Rapport
You know that harmony, fluidity, synchrony, flow one feels when engaged in a good conversation with someone? Known formally as rapport, these features are prototypical characteristics of many successful interactions. Speakers seem tightly enmeshed in something like a dance. They rapidly detect and respond to each other’s movements. Tickle-Degnen and Rosenthal equate rapport with behaviors indicating positive emotions (e.g. head nods or smiles), mutual attentiveness (e.g. mutual gaze), and coordination (e.g. postural mimicry or synchronized movements). Numerous studies have demonstrated that, when established, rapport facilitates a wide range of social interactions including negotiations, management, psychotherapy, teaching and caregiving.
The Rapport Project explores the potential of virtual humans to establish rapport with humans through simple contingent nonverbal behavior. Such a system can generate positive feedback (e.g., nods) by recognizing and responding to vocal or behavioral cues of a human speaker. Though a series of controlled experiments, we have demonstrated that such simple contingent behaviors can make agents more engaging, promote fluent speech and reduce user anxiety. These effects can be subtle; many studies indicate the benefits of such feedback fall outside of conscious awareness in that people often show measurable impacts on their observable behavior without reporting significant differences when introspecting upon their experience.
Current research involves the use virtual humans to deconstruct the and then simulate the factors underlying rapport. This project seeks to deepen our understanding of those factors such as entrainment (the phenomenon whereby people adapt to each others’ behavior through the course of an interaction), cultural variability in nonverbal feedback, and the link between feedback and social anxiety.
Our research has potential to significantly enhance basic understanding of human-centered computing. It can inform our understanding of the critical factors in designing effective computer-mediated human-human interaction under a variety of constraints, (e.g., video conferencing, collaboration across high vs. low bandwidth networks, etc.) by helping to identify crucial factors that impact social impressions and effective interaction. It facilitates the development of a novel method to support and enhance social interaction, which may promote the establishment of rapport and trust in distributed environments. As rapport is argued to underlie success in negotiations and conflict resolution, improving worker compliance, psychotherapeutic effectiveness, improved test performance in classrooms and improved quality of childcare, our work in this area could beneficially impact effective learning and problem solving in wide variety of domains.
This project differs from others in the following ways:
Rapport (and related concepts) have been extensively studied in social psychology but using autonomous agents to study and reproduce the phenomenon is a novel innovation. Some related wok (e.g. Bailenson) has explored the use of avatars (virtual characters who’s movements are determined by the behavior of a human listener)
Tags: behavior, conversation, interaction, negotiation
Become a fan.
Follow us.
Read our blog.
Watch our videos.