Team cognition can be observed in the flow of communications among team members. This is shown in the context of a simulated unmanned aerial vehicle ground control station. Automatic measures of low-level team communication flow were used to assess high-level constructs of team cognition. Measures show support for the expected results of manipulations in this task. Co-location and channel degradation effects were successfully predicted by CHUMS, ProNet, and a cross-correlation function-based Dominance measure. Results grant concurrent validity to the measures, and highlight substantive effects of the manipulations. In particular, in geographically distributed teams, communication patterns are less stable, and the route planner exerts less communicative influence. Some co-location effects drop with task experience. During a mission containing a five-minute one-way communication channel cut, all teams communicate more like distributed teams, and team members do create alternate pathways to retain information flow.
[1]
J. Walther.
Computer-Mediated Communication
,
1996
.
[2]
D. Budescu.
Tests of lagged dominance in sequential dyadic interaction
,
1984
.
[3]
Michael J. Smith,et al.
Usability evaluation and interface design : cognitive engineering, intelligent agents and virtual reality
,
2001
.
[4]
Peter W. Foltz,et al.
Evaluation of Latent Semantic Analysis-Based Measures of Team Communications Content
,
2003
.
[5]
Nancy J. Cooke,et al.
Procedural Network Representations of Sequential Data
,
1996,
Hum. Comput. Interact..
[6]
Peter W. Foltz,et al.
Some Promising Results of Communication-Based Automatic Measures of Team Cognition
,
2002
.
[7]
Nancy J. Cooke,et al.
Automating Measurement of Team Cognition through Analysis of Communication Data
,
2001
.