Toward a User Emotional State Taxonomy for Human-Computer Interaction

Emotions play a critical role in shaping user experiences with interactive technologies, yet the HCI field has lacked a comprehensive, validated framework that captures the full spectrum of user affective states. This study addresses that gap by developing a PAD (Pleasure–Arousal–Dominance) model–based taxonomy specifically designed for human–computer interaction. Using a rigorous multi-stage methodology, the research synthesized findings from a systematic mapping of over 100 user emotions in the HCI literature, identifying 21 core emotional states based on frequency, relevance, and generalizability. A Delphi expert review refined the set, while each emotion was assigned PAD dimensional values ranging from –1 to 1. User studies in healthcare and gaming contexts validated the taxonomy’s ability to represent emotional trajectories, with new transitions informing iterative improvements. By integrating data-driven synthesis, expert consensus, and user-centered evaluation, this research delivers an evidence-based taxonomy to guide the engineering of emotionally aware interactive systems.