Students' structured conceptualizations of teamwork in multidisciplinary student teams using concept maps

Higher education increasingly prepares students for teamwork in professional contexts, often around complex problems requiring cross-disciplinary collaboration. Despite widespread use of multidisciplinary and interdisciplinary teamwork in curricula, there is no consensus on how best to teach or learn teamwork competence in multidisciplinary student teams (MSTs). Existing research often measures student ratings on predefined teamwork aspects rather than capturing students’ own frameworks. Given students’ experience with teamwork but limited explicit training, their conceptualizations may be incomplete, potentially hampering performance and learning. The paper aims to uncover students’ own conceptualizations of teamwork in MSTs via concept mapping. Using Repko and Szostak’s definitions, the study focuses on student teams with differing disciplinary backgrounds (multidisciplinary teams) tasked with integrative, interdisciplinary research. The central research question is: How do students in higher education conceptualize teamwork in MSTs?

Contemporary teamwork models often adopt Input–Process–Output (I-P-O) frameworks tracing back to McGrath (1964). Hackman and Morris (1975) and Tannenbaum et al. (1992) elaborated these models by detailing inputs, processes, outputs, and feedback. Marks et al. (2001) distinguished team processes from emergent states; Kozlowski and Ilgen (2006) emphasized dynamic, multi-level interactions among inputs, processes, emergent states, outputs, and environment. Van Woerden (2023) further specified an I-P-O framework for teamwork. MSTs typically face task uncertainty, dependency, and cognitive/disciplinary diversity, yielding both opportunities (e.g., creativity) and challenges (misunderstandings, conflict, difficulty in shared mental models, brief collaboration windows, leadership ambiguity, and assessment constraints). Studies of students’ own perspectives show they emphasize communication, equal division of work, shared goals, trust, commitment, responsibility, organization, and time management while struggling with leadership, conflict, and diversity management (e.g., Hirsch and McKenna, Mayne, Varhelahti et al., Örnekoğlu-Selçuk et al., Meehan and Thomas, Neumeyer and McKenna). However, how students conceptualize interrelations among teamwork aspects is under-researched. Concept mapping (Novak and Gowin; Trochim) offers a structured way to elicit and visualize participants’ conceptual systems and interconnections. The present study uses Trochim’s cluster approach, enriched with relational mapping, to investigate students’ structured conceptualizations in MSTs.

Design: Qualitative concept mapping study using Trochim’s (1989) six-step method with adaptations (predefined concepts; manual analysis without specialized software; inclusion of relational links). The approach combines cluster and relational mapping (after Curșeu et al., 2007). Participants and setting: Dutch second-year undergraduates in an interdisciplinary bachelor program conducting an interdisciplinary research project in teams of 2–4 during fully online instruction (COVID-19). Of 136 enrolled (56 in block 3; 80 in block 4), 108 consented. Concept map session attendance totaled 53 students across teams; maps were included only when all map-making members consented. Twenty team maps were analyzed. Materials: A predefined list of 43 teamwork-related concepts adapted from Curșeu et al. (2007), with three added to reflect MST context and interdisciplinary tasks: different perspectives, complexity, develop. Concepts were translated into Dutch. Students could use only concepts they deemed important, creating a 0/1 inclusion per concept. Procedure: In week 8, students completed a 45-minute Miro-based concept mapping assignment: (1) cluster relevant concepts; (2) draw connections between concepts using lines (non-causal) or arrows (intended causal). The prompt focused on organizing concepts based on current team collaboration in MSTs. Teams also noted any missing concepts. Data collection: Teams worked on dedicated Miro boards during the final course lecture. Presence per team varied (1–4 members). Two maps were excluded due to no activity. Analysis: Steps 3–5 of Trochim’s method were implemented. For each map, all links were recorded: (a) co-clustering of concepts (score=1, considered weakest relation), (b) non-causal lines (score=2), (c) arrowed (intended causal) lines (score=2). A group interrelation matrix of all concept pairs aggregated scores across maps. Total connection scores were binned: ≥11 (17 connections), 8–10 (26), 6–7 (41), and ≤5 (819). Connections ≤5 were excluded as too weak/rare. Because many maps inconsistently used arrows vs. lines, directionality was not analyzed. Researchers constructed point and cluster maps directly from the matrix, drawing strongest connections first and weighting maps equally regardless of number of students present. The final cluster map visualized connection strength via line thickness. Step 6 (utilization) is discussed as implication rather than conducted empirically.

