Post-Temporary Accounts English Wikipedia patroller survey
Survey targeting users who do antivandalism work related to temporary accounts on English Wikipedia, to capture sentiments regarding temporary accounts and assess the audience's perceived ability to handle vandalism from temporary accounts
https://commons.wikimedia.org/wiki/File:(Public_release)_WE_4.8.3_Post-TA_enwiki_patroller_sentiments_-_Findings.pdf
May 2026
Claudia Lo for the Product Safety and Integrity team

Broader share-out scheduled for next Wednesday; I'm marking this particular ticket as "resolved".

Update: The report is finished and available internally for review. I am working with Zaree to coordinate a share-out meeting for the broader team.

Update: The report is complete and has been shared to stakeholders. Waiting on a final meeting about next steps, scheduled for next week, before closing this ticket.

Update: Analysis ongoing.

Update:

• I have decided to close the survey early, with 676 total responses (exceeding my target of 400)
• Analysis is under way

Update:

• Survey opened on Apr 14
• So far we have targeted the most active users in our recruitment pools; response rate so far has been very high, with about ~2000 users contacted and 591 total responses
• Given the rate of response we may be able to close this survey sooner than anticipated to give more time for analysis, especially the open-text answer question

Update:

• For this project I have completed 4 interviews with 3 more scheduled, so we are on track to complete 7 interviews this week
• Follow-up emails to two other participants who are partially through the interview process
• Recruitment closed for now as we've hit our targets for both populations

Update:

• Discussion guide and outreach messages complete
• We have started recruitment on Discord, with seven users contacted and three scheduled interviews for the next week
• We have coordinated prototype/stimuli creation. Current aim is to provide 2-3 days lead time for recommendation creation
• We may extend the interview by testing a second prototype during the session, using the same protocol. I don't expect this to require significant alteration of the discussion guide, but it's dependent on the progress of that prototype's creation. This is also a "nice-to-have", if it is unavailable we should proceed with the sessions regardless.

Belated update: I am closing this ticket as Q3 is over. The Suggested Investigations user interviews resulted only in two interviews, presumably from lack of interest; development on the tool wasn't blocked by the outcome of this study, but its findings were considered complimentary to ongoing development. I have created two annotated transcripts (edited for clarity, with anonymized illustrative screenshots inserted) and a short summary of the sessions to serve as the final deliverable (WMF internal document).

Update:

• I met with Sam and talked to Diego asynchronously to provide more details to the main ticket.

Update:

• Anticipated survey deployment date is week of April 13
• Assuming we keep the survey open for 2 weeks, and we're aiming for a target of ~400 responses total
• Recruiting pool:
  • Users with temp account IP viewing rights (~800 total)
  • Users who have reverted an edit made by a temporary account in the last month (querying this for Jan 2026 turned up ~12,000 users who had done so). We're sampling 3371 such users out of this pool, assuming that in March 2026 we also have about 12000 users who have reverted an edit made by a temp account.
• Users will be contacted via Special:EmailUser
• We have obtained a privacy release for this survey
• Survey is implemented in Qualtrics

Made this sheet (internal use) documenting vandalism-related userscripts and gadgets in use on English Wikipedia. The categories are very much a work-in-progress, they are my first pass at trying to sort them based on description. Largely based on information and links coming from Special:GadgetUsage and WP:User scripts/Most imported scripts.

"Users with extended rights", meant to capture all the various "admin+" level user groups that exist across different wikis. From Wikimedia Foundation Annual Plan/2025-2026/Global Trends/Users with extended rights:

For situations that require tools and access — such as the ability to block a user or delete a page — Wikimedians elect trusted members of their communities to carry out advanced tasks and implement policies, guidelines, and consensus. We have begun using an umbrella term, users with extended rights, to describe these users: administrators, CheckUsers, stewards, Arbitration Committee members, and many more.

Patrolling metrics on SPI/AIV report deck is available now (WMF-usage only for now). I will be updating this ticket shortly to reflect a continuation of this work. (This took way longer to put together than expected, including several rounds of additional requests for computed metrics and graphs.)

We have continued work on metrics from English Wikipedia's Sockpuppet Investigations (SPI) board, and obtaining baseline metrics from Administrator Interventions against Vandalism (AIV). All data and scripts used can be reviewed at T416469.
SPI data

• Case durations. Overall, case duration (how long it takes from when a case is first filed, to when it is marked closed in some way) is very variable. For the years 2023-2025 (where we have full-year data), the median case close time is approximately 6 - 15.5 hours. It is relatively consistent year over year. Case durations overall range from a minimum of 0.0378 hours (a little over 2 minutes) to 2594 hours (108 days). Thus, we can say that case durations exhibit extreme variance, with the widest range in the year 2024. This likely prompts an intentional effort to clear the SPI backlog at the end of that year.
  • Median case duration goes from 15.43 hours in 2024, to 5.74 hours in 2025. The post-TA case duration median is 5.33 hours, but there is such high variance that it's hard to attribute this to the introduction of temporary accounts.
  • In conclusion there's no affirmative evidence proving temporary accounts affected case durations at SPI. We can say that there was a sharp drop in case duration metrics from 2024-2025, but this is almost certainly due to an intentional backlog clearing drive.
• Number of checkusers over 2022-2026 at SPI. The total number of individual checkusers involved in SPI is 60, over the entire period; this number includes now-retired checkusers. The number of checkusers per case, over the entire period, varies between 0 - 3, with the vast majority involving 1 or fewer checkusers (mean 0.65, std 0.56)
• Number of opened vs. closed investigations. After resolving a data-cleaning issue where cases could be reported as "closed" without ever having an opening date, we find that the number of closed and opened cases are almost 1:1 over the entire dataset. If checkusers at SPI were overwhelmed, we might expect to see more cases marked opened than closed. However, we don't see any significant periods of time where this is the case.

