The legitimacy of community consensus as identified by our rather ad-hoc votes has been questioned, and will bring us no end of disputation. For instance, there are those who call the recent two-year ban on MarcusCicero as having arisen from illegitimate voting procedures. Even the recent Community Standards user talk page vote has been blasted for being unfair or illegitimate in its procedures, despite the care its proposer took in crafting them. The trouble boils down to this: what is the point of having votes if anyone can turn around and call them nonbinding because they perceive them to be illegitimate? It has been said that we need "specific, not general, policies for categories of votes." I agree. There is no reason why we can't set up a "standard for voting."
But "categories of votes" are the the most general policy we could craft, the tip of the voting iceberg, as it were. Within the categories, the composition of the franchise, what exactly gets a vote passed (levels of plurality/majority/etc.), how to deal with sockpuppet votes, vote duration, and a few more concerns become very important issues to deal with. In this essay, I will try to develop some preliminary thoughts on how to address the "standard for voting." As I move forward, I'll welcome input from anyone, with our ultimate goal being sitewide adoption of specific voting procedures.
- Note: Any indented sections labeled "proposal" are only meant to provide an idea of what the eventual standards proposal might look like. They are not hard-and-fast.
Categories of votes
Some votes are more important than others. Some are more far-reaching than others, more participated in than others, etc. I see these major types of votes, among each of which procedures will necessarily vary:
- Votes to alter the Community Standards
- Votes to remove a user from the site, or retract a ban
- Votes to penalize a user by anything less drastic than the above, e.g. removal of user rights, or restoration of rights
- Votes to alter site policy unrelated to the Community Standards, e.g. namespace remit changes, category organization changes
- Votes to delete or restore an article
Each of these types of votes carries a different weight, and I have put them in the order of importance I perceive. In the following subsections, I'll outline the major issues for each category, and get into more detail in sections further below.
Category 1: Community Standards changes
As the closest thing we have to a coherent set of policies, even if those policies aren't strict rules, changes to the Community Standards generally carry the most weight of any vote. As such, the legitimacy of these polls is paramount, just overtaking the importance of the category below. Votes in this category must therefore have the longest duration, strictest franchise requirements, and higher thresholds for passing any presented option. The ability for moderators to bootstrap any unique procedures for these votes should also be very limited.
Category 2: Penalization of users
Votes of these type tend to attract much more drama and attention than the above, even if their importance is eclipsed by further-reaching standards votes, as these tend to be questions of how to apply or interpret the standards rather than introducing entirely new standards or changes thereto.
That said, there is a major difference between the internet's version of capital punishment, a complete ban, and lesser penalties such as short-term blocks and user rights changes. For instance, users have questioned the legitimacy of a two-year ban passed by a vote of 20-14, but not much was made of a vote to remove bureaucrat rights that failed 33-34. Because of this, a vote to remove a user from the site requires slightly stricter eligibility rules, a slightly longer duration, and a slightly higher threshold for passing than votes to remove user rights. I'll discuss the implications of the forced removal of moderator rights later, as that's a very nebulous area.
Category 3: Editing policy changes
This category is a catch-all for votes that don't fit into the other three categories. It can include anything from proposed votes on changes to the manual of style, changes in namespace policy (most recently, funspace has taken centerstage in that area)... anything that's not a direct change to the standards or a moderation/chicken coop case, really. These votes tend to carry more weight than simple article deletion, however.
Category 4: Article deletion
Votes to delete (or less commonly, restore) an article are the least-important votes that are still meaningful. The moderators should have the most leeway in bootstrapping specific procedures for these polls, and the requirements for enfranchisement and duration should be reasonably low.
Across the four to five categories of votes, the franchise becomes more important as the votes become further-reaching. There are three main components of a wiki franchise requirement:
- How long an account has been registered;
- How many edits an account has made;
- Whether or not an account is an admitted sockpuppet, and subsequently, whether or not the sockpuppet account owner has also voted.
In general, in any meaningful vote, if both an admitted sock and its owner have voted, the sock vote is stricken. However, if the owner has not voted, the vote is usually kept; there are a wide variety of reasons why a voting by sock proxy can be acceptable. I'll focus the discussion on the first two components, then.
In the most recent category 1 vote, there was a kerfuffle over the proposer trying to change the franchise requirement after the vote had begun, which would have had the effect of throwing out several votes already in place. The need to have hard and fast eligibility requirements for these types of votes is very important, not only to avoid these types of delegitimizing incidents, but to increase the probability that the results will be adhered to.
- Proposal: In order to participate in a category 1 vote, an account must be at least two months old and have made at least 30 edits.
2A: Ban votes
As far as the franchise goes, votes to ban users ought to have a similar requirement to category 1. It is especially important that the results of these votes are accepted as legitimate, the chance of which increases as fewer suspected sockpuppets or newbies participate.
- Proposal: In order to participate in a category 2, type A vote, an account must be at least two months old and have made at least 30 edits.
2B: Other penalties
In a world where any sysop can confer or remove sysop rights, it is very important that a community vote is respected, because so many have the power to overturn it.
- Proposal: In order to participate in a category 2, type B vote, an account must be at least one month old and have made at least 30 edits.
Categories 3 and 4
For these less important votes, it should be within the power of the moderators to set any specific requirements. If moderators do not use that ability, we can and should stick to the policy of common sense that has served us well for years. There have been few high-profile, controversial votes in these categories.
