Talk:Training modules/Online harassment/First draft

General comments

Is there a reason the font for the test modules is so small? It's very inconvenient to keep changing the display size and then try to scroll to find your place in the text. —Neotarf (talk) 21:26, 16 January 2017 (UTC)

That one is my fault! I designed those templates on the fly (we may not even use them in the final modules, depending on what design we end up using for the modules) without putting too much thought into final appearance. Would it resolve your issue if I set them to the same font size as the rest of the page, or do you also feel like you need some kind of navigation/page anchor functionality added, as well? Kbrown (WMF) (talk) 14:52, 17 January 2017 (UTC)

Please yes, if you want people to look at them. My bifocals are not up to the challenge. I was going to either boldly change the font or try to read it in situ, until I saw the template. —Neotarf (talk) 23:15, 18 January 2017 (UTC)

Done! How do they look to you now? You might need to clear your cache to get your browser to pick up on the changed formatting. Kbrown (WMF) (talk) 23:43, 18 January 2017 (UTC)

Yes, much better, will try to look at them later, after I find the refs you asked for. —Neotarf (talk) 00:15, 19 January 2017 (UTC)

...(much later) Wow, sorry but my eyes are just glazing over reading this, and considering how many arbcom cases I have read, I have a pretty high tolerance for closely written text. You may want to try the kind of format we ended up with the simple MOS. Maybe a simple paragraph followed by a short quiz with the answers hidden and links to a longer discussion to see the rationale for the questions missed, i.e. "The types of harassment are x,y,z." "Which is not a type of harassment: a,x,b,y,z?" I would aim for maybe a 10th grade level.

A couple other comments:

• The common meaning of "actionable" is that something can lead to litigation, i.e. "giving sufficient reason to take legal action" or "subject to or affording ground for an action or suit at law"[1] It's an odd word choice given the WMF and admins' hair-trigger sensitivity to legal topics, see WP:LEGAL; I have seen users blocked just for typing the word "defamation". If by tellng someone something is "not actionable" the meaning you are trying to convey is "we do not have a solution for this and we are not willing to work with you to solve it", and slam the door in someone's face, it conveys that pretty well. A more consumer-oriented approach would be something like, if you can't solve their problem, then refer them to someone who can, and walk the problem over to that area, to make sure the ball doesn't get dropped.
• Is there some way to refer to editors who have been targeted for harassment as something more dignified than "victims", and convey the idea that they are valued? Maybe "targets" or even just "users" or "editors"? They are not damaged goods, they are still your customers in that they are the ones who are providing the free services that keeps the Wikipedia running.

Thank you for doing this, IMHO it will take a lot of work but is worth doing. —Neotarf (talk) 00:47, 29 January 2017 (UTC)

Deja vu: whatever happened to hiring an expert?

Whatever happened to the discussion at Grants:IdeaLab/Community discussion on harassment reporting? Didn't we just have this same conversation two years ago? How many of those people are still around? Four are gone; one, Kevin, is dead.

The page is really a compendium, curated by several librarians, of the ideas dealing with harassment that came out of the initiative. And Lila herself, who was then ED, answered this here. IIRC, at the time, the budget was under discussion, and from this comment, it looks like she intended to budget for a harassment expert, and she probably did so. (I suspect that salaries are not paid out of a grant budget but come out of the general budget.) Does anyone know what happened to this?

At the time I also proposed Comprehensive harassment and privacy policy which was basically a blueprint of elements for a successful anti-harassment policy and was based on an out-of-print book by an HR professional. (Ping me if you want more info about the source.) And SV proposed Hire a harassment expert.

This is not rocket science, harassment is a actual field of study, although the exact name for these departments has changed with time. Every time this comes up, it seems very evident there are a number of WMF employees who do not know what a standard harassment training for new hires looks like, and that this is not part of the WMF onboarding for new people. How is this not standard in California? Has WMF just grown so fast they have not been able to keep up with standard HR practice? And how can you have a conversation about this type of training with staff members who don't know what it is. —Neotarf (talk) 23:26, 16 January 2017 (UTC)

"Correct answer"

Hi, thanks for all the work you put into this! I just have a quick remark: You are using the term "correct answer" throughout the tests; I think this sounds a little harsh and probably a bit patronizing. How about using a term like "most appropriate" or "suitable" answer? Cheers, --Nicole Ebber (WMDE) (talk) 15:05, 26 January 2017 (UTC)

