Mental Models and the Acquisition of a Complex Skill across Individuals and Teams: A Multilevel Study

A mental model reflects the structural relationships between concepts within a specified knowledge domain. Measuring the structure of knowledge is important because it offers the possibility of capturing expert knowledge which is often difficult to assess using traditional declarative knowledge measures. The concept of mental model has been extensively studied over the last decades and it is often acknowledged in the training literature as one of the key antecedents of performance in complex tasks, particularly in the context of teams where the construct of shared mental models has received ample attention. Whereas the training literature has established the validity of mental models for predicting individual and team performance using single-level studies, the extant literature has not yet tested the validity of mental models as a multilevel construct. Consequently, the purpose of the present study was to assess the extent to which the relationships between mental models and performance generalizes across individuals and teams, that is to test a homologous multilevel model.

Participants in this study completed a dynamic, networked computer-based simulation. Three-person teams operated the simulator collectively (through specialized roles) and as individuals (performing all roles simultaneously) over the course of a 2-day 48-hour-interval protocol. The sample consisted of 243 individuals nested in 81 3-person teams. Consistent with the multilevel nature of the problem under study, multilevel analyses were conducted to test the study hypotheses.

Consistent with theory and previous research on individual and team cognition, it was hypothesized that stronger relationships between mental models and performance would exist at the individual level compared to the team level. In essence, processes occurring at the team level were expected to attenuate the relationship between mental models and performance compared to the individual level of analysis. Contrary to this expectation, the magnitude of the relationship between mental models and performance was similar across levels of analysis. Additionally, consistent with previous research on the effectiveness of declarative knowledge measures for predicting complex performance, the present results indicated that declarative knowledge was more predictive of individual performance than team performance.

In addition to performance, an objective measure of behaviors was utilized to further understand of the processes through which mental models translate into effective individual and team performance. It was hypothesized that the relationship between mental models and behaviors would be stronger for individual tasks than team tasks as a function of the additional interaction requirements associated with team tasks. However, contrary to this expectations, mental models and behaviors were more strongly associated at the team level than the individual level.