Harassment workshop

Translate this page

The Harassment workshop is the second in a planned series of discussions on the topic of harassment on the Wikimedia projects. The first consultation, launched in November 2015, collected some strong ideas and conversation on possible solutions and projects.

We have identified four major topic areas from the previous consultation and have presented them here along with relevant discussions and proposals from the past. This stage is meant to be a workshop on practicalities and approaches for specific solutions, with the goal of distilling ideas into actionable plans which could then be implemented by community members or the Wikimedia Foundation, as resources allow. We have listed some ideas for discussion under each topic area; please feel free to give your opinions in the sections below, or, if you don't see something listed that matches what you're thinking, contribute it to the open discussion section at the bottom of the page.

We ask that participants observe the Friendly spaces expectations when discussing these topics.

Technical

Background & context

Wikimedia projects use several varieties of machine-learning, detection, and edit filtering and flagging to assist contributors in detecting vandalism, spam links, and other unwanted content. There is potential to apply this concept to user talk pages and discussions to flag abusive posts. Relevant pages and discussions:

Modeling Talk Page Abuse
English Wikipedia: Village Pump discussion on edit filter for talk pages
Discussion on the Community Wishlist survey
Outside Wikimedia: "Doing Something About the ‘Impossible Problem’ of Abuse in Online Games", Jeffery Lin, Riot Games

Another potential improvement in the technical area is improving the tools available to the administrators who work to block harassers and abusive accounts. Relevant pages and discussions:

Discussion on the Community Wishlist survey

Existing initiatives

Contributors had a number of ideas for technical solutions in the last round of consultation, including:

Please share your thoughts about how any or all of these could be designed and implemented.

Discuss technical

Comment in this section

Let's not pretend that technical measures will stop harassment. But even if they reduce the problem by 1% it's still worthwhile because:
- An obvious first step is to knock out some of those Phabricator tickets. There is some real low hanging fruit that the WMF can tackle right now -- to the tune of a few days of developer time.
  - Which tickets? Are you referring to your links in 2015_Community_Wishlist_Survey/Moderation_and_admin_tools#Improve_MediaWiki.27s_blocking_tools? Nemo 08:45, 1 March 2016 (UTC)[reply]
    - Yep. MER-C (talk) 07:42, 2 March 2016 (UTC)[reply]
- Some technical measures will work across the board, helping to remove other unwanted content and users.
I would leave the machine learning stuff for now because it is difficult and, hence, expensive. Limited resources are probably better spent on hiring a harassment expert. They should know what the state of the art is and who to ask as well. MER-C (talk) 06:03, 19 February 2016 (UTC)[reply]
The biggest problem in my opinion is that it is not possible to register a confidential complaint about harassment. I would like a technical solution to this. The solution would be that if someone sends an email somewhere, then the email goes into something like OTRS with restricted access and managed control with a paper trail of registering the complaint. In my mind, being able to note harassment is a pre-requisite of addressing it, and right now, there is significant pressure on the community to prevent the community from registering complaints. The major pushback is that no one wants responsibility for acknowledging the existence of complaints, and because of the presumption that if complaints are made then someone becomes responsible for the impossible task of fixing all problems reported. I proposed this at Grants:IdeaLab/Centralised harassment reporting and referral service. I know that it would freak everyone out, but I would be comfortable if WMF would setup a locked email account without a plan to respond to emails just to get counts of how many people have complaints. I have trouble imagining a path forward that does not include a way to register a complaint that "help I am experiencing harassment", yet most plans tend to presume that allowing people to register complaints will never be possible. I do not care whether registration happens on or off wiki, so long as the complaint repository is held by an honest broker. Blue Rasberry (talk) 18:19, 29 February 2016 (UTC)[reply]

+1. A pile of examples will also help us craft a definition of harassment that is specific to our circumstances and will help with onboarding a harassment expert. MER-C (talk) 07:35, 1 March 2016 (UTC)[reply]

Training and education

Background & context

Currently, there is no formal training on the topic of harassment available for Wikimedia's front-line functionaries and admins. Although some training has been offered at past Wikimanias, neither the WMF nor the communities have existing structure or providers to do this on a larger scale. Relevant pages and discussion:

Gender Gap Allies training

Existing initiatives

Contributors provided a number of ideas for training/educational solutions in the last round of consultation, including:

Please share your thoughts about how harassment training could be designed and implemented.

Discuss training & education

Comment in this section

First, I don't think this "clickable button" is working...at least, it just brought me back to the same place I was before I clicked it.