Thresholds and runoffs
The issue of what results get a proposal legitimately passed has been largely ignored or dealt with on an ad-hoc basis since the beginning of the site's modern era. Recently, there have been more visible cases of passing thresholds causing a problem, such as the two-year ban of MarcusCicero. It was passed 20-14 (or 20-16, counting "goats" as "nays," but we'll get to that later), and those attempting to enforce this ban as community consensus have met with a fair amount of community resistance. It is the de facto truth that this ban is not widely held as legitimate, and the only reason for this is the low threshold by which it was passed. In all categories, not just 2B, thresholds are the number one factor in how the legitimacy of a vote is perceived.
In any vote in which more than two options are presented, one key way of legitimizing the final result is through a runoff. The recent Community Standards talk page vote includes a provision for a runoff between the top two vote-getters among the five options, for example.
A strict threshold for passing is the best way of making sure a standard change or new standard is quickly accepted. One that passes 30-28, for instance, will be impossible to enforce.
- Proposal: An option in a category 1 vote must accrue at least 75% of the vote to pass.
If, after the vote concludes, and no option has reached this threshold, then all of the proposed options will be held to have failed. The exception to this is multifaceted votes, in which three or more options are presented (not including "goat"). In that case, the threshold should not be applied until after a runoff has taken place between the top two vote-getters.
As we have seen with the MarcusCicero vote, a ban will not receive the resounding community support necessary to enforce it (especially in a world where most of the userbase can overturn it) if it is not passed by a significant margin. A simple majority is not effective.
- Proposal: In order for a user to be banned for a significant amount of time, they must have at least two-thirds of the vote against them.
For these votes, even though the result can be overturned by as many people as type A, the proposals tend to be less serious. For this reason, a simple majority should suffice for the vote to be binding. A plurality, however, should not be enough to pass a particular option; if no option achieves a majority, the top two options should be voted upon again in a runoff.
As they are the least far-reaching of the three policy-related vote types, a simple plurality should be sufficient for an option in this category to pass.
Article deletion votes present a special case, because thresholds only became an issue when people have attempted to delete articles too hastily.
- Proposal: In order for an article to be deleted, there must be at least four more "yeas" than "nays."
The length of votes is what truly separates the four categories I have mentioned. Duration is another essential part of legitimacy; a vote to ban a user that only lasts two days is obviously illegitimate, for instance, while a vote to delete an article should not be required to last two months to be legitimate.
Moderators should have the power to set the length of votes in any policy-related category, but they should be constrained by certain time frames. If they do not step in, then the lower end of the time frames should serve as the legitimate vote duration.
- Proposal: Category 1 votes must last between two and four weeks; category 2 votes must last between one and two weeks; category 3 votes must last at least one week; category four votes must last at least five days.
Article deletion votes tend to go on for a very long time, if only because people rarely check on Category:Articles for deletion. That there must be at least four votes' difference between "yeas" and "nays," combined with the at-least-five-days stipulation, is sufficient to ensure that article deletion happens legitimately.
The common-sense policy we've had regarding quick deletion of spam or personal information should continue, of course.
The omnipresent "goat" option has been interpreted in many different ways. It has been called an abstention, a extension of the "nay" vote, and a "present" vote, or combinations of those three. We need to make a decision on how to include "goat" votes in determining which option in a vote is passed. For instance:
- Proposal: "Goat" votes must be counted toward the vote total, but are neither part of the "nay" group nor the "yea" group.
In this scenario, "goat" is not a distinct option to vote for, but a glorified "present" vote. Because it is counted toward the vote total, however, the "goat" vote may actually prevent an option from accruing a three-fourths, two-thirds, or simple majority (for categories 1-3), though it wouldn't affect a category 4 vote.
The role of moderators
I have mentioned how the "voting standard" should complement the institution of moderators. This fits with the moderators' remit, specifically:
“”The second function of the moderators is to help boot-strap certain policies or procedures that we struggle with. A great example of this is how to vote for something, how long does a vote last, when does it pass or fail, what should even be voted on? There is no way to address these issues without some people having the power to set a certain amount of basic policy.
Experimentally, we have seen that this does not work on a fully practical level. We need the "voting standard" to supplement moderators' decisions, and in higher-level votes (such as category 1 and category 2), the less invention of procedures by moderators, the greater the vote's potential legitimacy.
How might we pass the standard?
Simply put, it's paradoxical to have to vote on a standard for voting. Nevertheless, we will need to find a way to pass voting standards, with as much legitimacy as possible and through some method of voting. We could have five moderators sign off on a procedure that involves very strict passing thresholds, for example. We could even use a secret ballot. It will certainly be difficult.
Exhortation to take this essay in good faith, to work together, and accept compromise
This essay is not a manifesto bent on destroying RationalWiki. Despite what the noise might be today, John Keynes was no Karl Marx - Keynes wanted to save capitalism. Similarly, I am not trying to destroy RationalWiki as we know it - I truly believe that the "voting standard" will help us overcome some of our most divisive issues. If we continue in a place where votes can mean everything and nothing at the same time, we will collapse as a community. Some have argued that this has already happened, but I still retain hope that the community can heal itself. Nobody is going to tell us how, not I, not moderators, not Trent - we are going to have to figure it out together.
One thing is certain: compromise will be key. Taking a hard-line stance on any issue we face and refusing to budge is counterproductive in the extreme. That's why I haven't proposed anything "hard-and-fast" here; I have only outlined the proposals needed to comprise the "voting standard." Refusing to admit that any perspective other than your own might be acceptable is destructive to our common purpose. It goes without saying that we have all been guilty of that, including me.