03:53, 15 August 2005 (UTC)[reply]
  • I am opposed to this proposal "tactical voting isn't a bug, it's a feature..." (yeah right). --Mysidia (talk) 23:25, 14 August 2005 (UTC)[reply]
  • Plurality-at-large (a more accurate term for this approach) is a terrible choice because it eliminates minority view representation. If a small majority (e.g. 51%) votes in a consistent way, then that half of the electorare will elect 100% of the committee. The other 49% gets no representation at all. The simplest approach that solves this problem is just to give everyone one vote in stead of X votes. This will at least tend toward proportionality. The technical name for that is single non-transferable vote. A slightly more complicated, better approach would be cumulative voting, which allows voters to vote X times, but they may stack multiple votes on a single candidate. Best of all in my opinion would be proportional approval voting or Single transferable vote.
  • I prefer approval voting. It's even simpler to understand -- you don't need to keep track of how many votes you've cast so far. --Carnildo 06:23, 15 August 2005 (UTC)[reply]
    • The real issue with approval voting is that, despite its name, its actually a negative vote - rejection voting would be a more appropriate name. You select the candidates you can't stomach and vote for the rest, jguk 06:30, 15 August 2005 (UTC)[reply]
  • Oppose. The problem with First Past the Post is that it tends to polarize the electorate. For a community that functions only when there are sufficiently many volunteers who are willing to work for free, a polarized electorate is the last thing we want to have. Therefore, I suggest that we should keep approval voting or even change to a Condorcet method; these methods have a calming effect on the electorate. Markus Schulze 15 Aug 2005
  • Oppose. The statement the only thing you can do to help your candidate win is to vote for him/her! is always true in almost every method. In Approval voting, unlike FPP, one always can vote for one's favorite without sacrificing a (possibly more viable) compromise candidate. The claim Although the current method is called "approval voting", it is in effect a negative vote - you select which candidates you cannot stomach and vote for everyone else. seems to be nothing but a "glass is half-empty" perspective. I don't understand why that claim is being made, or how Plurality is an improvement on this! KVenzke 13:08, August 15, 2005 (UTC)
  • Oppose. FPTP suffers from the terrible problem of vote splitting. A bad candidate can be selected if there is a bloc of voters that agrees on that candidate, even if the vast majority of voters dislike that candidate. An example of this is Scott Savol in the American Idol contest, who hung on even though most voters thought he was terrible. Vote splitting has also been an issue in recent presidential elections. Remember Ralph Nader? The value of approval is that it doesn't force you to choose between, say, Gore and Nader. andrew_myers 15 Aug 2005
  • Oppose. This is the wrong remedy for the problem. If the problem is irrelevant candidates polluting the ballot, then propose a change to the nomination process. Switching the voting system to FPTP simply lowers the amount of information you collect from voters, and will increase the likelihood of tactical voting. Approval is easy for voters, is easy to tabulate, and provides more information than FPTP. Please keep it. James P. Cooper 15 Aug 2005
  • Oppose. I don't see a problem with approval voting. Theresa Knott (a tenth stroke) 16:05, 15 August 2005 (UTC)[reply]
  • Support. --Merovingian (t) (c) 17:01, August 15, 2005 (UTC)
  • Oppose. We may not have parties but we do have groups of users who are more likely to get elected (long term admins, previous arbcom members, medators the like) tactical voteing will garentee that the committe is made of these indivduals.Geni 00:06, 16 August 2005 (UTC)[reply]

The approval-at-large system currently used to elect the Wikipedia Arbitration committee is not a good system. The proposed change to plurality-at-large (FPTP) would replace this with a worse one. The membership of Arbitration committee should reflect the full range of opinion amongst Wikipedia users not just the opinions of a majority or a plurality. The most appropriate method to use to elect the committee would be a candidate based system of proportional representation such as the Single Transferable Vote (STV). This would ensure that all shades of opinion are represented on the committee.

