C3Fire is a highly configurative system. The researchers configure the system to meet their research goal. The appearance and function of C3Fire differ between projects.

Bridging Cultural Barriers

GIS effects on
command and control


Network Warfare

Bridging cultural barriers to collaborative decision making in On Site Operations Coordination Centers

This research program's goal is set to contribute to the training of the UNDAC (United Nations Disaster Assessment and Coordination) team by identifying clusters of expectations and behaviours that vary systematically across cultures and have the potential to raise barriers to collaboration and decision-making.

The UNDAC team is a stand-by team of disaster management professionals and can be mobilized within hours of receiving a request from a country affected by an emergency or disaster. When invited, the UNDAC’s immediate mission is to carry out a rapid assessment of priorities and needs and to support the local authorities and the United Nations Resident Coordinator to coordinate international relief on-site so that their work is not a burden on the affected country’s resources.

The UNDAC team are responsible for ensuring that the humanitarian relief provided is effective and well coordinated.
Coordination of international relief teams is not an easy task. The teams that arrive at the affected area talk different languages, have different backgrounds and training, and bring differing numbers of people and types of resources to the site. In fact, coordination is so difficult that UNDAC has created a coordination concept called the OSOCC (On Site Operations Coordination Center). The OSOCC is the actual physical location where the UNDAC team does its work.

UI Examples

The relationship between the UNDAC team and the LEMA (Local Emergency Management Authority) can strongly constrain the activities of people working in the OSOCC. Because OSOCC personnel play a supporting role, their interactions with the LEMA must conform to local social conventions. Like all international personnel, the OSOCC personnel need to be sensitive and to follow local norms. Differences in norms for behaviour, collaboration, and decision-making can pose a challenge to the efficient coordination of humanitarian relief activities.

The inherently multicultural composition of an OSOCC makes it a natural site for cross-cultural disturbance. Cultural differences may pose barriers to collaboration, decision-making, and performance in an OSOCC.
The directive of this research is to identify clusters of expectations and behaviours that vary systematically across cultures and have the potential to raise barriers to collaboration and decision-making. Awareness of systematically cross-cultural expectations and behaviors grant a better ability to work with and around them.

Experiment design

Experiment Constraints
As a Joint Cognitive System, OSOCC operations place strict constraints on the conduct of the experiment. OSOCC teams are formed ad-hoc and on-site. Team members may or may not know each other. Because there is no time for team-building, they get to know each other as they work. As they get to know each other, their way of working together is likely to evolve.

These considerations led us to design the experiment to meet three sets of constraints:
- We need to elicit and capture spontaneous but collaborative emergency-services decision making in response to a simulated emergency.
Our experimental platform is the C3Fire micro world simulation of an OSOCC-like setting.
- We need to emulate the ad-hoc nature of OSOCC team formation and to capture the actions, words, expectations, and norms for decision making that guide the development of teamwork in the simulated OSOCC.
We use the C3Fire network of computers to form ad-hoc teams of four individuals (either Bosnians, Indians, Iranians, or Swedes).
- We need to gather individual self-report information about values, personality, and preferences.
Each participant completed a series of six questionnaires including the NEO-FFI and the 57-item Schwartz value scale.

Experiment Procedure
The participants reported to the laboratory in ethnic groups (either Bosnians, Indians, Iranians, or Swedes) of eight. In the laboratory, the eight were randomly and anonymously assigned to two teams of four decision makers. The purpose of the random and anonymous assignment to teams was to minimize reputation effects and to emulate the ad hoc nature of OSOCC team formation.

The teams’ task is to work together to control and extinguish a series of four simulated forest fires. This takes half a day and is repeated twice. Altogether eight series of simulated forest fires are done.

The structure of the experiment was a cycle of three sets of activities corresponding to the three constraints.
- The first activity in each cycle is team play using the C3Fire micro world.
- The second activity is team discussion of their play.
- The third activity is individual responses to a series of questionnaires including the Schwartz value survey.

The two series of four sessions followed the same pattern.
After the first session, the entire group of eight participants, the two teams, gathered together to engage in a group discussion. Then the participants were randomly and anonymously assigned to new teams of four.
After the second session, the newly formed team of four gathered together in a team discussion. At this point they were introduced to each other. The team that was not engaged in an after action review session filled out a questionnaire.
The teams formed worked together again in the third and fourth trials. After action reviews and questionnaires were held after the third and fourth trials as well.

First activity, C3Fire micro world

The C3Fire sessions were conducted over a server-client network in a laboratory with two separate rooms.
Each of the eight participants sat at a client computer and was linked to his teammates by the C3Fire software. Their only mode of communication was the email system provided by C3Fire.

The two teams of four were connected to different server computers. The two servers independently ran the same C3Fire scenario concurrently. Each trial lasted until the participants completely suppressed the fire or 20 minutes elapsed, whichever came first. The teams encountered a total of eight different scenarios.

All event data and email communication during the trials are stored in C3Fire for later analysis.

The group discussion and the team discussions were held in the server room. The sever room was provided with audio and video equipment. And all discussions where recorded.

The questionnaires where filled out in the computer room individually by the team that, at the moment, was not having team discussion.

The user interface of C3Fire.

The speed of burning and spreading of the fire are functions of vegetation, terrain, the presence of buildings, wind direction and speed, and are pre-set by the experiment manager.
Participants can extinguish the fire by directing fire trucks. The trucks are constrained by limits on the rates with which they drive, deploy, and fight the fire. The experimenter sets these limits.
C3Fire include four classes of fire-fighting resources. Two of the classes of resources are water and fuel trucks. Water trucks supply fire trucks with water, and obtain water from water stations. Fuel trucks supply both fire trucks and water trucks with fuel and obtain fuel from fuel stations. Both water trucks and fuel trucks are fully mobile but their tanks have a limited capacity.
The third and fourth classes are water and fuel stations. Stations cannot move but have an unlimited supply.
The addition of water and fuel trucks and stations makes the logistics of the simulation emulate some of the complications experienced by members of an OSOCC.

