CASE STUDY | WHATSAPP
Misinformation Pandemic
UX Research, UX Design, & Usability Testing
The spread of misinformation presents a complicated problem for social media platforms. And when it comes to COVID-19, misinformation can be a life-and-death matter. While several solutions have been suggested for public-facing posts (i.e. Twitter, Facebook), additional complexities are introduced in platforms that are used primarily for one-to-one messaging.
WhatsApp has become one of the fastest growing social media platforms, eclipsing 2 billion users in 2020. A plan for handling misinformation has the potential to significantly increase health behaviors and thus blunt the impact of the COVID-19 pandemic. WhatsApp presents an additional challenge for misinformation control because of end-to-end encryption of the users' messages.
Team
This team was a part of the University of Washington Human Centered Design & Engineering program. Despite working entirely remotely on this project, our team was able to form a collaborative and productive workflow online using tools such as Slack, Google Drive, Miro, and Zoom.
Cassidy Trier
Saakshi Goel
Florin Mehedinti
Vijayalaxmi MP
Ben Buck
Design Question:
How might we prompt WhatsApp users to think critically about COVID-19 information?
RESEARCH
Methods
Multi-Angled Approach to Understanding the Problem
We recruited users of multiple social media platforms for interviews (n=5) and survey methodology (n=64), and drew upon multiple IDEO design methodologies including:
RESEARCH
Conclusions
Key Observations Driving Design Requirements
-
Many users identified the importance of sources that they viewed as "neutral" or "lacking political agenda," (e.g. the CDC, academic journals, or colleges & universities)
-
Users also identified how brief and digestible forms of media were most memorable and clear.
-
Both existing literature (e.g. Chan et al., 2017) and users suggested that users were best suited to disbelieve misinformation when they were empowered to gather their own information.
-
Users were more likely to identify a sense of embarrassment or perception of bias when platforms aggressively police (e.g. disabling sharing) content users encounter.
Design Requirements
1
Invoke academic or apolitical sources to debunk COVID-19 misinformation
2
Ensure debunking is maximally digestible (e.g. visuals, bullet points)
3
Provide users tools to engage their own critical scrutiny
4
Empower the user rather than "call them out" as wrong
PERSONAS
Understanding the User
Research-Driven User Understanding
Our team created personas and storyboards in order to understand our primary and secondary users and their goals and frustrations. The story board shows how the personas might interact with the various WhatsApp misinformation solutions.
Primary Persona:
Secondary Persona:
IDEATION
Sketches
Working as a team to generate ideas
To begin the ideation phase of this work, our team first worked individually to generate fast sketches. During this phase, we focused on quantity of ideas. Once we had amassed a sizable amount of sketches individually, we each submitted our 6 best ideas and plotted them on two axes using a Miro board.
IDEATION
...and More Sketches
Cleaning Up & Moving Forward
Once we had our sketches plotted, we could better understand how they related to one another and what the strengths and weaknesses of each idea might be. We looked at common solutions that arose and cleaned up our favorite sketches.
.jpg)
IDEATION
Checking in with Personas
Using Persona Cases to Evaluate Ideas
With our sketches in hand, we decided to "check in" with our personas so see if these solutions we had drawn would meet their goals. We created scenarios in which our personas would interact with various ideas.
DECISION MAKING
Rational for Design Solution
A debunking strategy
Due to the constraints placed on the project due to WhatsApp's end-to-end encryption, we ultimately decided corrective information should not be delivered in response to individual user activity.
Our solution will deliver "prebunking" (Roozenbeek et al., 2020) information with the goal of getting correct information out into the public before misinformation can circulate. We decided to create a misinformation alert which would "pre-bunk" misinformation that has been identified as circulating in the users' area. The misinformation will be gathered from the rest of the Facebook family and will be regionally-specific.
PROTOTYPING
Medium Fidelity Prototypes
Finding Placement
We developed two prospective prototypes in order to find the best placement for the Misinformation Alert we wanted to implement. We ran this A/B during our usability testing. Both prototypes assume capacity within WhatsApp/Facebook to identify rapidly spreading misinformation in their family of apps, but test differing means of delivering these warnings.
Fixed Alert: This screen would be populated with misinformation that had been flagged as spreading in the Facebook family of apps, and it would be targeted for pre-bunking misinformation control.
PROTOTYPING
Medium Fidelity Prototypes
Finding Placement
We developed two prospective prototypes in order to find the best placement for the Misinformation Alert we wanted to implement. We ran this A/B during our usability testing. Both prototypes assume capacity within WhatsApp/Facebook to identify rapidly spreading misinformation in their family of apps, but test differing means of delivering these warnings.
The screen shown in these two prototypes would be populated with misinformation that had been flagged as spreading in the Facebook family of apps, and would be targeted for pre-bunking misinformation control.
Importantly, both prototypes included links to a straightforward COVID-19-related infographic.
Pop Up Alert
Fixed Alert
INFORMATION HIERARCHY
Content & Modality
User-Driven Decision Making
During our usability tests, we also asked users what kind of content they would like to see on on infographic on WhatsApp about COVID-19. We asked them about the modality they prefer for receiving information as well. We used Optimal Sort to streamline this process.
Content Sort
Modality Sort
PROTOTYPING
Usability Study Results
Refining our Prototype
After the usability studies, our team compared notes and compiled a list of recurring themes, recommendations, and gave each change a priority level.
PROTOTYPING
Final Design
Putting It All Together
Our team responded to the user feedback and adjusted the prototype.
In order to demonstrate the local-responsiveness of the information, we translated the infographic into English, Spanish, and Hindi, with different debunking information in each area.
SHARING
All of this information can be easily shared with other WhatsApp users via the sharing function.
Dismiss alert with clearly placed button
Concise explanation of alert purpose
Clear call to action linking to more COVID-19 information via infographic (below)
Horizontally scroll through recently identified misinformation. Link to infographic
Statically fixed container that moves with scrolling direction for minimum disruption