Week 9 marked a major transition for our team as we moved from prototype development into real-world playtesting. We conducted our first sessions at the CMU Children’s School, tested both prototypes with children, and gathered critical behavioral insights. Through direct observation and feedback from advisors, we identified key usability challenges, unexpected interaction patterns, and important directions for iteration. At the same time, we began formalizing our research paper direction, aligning our design work with academic framing. This week shifted our process from assumption-driven design to evidence-driven iteration, laying the foundation for refining both our game experience and research approach.
Playtesting at Children’s School
This week marked a major milestone for our project as we officially began playtesting at the CMU Children’s School. Due to scheduling conflicts with the regular playtesting time, our team decided to take turns attending the sessions. This week, our producer Min took the responsibility of going to the Children’s School and conducting the playtesting session on behalf of the team.
During our first session, we tested with around 10 children in a classroom setting, observing how they interacted with both of our prototypes in real time. Seeing multiple children line up and engage with the game provided us with our first real glimpse into how our design performs outside of controlled assumptions. This shift from internal testing to live observation allowed us to capture authentic behaviors, unexpected interactions, and patterns that could not have been predicted beforehand.
One of the most exciting outcomes of this session was observing emerging behavioral patterns in how children interacted with the game. During character customization, many children created avatars that reflected themselves (e.g., matching hair color), while others approached it more freely. We also observed identity-based preferences, such as selecting same-gender characters.
In another key interaction, multiple children independently chose the same gift-giving target, suggesting that their decisions may be influenced by factors such as layout, positioning, or visual hierarchy rather than intentional choice.
These observations highlight the importance of measuring implicit behavioral signals, which is central to our research approach.
Interaction & UI Challenges
Playtesting revealed several critical usability challenges that we need to address in the next iteration.
A major issue was the interaction model, particularly the distinction between dragging and tapping. Many children did not intuitively understand dragging until it was explicitly introduced.
In addition, UI clarity and navigation flow posed challenges. Buttons such as “Next” were often unclear, and children sometimes skipped over critical steps unintentionally—either by mis-pressing the “Next” button or accidentally swiping across the “Finish” button, leading to premature completion of tasks.
Advisor Feedback & Design Direction
To address this, our adivosr Jessica Hammer and Heather Kelley suggested structuring interactions into more explicit, step-based actions, ensuring that each user input is intentional and clearly registered by the system. For example, instead of a single tap to finish, children could be required to press and hold a button, allowing a visual indicator (such as a loading circle) to complete before proceeding. This helps prevent accidental actions and makes the interaction feel more deliberate rather than abrupt.
Additionally, some features, such as the decor section, were underused or not discovered, suggesting that our interface does not yet effectively guide attention or communicate interactivity.
Another key insight was the potential influence of positional bias (e.g., selecting the rightmost option), leading to the recommendation to randomize character placement in future playtests.
They also emphasized the importance of adding “juice”—visual, motion, and sound feedback—to make interactive elements more noticeable and engaging. For example:
- Highlighting interactive elements with color
- Adding motion cues when screens load
- Using clearer progression indicators (e.g., checkmarks instead of arrows)
Additionally, they encouraged us to focus on polishing problematic interactions first, ensuring that the most confusing parts of the experience are improved before refining already functional areas.
Iteration Priorities Moving Forward
Based on our findings, we identified several key priorities:
- Improve UI clarity and prevent accidental inputs
- Introduce interaction patterns more clearly (e.g., drag vs tap)
- Implement step-based interactions to reduce unintended actions
- Add visual and auditory feedback (“juice”) to highlight affordances
- Randomize layouts to reduce bias in behavioral data
- Simplify flows to ensure one clear action per screen
- Encourage exploration of underused features
We will continue iterating on both prototypes based on these insights.
Research Paper Progress
In parallel with design and playtesting, we began actively developing the structure for our research paper.
We outlined the core framing of SoKids as a game-based assessment system that captures implicit behavioral data, positioning it as an alternative to traditional methods such as interviews or sorting tasks.
Our current outline focuses on several key components:
- Defining the research gap in measuring early social categorization in children
- Presenting SoKids as a non-intervention-based system that avoids influencing behavior
- Identifying measurable behavioral signals such as interaction patterns and decision latency
- Designing a methodology that integrates playtesting data with validation measures
- Addressing ethical considerations, including consent and data handling
This process helped us align our design decisions with research goals, ensuring that our prototypes not only function as games but also serve as data collection tools for studying early social cognition.
Reflection
Week 9 represents a turning point in our project.
Rather than relying on what we expected children to do, we began observing what they actually do, and the gap between those two revealed the most valuable insights. Small design decisions, such as button behavior or layout positioning, had significant impacts on user behavior.
Moving forward, our process will be increasingly shaped by real user interaction data, allowing us to refine both our design and research methodology with greater confidence.
