32

How can you have a self regulating group of moderators in a community?

Moderators regulate the community, but there is not an official way to moderate the moderators themselves.

Here in Stack Exchange you have the community managers that could remove a moderator if it was necessary, the same in Reddit with admins.

How can this be handled in a community that doesn't have that that extra level above moderators?

3
  • 2
    and this is the first question, where the debate should be in both main, and meta....
    – rolfl
    Jul 29 '14 at 18:23
  • This does mostly sound like a main site question, at least with regards to a moderator team (or if you're in charge of others).
    – Jamal
    Jul 29 '14 at 18:26
  • I don't think there's enough information in this question, as it depends heavily on the medium of communication. If, for instance, a rogue moderator has the capacity to remove other moderators, the situation changes.
    – user35
    Jul 29 '14 at 18:31
12

How can this be handled in a community that doesn't have that that extra level above moderators?

It is difficult to imagine something like that. There is always a superstructure that gives the power to the moderators in the first place. That superstructure is the one that ultimately should deal with problems of the Moderators themself.

To put a more practical example: How do you moderate the executive government? Easy, you create another power that regulates them. Both powers in balance. In the example you provide, this is not possible, unless the user base is large enough to provide the means to build the community itself, as a nation could do. In those cases the people are the ones that shall deal with problems of moderators. This normally ends in creating a special forum/commission/group to evaluate and have the final word on the people with power.

As you can see, the one that gives power is always above moderators, the community itself can remove them if they deem necessary via the same method that put them in place.

5

The answer to "who moderates the moderators" is: somebody moderates the moderators, and if they don't, the community is headed in a difficult direction. There's not much anyone can do about this.

The backbone of any community is a level of users I'll call the "administrators." (In many communities, this is literally what it's called.) The administrators set the core rules in place, perhaps with community guidance, and determine a method for appointing moderators. It is the administrators upon which the community falls when the moderators fail in their duties, and if the administrators are not responsive, the community has lost its support.

Of course, it is fully possible for moderators to moderate each other. However, there are a couple risks associated with this behavior:

  1. The moderators may not know exactly how to respond to each other. This isn't something most moderators have had much or any experience in, and with new experiences (especially with something as trying as moderator moderation), there are a plethora of new issues.
  2. The moderators' self-moderation will spill over into the community, and will possibly poison the community. If moderators aren't careful, the methods they're using to work through problems with another moderator will become transparent to the community as a whole, as certain moderation actions flip-flop or are unclear. This has the potential to confuse users and even poison the atmosphere of a community, which should actively be avoided.

Without an administrative backbone, it is possible to self-moderate, but I would not recommend it. I have only seen this fail.

4

Moderators regulate the community, but there is not an official way to moderate the moderators themselves..... How can this be handled in a community that doesn't have that that extra level above moderators?

The answer is quite simple: it is the stake-holders of the site that moderate the moderators.

Definitions: For the sake of this answer I will call those stake-holders or their representatives the "owner". The stake-holders are those that pay the bills, or the persons who bear legal responsibility for the site.

A site is not created spontaneously with no "owner". It is that owner that sets the boundaries and editorial direction of the site. The owner should retain the ultimate administrative privileges, and moderators - while powerful - should not be able to override the administrative accounts of the owner.

If a situation arises where a moderator has gone rogue or is failing to enforce the editorial direction then the owner takes remedial action against that moderator. What remedial action is taken is beyond the scope of this question, but the owner must always be in a position of being able to enforce the ultimate sanction - which is the removal of rights from the moderator.

What if the stake-holder has no active contribution to the site? (What if they die? Become incapacitated? It's a free hosted site and they walk away?)

In this case it is up to the moderators to self regulate. If they can't then the community will eventually devolve until it either becomes stagnant, the hosting provider shuts it down, or the law steps in.

You could argue that any n number of moderators as a group can take remedial action against a specific individual moderator, but that is making the huge assumptions that:

  • the stakeholder left the moderators with sufficient powers and functionality to assert that action
  • you are living in a utopian world where that group taking action hasn't themselves gone rogue

So in this situation nobody can really moderate the moderators.

Can you just set up a site with some moderators and then walk away allowing it to self-manage?

I think that due to human nature it is ultimately impossible for a community to survive forever in this state without the site devolving, so the best option is to have a fresh stakeholder assume responsibility (if possible).

3

Slashdot has this implemented in interesting way.

There are randomly chosen moderators - a system chooses random users using a weight function that considers user's reputation (karma), visit frequency, and some other factors, including a hidden factor of "moderator reliability". A chosen moderator receives a pool of "votes".

The moderators moderate by assigning positive or negative vote to posts (along with tags like "Insightful", "Informative", "Funny" for positive, or "Troll" or "Offtopic" for negative), and users can filter off the chaff by score if they don't feel like reading troll posts, or assign extra weight to tags, e.g. bump "funny" up if they look for some laugh.

