Long-duration missions depend on flight crews’ abilities to make decisions under highly stressful conditions. However, crew members’ decision-making capabilities are affected by factors such as stress, gender and cultural background, and communication medium. Dr. Judith Orasanu is examining how these factors affect team interaction and performance. She is developing a system to monitor the state of crew interactions via physiological stress/arousal data, emotional responses (via facial affect), and team performance measures.
Overview
Distributed Team Decision Making in Exploration Missions
Principal Investigator:
Judith M. Orasanu, Ph.D.
Organization:
NASA-Ames Research Center
Technical Summary
Since the beginning of human space flight, crucial incidents relating to the psychological, behavioral, and interpersonal aspects of crew performance have jeopardized crew safety and mission success in both U.S. and Russian programs (Kraft, 2003). While crewmembers are highly selected and technically skilled, "the history of space explorations has seen many instances of poor interpersonal relations and faulty decision making" (Committee on Space Biology and Medicine, NRC, 1998). Interpersonal relationships will be of particular concern as crewmembers become more diverse in terms of culture, gender and professional backgrounds (Connors, 1985; Kanas, 2001; Santy, 1993; Kraft, 2003).
Two major goals drive our research. The first is to understand the effects of task and interpersonal stressors on team performance in challenging distributed decision-making situations, and to identify effective strategies for coping with these stressors. The second objective is to develop non-invasive technologies to detect low levels of stress in individual team members to allow the introduction of countermeasures before team dynamics deteriorate to the point of threatening mission success. Findings will provide a basis for designing or revising crew selection, training practices, and procedures to monitor and support team performance.
A simulated task environment was developed to address our objectives. The task, a dynamic computer-based search and rescue mission set in Antarctica, provides an ideal environment for studying team interaction and problem solving because its underlying cognitive demands reflect those of many real-world team tasks. A four-person crew must develop plans, manage resources, and collaborate in order to find a lost party, accomplish its mission (repair a communications antenna), and cope with unexpected demands, such as clearing blocked terrain, rendering medical aid, and helping to repair broken-down vehicles. Graphical displays of the terrain and evolving problem scenarios are presented via linked computers; communication among team members is supported by e-mail and voice communication. Three crewmembers engage in the search using virtual snowcats, while a fourth remains at the base station to coordinate search activities. Two levels of difficulty were created for each of three scenarios by varying the diagnosticity of cues and time pressures. Team cohesiveness was manipulated by task goals: half of the teams were induced to work in a fully cooperative manner; the other half received instructions and feedback designed to induce conflict between team members.
All actions taken during the game and various outcome measures are automatically recorded by the game software. While participants work on the task, their physiological responses and facial expressions are recorded and monitored for signs of mental and emotional stress. Physiological responses, facial affect, and personality measures will be correlated with individual and team behaviors and overall task performance to build models of effective team composition and performance.
Year one of this project was devoted to establishing a distributed team decision making laboratory; the development and pre-testing of the simulated task environment, training materials and six experimental scenarios; and the selection and pre-testing of a physiological monitoring and recording system.
In year two we selected a psychological assessment battery and ran a baseline experiment to assess the adequacy of the problem solving scenarios, the task manipulations, and the stress monitoring technologies during actual team interaction. Twelve teams each consisting of four U.S. males (at least second generation American) served as our baseline group. Team conflict varied between groups and task difficulty varied within teams.
Participants worked together over a period of four days in the laboratory. Day one consisted of training on the computer-based task and an introduction to the physiological monitoring devices. On days two, three and four, participants worked through six scenarios, one easy and one difficult scenario per day. Order was counterbalanced across and within days. Both individual and team performance measures were recorded for analysis. Personality measures were administered between tasks on Day two and three; state and group dynamic measures were taken at the end of each task day.
