Measurement or induction? Methodological challenges in studying empathy and compassion with technology-assisted tools

Empathy and compassion are central to prosocial functioning, yet their measurement and experimental induction remain methodologically challenging. Traditional self-report approaches are limited by retrospective bias and social desirability. In this narrative review, we examine emerging technology-assisted methods—including video-based paradigms, virtual reality, and artificial intelligence—as tools for assessing and inducing empathic and compassionate responses. We propose a key methodological distinction between approaches designed for measurement or induction and those that simultaneously measure and induce socio-affective states. This distinction may support study design in HCI, particularly when technology serves as both a measurement and an induction tool. While ecologically valid paradigms such as the Socio-affective Video Task (SoVT) enable the capture of dynamic, state-sensitive responses, they also introduce interpretive challenges, particularly in distinguishing compassion from empathy and in linking emotional responses to adaptive outcomes. Overall, we argue that no single method adequately captures the complexity of empathy and compassion. Future research should adopt integrative, multimethod approaches that combine subjective, behavioral, and computational measures, while explicitly addressing the interplay between measurement and induction. This perspective provides a framework for advancing more precise, ecologically valid, and theoretically grounded research in HCI, Human-Centered Computing, and socio-affective science.