Designing social interactions with animated avatars and speech output for Product Recommendation Agents in electronic commerce

• Main Author: Qiu, Lingyun
• Language: English
• Published: 2010
• Online Access: http://hdl.handle.net/2429/18488

Product Recommendation Agents (PRAs) and other web-based decision aids are deployed extensively by online vendors, to provide virtual advising services to their customers. While the design of PRA functionality has received increasing amount of attention in academic studies, the social aspects of human-PRA interactions are comparatively less studied. This dissertation investigates the potential of enhancing users’ social experiences with PRAs by developing and analyzing an anthropomorphic interface, which has humanoid embodiment and voice output. This dissertation first investigates the importance of choosing appropriate demographic embodiments for a humanoid PRA. The two demographic variables that have been assessed are ethnicity and gender. As suggested by similarity-attraction theories and social-identity theories, results of a laboratory experiment have revealed that users apply similar social stereotypes in human-human communications as they apply to evaluate humanoid agents. PRAs that match the ethnicity of users are perceived by the users as more sociable, more competent, and more enjoyable to interact with than PRAs that do not match users’ ethnicity; as well, same-gender PRAs are perceived as more competent and more honest than opposite-gender agents. In addition, the "match-up" effects of ethnicity appear to be more significant among female users than among males. Two interface components are also empirically investigated in this dissertation: (1) presence of a humanoid embodiment and (2) output modalities (text, computer-synthesized voice, or human voice). Results from a laboratory experiment demonstrate that humanoid embodiments increase consumers’ perception of a PRA’s social presence, their beliefs in its competence, and the enjoyment they derive from interaction with the PRA. A human voice also appears to be significantly more effective than on-screen text and computer-synthesized voice in improving the PRA’s perceived social presence and enjoyment. Furthermore, the important role of social relationships in influencing user adoption of agents is tested by integrating social presence, trust , and perceived enjoyment with the Technology Acceptance Model (TAM). Social presence appears to be a common antecedent of both trust and perceived enjoyment. Trust exerts a direct impact on user intentions to adopt PRAs, as well as an indirect impact via user perceptions of PRAs’ usefulness. Perceived enjoyment also influences adoption intentions through perceived usefulness and perceived ease of use. === Business, Sauder School of === Graduate