David Gamble

  • That is precisely what we don't want on the Arbcom. The effect will be the LaRouchites and the Scientologists (etc) organising to make sure they get their own arbitrator to swing cases their way. In my opinion the electoral system should make sure that all candidates with 'shades of opinion' should be kept off the committee. David | Talk 17:46, 17 August 2005 (UTC)[reply]

I don’t see the problem. The rules state that:

“Each part will be subject to a simple-majority vote amongst active non-recused Arbitrators - the list of active members being that listed on Wikipedia:Arbitration Committee.”

If the Scientologists get one member on the committee then that member cannot block a decision.

The rules also state:

“If an Arbitrator believes they have a conflict of interest in a case, they shall recuse themselves immediately from participation in the case. Users who believe Arbitrators have a conflict of interest should post an appropriate statement during the Arbitration process.”

So an arbitrator who was also a Scientologist would not be able to participate in a ruling on something like OT III or Xenu.

David Gamble

  • Oppose. It seems that multi-winner approval is a fairly appropriate voting method for this situation. In my opinion, block voting is inferior. (By the way, I believe that it is "block voting", not "bloc voting". See this link.) Although I love STV in general, I'm not sure that proportional representation is right for this situation. --Hermitage 22:21, 21 August 2005 (UTC)[reply]

"Single transferable vote" is a method of selecting someone who is disliked the least. "Approval voting" is a more straightforward method of selecting people or are disliked the least. The difference being that under STV you order people in the way you dislike them, whereas in approval voting you just vote like/dislike. Let's just keep things simple (complicated just ends up confusing people or pissing them off and probably won't make much difference in the result anyway).

Incidentally, the reason why I propose FPTP is because it is the most straightforward positive method of voting. At present we have the most straightforward negative method of voting. The straightforwardness I completely agree with - it's just the positive/negative thing, jguk 18:22, 7 September 2005 (UTC)[reply]

STV isn't a method with anything like the characteristics you ascribe to it; it's not in any significant respect similar to approval voting, and your characterisation of it as a "negative" voting method seems to be entirely ad hoc. (Anything that isn't plurality-based?)
You're certainly playing "simplicity" to the hilt, but I don't think it stands up. For it to be a pressing argument, either the electorate must be too "unsophisticated" to successfully cast preference-style votes of any kind in a manner consistent with their desired outcome (but are on the other hand, capable of voting "tactically" in a plurality election in a way that does so); or the "officials" choosing and/or running the system are so incompetent or corrupt that a result inconsistent with those preferences can somehow be conjured (as per your accusations regarding Condorcet criterion). Both sound pretty far-fetched to me. Alai 23:23, 7 September 2005 (UTC)[reply]
Many people don't understand the simplest methods - as the complexity of the method increases, the greater the number of people disenfranchised by it. I've got news to you, most people aren't psephologists! jguk 23:39, 7 September 2005 (UTC)[reply]
Entire nations "muddle through somehow" with STV and other preference-voting options; several are deliberately "backwards compatible" with single votes cast plurality-election-wise. Not only is wikipedia is a self-selecting electorate, votes "mis-cast" can be corrected while the election remains open. Just how many likely voters, in such circumstances, do you think a procedure like "rank your favoured candidates in order of preference" is likely to "disenfranchise"? This really does sound increasingly like the IBM sales pitch. Alai 00:52, 8 September 2005 (UTC)[reply]
Those elected tend to work independently, yes? Then proportional representation seems likely to be harmful for this purpose. KVenzke 20:37, September 8, 2005 (UTC)

Oppose as a solution in search of a problem. Filiocht | The kettle's on 08:09, 7 October 2005 (UTC)[reply]

Proposal 3

If there are any vacancies in the Arbitration Committee at 31 May 2006, an election will be held in June 2006 to fill those vacancies from 1 July 2006. The rules for that election will be the same as for the December elections, but with a slightly shorter timetable. Candidates will be elected to fill the remainder of the terms of the Arbitrators they replace.

Rationale There are always going to be drop-outs from the Arbitration Committee, it makes sense to have a process to replace them.