While admins, functionaries and arbitrators are more often targets of harassment because of their visibility and the fact that their decisions can affect other editors, I think that there needs to be more education for both admins and editors as what constitutes "harassment", "personal attacks" and "bullying". I often see incidents where Editor A makes an big statement, several editors disagree with their comment and Editor A says they are being bullied. Meanwhile, Editor B can make continual belittling comments to Editor C and the admin reviewing the situation sees it as normal internet belligerence that doesn't merit a block for personal attacks or harassment. So, in one case, the bar for bullying is too low (merely expressing disagreement) while the one for harassment is set way too high. Often personal attacks aren't seen as harassment until some form of outing occurs and that is way, way too late for action to take place. So, I think these terms need to have well-defined meanings that admins and editors can easily understand when behavior crosses the line into abusive.

Here's an example from this morning. A new editor posted "Would like to spent a night with you" on my talk page. Clearly unwanted and probably posted because I'm female. So, I reverted the edit. Another editor was alarmed and upset with the editor but this is not a comment that would trigger an edit filter or could be seen as a personal attack. As far as I was concerned, if I received repeated comments like that, it might constitute harassment but an isolated comment like that is just trolling. But I've seen far, far worse comments posted to editor and admin's talk pages and there is even worse stuff that gets sent via email. Having clear definitions on what is harassment and getting the word about this to editors and admins of all Wiki projects is a step in the right direction. Liz (talk) 21:34, 18 February 2016 (UTC)[reply]

@Liz: The buttons are sort of clunky, but I suspect you encountered the intended(ish) behavior rather than a bug. Because of some funkiness in the way mediawiki works section headers and the translation extension, I can't link the button directly to editing the relevant talk page section (because doing that uses section numbers, and as soon as someone creates another sub-heading or something on the talk page, that number would have a good chance of pointing to the wrong place) and I can't rely on the usual "edit section" links (because a translatable page has section editing disabled), so a button taking you to reading the relevant talk section is the best I can do with what I know. If you click one of the buttons from the main Workshop page, it should take you to the discussion section on the talk page for the topic whose button you clicked, so that you can edit that talk section as you usually would. If you click them while already on the talk page, it takes you nowhere. Please let me know if your issue was clicking the button on the Workshop page and ending up nowhere (because that would be a bug I need to fix), as opposed to clicking it on the Talk page and staying where you were (which is just sort of a side effect I can't get rid of). Or if you know a workaround for my workaround - definitely let me know of that too, because it pains me to use inelegant solutions like this :) Kbrown (WMF) (talk) 22:43, 18 February 2016 (UTC)[reply]

One complication is that what is acceptable depends on cultural and personal background, and hence the blurry line will be in different places across Wikimedia projects.

It's a long standing adage on en.wp that when someone screams "admin abuse", they are likely abusing adminstrators. That said, I agree training will help deal with potential cases of harassment. MER-C (talk) 06:36, 19 February 2016 (UTC)[reply]

Some general observations:

How much resources are available for this? The number of admins on all projects numbers into the thousands. Making it compulsory for stewards, arbitrators, checkusers and oversighters and optional for admins and OTRS members may be a good first step.
Training must be designed and continually updated in consultation with the community so that it reflects the circumstances we face. The development of training also requires the WMF to hire a harassment expert so that the training reflects best practices in the field.
I successfully proposed in the Community Wishlist survey for a user watchlist, especially to track problematic users. Should it be necessary to get this training in order to use this tool?
Training a whole bunch of admins isn't going to do anything to stop the next GamerGate. MER-C (talk) 06:36, 19 February 2016 (UTC) (revised 04:45, 20 February 2016 (UTC))[reply]

You're right that training admins and arbitrators can't prevent a social conflict from occurring but they would have better sense of how to respond when sorting out the mess that inevitably spills on to Wikipedia. I think it is partly due to being a little uncertain of how to deal with Gamergate type events that resulted in only a handful of admins overseeing the related articles. Usually, only one or two admins at a time will have anything to do with the subject area unless a case is brought to AE...most administrators won't go near the dispute with a 10 foot pole. Training could result in more admins (and editors) responding to future conflicts which will certainly happen. But first we need agreed-upon definitions on what behavior constitutes harassment and personal attacks instead of a vague "I'll know it when I see it" outlook that many have today. Liz (talk) 20:22, 20 February 2016 (UTC)[reply]

