Cognitive representations, control, and understanding of complex systems: a field study focusing on components of...

Users' mental models of systems are frequently asserted to be important in their interactions with those systems. Although there are few explicit definitions of mental models in the literature, mental models are assumed to be important and researchers have seemed to accept and understand the concept intuitively. In our research, we have extracted three themes from the definitions of mental models including: (1) the relations among features or components of a system; (2) the capability of the mental model to aid in the control of the system; and (3) the capability of the mental model to enhance the user's understanding of the system. The relationships among the components of users' mental models, including procedural and declarative knowledge, perceptions of the use of system features, and the relations among the similarity of use of system features, were investigated in a field study. Trained and untrained novice users of a new business phone system rated the similarity of use of nine standard features on the phone. These ratings were used to derive cognitive representations of the relations among the system features. Users' ratings of features on univariate scales about the use of the features and scores obtained on a test of procedural knowledge were related to these representations. The two novice groups' mental models were similar except for the perception of one feature; it was more accurately depicted in the trained users' representation than those who did not attend a training programme. Two types of experts' similarity of use ratings were evaluated to determine an 'appropriate' representation of the use of the system features; this was then compared to the novices'. This comparison suggested that there were deficiencies in the mental models of novices. Designing training programmes and instructional aids for systems using the discrepancies between novices' and experts' mental models are discussed.