August 4:

• Met with the WMF Board in the afternoon to host a conversation as well as listen in on other conversations. I shared back the learnings with the staff as relevant.

August 5:

• Participated in a full-day event for Users with Extended Rights. Some learnings and questions from the the day most relevant to the Research team:
  1. More than 100 users with extended rights attended the event. It was very good to see that one of the recommendations from the admin research came to life with this event (WMF organizing spaces for the exchange of ideas).
  2. I facilitated a session where I presented very high level learnings from the admin research and encouraged the audience to specifically focus on "barriers" for becoming an admin and sharing with one another (in small groups) 1 idea they have to make a change towards one of the four barriers (slide 6) or share something about what their community has already done. The focus was to share ideas or learnings. Sample ideas/learnings that came up:
     • no and sv wikipedias have re-election every 2 years and they have found that "termed" approach helpful.
     • ukwiki and temp adminship.
     • multiple folks brought up the idea of breaking down adminship to smaller tasks to reduce the burden of responsibility and create a ramp towards full adminship
     • enwiki shared that the experiment with the updates in the RFA process (including local securepoll for private voting) has been successful and they are seeing a change in the downward trend. there is a Friday session on this topic that I will try to attend. @Easikingarmager (not urgent) please check the trends for enwiki and see if we can verify this change in trend. This is important to be aware of.
     • tutorials for new admins (default content that can be adjusted by wikis to their needs)
     • another question that came up: in the plwiki study, it would be interesting to look at the users we surveyed or interviewed to see if they are from the affiliate or the community (sometimes users have double hats) and whether their responses to the survey were significantly different.
  3. NPOV related discussions with this group.
  4. A question that came up:
     • For the admin research, did we do a drilling on sub-rights? For example, it can be that in a project such as enwiki admins are declining but the community is giving more rights to other roles. @Easikingarmager can you follow up on this and let us know here the answer. (No additional work is needed. Only what we know from the research already concluded.)
  5. users with extended rights safety was a theme that was brought up throughout the day in different conversations.
     • We currently call out "interpersonal conflict" as a barrier for potential admins to become admins. It is worth considering if that category needs an expansion to include safety, or safety is a dedicated barrier we can consider adding in future communications. This concern was particularly acute for users from specific countries.
     • Naturally each time the topic came up, there were also ideas about how to solve for it. I'm not going to list the plus and minus of each possible solution that was brought up but will list them so we have them in mind:
     • Separating the "privileged rights" account from regular editing account and allowing the former to be anonymous.
     • Having a "shared" privileged right account that everyone who receives that right should use.
     • legal support

Section 3.4.2 (https://upload.wikimedia.org/wikipedia/commons/5/5f/%28Final_Report%29_Administrator_recruitment%2C_retention%2C_%26_attrition_%28SDS1.2.2%29.pdf) is probably the most relevant to your question, but I don't think we examined rights distribution changes over time if I'm understanding the question correctly (@CMyrick-WMF and @cwylo, am I forgetting about anything that would address this question about sub-rights?)

In T401012#11065353, @Easikingarmager wrote:

Section 3.4.2 (https://upload.wikimedia.org/wikipedia/commons/5/5f/%28Final_Report%29_Administrator_recruitment%2C_retention%2C_%26_attrition_%28SDS1.2.2%29.pdf) is probably the most relevant to your question, but I don't think we examined rights distribution changes over time if I'm understanding the question correctly (@CMyrick-WMF and @cwylo, am I forgetting about anything that would address this question about sub-rights?)

That's correct. We examined the current dispersion of administrative rights across wikis, summarized in 3.4.2; but not over time. The over-time trends that we report in 3.4.1 look at admins as a whole (i.e., sysops and other user groups with administrative rights together.)

Thanks, Claudia and Eli. Having looked at section 3.4.2 now, can you confirm that the Figure 1a in the report is for admin trends and not sysop trends?