Update: I am currently supporting the PSI team by running Suggested Investigations user interviews (WMF-only link). In brief:

Update:

• Tran has scraped English Wikipedia's SPI table by looking at revisions made by Mz7 (bot), and shared with me initial datasets of revisions, active case count, and cases by status from 29 Mar 2022 to 28 Jan 2026
• I have begun some early grouping of the data, viewable in this (Foundation only) sheet
• We are assessing the viability of AIV and how we would want to collect relevant data from that venue

Quick summary of progress over the past week:

• We have started by analyzing enwiki SPI, specifically, looking at revisions authored by Mz7 (bot), the bot which performs most archival and documentation tasks to the main table of SPI cases. This allows us to get case summaries for all cases dating back to 29 March 2022 (prior to that, a different bot called Amalthea (bot) performed an analogous function).
• We plan to get the following statistics for SPI:
  • Number of cases filed to SPI from 29 Mar 2022 to present (where "case" is identified by the unique username of the "main" reported account)
  • Number of checkusers involved per case
  • Number of cases by filing status over time
  • Time between the opening of a case and its closure by a CheckUser; this is complicated by the fact that a single case can be repeatedly opened and closed as new information is found, or as investigations are re-filed under other names.
• We are also investigating the suitability of AIV for this kind of metric gathering. AIV reports follow a very different format, but we are confident that relevant information can be gathered from the edit summaries for this page.

@leila To be declined.

Update:

• Reviewed final deliverables for T407493
• Reviewed analysis for UX tests (Userlytics panelists) for IRS non-emergency MVP

Update:

• Worked with Bethany and Katie to prepare UX tests for IRS non-emergency MVP, testing with Wikipedians
• Partially designed and then implemented screeners and a testing protocol for IRS non-emergency MVP, testing with Userlytics panelists (so non-Wikipedians), screened for familiarity with Wikipedia. We did so in order to produce quick-turnaround results on the UX test, and they did come back very quickly as we hoped. Kieran is leading the analysis of these tests at the moment, though I may be tapped to provide assistance if necessary.
• Minor assistance with T407493 analysis, now complete

Flowcharts have been completed. I have also mapped them to IRS categories, in order to facilitate design work on the non-emergency IRS workflows.

We have created flowcharts for various global support processes, which we expect wikis without admins (or, for cases of doxxing, wikis without local oversighters) to use when handling incidents of abuse.

Update:

• Reviewed screener survey, test protocol, outreach messages for IRS non-emergency MVP UX test (for invited participants)
• Helped design + port IRS non-emergency MVP UX test on to Userlytics platform, in preparation for testing with Userlytics panelists
• Minor assistance with T407493 analysis

Update for the last week:

• I've agreed with @KColeman-WMF that the desired deliverable for this is a flowchart of global support structures. This makes sense as global support requests are quite extensively formalized on Meta-Wiki
• Started work on such flowcharts

Update for the week of Sep 29 - Oct 3:

• Assisted in survey design for T402277 (wording, survey flow, survey logic)
• Quick-turnaround analysis for closed-choice questions in T402277

Belated close - The final report explaining findings and our decision to close the KR is completed (link for internal use). It contains a summary of findings, a longer writeup, and examples of generative AI misuse on English Wikipedia with annotations and explanations of context.

After consulting with the team and broader KR owners, we have decided to close this hypothesis.

Mapping non-emergency support structures
Documenting local help and support structures, especially for conflict resolution, across four wikis to inform recommendations for the Incident Reporting System.
Online Social Behavior
N/A (internal use)
Sep 2025
Project lead: Claudia Lo with Katie Coleman, for the Product Safety and Integrity team

Final presentation deck is available for internal use. Closing this ticket.

Pinging @DKumar-WMF to review for close.

Some final updates on elements brought up in this study's research brief:

• Attempts to garner specific comments and feedback on Temp Accounts largely failed, despite efforts from MoveComms and outreach to specific editors
• In the research brief, we initially considered sending out a short survey to gather opinions about the Temporary Accounts rollout. T402277 instead fulfilled this role, so this study did not incorporate its own survey so as to avoid duplicating effort.

Update on findings. @nettrom_WMF has retraced T395618 for the major project rollout wikis, looking at a 30 day window prior to first deployment week compared to a 30-day window after the last deployment (see "Methodological details" below for more.)

I've produced a short writeup for Katie's use, and am now working on a presentation deck for next Tues to share out to the team. On track to finish by listed due date.

We've begun drafting the discussion guide for this project, with a kick-off meeting hopefully scheduled for next week.

We've finished our cataloguing of support structures from Portuguese, Japanese, Turkish and Vietnamese Wikipedias. We are now synthesizing some insights to aid Katie in upcoming prototyping work for the IRS.

Update: We've catalogued 14 more support venues from Japanese and Turkish Wikipedia. At this point, I believe it might be more impactful to work through Vietnamese Wikipedia as both French and English Wikipedia support structures have had prior mapping exercises; I believe that getting a broad spread of wikis is likely more helpful for IRS design work, but I'll be confirming this with Katie in the upcoming week.

Update: We're still waiting for July's data to be available. In the meantime, I have compiled a table of the 31 wikis where Temporary Accounts are enabled, and policies on the temporary-account-viewer usergroup. 60% of the wikis do not currently have a local page for the usergroup (the usergroup page is a red link or redirects to the Meta-Wiki policy); 75% of them use the WMF's local access minimum thresholds for the group.

Update: We've run 7 support venues from ptwiki through our form, and are now working on Japanese and Turkish Wikipedia support venues.

No major update this week.

Update: We've finalized our assessment framework, which uses a form that we will fill out for each support structure we identify. We are now identifying support structures across Japanese, Turkish, English and Portuguese Wikipedia, largely using existing community portals as a starting point.

Update: We're still waiting for more time to pass before we begin quantitative data collection and analysis. On the qualitative front, I have been working with @sgrabarczuk to reach out to ambassadors and find good recruiting venues to get feedback.

Update: We have finished our preliminary selection of wikis, and will begin by mapping support structures on English, Portuguese, Japanese and Turkish Wikipedias. Our selection process only looked at wikis that had more than 10 monthly active admins, on the assumption that these were the ones most likely to have developed bespoke non-emergency support systems. These four wikis were chosen because:

• They cover a wide array of admin capacity (looking at number of admins, number of monthly active admins, and the ratio of monthly active admins to both monthly active editors, and total monthly edits)
• They represent a range in overall size of wiki
• Japanese Wikipedia has a particularly high number of anonymous users, and so are particularly interested to understand how their support structures might differ as a result

Update: We've concluded our research brief alignment and will begin work in earnest, starting by creating an assessment framework for support structures as well as finalizing our chosen wikis for this study.

Weekly update: Research brief alignment meetings concluded, we can begin in earnest. Morten and I will work to figure out the details of our approaches. We've also begun soliciting examples where temporary accounts has helped or hindered attempts to stop bad behaviour from the community, using existing comms channels for the Temporary Accounts project.

Weekly update: given the prioritization of incoming request tasks this quarter assigned to me, this is likely going to be picked up and started later in Q1 (Aug through Sep).

Update: research brief discussion meeting set up for next week.

Weekly update: research brief review meeting scheduled for next week.

Update: given the collaborative and iterative nature of this request, I believe it makes sense if both I and @KColeman-WMF jointly work on this research. I will take lead on wiki selection and identification of support structures, while she will be documenting these different support structures.

After discussion with @nettrom_WMF , I believe this task would be better scoped with this set of primary research goals:

1. Understand impact of deploying Temporary Accounts (TA) on pilot wikis, to inform development of anti-abuse features
   • Is the loss of IP address info at large, significantly detrimental to community anti-abuse efforts?
2. Document uses of TAs for abuse, to inform development of software-based mitigations
   • Are there documented cases where Temporary Accounts have broken existing anti-abuse workflows?
   • If yes, what are potential product recommendations to address these scenarios?
3. Document how IP reveal rights are being rolled out, as they now must be manually granted

@DDeSouza Please add the following project:

Our Meta-Wiki page is now updated and we have completed our share-out. Moving to sign-off column.

After further discussions with the team and with Isaac, we have settled on these additional points:

• The primary purpose of this research is to describe already-occurring phenomena rather than trying to (quantitatively) estimate their prevalence
• We are particularly interested in the ways in which generative AI is negatively impacting the social experience of editors, for instance, accusations of LLM use based on assumed English fluency versus contribution styles

I will also edit the ticket with further details about estimated effort, priority, and the use cases of the research outcomes.

Updated task description after meeting with requesting team. I have some additional notes around scoping and selection of wikis here:

Updated task description, especially around estimated effort and priority, after more discussion with stakeholder group.

Update: we are nearing completion on the written deliverables as agreed upon at kickoff. We're hoping to schedule a share-out sometime in the next week and to update our Meta-Wiki page at that time as well.

We have successfully completed all participant sessions!

Update: Our final two sessions are scheduled, and we have begun analysis of the sessions. The Userlytics workaround we developed (creating an unmoderated version of our testing protocol, conducting the interview portion via Google Meets) worked very well, so we may want to keep this in mind in the future for participants with no webcam.

Synthesis was completed and sent to @DKumar-WMF and Kadeem Khan for review on 16 May. Debra, please review for close?

Update: We're at 3 sessions complete, 3 more scheduled; if all goes well we will have hit our target session count.