Distributed Team Decision Making for Exploration Missions (NSBRI

Assessing Team Interaction for LongDuration Space Missions Judith Orasanu NASA-Ames Research Center Ute Fischer Georgia Institute of Technology Acknowledgement: This research is supported by NASA cooperative agreement #NCC 9-58 with National Space Biomedical Research Institute. 1 The Team Norbert Kraft, M.D., SJSU/NASA Ames Yuri Tada, Ph.D., SJSU/NASA Ames Lori McDonnell, MA, SJSU/NASA Ames Roberta Bernhard, MA, SJSU/NASA Ames Ron C. Miller, QSS/NASA Ames Heidi Binder, SJSU/NASA Ames Consultants – Sybil Carrere, Ph.D., University of Washington – David Matsumoto, Ph.D., UCSF Contractor: Aptima, Inc. 01/13/04 Orasanu.CKM 2 Rationale for Project While crewmembers are highly selected, technically skilled, and well trained, “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). 01/13/04 Orasanu.CKM 3 Overview Naturalistic Decision Making – What is naturalistic decision making? Team processes – – – – Taskwork and Teamwork Effective Team Decision Making Role of Communication Stress Effects NSBRI study: Distributed Team DM 01/13/04 Orasanu.CKM 4 Theoretical Framework Naturalistic Decision Making (NDM) Decisions embedded in task context Situation assessment a part of decision making Grounded in decision maker’s knowledge Dynamic, time-pressured situations Incomplete information Multiple, frequently conflicting goals Team decision making 01/13/04 Orasanu.CKM 5 Team Decision Making Two heads better than one – More information and knowledge – Metacognition: Questions, checks – Distribute tasks But – Process cost – Social loafing – Interpersonal tensions 01/13/04 Orasanu.CKM 6 Team Processes Taskwork – Shared models for TASK, GOALS, STRATEGIES, TEAM MEMBERS’ ROLES Teamwork – – – – Getting along Team cohesion, positive affect Trust and respect Leadership / Followership 01/13/04 Orasanu.CKM 7 Role of Communication Taskwork – Build shared understanding of task context – Coordinate individual actions – Mitigate errors Teamwork – Define and affirm relationships – Metacommunicative 01/13/04 Orasanu.CKM 8 Effective Teams Decision Making Thorough situation assessment Sensitive to task and time constraints Plan ahead Good task and resource management Communication Relevant information shared Others’ needs anticipated Normative communication patterns Crew-oriented error mitigation strategies 01/13/04 Orasanu.CKM 9 Stress and Team DM Decision Making Reliance on insufficient information Oversimplification and inflexible strategy use Reactive, run out of time Poor leadership Communication Relevant information not shared Imprecise language use and miscommunication Violation of normative communication patterns Verbal aggressiveness 01/13/04 Orasanu.CKM 10 Distributed Team Decision Making for Long-Duration Missions (NSBRI) Research – Stress and team decision making – Gender and cultural factors – Effective team interaction strategies – Ideal team composition Technologies - Monitor physiological arousal levels - Monitor task performance & team interaction patterns - Develop countermeasures - Model/Predict team dysfunction 01/13/04 Orasanu.CKM 11 Research Questions How are crew interaction and decision making affected by – Task and team stressors? – Team members’ gender and cultural background? What are markers of emotional stress? – Physiological / facial affect – Similar across gender and national culture? What team interaction strategies are associated with effective performance? – In mixed culture and gender teams? Is there an ideal team composition? 01/13/04 Orasanu.CKM 12 Method Distributed Dynamic Decision Simulation – Antarctic search and rescue mission – 3 search crew members; 1 base camp coordinator – 6 computer-based scenarios over 3 days Cognitive Team Processes – – – – – 01/13/04 Plan search strategy Share information Coordinate actions Manage limited resources Complete task within time window Orasanu.CKM 13 Screen View of Scenario 01/13/04 Orasanu.CKM 14 Crew Member’s Workstation 01/13/04 Orasanu.CKM 15 Design Manipulate Sources and Levels of Stressors – Task difficulty (within-crew variable) • Time pressure; ambiguity & reliability of information; task load – Interpersonal stress (between-crew variable) • Competitive vs. Cooperative goals Manipulate Crew Structure – Homogenous versus heterogeneous • Culture: Russian, Japanese, European, U.S. • Gender 01/13/04 Orasanu.CKM 16 Hypotheses Effects of task- and team stressors – Broad performance decrements re. information search and integration, strategy adaptation, decision quality – Disruptions in team interaction • Less information shared, more miscommunications • Irritation, increased team tension and conflict Effects of team composition – Heterogeneous teams less successful than homogeneous teams – More misunderstandings due to different communication styles (direct versus indirect) – Less positive team climate and cohesion due to differences in competitiveness 01/13/04 Orasanu.CKM 17 Arousal Effects on Team Performance Physiological arousal factors + facial expressions = Lead indicators of micro-social processes (Gottman and Carrere) – Predict breakdown of dyad in future – Different predictors for males and females Facial affect coding – Based on Emotional Expressive Behavior Coding System (J. Gross) – Irritated, Worried, Relieved, Happy, Puzzled, Surprised, Tense, Bored 