Comments

  • Seems a sensible thing to do, jguk 08:16, 14 August 2005 (UTC)[reply]
  • Agree. Robert McClenon 16:28, 14 August 2005 (UTC)[reply]
  • Good idea. As sometimes it may be the case that elections to a six month term and 18 month term happen simultaneously, I would say explicitly that the top vote-getter can chose which term they take (which is probably implicit). David | Talk 16:39, 14 August 2005 (UTC)[reply]
    • That's standard practice, yes. As for having elections, well, in our experience they've been horrendously divisive, nasty experiences that end up being popularity contests (which is not any way to select a group of people who are meant to be trusted by all – or at the very least, as many as possible – of Wikipedians). This also bears reflection on the choice of voting system, etc.. James F. (talk) 10:11, 15 August 2005 (UTC)[reply]
  • I'm neutral on this. Elections,. or jimbo apointees - doesn't really matter. Theresa Knott (a tenth stroke) 16:07, 15 August 2005 (UTC)[reply]
  • I'd rather have Jimbo appointees. --Merovingian (t) (c) 17:02, August 15, 2005 (UTC)
    • It would be useful to codify which method we use. I'm tempted to suggest putting this proposal up against a direct alternative stating explicitly that Jimbo may fill casual vacancies, which would be filled on a permanent basis at the next annual election, and seeing which of the two commands greater community support. Personally I'd prefer to eliminate Jimbo from the procedure - solely because WP is now at such a stage that as far as possible WP should position itself so that it can operate normally without him. I sincerely hope that we have him around WP for many years to come, but it makes sense not to be overly reliant on any one person, jguk 20:23, 15 August 2005 (UTC)[reply]

Proposal 4

Expand the numbers on the Arbitration Committee to 24. However, only 9 Arbitrators will hear any one case that has been accepted. The Arbitrators hearing any case to be decided on a "taxi rank" system from amongst those Arbitrators indicating that they are currently willing to hear cases. When an Arbitrator re-adds him/herself to the active list, they will go to the end of the taxi rank. Any active Arbitrator may, at the start of a case, choose to step down from hearing it, and go to the end of the taxi rank queue.

Rationale Having more Arbitrators is likely to reduce burn-out, and by having more Arbitrators, each Arbitrator will have fewer cases to deal with. Allowing Arbitrators to step aside from a particular case without recusing would also help burn-out. A taxi rank system would prevent individual Arbitrators from being accused of "picking" their cases - as they can opt out of a particular case, but not into it.

