An experiment undertaken with engineering undergraduate students at the University of Nottingham involved 26 groups of three being filmed during a study using a virtual-reality-based problem-solving exercise. After the exercise, each individual filled in a questionnaire relating to the exercise which allowed them to score themselves and their peers for contribution and overall grade. The comparing of video evidence with perceived contributions made it possible to observe patterns of behaviour based on temperament dominance. This ‘dominance’ was based on two simple parameters extracted from an electronic version of the Myers-Briggs test: first, the time taken to complete the study, called ‘decisiveness’, and secondly, the degree of Extroversion/Introversion. The more decisive subjects received higher marks from their peers, despite the absence of any video evidence that they had actually contributed more than their peers. The most dominant extroverts appear to ‘do more’ with respect to the physical operation of the mouse/keyboard and interaction with the visual simulation during the virtual-reality exercise. However, there was no link with these simple temperament measures with the degree of enjoyment of the tasks, which appeared to be highly consistent. The authors do not argue that visual-media tools, such as the virtual-reality environment described in this article, might offer solutions to problems associated with group work in engineering, but rather that information regarding the character traits of the participants may help to create more effective teams and to help understand the inter-personal dynamics within teams undertaking such tasks.