Sprint 2: SOS! Abducted Researchers in Grave Danger
As we come to the end of the 2nd project sprint of the Summer months, team WICked Nutrition has made a lot of progress that I, Onye, have agreed to act as the team’s scribe for this week and put it all in writing.
Research — What promotes behavioral change?
As studies have shown…
For these past two weeks, we put our pencils to our papers and have crafted, drafted, sketched, and brainstormed various design solutions with aims to approach our design research question for this Sprint 2: What design patterns promote behavioral change? As a common design research question that has been probed and studied and that still, today, stands as an open question, we began our deep dive by conducting background research. From Alqahtani’s (et al.) research paper (2019) where they explore different mobile applications and analyze which aspects and features contribute to behavior change and motivation, each team member chose up to four apps and used them for a week. In our competitive analysis, we were looking for four specific design aspects that we identified as requirements for our own design solution: (1) reminders, (2), self-monitoring, (3), praise and reward, and (4) personalization.
We asked the experts
Fortunately, studying at CMU has given us many opportunities to talk directly with academics who have dedicated their time studying this very concept of motivation and behavior change through digital mediums. We spoke to Julian Ramos, PhD candidate at CMU Human-Computer Interaction Institute, about the degree of personalization and other motivational factors affecting behavioral change. From Ramos and Robert E. Kraut Assistant Professor School of Computer Science, Geoff Kaufman, we learned that the question of what kinds of design specs contribute to or cause behavior change has yet to be answered. Adjunct Instructor of HCII Anna Abovyan, Professor of Practice, Paul Pangaro, and assistant teaching professor Raelin Musuraca were also among the list of experienced practitioners we met with to learn more about testing the right “thing” with participants, chatbot design, the capabilities and implementation of a conversational user interface (CUI), and how behavior change can be measured.
Our team was also given the pleasure to meet the CEO of Mobile Benefits and point-person behind the development of the QuickWIC mobile app, Max Thayer, as well as Vittorio Banfi, the co-founder and CEO of BotSociety, a voice and chat design software. It was great to learn what types of design patterns, chatbot capabilities work and haven’t worked in the past in order to apply (or avoid) such patterns when it comes to our design.
Longitudinal Study Update
We discovered 3 key findings from our four-week-long longitudinal study thus far:
Finding #1: Participants don’t actively seek out WIC resources, but value information that is proactively sent to them
- Control Group (didn’t receive daily tips): 1 out of 6 participants (17%) clicked on the resources we provided them in an email
- Experimental Groups (received daily tips): 8 out of 10 participants (80%) implemented at least one tip in their lives
Finding #2: Participants express more interest and engagement in info that is personalized to them
- 6 out of 6 participants (100%) in the Experimental Group that received personalized tips with their names and child’s names responded back.
- 4 out of 6 participants (67%) in the Experimental Group that received general tips responded back.
Findings #3: These tips were “encouraging” to participants
- Participants liked the support and reminders even if they already knew the information
- 5 out of 14 participants (36%) expressed interest in milestones and needing help to reach their goals
“I need someone constantly saying like, Okay, you should eat this or don’t eat that would be helpful.”
“5 out of 14” or 36% percent may seem like a small amount, however, it is notable to mention these participants expressed such interest unprompted. If we were to compare this small representative to the whole (of PA WIC participants), and prompt them more specifically, it’s safe to presume that this would be a large, common need among participants.
Decision-making
We’ve done the research, we’ve validated our design solutions, and we’ve updated our clients. What have we decided? Team WICked Nutrition has decided to create a digital WIC companion, WICster*– a proactive chatbot, help assistant that uses nudging in the form of daily reminders, to educate and encourage PA WIC participants in reaching their family’s health goals. WICster comes in two forms, an SMS-based chatbot and a mobile application (responsive web app or native app, TBD). SMS-based WICster will be designed for those who do not have access to download a mobile app or who prefer not to. Mobile App WICster will not only have all the capabilities SMS WICster does, but will also come equipped with a library of nutrition tips, activities, WIC recipes, and more.
*WICster is a temporary name for our product
Sketching — “It is our pleasure to introduce…”
These are what we know — (1) Our final deliverable will be focused on a companion that is proactive and instills feelings of motivation in new mothers. (2) New mothers will have access to WIC’s resources at times, in sizes, and with content that is useful to them. To help define these knowns, we engaged in the Press Release design sketch activity and came up with a defined framework for our product.
We Were AbductIVE in Reasoning
Abductive reasoning really propelled us into a forward-moving dimension of our design, in short, it changed our world. As researchers, we have been “taught out” of such type of reasoning, often encouraged to use deductive and inductive tactics to arrive at a conclusion. However, abductive reasoning, defined as the act of making a probable conclusion from what you know in the case that the major premise is evident, but the minor premise and therefore the conclusion are only probable (Merriam-Webster), helped us to synthesize our interview and user findings into actionable insights that could directly be applied to our standing design iteration. For example, we discovered that (1) controlling the context in which mothers receive health tips has a direct implication on its associated value and (2) creating value in journaling/reflecting will require giving prominence to features mothers already use or have interest in tracking (i.e. taking pictures of meals).
As the iterative design process indicates, our next steps after concluding on these abductive insights were to sketch.
and More Sketching
We went back to the digital drawing board and used the Why, How, Prototype, Iterate (or WHPI, pronounced “whoopee”) approach to develop sketches of our SMS and mobile app chatbot solutions in a goal-based fashion. We also incorporated The Golden Path design sprint activity to define the various conversation-style interactions a user would be able to engage in when interacting with our product.
Outcomes
The voice and tone of WICster will be casual, respectful, and enthusiastic with occasional comedic antics. Even so, we are exploring the option to let users choose the type of tone WICster would use to interact with them. When it comes to guidance and nudging, different people may respond better to a more serious, coach-like tone while others may benefit from a more warm, nurturing tone of voice.
We were able to sketch and script the onboarding or first interaction of WICster for both SMS and mobile app platforms. Since personalization is one of our key requirements, it is important users are able to customize their settings, understand WICster’s capabilities, and immediately get acquainted with their WIC companion as soon as possible.
Tackling technical feasibility
In order to design a solution that is practical and ultimately usable by our clients and participants, we began researching and testing out various software, so that we could test functionality and design for our project hand-off. We began building this functional chatbot using BotSociety, Twilio, Firebase, and Google Sheets.
Retrospective
With a few iterations under our belt, we can’t help but notice that there are only a few more weeks left in our capstone project! What’s to come next? How will participants respond to a higher fidelity of WICster with various new features, resources, and nudging capabilities? Well readers, this about wraps up my scribing duties for this project. It’s been nice sharing with you. You can find the answers to these questions if you…
–Come back and read our next Sprint’s blog post!
Thank you for your time,
Onye ✌🏾