August 6:

• The day was light in session attendance for me and very much packed with hallway conversation and connecting with people on different topics.
• I attended PTAC's session and I appreciated the group's openness to share about how the past (almost 1 year) been for them as well as hearing the audience's questions who asked good questions about how PTAC had managed trade-offs and decision making (I especially appreciated the question that asked the group to give a specific example). A couple of things remained with me from that conversation: knowing that the group had faced some hard choices/trade-offs and that the group had access to data for decision making. I'm really appreciative of PTAC's willingness and commitment to try a different model of advising WMF for product and technology matters and look forward to see what the group decides about its future in the coming months.
• I had a couple of really good conversations with volunteers who are getting PhDs in topics squarely on or very related to the NPOV work. It was a good opportunity to share notes and brainstorm about how to approach reducing the complexity of research on NPOV given its very technical and layered definition. We also had some good conversations about the NPOV term (neutrality or impartiality or ... these each mean different things to different audiences in the world) and I connected relevant folks in the working group to the people interested to continue conversations.
• Fun fact: I was lucky to meet one of the researchers from India who has been following the Monthly Research Showcases for the past 12 months. It was a wonderful and serendipitous connection. Lots of appreciation and love from the researcher back to the team who organizes Research Showcases.
• Broken references on Wikipedia came up as a theme in multiple languages. The actual solution at a high level is not hard. In my conversations, editors already knew what sources had issues that they wanted to fix. so all they needed were two things: 1. a way to easily get a list of articles that have used the references in the languages of their choice (@fkaelin pinging you here given that you're looking at retrieval via LLMs, with higher tolerance for false-negative), 2. check if the link is broken, 3. give the alternative link (which they know it exists b/c the broken links they were interested in come from very large organizations that have moved content without redirects). I thought about this research proposal and will connect the volunteers to the research group to see if they want to pursue a collaboration.

In T401012#11070271, @leila wrote:

Thanks, Claudia and Eli. Having looked at section 3.4.2 now, can you confirm that the Figure 1a in the report is for admin trends and not sysop trends?

Correct. There's not a tidy/neat overlap between admin and sysop due to cross-wiki variation, and section 3.1.2 definitions goes over a few related details if folks are interested.

All clear. thank you.

August 7:

• Motivation: The team focused on improving the WikiLearn platform and its connection and impact is in search of the right way to motivate WikiLearn learners to learn the different skills through WikiLearn courses. Asaf hosted a workshop related to the topic.
  
  @TAndic I wonder if you have learnings from lit review and framing that you want to pass on to Asaf. @KinneretG the question of how and if to connect the grant making to "skills" came up. This can be something that the team experiments with. Their learnings can be interesting for how we handle WM Research Fund grantmaking (for this year we added "established Wikimedia researchers" as a requirement.)

• 5 Research Findings session: Eli and Isaac supported me with this session and here are my top level learnings from it:
  1. This is the first time I had 85-min for this session and I had a question whether people will attend b/c that is a very long time commitment (particularly b/c they were numerous sessions in parallel). My estimate is that we had around 70-80 people in the room (and some people unfortunately gave up and left b/c the room was at capacity). That's roughly 10% of the total in-person attendees. My conclusion is: A long time for this material is not a deterrent.
  2. More time available made the session more interesting. I covered 4 learnings (though I was ready for 5) and could meaningfully understand and discuss them. I think 4-5 learnings is the right amount to share for this session. My sense of the crowd is that they come for the diversity of topics and they still want to engage meaningfully with at least some of them in depth.
  3. I talked about readership data, editorship data and the challenge of decline in account registration and activation, admin data, and ongoing research on the topic of NPOV. We then had 3 full-room facilitated discussions by Mako, Chuck and myself about: "what can you do in your work to change the trend that you see" or feedback for ongoing research.
  4. Looking to the future: If we keep the format as long for the future years (which I hope we can), I think there is room to work on the questions that they work on and make these questions even clearer. For example, it would have been helpful if I would give a guidance such as "assume this is all the information that can you know from this research. You are asked to take an action with this information and all the other knowledge that you have about the projects to attempt to change this trend. What is the first step you will take? OR what would you do?". This is an easy way to not spend more time seeking more information (e.g., "how does the plot look like for my particular project?" which is a very valid question and it will delay thinking about what can be done or proposed already with the information we have).

