ExplorHer

user testing and lab study

In order to truly test the successfulness of the app and its integration of research and interactive design, I completed a study in the University of South Carolina’s Biometrics and User Experience Lab. With the help of Dr. Taylor Wen and Cindy Chen, we programmed the prototype into iMotions, a program that enables eye-tracking data and facial expression analysis. This data allowed me to analyze participant engagement levels and emotion types while interacting with the prototype and study fixation time on different screens. I began with collecting ten participants, women ranging from 18 to 23 years, and brought them in individually to the lab. After a short description of my study, purposefully providing little to no instruction on correct usage and recommended paths, I left the participant alone with my prototype for them to freely interact with for ten minutes, mimicking what a new user's first-time experience would be like. Simultaneously, while participants interacted with the app, a camera attached to the computer screen tracked eye movement and emotion, behavioral, and facial expression signals. After the ten minutes, the participants were asked to complete a written survey with both multiple-choice answers and written responses. The multiple-choice answers were based on a scale, providing insight surrounding language, ease of us, credibility, and education retention. The written responses provided a space for more open-ended feedback, with looser prompts asking about their overall experience and suggestions for improvement.

The eye tracking data provided insights into the effectiveness of my visual design choices. The heat maps provide a visualization of which parts of the prototype attract visual attention. Overall, the heat maps deemed the design as successful in capturing visual attention and encouraging engagement. 

The facial expression analysis data is captured through detecting small signals in facial landmarks such as brow furrow, eye widen, smirk, lip press, etc. These signals are then calculated into facial expression scores using emotion signals, categorizing the movement into emotions such as valence, engagement, confusion, joy, etc. For this study I focused on the following emotions: confusion, surprise, engagement, joy, and valence (pleasantness vs unpleasantness). I studied each recording in depth, noting spikes in engagement with coordinating screens. Peaks in joy surrounded small animations such as the “lesson completed message” while peaks in engagement and confusion were common during the quiz. The quiz required direct thought and choice while the lesson only required a click forward, explaining why engagement appeared to be higher during the quiz section. However, the drag feature in the lesson almost always caused a spike in engagement, breaking the norm of the interaction that had involved only clicking prior. Overall, engagement levels were highly variable. Attention rates ranged between 98-100% of the time spent in the interaction. I believe the data demonstrates the integration of research and design through this prototype as an efficient method of presenting research on the female hormone cycle to young women.

After the ten-minute interaction, participants filled out a written survey to reflect on their experience and provide personalized insight. Survey questions focused on the purpose of this study, asking about the ease of understanding language, trustworthiness of the app, content retention, and learning experience. The diagram to the left visualizes the results.

The free response-based questions provided a space for any additional comments, and asked for suggestions for improvement, what users liked about the app and experience, and for any honest feedback. The written feedback was overall positive, one participant wrote it was “a great experience,” she “loved how easy it was to navigate,” and the “illustrations [...] were so cute and made the learning more engaging.” Several participants commented they hoped the prototype would be developed into a full app one day, saying the content was “information everyone should know and look into” and asking to “please make it real.” I also received several suggestions to further strengthen the app. Recommendations included “more content on holistic ways to approach women’s health,” “a quick guided tutorial regarding app navigation,” and receiving clear “doctor, psychologist, or nutritionist approval” to solidify content trustworthiness. Several other trends in the feedback included praise for the visual aesthetics, options for different media in the featured resources, visibility of credible sources, logical organization of content, and easily understandable language.