Training and education are not isolated events where people complete an online course or attend an in-person class, but rather, they are part of community building and part of a larger plan to establish and strengthen inter-personal relationships among community organizers. Before anyone considers in-person events, consider that in-person Wikipedia management is a major hassle for community members. When WMF staff visit a Wikipedia group, they expend a huge amount of resources without thought. When Wikimedia community members want to meet, it takes ~150 hours to get a small grant that covers a year of coffee and cookie funding. If anyone wants to overlay training on top of community meetups, first secure the little things like making sure that internationally any group that wants milk and cookies for 2-hours meetings gets it. It can be more discouraging than doing nothing at all to see expensive training provided when the community has to scramble to make sure that routine human courtesies are provided in regular meetups. The money ought not even be the issue. People could buy coffee but getting little courtesies from the WMF goes a long way to set up community infrastructure and allow volunteers to work in-person to address harassment problems. Never presume that providing in-person training will be a default good. Make sure that any money spent on training matches the regular event budget of the group getting the presentation. Blue Rasberry (talk) 18:27, 29 February 2016 (UTC)[reply]

In Greece we have solved harrasment through education. We run Wikipedia School (long live training sessions open througout the whole year) where we teach Wikipedia in a way that has also elements of martial arts philosophy. We create love anti-bodies in trainees. Wikipedia School is really a Wikimedia School also, with high internal community health, as seen from willing to participation to our new Wikipedia Association in Greece, currently under formation.ManosHacker (talk) 05:20, 3 March 2016 (UTC)[reply]

This is extremely interesting, ManosHacker. Is there a link to learn more? Nemo 10:33, 20 March 2016 (UTC)[reply]

Hello, I think that the first step is establishing an anti-harassment policy that states which behaviour is encouraged in Wikimedia projects and which is disencouraged / forbidden. This document must be agreed by consensus with enough time to solve any disagreements.

The next step is to promote the policy widely in all Wikimedia projects, to encourage a healthy environment. Everyone must be aware of the importance of preventing and solving harassment, and know how to do that.

And the final step is to enforce the policy

I stress the importance of doing it in order. You can't enforce the policy if it's not written or agreed. --NaBUru38 (talk) 22:39, 19 March 2016 (UTC)[reply]

I think it should be noted that even after harassment has been defined, plans of action established and training is given - not everyone has a temperament for this kind of work. It requires a kind of empathetic neutrality that is rare. No matter the intentions behind a conflict between people, an inept arbiter can make a mountain out of a molehill, while the reverse is true of an arbiter with the skill to see the issue from both sides, without judgment. A great arbiter can create accord where there was conflict. Just my two cents - the performance of arbiters must itself be reviewed. Housiemousie (talk) 18:22, 18 April 2016 (UTC)[reply]

Improving dispute resolution

Background & context

The Wikimedia projects have varying forms and methods of resolving and mediating disputes between contributors. Unresolved problems can fester, and contribute to long-term animosities and battleground behaviour. Relevant pages and discussion:

Existing initiatives

Contributors provided a number of ideas for improving dispute resolution in the last round of consultation, including:

Please share your thoughts about how any or all of these could be designed and implemented.

Discuss dispute resolution

Comment in this section

I think those projects, which does not have its own dispute procedures and are missing arbitration comittee would fruit from Meta's AC.--Juandev (talk) 14:21, 18 February 2016 (UTC)[reply]