Overwhelming amount of content

A training module is something you should be able to get through without fatigue in one sitting. The amount of content in this one is overwhelming. When I do user testing with instructors using Wiki Ed's training modules, it typically takes 30-60 minutes to get through a module that is, I would estimate, 1/5th as long, and those people are highly motivated. This is likely several hours worth of careful reading. It will be more effective if you 1) cut it down as much as you possibly can (and edit it ruthlessly for easy readability), and 2) break it up into multiple modules on distinct subtopics.--Ragesock (talk) 20:20, 2 February 2017 (UTC)

Comments on specific sections

Comments on H2: "Basics"

I have a question about who this is written for. This section says: "The Support and Safety team is working to improve these processes on our end, but the majority of harassment complaints will be seen by you, the functionaries, first." This is needed for a much larger group than the functionaries. SarahSV ^talk 00:08, 7 January 2017 (UTC)

+1. Someone who becomes a functionary for life by fighting vandalism on Pokeman articles does not automatically have the temperament or skillset for this. For starters, the people who do this need to be accountable, perhaps NDA, and perhaps rotate or comp time, because burnout. Functionaries do need trained how not become harassers themselves - perhaps a different type of training. —Neotarf (talk) 23:48, 16 January 2017 (UTC)

As a matter of practice, a lot of reports on enwiki relating specifically to off-wiki harassment come to the functionaries or to arbcom, because the victim doesn't want to draw public attention to the problem and/or because part of the harassment involves revealing personal information. No doubt many people other than functionaries can benefit, but they're the ones likely to end up with private complaints. Also, functionaries do sign an NDA, though maybe not specifically with the scope you're thinking of. Opabinia regalis (talk) 00:46, 19 January 2017 (UTC)

And what if they refuse to follow it? What if the ones harassing and posting personally identifying information are the functionaries? —Neotarf (talk) 02:07, 19 January 2017 (UTC)

A comment on the "common forms of harassment" section: I think this should mention the publication of personal information, either on or off-wiki; this is one of the more common forms of harassment we hear about. "Online stalking" isn't quite the same behavior. Opabinia regalis (talk) 00:51, 19 January 2017 (UTC)

Comments on H3: "Handling harassment reports"

• I understand the point this is trying to make, but I'd suggest editing the example to sound a little more complex. A report of outing with a link to a specific diff is usually quick to investigate and suppress if needed; it'd be slower to write this kind of holding-pattern response than it would be to just deal with the issue and then give a more definitive response. Opabinia regalis (talk) 00:58, 19 January 2017 (UTC)
• In the "third-party report" section, there is a statement "your investigation and any outcomes will not be contingent on the target's approval". I think this should be weakened a bit; in some cases the most immediate risk is to a specific victim and it's not reasonable to expose them to that risk over their objections because of less immediate or more hypothetical possible risks to others who haven't yet been targeted. Making this statement so definitive might discourage third-party reporters for fear that they'll expose the victim to risks they aren't willing to take on. Opabinia regalis (talk) 01:02, 19 January 2017 (UTC)
• As with some of the concerns expressed below, the last paragraph about "considering the reporter's community status" needs some refinement. In a sense, it's absolutely true that people handling harassment reports will consider this stuff; if nothing else, someone might have a reputation for submitting bogus harassment reports or for harassing others by trying to get them in trouble. But the way it's current written, it sounds like the subtext is "be extra skeptical of reports from people who have a poor community reputation". Of course we know that in practice some people end up with a poor community reputation, and might even find themselves embroiled in dispute resolution or blocked from editing, because they behave erratically or appear to overreact, which they do because they're being harassed. Opabinia regalis (talk) 01:24, 19 January 2017 (UTC)

Comments on H4: "Communicating with victims of harassment"

• Regarding "Nothing we can do" - however empathetically it's phrased, this is often the result. One of the most common patterns observed this past year is a report of the form "someone has posted insulting things about me / exposed my personal information / impersonated me on another website". In most cases the only thing we can do about that is submit a report on the other website or encourage the complainant to do so. Often the other site is not one specifically focused on Wikipedia, but is a larger social media platform, and if the offending content taken out of its Wikipedia context isn't a violation of the other site's policies, nothing will happen. Opabinia regalis (talk) 02:05, 19 January 2017 (UTC)
• Regarding safety for the functionaries: I think most of the enwiki functionaries know this stuff, but this is a good place to link some good general web safety type information. Use an email service like gmail that doesn't expose your IP. Turn off automatic loading of images. Consider using a proxy to visit a personal website alleged to have harassing content on it. If you don't use your real name on-wiki, be very careful about using it in private correspondence. If you get involved in a situation on-wiki, e.g. by blocking a harasser, be aware they may target you next. This document is written as if the person being harassed is not themselves a functionary/arbitrator/etc, but historically people in high-profile positions on Wikipedia have been disproportionately targeted for harassment. Opabinia regalis (talk) 02:05, 19 January 2017 (UTC)