August 8:

• State of Wikimedia Research session was back this year where Mako and Tilman presented 5 research associated with 5 themes from the Wikimedia research community. Well-attended and it was nice to see research that was presented as part of the Monthly Research Showcases or the WMF-RAY made it to the list. (I expect both slides and the recording be posted in the link in the coming weeks) @KinneretG cc.
• I had a good conversation with the organizers of Wikimania 2026 (Paris) for a possible Research track as part of the program, more similar to what we organized in 2019 in Sweden. My pitch is on their table to consider. We also discussed the possibility of a full-day event (pre-Wikimania or not), something that Jerome had proposed and tested the waters about as part of Wiki Workshop 2025. I look forward to learn in the coming couple of months what may be possible. (fwiw, I communicated that I'm interested in us doing something as high quality and ambitious as we did in Sweden, at the very least, or not doing anything in particular at all because it will be additional work and it makes sense to do it only if the quality is high and we can have more impact than the usual research label/track that gets attached to the Wikimania sessions that are research related.) @KinneretG
• Wikimedia Armenia has been organizing campaigns to help researchers contribute to scientific content on Wikipedia. @KinneretG you may be interested to be at least aware of this work and session. Susanna and I explored different funding options (she was wondering if that work is fundable via Research Fund and we concluded the general fund is more appropriate because the focus is on editing. I did share with her this WMF-RAY winner paper (b/c they don't get a lot of edits, yet).
• Research Meetup: and I forgot to say in August 7 report that Dariusz, Iolanda and I co-organized an informal Research Meet-up (thanks Kinneret for booking a timeslot). ~10-15 people joined. We each shared what was top of mind for us, where we needed help, and where we could offer help to one another. (etherpad) Some of the questions we tackled in the room:
  1. where can I publish my research?
  2. what is most important for me to consider when publishing my research?
  3. I need help organizing a research event. Who can help me with some of my questions about event organization?
  4. I have done research with a focus on a contentious topic. I have identified some editors who may have done questionable or inappropriate actions. Can I mention their username in my paper?
  5. and more
• Two enwiki admins shared the result of the updates in the enwiki RFA process (session). @Easikingarmager let's watch this space. The editors showed already some improvement in terms of recruitment numbers. This is important change that is happening on the ground and it would be good to keep an eye on it and see if the change in trends are to stay and also whether they will sufficiently compensate for the attrition in enwiki. And consider checking out the talk. They did a good job presenting how they brought this change to life on the ground.

August 9:

• Admin panel: I'll start by saying that as much as I know how committed Wikimania participants are, after having a pre-day for users with extended rights, talking about admin trends as part of the 5 research findings session, and the night before in a shuttle with a group of folks I was questioning how much engagement we get with this panel. My estimate of the participation is somewhere between 30-40, a good room size, and a lively conversation. I had 3 panelists from enwiki, ukwiki and itwiki and we had a good conversation about their perspectives wrt barriers for potential admins to become admins as well as ideas for experimentation. Good engagement and questions from the crowd as well. User:ata made a call for more experimentation to the community that stayed with me, and User:OhanaUnited emphasized the importance of motivation and also focusing on retention. Key topics that came up (as part of this session and the previous ones):
  1. need for more experimentation
  2. appreciation for research and WMF recognizing that the biggest impact should happen on the ground and will likely be policy based.
  3. appreciation for the research and repeated ask for future reports to be to-the-point and targeted to the specific audience (editors/admins/...).
  4. some specific ideas: termed adminship (nowiki, svwiki), breaking down admin tasks (further), recognizing current admins for higher retention, mentorship (via good mentors) to help potential admins become admins, temporary adminship (ukwiki), ... session info along with link to etherpad and slides (recordings will be posted in the coming days/weeks).
• Wikipedia readers (as part of a relevant session): There are two key learnings for me from this session:
  1. A reminder that editors do not know about readers unless WMF studies readers and tells editors about them. They were very open to this fact.
  2. We must continue thinking about ways to communicate the relevant pieces of our learnings about readers to editors. They are wary of large research reports. One idea that came up in the session is: "every once in a while show us a did you know x about readers" window, don't overwhelm us but give us critical information we should know and learn about. This particular solution may or may not be something that we implement, however, as part of serving "editors" I would like explore more about how we help them have the key information they need to do their work best. @DKumar-WMF let's talk about this at some point.
  3. The topic of human-centered design for readers and how to do that so that editors can come along on this journey was one, perhaps unspoken theme, of the panel. At the end the reality is that the Foundation has a few people on this topic (including Design, UX Research, Applied Science, ...). So the question is: how can this group learn about readers globally and how they can meaningfully engage editors for designing for readers. I see a lot of opportunities for WMF and editors to work more closely to serve readers in different ways. @DKumar-WMF cc.

