Wikimedia Clinics/012

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search

Wikimedia Clinic call #012 - January 13th 2021

(Attendance: 7 staff; ~16 volunteers)

Topic 1: Growth team presentation[edit]

If you want to implement the Growth features on your wiki, please contact User:Trizek (WMF) and see this page.

Please check out the project page for our upcoming project around a task for adding images.

  • The goal is to grow the base of contributors to the movement - and the diversity of contributors
  • At the moment there are barriers to enter the movement: technical, cultural, and conceptual, communication processes, etc.
  • We are building a set of tools to help newcomers when they make their first edits and contributions:
    • Tool 1: welcome survey - understand why/how people create an account on Wikipedia.
      • The survey asks why the contributor wanted to make an account in their community.
      • Asks if they have already made any edits.
      • The survey is optional.
    • Tool 2: Help panel - give editors context to guide them in making their first edits
      • Newcomers struggled to edit, didn't know how to copy-edit.
      • We built how-to structure on making new edits.
      • Newcomers are informed how to make new edits or add new links to articles that needs editing; Artificial Intelligence makes some of the suggestions/edits, and there could be poor AI suggestions.
      • There is a mobile prototype.
      • There is a community conversation for new way to help newcomers to input photos.
    • Tool 3: homepage - where newcomers go to find centralized resources, impact module, direct access to a mentor, and suggested edits(providing tasks to newcomers that fit their skill level)
      • the tools are on the main page once account has been made.
      • one can see one's impact by the amount of edits that you have made.
      • there is a tool to directly talk to mentor.
      • main page (a "dashboard") on your account, allowing you to filter what topics of interest you want to see suggested on your page.
      • users can manage the level of edit fluency on their page.
  • Are these tools working? Yes!
    • 11.6% increase in likelihood of retention of newcomers (versus control group)
    • improved the amount of edits done by newcomers (+22%)

most wiki's there are Help Desks to ask questoins

  • The tools are on 17 Wikipedias currently (Jan 2021)
    • currently 500-600 newcomers are offered suggested edits
    • 55000 new edits attributable to these new tools and features from new users across the 17 different Wikipedias.
  • Reports on how many people are using the new tools are presented to the Foundation Product team.
  • Newcomer tasks page where you can see how the new features are doing in the 17 different Wikipedias

Some resources[edit]

Know more about Growth:

Try the tools:

On wikis where the Growth features are available, and also on the test wiki, go to your user preferences and then:

  • enable the Help panel in the Editing tab.
  • enable the Newcomer homepage in the User profile tab. Enabling the Newcomer homepage gives you access to Newcomer tasks.

Discussion[edit]