Comments on H6: "Investigating reports"

Hi Maggie and Patrick, this bothers me (bold added):

Regarding the victim:** Has your team dealt with this person before (whether as a victim, a reporter, or a harasser)? Were their reports and/or opinions reliable in any previous dealings you have had with them?** Does this person have a history of having been harassed, whether by the current alleged harasser, or by others? (If yes, you may, for instance, be dealing with a sockpuppet of a previous account.)** What is this person's reputation in the community, and are the events in this report uncharacteristic for that reputation? They may be known as someone superbly level-headed or as someone who overreacts; either of those being true will have bearing on how you interpret their report.

Something that is missing from these documents is an awareness of how harassment affects the victim. It can cause extreme anxiety and fear. The sensible thing for people who feel that way is to disengage from Wikipedia, but what often happens is the opposite. Victims are drawn to the source of the harassment, drawn to watch it unfold, and are at the same time absolutely repelled by it. As a result, they experience enormous anxiety every time they sign in. I've experienced this myself, and I've watched it happen to several others. It can lead to emotional interactions with other editors, which the harassers point to as evidence in support of their complaints—"See? This person is nuts. Look, they're attacking everyone."

The WMF's document seems to support the idea that there is a "perfect victim". A good harassment victim might be "superbly level-headed"; then again, it might be "someone who overreacts". This not only needs to be rewritten, but the thinking needs to be challenged. Imagine if a law-enforcement document were to make that distinction when discussing how to handle rape complaints.

One of the biggest problems we have had on the English Wikipedia is that harassment victims did not behave like "perfect victims", and therefore weren't taken seriously. Some even ended up in front of ArbCom, with material from the harassers being used as evidence against them.

The document does offer a caveat: "Context is not synonymous with rationalization or excusing." But it's too little. The "perfect victim" versus "bad victim" meme needs to be removed entirely. Someone who appears to have over-reacted for a long time might be doing so for a good reason. Or they might not. It's unknowable, and it therefore can't be taken into account without prejudicing the report. SarahSV ^talk 00:52, 7 January 2017 (UTC)

Pinging Smallbones. SarahSV ^talk 01:03, 7 January 2017 (UTC)

This is a tough call, if you want to make it tough. But I think, ultimately the only way to go is to enforce all the rules involved in the case according to the letter and spirit of the rules, without regard to personalities or previous history. The 3 worst harassers that I can think of on Wikipedia all were popular editors in the sense that they had a large base of fans and supporters. If the WMF staff wants to email me for their names, I'm sure they'll agree that we don't want to let these people back. In that sense it is simple - just enforce the rules as written.

There's an old observation about people whose legal defense is the 1st amendment to the US Constitution. "People who use this defense are not the type of people you'd want to introduce to your family." e.g. they are espousing unpopular causes, pornography, or wing-nut politics. So be it, the 1st amendment protects us all, and defending these "nuts" protects us all.

There is certainly some of this involved in Wikipedia harassment cases. Some editors will inevitably have unpopular views and perhaps be a bit pushy in trying to get them across. We need some folks like this as editors. What sometimes happens, though is that an unpopular editor is told in effect to leave the project, via harassment. We just can't allow that, no matter who the personalities are. Just enforce the rules in as straightforward way as possible. "Ignore all rules" has no place in our anti-harassment efforts.

Sarah's statement: "The sensible thing for people who feel that way (extreme anxiety and fear) is to disengage from Wikipedia, but what often happens is the opposite." I disagree with her characterization of this as "sensible", at least from the victim's pov.

If somebody comes to another editor and via harassment in effect says: "We don't want you here - no matter what the rules say - you don't have any right to edit here." I think a very rational response is "well, I can a) let them drive me away, but if I do that every time I run into opposition, I'll never be able to express my views anywhere; or I can b) stand up and fight, using whatever tools I have at my disposal." B) may not be pretty, but it is not irrational. I think that harassment victims who stand up and fight back deserve our respect, even if they sometimes make bad choices in how to fight back.

