While family informatics has been developed for monitoring and tracking family-centered health data, there remains a gap in understanding how family informatics can support families in reflecting on their social behaviors and emotional dynamics. We address this gap with SELaD, a system that captures and visualizes social-emotional data from daily family interactions using audio, video, and physiological sensors. In a semi-naturalistic study with 17 families ($n=51$), we investigated how this data facilitates reflection. Our findings reveal a process we term \emph{relational reflection}, where families collaboratively interpret multimodal data to deepen their understanding of conversational dynamics and emotional influences by recalling their shared history and expectation of good communication. This process was particularly enriched by emotional data from multiple sources that families could cross-reference and reconcile. This work presents SELaD as a technology probe and empirically grounds the concept of relational reflection, positioning it as a foundation for designing future reflective technologies.
ACM CHI Conference on Human Factors in Computing Systems