• Four internally cohesive but weakly interlinked clusters emerged from aggregated student concept maps:
  1. Interaction cluster: teamwork, group, interaction, meetings, cooperation, communication, time spent together. Communication bridges to other clusters.
  2. Trust cluster: trust, respect, friendship, involvement, being informed, cohesion, patience, understanding (often interpreted empathically). Represents desired/ideal emergent states; strongly interconnected internally; weakly linked to task-oriented clusters.
  3. Conflict/divergence cluster: two sides tightly linked through divergences/arguments/conflict. Productive side (divergences, different perspectives, debates, arguments) connects to consensus and compromise; destructive side links conflict to stress, frustration, inflexibility, defensive reactions. Indicates students associate divergences with both opportunities and risks.
  4. Innovation cluster: knowledge, develop, innovation; the only cluster directly connected to achievement and satisfaction; connected weakly via complexity to the conflict/divergence cluster; not directly linked to trust or interaction clusters.
• Inter-cluster connections are sparse and thin, suggesting students see distinct facets with unclear mediation among them. Only a few maps linked clusters directly; notably, two teams linked trust to innovation.
• Connector concepts identified include: complexity/innovation; consensus/communication; trust/communication; involvement/communication; involvement/time spent together; group/time spent together; meetings/time spent together.
• Two concepts, unit and influence, lacked sufficient connections to be placed in the final map.
• Quantitatively, 17 connections scored ≥11, 26 scored 8–10, 41 scored 6–7, and 819 scored ≤5 (excluded). Twenty team maps informed the aggregate.

The findings address the research question by revealing that students compartmentalize teamwork into definitional interaction, ideal socio-emotional states (trust), conflict/divergence processes, and task/output (innovation) with strong intra-cluster coherence but weak inter-cluster linkage. This suggests a relatively shallow or incomplete systems view: students do not clearly conceptualize how ideal emergent states (e.g., trust, cohesion) influence task execution and innovation, nor how interactional definitions connect to conflict management and performance. The conflict/divergence cluster’s dual nature mirrors students’ ambivalence about conflict (productive vs. harmful) and aligns with literature showing spillovers between task and relationship conflict. The limited linkage between trust and innovation contrasts with theoretical accounts positing emergent states’ roles in performance and viability. Identified connector concepts (e.g., communication, involvement, time spent together, complexity/innovation) correspond to widely recognized teamwork levers and can be focal points in instruction. Overall, the results highlight gaps between students’ compartmentalized frames and more integrated I-P-O models of teamwork from the literature, informing targeted pedagogical interventions for MST contexts.

This study contributes an aggregated, structured visualization of how students conceptualize their own teamwork in MSTs. Using an adapted concept mapping method, four distinct but weakly connected clusters were identified: interaction, trust, conflict/divergence, and innovation. The compartmentalization indicates students’ limited integration of socio-emotional states with task processes and outcomes, and an ambivalent view of conflict situated near both productive divergence and harmful dynamics. The work advances understanding of student frames, providing a basis for designing teamwork teaching that explicitly addresses observed lacunas, such as linking trust and communication to task execution and innovation, differentiating and managing healthy vs. harmful conflict, and leveraging connector concepts to bridge clusters. Future research could co-generate concept lists with students, include omitted aspects (e.g., leadership, task division), ensure full team participation, and further theorize/validate aggregation procedures for concept maps, potentially incorporating directionality and temporal dynamics.

• Predefined concept list constrained student input; some relevant concepts (e.g., leadership, task division) were absent. Only two additional concepts were suggested by students (structure; sense of responsibility).
• Partial team attendance and consent led to variable representation across teams; only 20 maps were analyzed.
• Aggregation assumptions: Combining 20 maps into a single cluster map may omit nuances (e.g., images/labels) and assumes comparability across teams; directionality of relations was not analyzed due to inconsistent arrow use.
• Contextual confound: Teams also participated in a reflection-intervention study; although measured effects were minimal, this may have influenced mapping behavior.
• Naturalistic setting limited control over presence/consent and full adherence to the original concept mapping steps (e.g., participant-generated statements).