Wikipedia Sucks!

Jan Schwanz wrote:Some of you may remember that Trust & Safety staff used to sign with their individual accounts when discussing Office Actions. Unfortunately, this is no longer possible due to safety concerns for Foundation employees, as in the past staff have been personally targeted for threats of violence due to their Office Action edits. I am taking the step of making this statement personally in this case due to extraordinary necessity.

The Garbage Scow wrote:I can relate to the wish to avoid being targeted by nutbag stalkers, but this sounds more like an attempt to deflect accountability for their decisions. State and federal judges have angry people threatening them, it doesn't mean they get to hide behind a screen while issuing judgements.

MrErnie wrote:The unformatted text runs about 200 pages, and there's about 107,000 words. Regarding the WiR tweet, I wonder if Fram now has a case to make regarding his "safety," now that a prominent Wikiproject is "attacking" him. "Safety" and keeping Wikipedia a "safe space" seem to be Jan's big points.

Jan Eissfeldt update (06/17/2019)
On the question of how many cases reaching T&S result in office actions, the answer is two-fold:
Roughly 90% of the outreach to T&S does not result in T&S cases. There are two big reasons for that: community self-governance and the hurdle for opening T&S cases being consequently pretty high. Much of the outreach we receive therefore can be routinely addressed by others and is being redirected - including to OTRS, ArbCom, other community processes. Last quarter, for example, the percent of T&S cases opened relative to outreach received was 8.1%, the quarter before 11%.
Within these ~10% that become investigations, T&S cases resulted in actions in 48.18% of all investigations conducted over the last four years. That number includes both types of office actions: secondary like a private conduct warnings, and primary, like Foundation global bans.
For historical context: T&S cases historically used to come mainly from the English language projects but that has steadily declined to less than a third of cases (again Q1 and 2 18/19 data). The main cause for the trend has been a consistent rise in requests from other language projects.
I know some of you have expressed concerns about the new reporting system and the universal code of conduct here and on ArbCom’s talk page. T&S staffer Sydney Poore, who has been pinged by several editors already, will be engaging directly about these initiatives in the conversations.
On questions on better communications of office action procedures: Going forward, news of all substantive changes to the office actions policy will be going out to all communities; just like technical changes already do. T&S will work with ComRel to make sure it follows the usual setup and feedback reviewed on the policy’s talk page on Meta. We are also reviewing, in line with Vermont’s suggestion from last week, whether to include individual public office actions, which is more complicated.
We have heard your concerns about fairness to Fram of the case as it proceeded. Balancing fairness to the accused party with the safety of the accusing party is something we have been working on for quite a long time, and it’s not something we’ve perfected. Generally, we will reach out to the accused party for information if doing so is possible without violating the privacy of the accusing - or other involved - parties, but our efforts do remain a work in progress when it comes to finding the right balance in each individual case.

The process T&S cases go through within the Foundation.
As far as the ability for others to avoid making mistakes and finding themselves unexpectedly sanctioned, unfortunately, we cannot publicly disclose details of this or any particular case, for all the reasons previously discussed. This means that, as much as we understand your wanting the information, we cannot tell you what specific behaviors by Fram brought about this action. We can, however, say that abiding by the ToU is required of everyone who edits a Wikimedia site. That includes refraining from behaviors described by the ToU, including “[e]ngaging in harassment, threats, stalking, spamming, or vandalism”. In cases where we believe a user may not be aware that they are violating expected behavioral standards, even repeatedly, we give conduct warnings prior to any action being taken. In Fram’s case, as noted on Commons, we did send more than one of those warnings/reminders before the most recent step.
There have been suggestions that T&S should have piloted the newer office action measures first before proceeding with a potentially controversial case like Fram’s. I don’t agree with that as I think that bending the selection of cases to cherry-pick a good “starter” case endangers the independent investigations approach T&S has to uphold.
There have been questions about the investigation process itself. As indicated on Friday above, my team has built a graphic to visualize the overall process to make it easier to navigate. Traditionally, it has been documented as a table on Meta and is always followed. I hope that the graphic puts the number I detail in the first bullet of this edit above into its context. ~10% go through the process visualized here and less than half thereof result in office actions taken. Jan (WMF) (talk) 21:53, 17 June 2019 (UTC)