And then there are the metamoderators - again, randomly chosen from "old, reliable" users, and they each receive a random bundle of moderated posts, and decide whether given post was 'fairly' or 'unfairly'. Each "fair" increases, and "unfair" decreases the "moderator reliability" (it's not a linear function; you can't accumulate tons of 'fair' and then start unfair moderation with impunity). Anyway, if you get enough 'unfair' metamoderations to your name, the randomizing system simply won't give you any votes to distribute.

One more point: you can't both vote and comment in the same thread.

3

They key for a team of moderators is mutual respect. With it, they can discuss actions and better learn from mistakes. Without it, there will always be somebody beyond reason.

This is why SE moderator teams generally work so well. In the Ubuntu community we also have things like community councils with elected members where users can take their grievances. It's that sort of oversight that keeps massive communities from spinning out of control.

But these are far from typical, most smaller communities are one-person operations. Consider small one-topic forums, bug trackers, fan sites. They're not hard to imagine because they exist everywhere. Mostly one person and it's their house; you have to follow their rules.

Even if jobs are farmed out, there's always going to be that administrator at the top of the food chain. If you have a problem with an underling, take it to them. If you have a problem with the sole admin directly, you're stuck.

2

You may wish to use a form of election or poll to see whether the public opinion still values a specific moderator's input. By seeing what the general consensus is, you may wish to adjust your actions accordingly. Alternatively, you could be effectively like a dictator and say when enough is enough, and remove moderators or suspend them as you see fit. This is only applicable if you are the community's owner - after all, you created it so you should call the shots.

Another way of looking at it may be to provide some kind of training. I created a Moderator Guide for my moderators, which covers the different kinds of issues, how to deal with them, how to act professionally, and what punitive actions to effect on members (and how they correlate to their misdemeanour). This way, everyone is on the same playing field.

2

There are often two levels to this. The primary level is generally self-moderation. Other than being able to spread work-load around, one of the primary reasons for having a team of moderators is that complex or controversial issues can be reviewed by more than one to establish consensus. If a moderator is out of line, then the other moderators can often reel them in.

Failing that, community owners are generally the primary direction setters for the moderation community. Someone pays the bills for the community and often has the final say on how things are done because of that. Ideally, they shouldn't interfere unless the general direction is going astray (for example, the moderators as a whole are drifting off target). They shouldn't have to take action often, if ever, but it is the last line of defense.

I suppose arguably the next step beyond that is the community itself. If the moderators and the owner take the community in the wrong direction so severely that problems with leadership can't be addressed, the community itself can simply leave and form another community where the problems are addressed.

2

Tl;dr:

If moderators of a community know that they are monitored well by the community and that the community can remove them at any time, the community does not need a group of special people to moderator moderators.


To answer how moderators can be moderated, we should consider the following questions:

  • How are moderators selected in a community?
  • Is there any superstructure above the moderator level?

Regarding the above questions, the following solutions are usually provided:

  • If the moderators of a community are elected by the community themselves, then they can monitor moderators actions so that they can remove some moderators in case they have abused their privileges.
    However, I think this solution cannot be applicable in all situations because accused moderators can claim that they cannot reveal all information because of some sensitive issues, and so the community cannot judge all moderators actions.
  • If a community has an owner or administrator team, then they themselves should moderate the community moderators because they own the community, they pay its costs, and they develop and administer the community platform.
    However, I think community administrators should not interfere in moderation or meta issues and make brute decisions in such situations. I think community administrators should not have special privileges in meta issues only because they pay the bills and run the platform; in fact, since they are the administrators of a community, they deserve making final decisions on the structure of the platform. But, the community members spend their time and energy to contribute to the community, so they should have the same right to judge meta issues as any other people have.
    Also, some community members may suspect that there may exist some friendship between the community administrators and moderators, so they may doubt whether prospective appeals will be treated fairly.

So, what is the best solution?

I agree that the community themselves should moderate the moderators; however, to avoid the possible problem explained above, we need to add the following features to the community:

  • Transparency
    Some people believe that some moderating actions should not be revealed because of some sensitive issues. However, there are two points questioning such a belief. First, if knowing some information may lead some community members to abuse their privileges, this means that the platform or the community structure has some bugs (holes), so the administrators should try to fix such bugs, instead of evading them. Second, any feature has its pros and cons. Even if there were no solution for the the problem mentioned in the first point, one should not neglect the benefits of "transparency". With transparency, community members can trust their moderators and administrators; trustworthy should be considered as one of the first priorities for each community.
    It is worth noting that it is not needed to reveal all moderation details; it is only needed in case there is some community member insisting that some community moderator has abused their privileges, and so by revealing the details, community members can judge related actions well.
  • Flexible Removal Process
    Moderator actions being able to be monitored, a community needs to have a process such that they can remove any moderator at any time; they should be able to remove a moderator in the same way as they selected them. Before such a process, the community can be informed of community members' opinions about a moderator from votes on a meta post, for example.
0

My solution would be to show the other moderators the moderation log and allow them to vote-undo the "rogue" mod's action. That way there's no "extra level" and everyone's on equal footing. Any mod's action can be undone by a simple majority of the other moderators. Of course, then you'd have to highlight the modlog to the other mods.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.