So, how to sort out these messy situations? It's impossible, unless you follow one simple rule - just enforce the rules in the most straightforward way possible, without regard to personalities or editing history.

Smallbones (talk) 06:05, 7 January 2017 (UTC)

Smallbones, sticking to the rules no matter who is reporting the harassment is important, I agree. But harassment can be subtle and open to interpretation, and the victims are carefully selected. There is a huge literature in criminology about victim selection and how vulnerabilities are spotted and exploited.

There is also a huge literature about the concept of the "ideal victim" (an older woman attacked by a violent stranger) and the "heirarchy of victimization". That's the equivalent of the "superbly level-headed" Wikipedian. But how do others—the less-than-ideal victims—acquire the status of victim? How can the Wikimedia Foundation make sure that the arbitration committee, functionaries and administrators do not stand in the way of the acquisition of that status when someone is being attacked? Our dispute-resolution processes have managed to get it upside down many times, so this is a key issue.

I strongly agree with "I think that harassment victims who stand up and fight back deserve our respect ...," but how many times have we seen that work out well for the victim? I can't think of a single example. But I can think of lots of examples where the community has turned on the victim for pointing it out. SarahSV ^talk 18:19, 7 January 2017 (UTC)

I won't argue, we're not going to see many "superbly level-headed" Wikipedians reporting harassment. The harassment itself almost assures that the victim won't be level-headed at that time. Combined with victim-selection by the harassers, you're going to find that the victims are often going to be unpopular or espouse unpopular positions even before the harassment starts. They'll also likely be from minority groups, or women, or groups that are under some form of pressure where they find it hard to assume good faith. I do think that I can remember 1 or 2 informal reports of people acting pretty well in the face of harassment. The only question I'd have for them is whether they felt "level-headed" at the time of the harassment. Smallbones (talk) 19:15, 7 January 2017 (UTC)

Hi SlimVirgin and Smallbones, and thank you so much for giving us your thoughts here. I wanted to leave you a quick note to say that we are reading everything said here and, especially for the delicate topic of how best to support victims, we are taking your thoughts to heart. We are going to be refining and revising the wording (and to a lesser extent, content) of the modules in the coming weeks, and this section is included in that. While the module content is ultimately being determined by the Support & Safety team, we would welcome any specific thoughts or wording suggestions you or others have about best practices for investigating harassment cases, what to do/not do when investigating, etc. We're focused here on presenting not only correct information, but correct information that's understandable, and by the widest audience possible. Kbrown (WMF) (talk) 21:35, 9 January 2017 (UTC)

Arbitrary break

Yes, this is extremely troubling. Re: Does this person have a history of having been harassed, whether by the current alleged harasser, or by others? (If yes, you may, for instance, be dealing with a sockpuppet of a previous account. It looks like the minute someone says anything about harassment, someone can start a dossier on them, in a place they cannot see, and cannot even know if it exists. Something like a credit report at least can be examined, and the customer has a chance to challenge any untrue statements, but this looks like a tremendous opportunity for an adversary to spread material about someone that is unsourced, or misrepresented. It also seems that this could be used as a rationalization for a checkuser fishing expedition, or "duck test", which has a terrible reputation for accuracy, and tremendous potential for abuse.

Also it appears that the admins, rather than being trained and educated on harassment, are being given a huge latitude to determine for themselves what constitutes harassment, using an unsourced statement from "Wikipedia's article" that defines harassment as "behaviour which disturbs or upsets". A lot of things can be disturbing or upsetting to Wikimedians, for instance women and other underrepresented groups editing Wikipedia, or editathons that try to fill in missing knowledge gaps about these groups. Sometimes it has been merely objecting to harassment--this definition has even been used by the arbcom. This kind of approach is telling users not to have the difficult conversation, not to rock the boat, but only what someone will approve of, even if it runs counter to the project goals. —Neotarf (talk) 22:22, 16 January 2017 (UTC)

Hi Neotarf. Couple quick comments to address your thoughts, though please keep in mind that these are not definitive and I and the rest of my team are continuing to read and consider feedback we see here:

1. With regard to your worry about opposition research: Our intention is not to encourage people to compile 'dossiers', but rather to encourage the people investigating harassment incidents to look at more than the single reported action; often past conflicts, experiences, or situations are relevant to understanding either the involved parties' perspectives or the environment that led to the harassment. I do understand your concern about how that kind of research could be done poorly or for the wrong reason, but the fact is that it does need to be done in order for an investigator to understand what they're looking at (consider, for instance, a situation where someone appears to fly off the handle with no provocation, but where deeper investigation shows that the issue they "flew off the handle" about was actually something that someone had been harassing them about long-term, and they interpreted the latest situation as a continuation of that. That's a different situation than one in which the person genuinely just decided they didn't feel like being civil for no reason at all. If you approach both on the surface only, they look identical despite being very different and quite possibly needing different types of handling). And this type of background research - really, nearly any type of documentation amassed during a harassment case - needs to be handled with due respect to privacy of those involved. While transparency is an important value of our communities, one can only be so transparent in cases involving harassment without re-victimizing those we are trying to help. The last thing a harassment victim would generally want is for a detailed description of the best ways to hurt them to be published on a noticeboard, or for functionaries to, in banning their harasser, point more people to where they can read the harassment.
2. With regard to the definition of harassment we use and how that leaves a lot of ground open: you're right, it's not great. But what we found while trying to put together this content is that it's very difficult to define 'harassment' in a way that is understandable (to those who might not be familiar with the concept), translatable (by those who may not be reading in English), and easily relatable to Wikimedia projects and the harassment that goes on on them (as opposed to extremely general definitions, which is the direction we edged toward in this draft). I would be really interested to hear how you would define it in this type of "basic training" module, or see links to any useful definitions/resources you know of that we could incorporate. Kbrown (WMF) (talk) 15:26, 17 January 2017 (UTC)

Re the first question: so, if you have a dossier on me, can I see it?

Probably the best way to find out this information is to ask. Interaction tools can only do so much. Will they show for instance, if someone is pinged or thanked? Or if someone has a conversation about someone or files an SPI against them without notifying them? Someone could also be stalked for a long time and they wouldn't necessarily know about it. And I can't imagine most people would not be reluctant to post such things openly or even initiate such a discussion without knowing they would be protected from retaliation. There was a recent Twitter thread that brought out some of these issues, the discussion is about in-person harassment; you have to click several times to see all the comments, but this is very similar to what I have seen people experience in WP behind the scenes. https://twitter.com/alicegoldfuss/status/818174763608641536 the money quote there being "...if you want to know about it, you have to prove your commitment to taking harsh action....why would I tell you someone harassed me if you're just going to ask him about it?....I'm only naming names if I know the perps will get fired and/or I am completely safe from them professionally". So it will take more than essays about how to listen with empathy. You need a structure behind it.

Re the second question: most of the research has focused on generic harassment with gender considerations being secondary, which is probably the approach that creates the least political blowback. But if you do not look at gendered harassment separately, you risk harassment being defined by men who have not experienced it and do not understand what it is. And you will not fix it. As far as sources, I would start with the list at “Privacy, Anonymity, and Perceived Risk in Open Collaboration. A Study of Tor Users and Wikipedians”, which is specific to Wikipedia, and which I did not participate in, but this really resonates with what I have observed. This will probably yield a list of the hardest problems, and end up in the "unresolved" section, but they're not going to get resolved until they're acknowledged. For background understanding, this explains characteristics of gender-based trolling. Also the paper “Defining Hate Speech”. Most of this is not written on a tenth grade level and will take some digging to extract the WP stuff, but it is mostly recent and I think on point. Here is one that is very readable, by someone who appears to be still active in the SF community, that I referred to for my harassment proposal, but it is out of print so I will post a temporary link. Less useful, but perhaps more readable, with WP I would also mention borderline creepy behavior, and using edit summaries as ways to conceal dog-whistle harassment; and there is some curious miscellaneous stuff here, like DARVO, a harassment strategy of "Deny, Attack, and Reverse Victim and Offender." A lot of this has already been on the mailing list, and is just what stands out for me at this point.

"Translatable", you may have to do a pilot program with enwiki first, the other wikis may have totally different cultural stuff going on. —Neotarf (talk) 01:39, 19 January 2017 (UTC)

Here's another one, from Ashley Judd's "speech project", useful for being very visual. http://wmcspeechproject.com/online-abuse-101/ —Neotarf (talk) 23:39, 28 January 2017 (UTC)

First draft feedback summary

Thank you to everyone who has given feedback on the first draft of these modules! We have compiled a summary of the feedback we received here, and over the next month or so we will be working on revising and incorporating the feedback we received into the next draft of the modules. Kbrown (WMF) (talk) 21:23, 30 January 2017 (UTC)