Suspending and reinstating collaborative activities

There is an interest in how people coordinate multiple activities with multiple partners, and in particular, how people deal with interruptions. Interruptions happen unexpectedly. They can cause errors and loss of productivity. Effects of interruptions have been studied in individual tasks, but not in collaborative tasks. When pairs deal with an interruption, they have to jointly suspend their activity, address the matter, and later jointly reinstate their activity. The main goal of the dissertation was to define a model of suspensions and to measure constrains on the suspension and the reinstatement steps. Pairs suspend momentarily without taking leave of each other. This requires the coordination of two important processes: Politeness and common ground. First, politeness is often involved in suspensions, because asking one's partner to wait while one does something else is facethreatening. Two factors affect politeness: The degree of personal responsibility of participants proposing suspensions and durations of suspensions. Second, continuing tasks requires reconstructing joint representations of the tasks (common ground). Several factors affect the reconstruction of common ground: The persons interrupted the timing of suspension and the availability of cues about the state of the task. Five studies were conducted. Study 1 used naturally-occurring suspensions in telephone conversations from a corpus data. Study 2 & 3 used suspensions triggered with a cover story in laboratory. Study 2 & 3 manipulated participants’ roles in conversations and durations of interruptions. Results revealed that participants were more polite when suspension lasted longer, and it took more collaborative effort to reinstate conversations. Also, initiators of suspensions were more polite when they were listening than speaking. Study 4 & 5 manipulated the duration and the timing of interruptions during a goal-oriented task. Additionally, Study 4 manipulated the participants’ role during interruptions, and Study 5 manipulated the visibility of workspace between participants. Results showed that participants took more time to reinstate tasks when interruptions lasted longer and when it happened in the middle of sub-tasks, compared to when interruptions were brief and happened between sub-tasks. Also, they took more time to reinstate when both participants were distracted during interruptions, rather than when just one was distracted, and when participants did not share their workspace, rather than when they did share a workspace.

Thèse de doctorat : Université de Neuchâtel, 2010 ; Th. 2133