The goal is to illustrate how visual representations of teamwork analytics can engage non-technical people (e.g. educators and students) and teamwork experts (e.g. academics interested in teamwork research) by making the mapping from low-level multimodal data to higher-order constructs (HOCs) open to human scrutiny. Functional prototypes and intelligent interfaces, tuned to the needs of different audiences (according to what has been envisioned by different stakeholders in Phase 1) will be trialled with participants, including those who participated in Phase 1 and new participants.
In a second feasibility study, we identified additional components (shown in blue in the figure below) to include parameters from the assessment criteria (A), used to generate a higher-order information structure, the Learner Model (LM); a structured representation of students’ performance, misconceptions or difficulties. The LM can be visualised as user interfaces (S) that communicate HOCs or provide feedback to different stakeholders via visualisations, charts or text. For example, the multimodal matrix (M) could be interrogated for the system to create a model of individual students or the group, to provide feedback about errors made by the team, affective states or positioning strategies. The figure presents one low-fidelity prototyped interface presented to nursing students. Feedback is provided about an error made by a team during a simulation, based on rule-based algorithms detecting the order and timing of actions performed by students.