  • Question: Does the feedback by the user also train the AI, or is there no feedback mechanism? What is the AI training mechanism?
    • The module decides which phrase or word is a link, and looks at how many times the phrase or word is a link in other articles. Researchers can use this to improve the model. There is no direct feedback mechanism.
  • Question: What is the effect on retention of newcomers?
    • When we deploy new features, we have 80% of new accounts receive the new experience ("treatment group"), and 20% do not get the new feature ("control group"). So we're always able to evaluate the impact of the new feature.
    • We evaluate all features together, but we believe the "newcomer tasks" feature is responsible for the most significant impact on retention. (Because earlier, other features alone did not have such impact.)
    • "article activation" means having made any edit at all
    • "constructive article activation" means having made an unreverted edit. This is significant, as it shows the edits were not just made (and quickly reverted), but "stuck" and are likely good.
  • Question: One of the most commonly requested features in newcomer surveys -- coming from free text questions, not a directing question -- is for a real-time chat system. Is this considered?
    • Yes, we've heard this a number of times, too. Asking on talk pages is of course frustrating in that a question on a talk page may not get answered for a while, and certainly not in real time.
    • But how do we make sure there are people ready to answer questions in real time? It would be a poor experience to offer real-time chat and turn out to not actually provide people with real time responses.
      • One proposed solution is of chat as an extention or tool - you could use notifications as a on-ramp to talk pages, with a notification reply feature, that saves to the talk page - or a skin improvement that recast talk pages as more chat-like "thread-like" displays for new users while leaving old skins unaffected. Directing requests to be posted at the help desk, where there are more people to reply, is also a solution.
  • Question: This tool that offers links in an easier way: Why is it only limited to newcomers? And thinking further: not just for links but other problems with articles in this "gamified" type contribution?
    • When experienced editors see the new tools for newcomers, they usually think it would benefit experienced users as well. The product team is concerned with the newcomers' experience. Additional team capacity would be needed to also focus on experienced users. However, anyone can turn the feature on via Preferences on wikis where they are available.
    • a volunteer comments: only selected experienced editors should have access to function as mentors in these mentoring channels.
      • Growth team responds: Experienced users who volunteer to share their experience add their names to a page where the Growth tools pick their names and assign them to newcomers. No filtering is done, and it is up to communities to ensure the mentors are appropriate (e.g. by protecting the page to control edits)
      • We also provide resources for mentors, so that they can improve their skills while facing newbies.
  • a volunteer comments: "As a mentor, I realized that large portions of newcomers are using mobile devices to edit Wikipedia, and I have to check this in each question, because my answer would strongly depend on which interface they are using. A label showing this information with the request for help would be useful, so it is not necessary to check it in the page history.
  • Question: What has been the qualitative feed back on the tool?
    • In product development we want to combine the quantitative and the qualitative. In addition to the stats we collect, we have contractors as ambassadors - helping us to gather qualitative feedback for Arabic, Czech, and Bengali Wikipedias.
    • We are only asking people who are using the tools, so are biased in this respect.
    • Users that try suggested edits say it helps them find work they are interested in or learn about things of their interest, and that it helped them learn policies on Wikipedia. People say they like adding links- it's fun and easy. We'll try to publish a summary of these responses.
  • Question: What is being done to counter problems like the ones that accompanied the introduction of Visual Editor?
    • Visual Editor was a revolution, whereas the new tools for newcomers are to be made for more accessibility.
    • There are people who are not thriving with the current editing tools. These new tools may help them integrate better on Wikipedia.
    • More structured tasks are needed.
    • Community members want to have newcomers quickly learn as much as they know - but too much information can scare away newcomers
    • The most valuable feedback is the one that helps us improve. Please find us and share your feedback and concerns.
  • volunteer from NLWP: Dutch Wikipedia is not very innovative. Adoption of these features may not be easy. Personally I would love to see it.
    • Growth team: what would be the concerns?
      • General mistrust of WMF.
      • Unleashing untrained newcomers without community capacity to train these new editors or check their edits. We have to be very careful of what we propose.
        • Growth team: We just offer the features to communities; it's up to them to decide whether they want to adopt them.
  • Question: Mentoring is difficult work, and requires a certain temperament. There is a risk of having the wrong kind of people as mentors. How is this handled?
    • The community will set the rules for people to qualify to be a mentor. People who want to be mentors should have a lot of experience but also have the desire to want to mentor people. The default rule is to have people with experience, and a lot of experienced people don't think they can reply to basic questions. Growth recommends to protect the mentors page to avoid newcomers misguidedly listing themselves.
    • Newcomers have found that certain mentors are not suitable to be mentors. People have asked mentors to remove themselves from the mentor list.

Topic 2: The COVID-19 pandemic's effect on Wikimedia activity[edit]

  • The Wikimedia Clinic program was created at the beginning of the Pandemic. We're interested to hear more about how the pandemic has affected Wikimedia activities: what ways have you dealt with the pandemic in the context of wikimedia work?
    • from Hungary: we miss the in-person meetups.
    • from Ghana: we hosted office hours, to help volunteers with problem they encountered editing, and held virtual trainings.
    • from Egypt: we leveraged the pandemic: we arranged a competition about knowledge, to raise the number of people participating in Wikipedia. The schools were closed for months; in Arabic Wikipedia - we couldn't reach a lot of people (readers) in the Middle East and there are always efforts to reach people and tell them about Wikipedia. People are staying at home and teaching from home, through Zoom and Google Hangouts; this can help recruit Wikimedians. We must be optimistic.
      • We also have input about the Growth presentation: Some Wikimedia Commons rules are not clear to some people. Some editors are curious about privacy and copyrights about photos and videos. It would be good to make it easier to correctly upload photos. Finding the correct license for uploading photos is hard for beginners.
    • from Turkey: We've done training courses. Many organizations are creatively thinking of ways to get people to edit. No face-to-face meetings, but online meetings, which helped include more people from across our country, as well as people who want to remain anonymous so don't show up in person.
    • from the Netherlands: I luckily already have my contacts in my wikis, I know many people personally, and that helps. But for newbies I can imagine it is more difficult to join without the opportunity to meet experienced Wikipedians in person. What I miss, what hardly happens, are contacts with GLAMs, and educational activities.
    • from Tatarstan: I had a lot of work cancelled from March through end of May, so concentrated on developing pillars for Wiki-Smart Tatarstan & it helped me to use the opportunity that came in July that ended up in everything I shared today in wikimedia-l https://w.wiki/uDi