Comments

More numbers on ArbCom, plus allowing Arbitrators to duck out of cases without recusing, will help reduce the workload, jguk 08:16, 14 August 2005 (UTC)[reply]
Good idea in principle, but doubling the current amount of arbs may be taking it slightly too far. I'd prefer maybe 20 of them. Also, this may be hypothetical, but supposing we have 15 open slots by december and we get only 15 candidates total, then it's possible that one of them is unsuitable and only gets a handful of votes, but is elected anyway because we feel we 'must' fill all slots. This probably won't become an issue but people should at least think about it. Radiant_>|< 08:45, August 14, 2005 (UTC)
That is an important point. Perhaps there could be some minimum vote requirement, or some minimum something to prevent this. Maybe even no election until there is a minimum number of candidates. Dmcdevit·t 09:15, August 14, 2005 (UTC)
Or we could just let Jimbo veto an unsuitable candidate. Or we could delay the election until we had (for example ) at least twice as many candidates as vacancies. Or we could reduce the number of vacancies to half the number of canbdidates. Theresa Knott (a tenth stroke) 16:22, 15 August 2005 (UTC)[reply]
I agree with the general principle, but I'm not sure of the specific numbers. Maurreen (talk) 09:05, 14 August 2005 (UTC)[reply]
Maybe require 30 percent? In Wikipedia:Arbitration Committee Elections December 2004, seven candidates ranged between 30 percent and 33 percent. The top three of them were elected; the difference with the next candidate was just one vote. Maurreen (talk) 10:16, 14 August 2005 (UTC)[reply]
most admins would be sutible and if you had a situation where it looked like a unsitible candidate was going to be elected automaticaly I suspect that other people would step up just to keep them out.Geni 10:23, 14 August 2005 (UTC)[reply]
Could we stagger the elections that enlarge the ArbCom—perhaps elect half the batch in December and the other half in March, or something similar? That would allow some additional good candidates to come forward the second time around, and also would avoid swamping the ArbCom with unseasoned members. TenOfAllTrades(talk) 21:35, 14 August 2005 (UTC)[reply]
  • I like this idea. I'm not convinced about the taxi rank system. Though it might work. Let the arbitrators work out the best way of dividing up the cases. But certainly the work load is very high at the moment and likely to get higher as wikipedia continues to expand. Theresa Knott (a tenth stroke) 16:14, 15 August 2005 (UTC)[reply]
  • Anything to streamline the system. --Merovingian (t) (c) 17:07, August 15, 2005 (UTC)
  • So, in order for a case to be open, 9 people have to actively affirm their participation (in serial fashion). This will probably tack on at least 2 weeks to every case (assuming everyone checks the roster every other day); 4 weeks is probably a better estimate. →Raul654 05:57, August 16, 2005 (UTC)
    • No, in order for it to be open there need to be 5 accept votes (ie same rule as before). The 9 Arbs next on the active list would be the Arbs to decide the case - assuming none of them actively removes themselves - no time would be wasted at all, and by lessening each Arb's workload it is more likely that cases finish quicker, jguk 18:09, 16 August 2005 (UTC)[reply]
  • Weak Support. Sure having more arbitrators would help lighten workload, but wouldn't this open up the debate of "this user doesn't like my opinion, I want someone else on my case or you all suck" or something equivalent? Luigi30 (Ταλκ) 03:05, 5 October 2005 (UTC)[reply]

Proposal 5: Run-off

Hold a runoff election. The number of candidates in the run-off could be set at twice the number of open seats.

Comments

My rationale is that the last election had, I think, seven open seats and 34 candidates. For voters who haven't already had interaction with any given candidate, that's a lot to study and research to make an informed vote. But I'm open about the specific number for the run-off. Maurreen (talk) 09:05, 14 August 2005 (UTC)[reply]

Proposal 6: Alternates

I'm not sure this is strictly about the election, but this seems like as good a place as any for this idea.

Designate alternates, as, say the top three candidates who are not elected. That is, if the top seven are elected, candidates eight through 10 by vote count would become alternates. If any member of the arbitration committee resigns, the alternate would fill in until the next election.

Comments

My rationale is that this keeps the committee at full strength without needing to do more than an annual election. I'm open about the number of alternates. Maurreen (talk) 09:05, 14 August 2005 (UTC)[reply]

  • This one kind of depends on what happens with Proposal 4. If we have a pool of 24 arbiters (as suggested in P4) then having the committee reduced by three or four (or even six or seven) over the course of the year isn't going to be a disaster. This is an excellent idea if the committee isn't enlarged. TenOfAllTrades(talk) 21:29, 14 August 2005 (UTC)[reply]
  • I think this proposal could be a good idea even if we have an enlarged ArbCom. All candidates that receive at least X% of the vote could be eligible as an alternate. When the first ArbCom member resigns, the first alternate would be the top candidate (in terms of votes) who wasn't elected. As additional members resign, move down the list of alternates in descending order of votes. All alternates would have their term end after the next election (unless they were re-elected). Carbonite | Talk 17:17, 15 August 2005 (UTC)[reply]
  • There was a theoretical dissucussion a while back on the mailing list about emergency arcom members if too many had to recluse. This migh also be helpful there. However I don't think it will be required with expanded arbcom.Geni 00:13, 16 August 2005 (UTC)[reply]
  • I would prefer this solution to premature vacancies, over a system that puts one person in charge of selecting replacements. I do think that more than one election annually is undesirable. Depending on what type of voting system is used, though, alternates should have to meet some minimal threshhold of number of votes received compared with the number of voters in order to be eligible to serve as replacement arbitrators. For instance, if approval voting is used, you might specify that any candidates who received affirmative votes from at least 30 percent of voters, but were not elected because other candidates received even more votes, would be retained on the list of alternates. Mamawrites

