Gaze data provide sensitive metrics of attention-synchronization in teams

Abstract

Teams of three can be conceptualized as nonlinear dynamic action systems with coordination activities reflected in attentional synchronizations emerging between the individual cognitive systems [1]. A team s oordination function might be impaired when the communication is prohibited or when there is conflict in the team [2, 3]. Successful communication is reflected in patterns of coupled visual attention [4]. It is tested if an integrated gaze-measure for N=3 generated by bivariate cross-recurrence analyses [4] is sensitive to established treatment effects, and thus can provide objective assessment of coordination processes within small groups. A 2 Communication [R/T on vs. off; within] x 2 social context [dissonant vs. non-dissonant; between] mixed design was chosen. In a Synthetic Task Environment (STE) [5] comprising detection and decision tasks, small groups have to react to critical events and decide about investments for remedy. During the detection task a group can draw simple inferences based on disjunctive information to reduce uncertainty. In the R/T-off condition the communication channel is deactivated during the signal detection task. For a dissonant social context a social dilemma plus competitive priming (by role description and incentive structure) is induced into the collaborative decision task. Gaze behavior of each participant is measured with an Eye-Follower® remote system (120 Hz). The coordination indicator percentage of integrated cross recurrence rate (%iCRR) is calculated based on the synchronized gaze-data utilizing [6]. Only measurements on a signal detection task performed in the context of a synthetic-task environment are reported. 144 participants (age: M=23.01; SD=5.03; gender: 51.4% female, 48.6% male) were grouped into 48 teams (15 ab-initio ATCo candidates, 33 students). During a 3,5h session 2 scenarios are performed, each comprising 48 signal-detection trials with 3 malfunctions forcing a collective decision. Experimental conditions were balanced. A Mixed-Model Analysis of the %iCRR values produced a main effect for R/T [F(1, 679.2)= 92,21, p> .05, ηp²= .0841]. The magnitude of %iCRR was higher when the communication channel was open (cf. figure 2). A main effect for social context [F(1, 679.2)= 6,69, p> .05, ηp²= .006] indicates lower behavioral couplings for groups performing under dissonant conditions (cf. figure 2). A significant interaction [F(1, 679.2)= 4,47, p> .05, ηp²= .004] indicates a mediation of the social context effect by communication (cf. figure 2). Resume and practical implications Indication is provided that with focal goals in conflict (dissonant social context) behavioral coupling decreases. We assume the metric of integrated gaze cross-recurrence to be a sensitive objective measure for the assessment of the macrocognitive process of coordination within small groups. Literature [1] J. C. Gorman, T. A. Dunbar, D. Grimm, and C. L. Gipson, "Understanding and modeling teams as dynamical systems," Frontiers in psychology, vol. 8, p. 1053, 2017. [2] C. Bearman, S. B. Paletz, J. Orasanu, and M. J. Thomas, "The breakdown of coordinated decision making in distributed systems," Human Factors: The Journal of the Human Factors and Ergonomics Society, vol. 52, no. 2, pp. 173-188, 2010. [3] N. J. Cooke and J. C. Gorman, "Interaction-based measures of cognitive systems," Journal of Cognitive Engineering and Decision Making, vol. 3, no. 1, pp. 27-46, 2009. [4] K. Shockley and M. A. Riley, "Interpersonal Couplings in Human Interactions," in Recurrence Quantification Analysis: Theory and Best Practices. J. Webber and N. Marwan, Eds. Cham: Springer International Publishing, 2015, pp. 399-421. [5] D. Schulze Kissing and C. Bruder, "Interactive Team Cognition: Do Gaze Data also Tell the Story?" Proceedings of the 19th ISAP, Dayton, Ohio: Wright State University, 2017. [6] N. Marwan, "Cross Recurrence Plot Toolbox for Matlab", Reference Manual, Version 5.19, Release 30.2 ed Potsdam: Potsdam Institute for Climate Impact Research, 2016.