Interdependencies among decision makers arise whenever different classes of fire-fighting units are assigned to different participants in the simulated OSOCC. For example, the locations and activities of water trucks and fuel trucks constrain the actions of fire trucks. If different people have control over these different resources, their actions are mutually constraining. This provides ample opportunity for conflicts to arise.

In this first activity of the cycle, teams of four worked together across a computer network to coordinate emergency services operations posed by the C3Fire micro world. C3Fire recorded the actions they took and their email communication. The recordings of the teams’ actions are the primary source for analyses of team behaviour. Transcripts of the emails are one of the primary sources of data about patterns and content of team communication. The unit of analysis for these data is the team.

Second activity, Team discussion

One of the salient social characteristics of an OSOCC is that its members may not know each other when they arrive on site. They get to know each other as they work. There is no straightforward way to capture this emergence from anonymity in an experimental setting in which participants may indeed know each other.
We crossed this hurdle by bringing eight participants into the laboratory at once and splitting them, randomly and anonymously, into two groups of four. Initially, no one on a team knew who the other three team members were. As they day advanced, they became better acquainted and had numerous opportunities to interact as a team.
Specifically, team membership is randomly assigned and unknown during the first C3Fire trial of each day. There was no opportunity for consultation before the trial. As in a new OSOCC, the team had little common ground (other than their shared ethnicity) but had to attack the emergency immediately. This situation was repeated for the second C3Fire trial of each day. Team membership was randomly shuffled and, once again, there was no opportunity for consultation. Participants found themselves in a new OSOCC facing a new emergency.

We, as experimenters, did not establish an organizational structure for communication and control that the teams were to follow. E-mail communication and truck control were not constrained by the experimenters or the C3Fire software. As a result, all participants could communicate with all other participants and command all trucks and override commands made by other participants. In short, all structure was left to the teams.

After the first session of each day, the entire group of eight participants gathered together to engage in a semi-structured interview about C3Fire and what they did in the first trial.
The conversations focused on technical issues related to the C3Fire interface and on strategies for fire-fighting and logistical support. Then the participants were randomly and anonymously assigned to new teams of four.
After the second session of each day, the newly formed teams of four gathered together to watch and discuss a replay of their second C3Fire trial. At this point they were introduced to each other. This conversation gave them the opportunity to get to know each other.
The teams formed worked together again in the third and fourth trials. After action reviews were held after the third and fourth trials as well. These sessions provided us with opportunities to observe the emergence of camaraderie and assess whether team behaviour changed as a function of familiarity.
The after action reviews took place in the laboratory immediately following second, third and fourth of the C3Fire trials. The team members gathered around the server, which replayed the recent trial on its monitor. After the experimenter asked a few warm-up questions, the participants were allowed to discuss the trial in any way they chose. Their discussion was recorded using a digital video camera and a pair of microphones.

In this second activity, the team sat around a computer monitor, watched a replay of their C3Fire session, and engaged in an open-ended conversation about their play. During these after action reviews, most teams developed an organizational structure, allocated responsibilities, and debated alternative strategies for dealing with the emergencies posed by C3Fire. Their conversation was recorded using both a video camera and audio equipment for subsequent qualitative and quantitative analyses. Once again, the unit of analysis of these data is the team.

Third activity, Individual responses

The team that was not engaged in a team discussion after a trial filled out individual questionnaires. Totally they completed six questionnaires. Sixty-three male participants who identify with four different national cultures, ages 19-56, have participated in our experiment. Thirty-two of the participants identify themselves as Swedish, eight as Bosnian, eight as Iranian Swedes, and 15 as Indian.

The demographic inventory is a 16-item self-report questionnaire that assess the participants personal, academic and work related background, international experiences, estimated language skills, experiences of military, emergency service or police work, and experiences of using computers.

The NEO-FFI is designed to measure the ‘Big Five’, five domains of adult personality: extraversion, agreeableness, conscientiousness, emotional stability, and intellect/openness. It will give us valuable insights in the participants’ personalities and will facilitate differentiation between individual and cultural differences.

We have created a conflict avoidance instrument consisting of 23 items that assesses a person’s disposition to react to conflict.
In some cultures, the wishes and welfare of the collective is more important than the wishes and welfare of the individual. Collectivist cultures seek harmony. In individualistic cultures, the focus is the welfare of the individual. There is likely to be more conflict between individuals in an individualistic culture than a collectivist culture.

Tolerance for Uncertainty People react differently to situations in which they feel unfamiliar or without complete information. They have different tolerance for uncertainty. We designed a 24-item instrument that assesses the degree of comfort in making decisions with incomplete information and in unfamiliar situations.

People from different cultures can differ in how far ahead they plan, their time horizon, and how much control they feel they have over their own lives. On the one extreme is the individual who believes he has full control over his life and plans far ahead. On the other end is the individual who believes that he is in the hands of fate/God/Allah/etc., who thinks that it is impossible to know what tomorrow brings, and who sees no point in planning ahead. Imagine the barriers to communication and planning when these two are forced to work together. We expect their norms for planning and decision-making will collide dramatically.

Finally the Schwartz Value Survey will provide us with a measure of what values our participants find important and gives us the opportunity to test if these values are considered unequally important across our national groups.

In this third activity, members of the team individually filled out a series of six questionnaires. The questionnaires were drawn from the literatures on social, clinical, and cultural psychology. The unit of analysis of these data is the individual.