Proposal 7

Expand the number of arbitrators to about 30. The ArbCom will be divided into three subcommittees, or "circuits", of about nine users each. Each circuit will hear cases individually, usually in a rotating order based on when they are filed. To prevent playing one circuit off another, no circuit may interfere with the operations of another circuit. To allow for the inevitably greater workload as Wikipedia continues to grow, more arbitrators and more circuits may be added at every election. There will also be a three-member "Grand Jury", so to speak, which will determine which cases are accepted, assign them to circuits and resolve any "jurisdictional" disputes about which circuit will or will not hear cases. The Grand Jury cannot hear cases, stop cases in progress or otherwise direct a circuit to act or rule in a particular way.

Rationale A proposal like this will help lighten the load of complex, time-consuming cases like /Climate change dispute and /Trey Stone and Davenbelle, which bog down the entire ArbCom and prevent it from handling other cases. If a single circuit gets tied up in one case, the other circuits will be free to handle other cases as they come along. Unlike Proposal 4, arbitrators can hear more than one case at a time, with a caseload divided more or less equally among the entire Committee. One problem I can anticipate is that if a lot of arbitrators leave a particular circuit, its operations will essentially stop. This won't be so much of a problem if Proposals 3 and/or 6 pass. Another possible concern is that 30 seems like a lot of arbitrators. However, Wikipedia has expanded a lot since the last election and there are probably enough good candidates to fill all slots should such a proposal pass.

Comments

  • Support; the AC should be able to multitask. --Merovingian (t) (c) 00:43, August 18, 2005 (UTC)
  • What is it about proposal 4 that would preclude having the ArbComm hear more than one case at a time? I understand that if the committee is expanded to 24 arbitrators, then more than 2 cases would not be possible simultaneously, but theoretically, 2 cases of 9 arbitrators each should be possible. If the committee is expanded to 30, 3 cases would be possible simultaneously, without implementing the proposal for circuits. Mamawrites
    • I was under the impression Proposal 4 wouldn't allow a single arbitrator to hear more than one case. I could be wrong, though. However, Proposal 4, like 7, would allow the ArbCom as a whole to hear more cases. I think the advantage of 7 is that it would allow for more cases to be heard because the workload could be broken down between circuits as needed. If one circuit is extremely busy with a complex case involving multiple users, like /Climate change dispute, another circuit could take two or three smaller cases. sɪzlæk [ +t, +c, +m ] 03:40, August 20, 2005 (UTC)
  • I'm hesitant about the proposal for circuits, because of the possibility of inconsistent rulings among the three circuits. Mamawrites 19:29, 18 August 2005 (UTC)[reply]
    • That is a possiblilty. But unfortunately, things like that happen with any judicial system with more than one deliberating body. Assuming the community continues to select arbitrators who respect community norms and consensus, the circuits will likely respect each others' precedents. On the off chance we ended up with a "rogue" circuit, Jimbo and/or the Board of Trustees could always step in. sɪzlæk [ +t, +c, +m ] 03:40, August 20, 2005 (UTC)

Proposal 8

The vote should be conducted on a Condorcet method. A Condorcet method should be used to calculate a complete ranking of all candidates. If there are X vacancies, then the X top-ranked candidates should be declared elected.

Rationale The main argument against Condorcet methods is that is was too complicated for voters to rank all candidates. However, it is not necessary that voters rank all candidates; voters can give the same preference to more than one candidate and they can keep candidates unranked. When e.g. a voter had approved the candidates A, B, G, and P under approval voting, then this voter can simply give a first preference to the candidates A, B, G, and P and keep all other candidates unranked.

This website can be used to calculate the winners.

