H. Artman and R. Granlund. Team situation awareness using graphical or textual databases in dynamic decision making. In Proc. 9th Conference on Cognitive ergonomics, pages 149--155, 1998. This paper reports on a experiment about how people share information within a team when controlling a dynamic system (a simulated fire : geographically large, require cooperation between several actors). Four people worked together on a joint task in a layered organisation :
- layer 1 is operative : two fire chiefs who are commanding two fire units each - layer 2 is supervisory: two persons that are to co-ordinate the fire chiefs. receives all information from the fire chiefs but has to construct an overall picture (situation awareness) of the area.
In order to conduct this experiment, they used a microworld (c3) that is model of the real task. There was two conditions :
- Textual : the operators share a textual database where they can and should register the incidents and what they have done about it. This database is accessible to all operators, and constitutes a kind of collective memory of the units actions. - Graphical : a Geographic Information System (GIS), which visualizes the ambulances location in the area for which the team is responsible.
General hypothesis : the team using the graphical database will have more a higher correspondence of SA with the state of the world AND hence will be more effective.
Procedure : Seventy-two undergraduate students, The task is to extinguish this fire with four fire fighting units. The problem is that each person in a fire fighting unit only can see a limited window, 3 X 3 squares. The whole area is 20 X 20 squares. Two supervisories (emergency co-ordination centre operators) that control two operatives (control all four fire engines as ordered by their superiors). They should try to save the houses as well as extinguish the fire.
They were given a survey about how they experienced the fire, the co-ordination between team members, some questions related to situation assessment as well as to the interface (4-graded Likert scale).
Results : - no significant performance differences between conditions. - participants in the graphical condition did better mapping the map with the real world, at least in the first session. - learning effect, especially for the textual condition - Team SA consist more of awareness of co-ordination than perception of system dynamics. About Mutual Modeling : the subjects think that they need to know what others are doing. The more successful teams are more likely to do not think it is as necessary as the less successful teams. The more successful teams think they know what others are doing without others telling them than is the case with less successful teams. Thus, successful teams seems to care less about what others are doing, or / and can infer this from other information. It could be that successful teams rely more on silent co-ordination.
Two key concepts :