01/13/04 Orasanu.CKM 18 Dependent Measures Affective measures – Physiological responses – Facial expressions – Communication content Task performance and process measures – Search and rescue task outcomes (individual/team level) – Problem solving/decision making – Crew interaction strategies Correlations and Sequential Analyses – Psychological measures, affective measures and crew performance 01/13/04 Orasanu.CKM 19 Psychological Measures Trait measures – Neo FFI – Toronto Alexithymia Scale (TAS) State measures – NASA TLX (workload) Group climate measures – – – – – Daily feedback Survey of Group Development Behavior Frequency Observation Guide Impressions of Other’s Behavior Level of Acceptance Questionnaire 01/13/04 Orasanu.CKM 20 Physiological Monitoring ECG: Resp: EMG: PPG: SCL: Electrocardiogram Respiration Electromyogram Pulse plethysmograph Skin conductance level 01/13/04 Orasanu.CKM 21 Performance and Process Measures Search and Rescue Task – Mission completion; time to completion; points accumulated; resources used; distance traveled; number of seismic monitors processed; number of coordinated actions Communications during Planning and Task – Number of contributions per crew member – Structure and function of communication – Relational aspect (both positive and negative) Decisions – Task prioritization – Search strategies 01/13/04 Orasanu.CKM 23 Analytic Approach Affective measure to event – Relate significant change in physiological response / facial affect to task context – Example: increased heart rate occurred when crew member’s vehicle ran out of fuel Event to affect – Relate stressful events to physiological and facial responses – Example: at end of game crew members’ heart rates, respiration, EMG and SCL levels increased as they were trying to complete final task Communication to affective and performance measures 01/13/04 Orasanu.CKM 24 Successful Communication G: So we found two of the waypoints? Performance Assess R: Three PA; Answer G: There’s one at 3,26. . .another’s at 5,21 and the last one’s atPA; Elaboration R: 5,15 PA; Completion G: 5, 15, Acknowledge and Red, where do you think they ended up? Projection; New R: I think they came out of 6,7 and headed east or even southeast Projection; Answer G: OK Acknowledge R: Green, so where have you looked? PA; New I can see where you are now, Location; New but have you gone through that middle section there? PA; New G: Which coordinate? Information; Question R: The middle- between 9 and 13Information; Answer OK, I’ve found the other seismic monitor with prerequisites Self Report; New G: Nice job. . . Acknowledge; Praise ok let’s pool together the information we have for the moment. PA; Old 01/13/04 Orasanu.CKM 25 Unsuccessful Communication B: So Green you’ll take that row straight up G: Yup. B: and then did somebody get that row to the left of that? Oh yeah, you’re right there, Red, so right you can cover the row straight up above you. R: What row? B: Red, you’re at 16, 23? R: yeah. B: Alright, I guess, you know where you’ve covered, so R: But where, where’s everyone else? I thought we were sticking together this game? B: Yeah, we’re going to cover the lower seg, … we’re going to stay below row 22 for now, cover this bottom section, then rejoin R: wha, where are we joining up again? (…) B: Purple refuels at FD1, Red and Green will refuel at FD2 R: Where’s FD2? B: FD2 is at 7 by 18 R: Ok, uh, and where are we meeting up again? (…) 01/13/04 Orasanu.CKM 26 Unsuccessful Communication P: Again, we’re getting caught up on these emergencies B: ‘cause the emergency was right there. At a time you were heading to where you were going. P: No . . . I’m sorry . . . I . . . you know, I . . . I don’t have the men. ( . . .) B: You had the men, if you didn’t process something else on the way Performance Assess; Non-support; Criticizing P: Well Mister Blue, I mean, I apologize for that. 01/13/04 R: ’m not going to make that. . .5,7 emergency . . . G: Why?! R: I’m trying to get there but- . . . . . . . . R: Yep, it’s gone . . . damn-it. G: That’s 75 points, Red. B: Red can you get the (emergency) at 15, 1? . . . . . R: Okay, I’ll get that one . . . that’s not a problem G: That’s what you said with the last one . . . . PA; Non-support; Criticizing R: I know . . . (loudly) hey, at least I haven’t run out of gas yet ok! PA; Non-support; Criticizing B: Whoa, whoa, whoa . ( . . . .) Mediation Orasanu.CKM 27 Baseline Study Findings Task very sensitive to teamwork Identified positive & negative behaviors associated with performance Relation between team dynamics and performance – Two factors predict performance • Social Support: Trust, Task Facilitation, Cooperativeness • Instrumentality: Leadership, Task Orientation, Openness Physiological measures being analyzed – Correlated across players – Related to task stressors 01/13/04 Orasanu.CKM 28 Next Steps New study - Gender / culturally diverse teams Develop predictive model of effective team performance – Use individual difference factors (personality, ability, self-report) as covariates – Team climate, team dynamics, team interaction as mediating variables – Task difficulty, Team/individual stress variables Develop noninvasive technology to monitor team psychosocial state Develop countermeasures Hand off to crews to use themselves! 01/13/04 Orasanu.CKM 29