Abstract
Today, we are surrounded by devices that collect data about us. Think of your smartphone tracking screen time, phone pickups, and app usage. Or consider your smartwatch measuring your resting heart rate, stress levels, and sleep score. In academic literature, these are called Personal Informatics (PI) systems. These support people
... read more
in gathering personally relevant information for self-reflection and self-knowledge.
Self-reflection offers benefits like enhanced self-insight, motivation for behavior change, and support for life changes. However, reflection often does not happen automatically and must be encouraged. Some people are more inclined to reflect, and it develops with age or training. Therefore, how users can be supported in reflection should be carefully considered when designing such systems.
The benefits of reflection are widely recognized in Human-Computer Interaction (HCI) research. Many studies have been conducted to understand how technology can promote reflection. However, at the start of my PhD research, questions remained: What strategies and techniques can we use to encourage reflection? How do people reflect on the data they collect in their daily lives? How can we measure technology's influence on reflection? These questions guided my PhD research.
As a starting point, we needed an overview of strategies to encourage reflection. Many prototypes have been developed to support users in reflecting on their health, productivity, and work-life balance, using techniques like data visualizations, questions, and leaderboards. Through a systematic literature review, we created an overview of all these strategies, resulting in a taxonomy with 11 strategies and 74 techniques for reflection.
Next, we need to understand how PI users reflect on their data in daily life. Interviews with fitness tracker users gave insight into how they reflect on their health and well-being. Based on these interviews, we developed a model describing how a system can create a temporary state of reflection. Lacking an instrument to evaluate systems, we developed a validated scale to measure whether technology supports reflection. This model and scale provide a theoretical foundation for future research into reflection in PI systems.
In the final part, we explored the impact of metrics on reflection. The earlier studies showed that reflection is promoted only if the data collected is relevant to the user. To investigate this further, we conducted two additional studies. The first examined how a metric's abstraction level affects reflection using a fictional "health score," leading to three design guidelines: metrics should form a coherent story, help users connect data, and explain how they're measured. The second study explored how data physicalization (making data tangible) can promote reflection on complex metrics like blood pressure. We found that tangible data can encourage reflection but requires guidance.
In conclusion, this thesis contributes to the understanding of technology-supported reflection through a structured literature review, a taxonomy of strategies, a user reflection model, a validated scale, guidelines for designing reflective metrics, and an exploration of the use of tangible data to promote reflection. As such, this thesis lays the foundation for developing technology that supports data-driven reflection.
show less