Overall key themes from the event:

• AI:
  1. Confirmation that the AI strategy for editors is focusing on what this group (for the most part) wants to see.
  2. Clarity in the space of AI and readership will be helpful. (something we knew we should work on after the first piece was out.)
  3. It is very important to engage with the community on this front. I see a lot of value for creating spaces where people can come together, share different perspectives and talk. My read of the different rooms on this topic is that we will benefit from creating such spaces, sharing and hearing each other.
  4. Having been exposed to all the hopes, fears, and needs of the community during Wikimania, I am even more eager for us to focus on implementing the AI strategy for editors. Editors need help and the areas of retrieval and moderation support are high priority for them.
• Users with Extended Rights
  1. Appreciation for creating the spaces for these users to connect, for the first time in this way.
  2. A nice suggestion by someone that perhaps next year we can bring UWER along with "potential UWER" to one place and help the two groups interface with one another, learn from each other, and support each other.
  3. For every community that aims to do an intervention on the admin front, it is important for us to invest in supporting them with data analyses so they can track their numbers. Even better, if they can run the codes themselves, we can create an "action package" and offer links to codes as part of that so they can run with it. I offered that all folks who want to take action to feel welcome to reach out to Eli, Claudia or Leila for an office hour conversation.
  4. My overall assessment is that editors on the ground (at least in some Wikipedias) will act on this research. it is on us to continue supporting them in ways that is appropriate for our roles.
• NPOV
  1. It was really good to be able to get a sense of folks on the ground about the NPOV work by the working group.
  2. After a few conversations, Steinsson's paper is a piece of research I'm actively thinking about.
  3. (this part is not research related and I have passed on the feedback to relevant folks in the working group) there seems to be a need for making some terminology clearer in the NPOV definition or a standard policy related to it. Terms like impartiality, fairness, unbiased, reliable, neutral, etc. can have different meanings in different fields and it's okay for us to use them, of course, but then we should say what we mean by them in the context of WP.
  4. Editors communicated frustration (mildly put) as they have had to deal with NPOV on contentious topics repeatedly over the past few years. The pressure on those on the ground is very much felt in these conversations.
• Readership:
  1. WMF to communicate relevant learnings from research on readers back to the communities.
  2. How can WMF and editors work together to build for readers? what are the roles and responsibilities of each group? This requires some thinking and designing for.
• Communication: within the Research team, there is opportunity for exploring our ways of communication to the different WP projects. There is a lot of love and ask for more research, and at the same time repeated requests for information about findings to be transferred to the communities in ways that they can use them in their work. With arwiki editor we explored the possibility of sharing some of the arwiki findings as part of Wiki Arabia, there was also asks for nuggets of findings to be shared throughout the year. It is important that we think about this and figure out if there is something new we want to experiment with.
• Global trends which inform annual planning (which Research contributes to but does not lead): I have some feedback that I'll pass along to the relevant teams.

This is the end my reporting. Happy to answer questions here or offline. Closing the task.