Distributed Collaborations and the Effect of Sociometric Feedback

Introduction

Recall the last time you were assigned to a new group. You met your group members for the first time, introduced yourself, and then your group members started to talk about the tasks and roles of the group. But as the conversation developed, content of the conversation was not the only thing that you were focusing on: your senses were intensely observing the signals that are underneath the spoken language. Information such as who is sitting next to whom, who talks the most, who sounds the most confident, who asks questions, and who replies are all information that gave you a sense of how the group works and how you should behave in the group. Now imagine that this first encounter with the group was through a conference call. Your ears and brain had to work even harder to get a sense of the other group members: are they paying attention, are they expecting you to say something, or did you just say something stupid? You put in extra effort but it was much harder for you to get a clear understanding of what was happening and how the group works. You noticed that your behavior is somewhat different from what you normally do when you are in the same room with people: you either talked more or less than usual; you didn’t know when you are supposed to speak; and sometimes you were less confident about what you just said since you couldn’t see the responses of the other party.

As you can recall from your personal experience, distributed collaborations¹ are quite different from co-located collaborations. Researchers in organizational behavior have also confirmed the difference. Hinds and Bailey (2001) have demonstrated that distributed collaborations have very different dynamics compared to co-located collaborations, and that these differences often lead to poorer performance. Rocco (1998) and Bicchieri and Lev-On (2007) found that trust between group members breaks down in computer-mediated communications (CMC). Bouas and Arrow (1995) found that group identity in distributed groups was much lower compared to co-located groups. This makes it extremely difficult for distributed groups to establish and maintain trust and cooperation, which brings down the overall group performance.

Then why are distributed collaborations so different from co-located collaborations? It is surely not the content that is missing, as we have no problem sending and receiving content using voice or text. Then what is the missing piece that makes distributed collaborations so much harder than co-located collaborations? We believe social signals are the answer. Many of the social signal exchanges that provide context to the interactions are lost in distributed collaborations. This makes it harder for group members to understand how the group works, and how they should behave. Additionally, the lack of feedback from one another makes it difficult to reflect on their own behavior. The common use of question marks, exclamation marks, capital letters, and smiley faces are all examples of people trying to augment social signals in text based communication. The advance in communication technology has enabled some level of exchange in social signals. State-of-the-art video conferencing systems allow the exchange of richer visual signals, however there are still many social signals that are lost during the communication. Sense of proximity, back channel conversations, and eye contact are examples of social signals that are still not communicated in CMC.