wemar
A conversational social robot intended to check the mental state of teens and lead them through mindfulness exercises and stress-relieving activities.
Project for Designing for Kids course, sponsored by Team EMAR at the University of Washington.
Role: UX Designer
Team: 2 researchers, 1 developer, 1 designer
Background
Project EMAR is an ongoing interdisciplinary effort from UW Seattle and Tacoma campus researchers to measure teen stress by way of a social robot. The goal of Team EMAR, which stands for Ecological Momentary Assessment Robot, is to be engaging and trustworthy for students to feel safe while confiding stress and mood information.
Participatory design sessions and interviews conducted with high school students in local Seattle schools have helped shaped the friendly design of several EMAR iterations. However, current iterations of EMAR remain limited in dialogue options when interacting with users, and beyond measuring adolescent stress cannot offer methods to relieve it.
Design Question
How can we design a more interactive and conversational social robot that can introduce teens to the concept of mindfulness and alleviate stress?
Research
Competitive Analysis
We started by conducting a competitive analysis on existing mindfulness apps and programs to draw observations on effective stress-relieving techniques.
In particular, we focused on the Learning to Breathe (L2B) mindfulness curriculum, which had the backing of multiple research studies and specifically focused on a high school audience. The six tenets of this program (body, reflections, emotions, attention, tenderness, habits for a healthy mind, and empowerment) were represented by the acronym BREATHE; a pilot study of the program indicated that participants experienced fewer negative moods, improved calmness, and greater self-acceptance compared to a control group.
Another existing service that we reviewed was the mobile app Headspace, which claimed that habitually using its meditative mind-training sessions for 10 days reduced user stress by 14%. Despite Headspace utilizing a paid subscription method that would likely be too costly for our teen users, we found particular value in the various duration times available for exercises, as well as its clean and attractive visual design.
Interviews
Our team then conducted semi-structured interviews with three local high school students. These interviews aimed to better understand the following:
Common stressors and stress relievers
Role of technology in causing/dispelling stress
Familiarity with mindfulness in a mental health context
From these interviews, we determined that we should design for a teenage audience that wasn't familiar with mindfulness but utilized technology as a way to distract themselves from life and school stressors.
Design Requirements
Based on our research findings and the background research gathered from the EMAR team, we outlined a set of design requirements:
Help high school students relieve stress in a non-judgmental and non-disclosing manner
Be easy to use within a short amount of time within a high school context (between passing periods or during lunch)
Appear and interact with high school students in a fun and friendly manner
Introduce mindfulness as a method of reducing stress
Ideation
Co-Designing with Teens
Once we determined our design requirements, we conducted a co-designing session with three different high school students. Having held a separate team session to divergently sketch our own ideas beforehand, we gave our teen participants 10 minutes to draw/write down what methods they used to relieve stress; after briefly introducing EMAR and our own sketches, we gave them another 15 minutes to drawn their own takes on how EMAR could be improved to better measure and reduce stress.
Storyboards
Following this session, we conducted a follow-up team session to converge our ideas and our participants' ideas together. From these converged sketches, we picked three of these ideas from which to construct storyboards to guide us through potential user flows:
The ability for WEMAR to play music, and have students be able to listen via headphones
Attaching stress balls that students could squeeze to relieve muscle tension and stress
The ability for WEMAR for students to watch funny videos using WEMAR
Information Architecture
Having decided to convert these three ideas into our WEMAR robot's three principle features for providing stress relief, we drafted out an information architecture flow.
Prototyping
Lo-Fidelity Prototype
Because we now wanted to test and tweak with WEMAR's interaction flow based on our information architecture and features, we designed a low-fidelity cardboard prototype that resembled the design of the original EMAR robot using sticky notes imitating screens to Wizard of Oz our way through interactions. We also used a combination of styrofoam and tape to crudely craft arms and a little cup to hold the stress ball.
We asked our classmates and instructors to give us feedback on this iteration during a class critique session. From this session, we decided on the following changes to future iterations:
Place the stress ball and its holder separately from the robot
Use qualitative options when assessing mood, rather than a quantified bar
Require users to insert their student ID in an attached card reader for security purposes, emphasize this reader will NOT take their data
Revise WEMAR's responses to sound conversational and more empathic overall
Hi-Fidelity Prototype
For our hi-fi prototype, we wanted to improve both WEMAR's internal and external capabilities. For this iteration, we used Rhino to design a 3D model and then laser-cut wood to build a new body; we included a slot for an iPhone to display interaction changes in WEMAR's eyes that we manually rigged using Keynote, and a slot for an iPad to display the screens that we designed in Framer X. We also included a 3D-printed card reader and stress ball holder that would sit at WEMAR's base.
We used Framer X so that more complex logic between interactions (returning to a previous screen, conditional variables and input) were implementable in this iteration, although some screens still proved to be buggy on some attempts but not others. We also used a freesource text-to-speech site to voice WEMAR's lines and added these sound files into Framer X for each screen.
Conclusion
Exploring this project came with its own hurdles: it was a challenge to work with a sensitive topic like mental health and mindfulness, especially when designing for an emotionally vulnerable user group like high school students. Because we were informed by the EMAR team that extensive approval would need to be sought if we wanted to obtain user information in the same capacity as their team did in visiting local high schools, we had to tread carefully in designing WEMAR with those privacy, legal, and moral constraints in mind.
Through WEMAR, I figured out how to use Framer X for the first time (even if it wasn't really my favorite prototyping tool) and learned how to be a better interviewer for research - part of the fun was just sitting inside a Starbucks and casually talking to students about college/life advice, or even about topics like video games and music. I've learned it's critical to build rapport with your participants for more honest feedback, and adjust your interviewing style to be more flexible depending on the person or group.