Page MenuHomePhabricator

Conduct a control test of as-is reply workflow
Closed, ResolvedPublic

Description

In an effort to understand, qualitatively, what contributors' experiences replying to specific comments on talk page is like, we are going to run a usertesting.com test of the "replying to specific comments" workflow on desktop, as it currently exists.

These test results will also be a helpful baseline to compare against those in T236921.

"Done"

  • Draft testing goals
  • Run test(s)
  • Summarize test findings
  • Draft improvements to address test findings - this can be tracked in T236921

Usability Test

Test Script: Google Doc

Testing Goals

Talk pages are integral to many workflows including receiving feedback on edits from other contributors. This year, the Editing team’s goal is to partner with communities to trial features that make contributing to Talk pages, and understanding the information and activity inside them, easier and more intuitive. Our first iteration on Talk pages is tackling the workflow for how contributors respond to comments. There are many obstacles for both junior and senior contributors in the current workflow, including figuring out where to respond and how to sign and indent their responses.
The primary goals of this test are to gain insights on discovery, composition and view history of Discussion pages.

Research Questions

Discovery

  • Are users able to successfully select the reply button without any guidance/onboarding?

Composition

  • [Junior Contributors] Is the replying workflow – in its entirety – intuitive and easy to use?
  • At what points – if any – are participants not sure what to do? At what points – if any – are participants not able to predict what impact their actions will have?

Revisions

  • Are users able to successfully select and compare revisions from the History page?
  • Are users able to tell what was added to the page?
  • Is the information presented presented in a way that is understandable to users? Are they able to tell what was removed or added?

Event Timeline

iamjessklein updated the task description. (Show Details)
iamjessklein updated the task description. (Show Details)

I ran the first test on usertesting.com on November 27, 2019 - [CT 1-5] This is a control test. We tested a discussion page on beta that had no prompts, just discussion topics to reply to. The detailed findings can be found on limited access test log.

  • 5 tests were conducted
  • 2 participants were female ;3 participants were male
  • 4 participants were desktop web users; 1 participant was a mobile web user
  • All 5 participants were screened to ensure that they were technically advanced web users who have used Wikipedia in some capacity.
  • 2/5 participants were ESL taking the test in English.

Here are the findings:

Identifying section heading
✅5/5 participants identified the section heading
✏️ Many of the participants could not differentiate between the title/heading and initial comment.

Replying to comment
✅ 5/5 participants identified how to reply to and write a comment
❌ 0/5 participants used proper indentation or signing practices
✏️ It was challenging for participants to equate the edit link with a reply
✏️ Several participants didn't consider their edit a reply. One participant said, "I never figured out how to make my own comment, I just figured out how to add content to the existing comment on the topic"

Publishing reply
✅ 5/5 participants published their reply
✅ 3/5 participants wrote an edit summary

Locating published reply on article page
✅ 5/5 participants located their reply
✏️ Several participants took several minutes to find their reply on the page

Cancelling out of task mid-edit
✅ 5/5 participants cancelled out of task mid-edit
❌ 0/5 participants anticipated that their actions would cancel them out of the edit appropriately

Viewing page history tab
❌ 1/5 participants gave up after struggling with this task
✅ 4/5 participants found the page history tab successfully
✏️ One participant said "one thing that confused me was the timestamp (im not sure if this is my timestamp - it's 2:07 here and that doesn't correspond to the military time here).

Locating reply/summary in page history
✅ 3/5 participants were able to locate their reply in the page history
❌ 0/3 participants who located their reply did it with ease
✏️One participant said "there are a lot of shortenings, colors, symbols. Everything together surprised me."
✏️Even though one participant successfully found it and clicked on it, she didn't think that the edit link took her to what she edited.
✏️3/5 participants commented on the look and feel of the page using words such as " confusing" "messy" and "verbose"
✏️two participants suggested reordering the information on the history page so that the content summary appeared first.


Some notes on the test strategy:

  • Users did not log in and we should test in the future a version where participants log in.
  • I think that we should run another test with lots of comments populating the discussion
iamjessklein renamed this task from Workflow as-is: conduct usablity testing to Conduct a control test of as-is reply state.Dec 4 2019, 2:54 AM
iamjessklein updated the task description. (Show Details)

The next step here is to publish the second iteration of the control test on usertesting.com (with lots of comments populating the discussion) and to then log and synthesize the findings here.

iamjessklein renamed this task from Conduct a control test of as-is reply state to Conduct a control test of as-is reply workflow.Dec 4 2019, 2:57 AM

Update: v2 of the control test is running. I will post the findings here.

I ran the second test on usertesting.com on November 28, 2019 - [CT 6-10] This is a control test. We tested a discussion page on beta. In contrast to the first test, this test had pages filled with lots of comments from other users. The detailed findings can be found on limited access test log.

5 tests were conducted
3 participants were female ; 2 participants were male
5 participants were desktop web users
All 5 participants were screened to ensure that they were technically advanced web users who have used Wikipedia in some capacity.
1/5 participants was ESL taking the test in English.
Here are the findings:

Identifying section heading
✅5/5 participants identified the section heading
✏️ Many of the participants could not differentiate between the title/heading and initial comment.

Replying to comment
✅ 5/5 participants identified how to reply to and write a comment
❌ 0/5 participants used proper indentation or signing practices
✏️ It was challenging for participants to equate the edit link with a reply
✏️ One participant did not edit using the "edit" button, but instead used the "create new section" tab. "It was hard to find an actual reply button so the only option was to create a new subject and tag the user or edit his subject with my answer."

Publishing reply
✅ 5/5 participants published their reply
✅ 0/5 participants wrote an edit summary

Locating published reply on article page
✅ 3/5 participants located their reply
✏️ Several participants took several minutes to find their reply on the page
❌ One participant gave up at this point (after clicking around she ended up on the revision history of wikipedia:automatic edit summaries)

Cancelling out of task mid-edit
✅ 4/5 participants cancelled out of task mid-edit
❌ 0/5 participants anticipated that their actions would cancel them out of the edit appropriately

Viewing page history tab
✅ 4/5 participants found the page history tab successfully
✏️ One participant said ""for me, all i see is a jumble. i can make it out but I can't really see what it is."

Locating reply/summary in page history
✅4/5 participants were able to locate their reply in the page history
❌1/4 participants who located their reply did it with ease

Other:
✏️"I've edited something once or twice and so discussions is something I wouldn't ever go on." - Why?
✏️Several people mistook the undo icon as a back button
✏️ 0/5 users signed their comment, one participant said "I probably should have put my user name, but i didn't" after noticing other users put their username in the discussion comments.

iamjessklein updated the task description. (Show Details)
iamjessklein subscribed.

@ppelberg these two control tests are concluded. I am reassigning this ticket over to you - to make sure that the findings go on wiki.

Cancelling out of task mid-edit

Which editing environment were these people using?

@Whatamidoing-WMF this is the current state (no changes - no reply button) and is a mix of mobile and desktop web users.

JTannerWMF removed a project: Mobile.
JTannerWMF moved this task from 📢 To Do to 👁Needs Engineering Input on the Editing Design board.
JTannerWMF subscribed.

@ppelberg to create a new task for testing on mobile and close this ticket out. This ticket should be referenced in the new task.

@ppelberg to create a new task for testing on mobile and close this ticket out. This ticket should be referenced in the new task.

T245798