Comments

  • Absolutely object in the strongest possible terms This would be an absolute disaster. This method has only led to bitter arguments in Wikipedia in the past and is completely contrary to the way WP works - and it only appears to have been used in practice when politicians wish to achieve a result chosen beforehand. It is also extremely complicated (remember - some people even get first past the post, the simplest of all systems wrong - this one I can't even understand with a first-class maths degree!). It gives extreme undue weight to those voters who understand it (who are usually the same people that promote this method. Voting on WP can get bitter as it is. However, if we ever saw this method get adopted widespread in WP, it would make the past problems seem like a love-in, jguk 15:37, 20 August 2005 (UTC)[reply]
  • Strongly Oppose. I agree with Jguk. We need a voting method that's simple and can be understood by the vast majority of voters. It shouldn't be necessary to use a website to calculate the winners. Carbonite | Talk 15:49, 20 August 2005 (UTC)[reply]
  • Support method, object to overall proposal - I strongly support Condorcet methods for elections, and strongly encourage the Wikipedia community to get educated on their use. Arguments against Condorcet methods are often luddite in nature; Condorcet methods are complicated, but only as complicated as they need to be. Fairness is more important than complication, and Condorcet methods are much fairer than other methods. However, given that there's no support for any Condorcet method in the MediaWiki software, I think it's too soon to suggest its use for a Wikimedia election. -- RobLa 22:08, August 20, 2005 (UTC)
    • "Complicated" always means that the votes of those that understand the voting method are much, much more valuable than those that don't - and "complicated" also means that very few understand them (and it tends to be those that understand them that argue for them). I'm willing to be outvoted if the vote is fair - but I do strongly believe that every person's vote should carry equal weight. In practice condorcet voting is entirely anathema to that, and that, coupled with the divisiveness I've already seen when some people have tried to implement such methods on WP so far, mean that I object to this in the strongest possible terms, jguk 22:26, 20 August 2005 (UTC)[reply]
      • Beware of words like "always" in arguments. Voting strategy in Condorcet is simple: rank all candidates based on your sincere preferences. This almost always works, so much so that with the right set of rules in place, the risk of a weird outcome is negligible. Election methods geeks like myself like to argue about those weird cases where it's not a good strategy, and what systems mitigate the strategic problems in the best way, but those cases are really rare. However, plurality elections often lead to strategic problems for voters. Condorcet methods represent the state of the art in election methods. Yes, they are complicated, much in the same way the car you drive today is a lot more complicated than the cars people drove 50 years ago. -- RobLa 22:41, August 20, 2005 (UTC)
        • I have a first-class maths degree from Oxford, and whilst I admit I have never sat down for a few hours to work through the algebra, I really have no idea how best to vote on a condorcet method if I wish my preferred outcome to happen. If I cannot understand it, I'm fairly sure that most others cannot understand it - all I can understand is that those who promote such a method tend to end up celebrating their victory once the vote has closed. One truth is self-evident: condorcet voting is the antithesis of one-person-one-vote, jguk 22:59, 20 August 2005 (UTC)[reply]
          • all I can understand is that those who promote such a method tend to end up celebrating their victory once the vote has closed. Should this be interpreted as hypothetical, or do you have some basis for saying this? KVenzke 13:07, August 21, 2005 (UTC)

It's not hypothetical - and examples of organisers of elections picking a method such that is most likely to give the result they want abound everywhere, jguk 13:49, 21 August 2005 (UTC)[reply]

Could you give me an example of when the condorcet method has been selected by organisers who could predict the results of the election using that method? I've never heard of such an example, and you say they abound everywhere, so I'd be grateful if you could suggest where I could start looking for such an example. Or were you talking more generally about election organizers showing bias in the selection of other types of voting systems? Mamawrites 16:06, 21 August 2005 (UTC)[reply]

Like RobLa, I support Condorcet methods in general (especially Smith-efficient methods), but I am not especially enthusiastic about this proposal. Condorcet methods are IMO the best systems when the goal is majority rule, but that is not the goal here. I find Jguk's objections to Condorcet to be quite weak. He refuses to learn about how Condorcet works, but he continues to make arguments that he cannot support. (By the way, if we were talking about the Borda count, I would agree with many of Jguk's comments. I wonder if he has the two confused?) --Hermitage 22:35, 21 August 2005 (UTC)[reply]