(Spanish) Creo que sería francamente positivo que se pusiese en marcha una comisión de arbitraje y resolución de conflictos a nivel global para que éste dirimiera las discordias que continuamente se acumulan irresueltas en requests for comment para aquellos proyectos que no cuenten con mecanismos de resolución de conflictos o de comisión de arbitraje. Obviamente el método RfC no funciona, porque nadie ostenta la autoridad necesaria para resolver, incluso de manera forzosa, los conflictos que se generan en el seno de wikis pequeñas y medianas; con la consecuencia de generarse más conflictos, más tensión y desde luego menos confianza mutua. No podemos seguir así. No podemos continuar pidiendo a los usuarios que rellenen RfC's para que nada pase luego, ni apelar a que ello sería añadir más burocracia cuando en realidad supondría dotar de garantías a un proceso. Los stewards no somos árbitros, esto tiene que quedar muy claro y por ello es necesario la creación de un órgano resolutorio donde personas interesadas y capaces para la mediación y el arbitraje puedan trabajar en dirimir estas discordias que, como digo, inundan Meta aguardando una respuesta por alguien o algo que hoy en día no existe. Los stewards, excepto en los casos de flagrante y evidente abuso o emergencias no podemos intervenir, ni se nos puede forzar a asumir un rol que no está contemplado. Pero hasta llegar a ese extremo -de emergencia- hay escalas intermedias que de igual modo necesitan ser atendidas. Necesitamos un órgano especializado con un procedimiento reglado y tasado, a la vez que sencillo. Y ágil a poder ser. Gracias. —Marco Aurelio 15:40, 18 February 2016 (UTC)[reply]
It is necessary to create a Global arbitration committee.--6AND5 (talk) 16:30, 18 February 2016 (UTC)[reply]
I agree there should be a global arbitration committee. Global dispute resolution procedures should have an opt out for larger projects. MER-C (talk) 06:41, 19 February 2016 (UTC)[reply]
Talk:Harassment consultation 2015/Ideas/Blocked indefinitely#Meta arbitration--6AND5 (talk) 23:31, 19 February 2016 (UTC)[reply]
- On the other hand look at the statement of user Vituzzu. I hypothesize that on the big projects community is strong enough to deal with power abuse, so they probably dont need AC, but smaller projects definitelly do.--Juandev (talk) 09:01, 20 February 2016 (UTC)[reply]

We need a less corrupt AC on the Kabardian Wikipedia. 96.237.27.238 13:11, 28 February 2016 (UTC)[reply]

I have found that there is an alarming ignorance among Wikipedians (including administrators and arbiters) regarding what rights an editor might have. The attorney, Fred Somers discusses the matter of expulsion of members of private clubs on his webpage. In his concluding remarks in the case Pinsker v. Pacific Coast Society of Orthodontists, the judge ruled that an "obligation flowing from this 'public service' status is that defendant organizations may not reject an application [to join the organization] without affording the applicant a fair opportunity to answer the charges against him."

The structure of the ANI procedures in the English Wikipedia, unlike the structure of Arbcom hearings, denies the accused a fair hearing. In the Arbcom structure, each individual has their own subsection where they can make comments, but in an ANI hearing, all postings by everybody including the accuser, the defendant and passers-by are all interlinked (which creates a fertile ground for bullying). I believe it mandatory that an Arbcom-like structure should be implemented for ANI hearings in the English Wikipedia, especially after one administrator told me "Wikipedia does NOT have a formal justice system - we have WP:CONSENSUS… Maybe 'mob rule' does equal WP:CONSENSUS ... "

One of the consequences of this statement is that I was blocked for trying to defend my rights. As a result of ignorance of basic legal principals by various administrators coupled with malicious mistruths published by other editors (which I was unable to refute due to my exclusion from the relevant pages on Wikipedia) led to sanctions against me being increased to the extent that my e-mail facility was blocked. When I quoted the above law in a UTR application for readmission, the administrator who handled the application wrote to me telling me that Californian case law does not apply to Wikipedia. This statement suggests to me that this administrator has never opened a law book, attended a course on law or even looked up the Wikipedia article en:Case Law. This was followed a few days later by another administrator (who appears to me to be similarly ignorant about legal matters) blocking my UTR access for three months for so-called "wikilawyering".

I propose therefore that the section "Dispute resolution" on the English Wikipedia page "Civility" should require that the defendant shall, in accordance with the Law of the State of California (Pinsker v. Pacific Coast Society of Orthodontists), be afforded "a fair opportunity to answer the charges against him".

I propose furthermore that any administrator who makes a ruling concerning legal matters (including what is and what is not "wikilawyering") must be prepared to publish, if requested, the extent of their legal training and/or knowledge. This will at least give the victim of bullying some respite should the bullies overstep the law.

Martinvl (talk) 12:02, 27 July 2018 (UTC)[reply]

Policy and enforcement

Background and context

The Wikimedia projects have many behavioral policies, and several different approaches to enforcing them. Some projects lack any policies at all, while some suffer from overlapping or unclear practices. Often, enforcement of behavioural issues is a "pain point", dividing communities and producing inconsistent results. Beyond the Wikimedia Foundation's Terms of Use, there is no movement-wide expectation of behaviour.

Relevant pages and discussions:

Existing initiatives

Contributors provided a number of ideas for policy and enforcement, including:

Please share your thoughts about how any or all of these could be designed and implemented.