The manipulation of task difficulty was successful as players were generally more successful in completing tasks in "easy" scenarios than in "difficult" ones. The team conflict manipulation resulted in significant differences between groups, but in the opposite direction than expected. Scores earned by individual players in the Competitive condition were higher than those in the Cooperative condition were. Further analyses revealed that members of Competitive teams collaborated on tasks significantly more often than members of Cooperative teams. The superior performance of individuals in the Competitive condition may arise from several sources: differences in performance feedback provided in the two task conditions, in team strategies associated with each condition, or in cognitive abilities of the individuals randomly assigned to conditions. All three possibilities will be addressed in research during year three.
Team interaction and communication are analyzed from transcripts made from audio- and videotapes of the teams during planning and mission execution. Our initial review of team communication found both effective teamwork and negative affect. Effective task performance was found to depend on team collaboration: a coordinated search strategy, communication about mission-critical information, task assignment and prioritizing, and plan adaptation based on effective strategies were evident in successful teams. Team members established a sense of team cohesion and positive affect by treating each other with respect, by backing each other up, and by providing positive reinforcement for successful efforts and objective feedback on errors. Negative behaviors associated with team failure included players being withdrawn and passive, engaging in competitive behavior and deceit, and instigating conflict through negative feedback, criticism, and public embarrassment or humiliation of others. These behaviors will serve as the foundation for analysis of behaviors in culturally diverse teams in year three of the current grant and, once validated, as a basis for training in our follow-on proposal.
Physiological responses were monitored while players engaged in the search and rescue tasks. The time-coded output were filtered and cleaned prior to analysis; statistical analyses are still underway, but initial observations revealed physiological reactivity at points of task pressures.
Our plan for year three involves replicating the baseline study of year two with multicultural gender-mixed crews. The goal is to determine whether task and team stressors have similar effects on these diverse teams as on homogeneous male teams and whether physiological indicators will be similarly related to performance. Before conducting this study, we will run a second baseline study with a modified task context and performance feedback to ensure a more systematic manipulation of team conflict.
Given that the Antarctic search and rescue task context is inherently cooperative, the manipulation of team conflict by competitive instructions and individual rewards was less effective than anticipated. To induce conflict, thereby enabling the study of stress effects on crews, a new competitive game has been created. While structurally identical to the original search and rescue game, it is inherently competitive as it involves the search for ancient treasures stolen from a famous museum, hidden in Antarctica. Converting this inherently competitive task into a cooperative team task is relatively easy, in contrast to making the inherently cooperative search-and-rescue mission into a competitive one. A second change concerns the performance feedback in the Cooperative condition to include individual as well as team scores. These players will receive the same information regarding the success of their strategies as players in the Competitive condition. A third change involves the inclusion of a cognitive abilities test to control for possible inter-individual differences among study participants.
U.S. males (at least second-generation American) will participate in the new baseline study. Six teams will be assigned to each condition (Cooperative and Competitive). Physiological, biometric, personality, and group dynamic measures will be collected, as in the first baseline study.
The second study will involve culturally diverse and gender-mixed teams. Participants will be sought who are from countries involved in the International Space Station. College graduates (25-45 yrs old) whose English language skills are adequate to read instructions and communicate with team members will be invited to participate. Non-US participants must have been born in their native countries and have lived in the U.S. for less than two years. Based on the concept of cultural collectivism/individualism (Hofstede, 1984), we hypothesize that certain team compositions will lead to more effective performance than others will. Twenty-four teams (involving male and female players) will be included, half in the Cooperative condition and half in the Competitive condition. These groups will be compared with the U.S. male baseline group.
This project addresses Risk #18 under Human Behavior and Performance: Human performance failure due to poor psychosocial adaptation. This study will determine the extent to which task and interpersonal stressors affect performance on tasks that require team collaboration for successful performance. Moreover, the study will examine the impact of personality factors and cooperative behaviors, trust, and team cohesion as mitigating factors that can overcome the effects of situational stressors. Tools for assessing stress at both individual and team levels wills serve as a foundation for introducing countermeasures.
Implications for future earth-based research: The study provides integrated physiological, self-report, behavioral, and performance measures from teams working together over a period of time on a common task. Models based on replication with variations in team and task structures will significantly advance our theories of team performance under stress and provide a basis for training in diverse earth- and space-based environments.
This project's funding ended in 2004