I have requested on Jguk's talk page that he substantiate his allegations with specific examples (which, after all, "abound everywhere"), or else withdraw them. Absent evidence, the claims seem too much like FUD. --FOo 18:27, 4 September 2005 (UTC)[reply]
Update: On his user talk page, Jguk has declined to provide any examples. This bolsters my belief that his claims are mere FUD. --FOo 21:37, 4 September 2005 (UTC)[reply]

Proposal 9

Drop the "The newly elected Arbitrators will be allowed to choose the length of their term after the election from those available, in order of most votes received." bit and change it to "Arbitrators may stay for up to three years.

Rationale It is pretty clear that people leaving to early is the problem rather than people staying to long

Comments

  • This potentially could lead to there being no elections in a given Decemeber if nobody wants to step-down then. I would support an addition along the lines of "One third of the seats will be up for election every december, if fewer than one third of the arbitrators wish to step down at the time of the elections, the seats held by the longest serving X abritrators will make up the balance (where X is the difference between the number of arbitrators stepping down and the number of seats to be contested). The X arbitrators may be re-elected, but may not serve longer than three years in total.". Yes this is very complicated, and it can probably be simplified, but I don't think it leaves any gaps. Thryduulf 20:51, 20 September 2005 (UTC)[reply]
  • Although it's certainly true that people leaving to early is a problem, I don't see how this proposal would help with that; unlike mid-term burnouts, after all, anyone who faces the end of a term limit is going to immediately be either replaced or reelected. If anything, it seems to me that having people standing for election more often and constantly cycling the people in the ArbCom serves to reduce burnout. Possibly, if we wanted to fix the 'midterm burnout' problem, we could have a system where the X people who came in just below the pole in the election can be automatically moved in to replace ArbCom members as they resign midterm (serving out the rest of their terms), or temporarily tapped in by ArbCom members who need to take a break for one reason or another. Of course, there are could be a host of other problems inherent in that, but it might be worth considering. This would avoid both the difficulties of the midterm election proposal above, and the complexities having extra 'midterm-resigned' seats up for election that are referred to in this proposal. Last election we had several people who came in just a few percentage points short of getting an ArbCom seat... It's a shame to let seats stay empty until the next election and to leave the ArbCom understaffed if those people are still there and willing to serve. Aquillion 21:24, 20 September 2005 (UTC)[reply]

Proposal 10

If any person is found to have used more than one account to vote, then all votes cast by all accounts they have used will be removed.

Rationale This is to remove all benefit from using sockpuppets. If User X and User Y both vote, and User Y is found to be a sockpuppet of User X, then the votes cast by both users will not count. This means that by trying to get 2 votes, actually ends up with 0 votes.

Comments

Proposal 11

All arbcom terms will become two years or until resignation, whichever comes first.

Rationale Much like the three year proposal, only with more frequent elections. Also, three years is an eternity - when the first arbitrators were appointed, three year terms were about as long as Wikipedia had existed. Honestly, one year would be ideal, but there should be some continuity ensured in the arbcom.

Comments

Proposal 12

Retain the current number of Arbitrators at twelve. Grant the Arbitration Committee the power to assign a three-Arbitrator panel for cases in which such a panel is thought to be appropriate. All three members of the panel will decide the cases and must agree on each vote in order to assure passage. The decision will be approved or overturned by a majority of the entire committee once the three-member panel has made a final decision. For cases in which the committee thinks it appropriate, the en banc system of having all non-recused Arbitrators will continue to be used.

Rationale
Will speed up process for relatively cut-and-dry cases that don't require a vote of the full committee. Retains the number of Arbitrators at 12, which doesn't make the committee too large or too small. Allows the committee to continue deciding cases as a whole in particularly controversial or difficult cases. Allows for checks and balances of the panel by the full court via the ratification vote. Neutralitytalk 04:04, 6 October 2005 (UTC)[reply]

Comments
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Arbitration_Committee_Elections_January_2006/Proposed_modifications_to_rules&oldid=25198365"