Discuss policy and enforcement

Comment in this section

Some harassment may exist because of missing project policies. What about to provide basic policy kit for all new and small projects. Than its up to projecets to change them regarding their needs. But If they will not do that, they can use and exists regarding that basic kit.--Juandev (talk) 14:25, 18 February 2016 (UTC)[reply]

Some general observations:

Trolls, advocates and other undesirable internet denizens don't care about fancy codes of conduct or covenants.
Enforcement is sub-optimal when one can trivially create another account and continue on as before.
Default policies should also be considered, but this will have to be a community initiative (bundle this with default dispute resolution?).
I support the WMF hiring a harassment expert urgently because they can help us figure out where to spend limited resources and if we're missing anything.
What's missing here is more resources for the WMF Legal department so that they can send C&Ds and take legal action against the worst cases. MER-C (talk) 06:37, 19 February 2016 (UTC)[reply]

Some Wikis have tiny communities or no community at all. They suffer from issues:

Admins or bureaucrats do not or hardly understand the language. That should be prohibited by a general rule.
Admins are active forever and cannot be decommissioned, because
- They themselves set the rules,
- Nobody dares to challenge them. Remedy: bureaucrat and admin power life times must be automatically limited from the outside.
Admins or bureaucrats indulge in illegal actions, some secretly delegate those to someone outside their jurisdiction. They cannot be identified as persons, they cannot be taken to court, and their destructive actions are so irreversible. The Wikiverse lacks a way to appeal. --Anonymous673 (talk) 11:45, 9 March 2016 (UTC)[reply]

@MER-C: I recently requested that the WMF Legal Department make a ruling regarding a comment by an administrator that I believed to be wrong in law. The WMF Legal Department replied that since their function was to represent the WMF, they would have a conflict of interest if they were to follow up my request. Martinvl (talk) 15:22, 31 July 2018 (UTC)[reply]

Open discussion

Contribute a new idea, or talk about meta-level issues - go for it!
Comment in this section

Programs being done by Wikimedia affiliates as mentioned in 2015-2016 FDC proposal

Several Wikimedia affiliates put plans to address on wiki conflicts in their Annual Plan and proposal to FDC. It would be good to follow up on these to see if they are happening. Includes gathering metrics. Sydney Poore/FloNight (talk) 18:47, 17 February 2016 (UTC)[reply]

From WMNL study even from the findings of previous harassment report performed by WMF is clear that Wikimedia community does not cover the natural sample of people if we look on motivation. It is cristal clear, that most of Wikipedia contributors are beaurocratical type of people (it means those, who prefer clear policies and ordered work, but who are not so interested in interpersonal communications). This in fact helped Wikipedia to jump as a test project very quickly to success, but I dont think it is something, which would help it for the future. Other "social" projects, created on the model of Wikipedia, where more open to other motive types and created much natural communities.

As a social type, I personally dont feel confortable in Wikimedia community and I was affraid that predominant beaurocrat types wont incorporate sociable and other people as they will argue "Wikipedia is not for socializing, but for contributing". But with the study of WMNL I am surprised that the community recognizes this as a problem and would enjoy "more women", who they define as "more sociable". But its not just about the women, there are also sociable man.

So definitely more sociable people, could help to decrese tensions between editors. Here I would link my idea of Environment for sociable people (profile sociable project linked to others). And that is my contribution to this consulation.--Juandev (talk) 22:59, 21 February 2016 (UTC)[reply]

Safe space task force

This was proposed some time ago, but I found it when I google the topic recently--it seems it was dropped somehow. This is a newish topic that has had some limited success but needs some tweaking. https://meta.wikimedia.org/wiki/Grants:IdeaLab/Friendly_space_task_force —Neotarf (talk) 23:36, 17 February 2016 (UTC) Pinging the previous participants: Econterms, Ftripodi, Slowking4, Missvain, Djembayz.[reply]

Food for thought

[1]. MER-C (talk) 01:40, 31 March 2016 (UTC)[reply]

Wikiversity

I have an ongoing research effort at Wikiversity: v:Dominant group/Wikipedia which has found a dominant group on Wikipedia that is either inadvertently or systematically discouraging women contributors to Wikipedia. The study of the test group is still under way. A second effort v:Dominant group/Wikiversity is studying the loss of an entire women's volleyball team or group from contributing to Wikiversity. It is still in progress but the model that best fits their actions points the blame at the group's advisor who made no effort to let Wikiversity know that his students had chosen Wikiversity to host their high school project. As a control group, the study uses v:Upper Limb Orthotics. --Marshallsumter (talk) 17:59, 10 June 2016 (UTC)[reply]