Within the area of Computer-Supported Cooperative Work (CSCW), there has been an explosion of interest in how recently developed network technologies might be applied to support the collaborative endeavors of widely distributed participants. Increasingly powerful systems for desktop conferencing, group meeting, and distributed design have been developed. Though the technologies applied in such systems vary widely, their underlying design goal is essentially the same: to support interactions that are functionally equivalent to face-to-face interaction.

This dissertation evaluates the extent to which currently available technologies achieve this goal by comparing the amount of communicative breakdown experienced by pairs of participants interacting in three communication environments: copresent, audio­mediated and audio/video-mediated. In all three environments, participants had access to a shared workspace, in which they used a graphical computer simulation to collaboratively explore the behavior of a simple cardiovascular system.

Videotaped interactions were analyzed in a series of three studies, intertwining the qualitative techniques of Conversation and Interaction Analysis with more traditional quantitative techniques to progressively refine understanding of the functional differences that exist between environments. Four categories of communicative breakdown were identified: failure to maintain shared conceptions of current topic, failure to establish shared reference, and failure to regulate access to the verbal channel and to a shared cursor.

Statistical results showed that copresent interactions were significantly less prone to breakdown than interactions in either of the two technologically-mediated environments; no significant differences in the incidence of breakdown were found between audio-only and audio-video interactions. A subsequent qualitative analysis showed that breakdowns in technologically-mediated interactions were related to a profound insensitivity to nonverbal displays like direction of gaze, deictic gesture and manipulation of objects in the task context. This result demonstrates that, though visual access to a partner is clearly vital for avoiding breakdown, the visual access afforded by a video image is fundamentally unequal to that afforded by physical copresence. More generally, there is a great deal of difference between technically making more communicative resources available in an environment and the practical utility of such upgrades to participants.