Page MenuHomePhabricator

Conduct pre-deployment usability testing of Multi-Check Phase 2 (References)
Closed, ResolvedPublic

Description

In T379131, we will be running an A/B test to evaluate the impact of lifting the constraint on the number of Reference Checks people can see in a single edit session.

This task involves the work of conducting a usability test of – what we're calling – the "Multi-Check (Phase 2)" experience to validate that the UX is in a good enough place for us to evaluate the impact of it.

Learning objective(s)

  • How do people respond to being confronted by, potentially, a significant number of additional steps before being able to publish the new information they just finished drafting?

Decision(s) to be made

  • What – if any – adjustments might we make to the Multi-Check (Phase 2) UX before beginning the A/B test of the experience?

Requirements

The usability test will be conducted on both mobile and desktop, using a scenario of an edit session where multiple paragraphs are being added. Upon clicking "Publish changes" (or tapping ">" on mobile), the user will arrive at the "Multi-Check (Phase 2)" experience where they will have multiple checks to complete.

Through this test, we seek to understand if the overall heuristics are functional for this prototype. The more conceptual learnings from having Checks and such will be evaluated through other tests once the feature is deployed.

For this test, we will be using the following prototypes on: Desktop and Mobile.


T349027

NOTE: consult with @ovasileva to ensure the research questions cover how people experience the "right-rail" Edit Check design and the Page Tools Vector (2022) introduced.

Event Timeline

ppelberg added a project: Goal.
ppelberg updated the task description. (Show Details)
ppelberg updated the task description. (Show Details)

@nayoub: the additions you've made to the task description align with what you and I had converged on offline. Said another away: this looks great and I think is in a position for work to begin.

With the above in mind, @JFernandez-WMF, I'm boldly assigning this over to you. Of course, please let me know if you notice anything unclear/ambiguous/etc. that Nico or I might be able to address.

Per today's offline meeting with @JFernandez-WMF, next steps are as follows:

  • Update protocol to:
    • Ensure Paste Check is not activated
    • Specify people add references in response to the Reference Checks they are shown
    • Included revised follow-up questions. See below.
  • Re-run test (desktop + mobile)

Follow-up questions

  1. Does this experience cause you to feel more or less confident editing Wikipedia? Why?
  2. What about this experience stood out to you?
  3. What do you think could be improved before this feature is offered to everyone?
  4. What did you think about when this feedback was presented to you?

Two rounds of unmoderated user tests were run in Userlytics — see deck for the first round here:

Second round was tested with 5 desktop participants and 5 mobile participants. For desktop testers, 75% of participants answered that they had edited Wikipedia once or twice, and 25% had never edited. On mobile, 60% selected 'Yes, once or twice', and 40% selected 'No, I haven't.

The task they were asked to do was to add a few paragraphs without citations to an article. They were then asked to work around the flow of actually adding a citation through Reference Check, and also work around the flow of selecting 'No' to adding a citation.

Main highlights

  • Overall, the tool was found easy to use on both desktop and mobile.
  • It boosted confidence in editing and in using Wikipedia in general
  • There was divergence in feedback re: number of prompts - most of the users didn't feel overwhelmed with the amount, but there were some folks that found them time-consuming and saturating.
  • Most of them said they wouldn't make any changes before testing on a larger audience.
  • May be worth looking into the potential of users feeling overwhelmed when facing a number of citation/referencing 'issues'

Here's the .pdf with notes: