Vividness and source of evaluation as determinants of social responses toward mediated representations of agency

Computer interfaces increasingly make use of characters or agents to represent human activity in the context of computer-mediated interactions. These representations may take on human-like physical forms, and may also exhibit human-like behaviors. Such similarities to human appearance and activity are important determinants of the extent to which users interact with these representations as social actors, that is, the extent to which they use social rules that are common in human-human interaction in situations involving humans and computers. In this laboratory experiment (N = 80) subjects used two computers as part of a tutoring/testing/evaluation procedure. Independent variables included the modality in which the tutoring and evaluation sessions were presented and the source of the evaluation in a 4 (modality: text only, audio with some text, audio with still images, real-time video) by 2 (source of evaluation: recording of same person that tutored subjects (self-evaluation)/recording of different person than tutored subject (other-evaluation)) between-subjects design. Dependent measures included a paper-and-pencil questionnaire measuring subject affect toward and performance judgments about the tutoring and evaluation sessions. Results replicate the finding that subjects do apply social rules derived from interactions with other humans to computers. Results also suggest that perceived socialness of the computer varies according to modality, with text and real-time video being the most well liked, most natural, and most social, followed by voice-only presentation and finally still images with sounds.