Identifying mental models of complex human–machine systems

Abstract

The notion of a `mental model' is widespread but ill defined. Its meaning can be clarified by noting both the cognitive processes of the worker and the context of the task. A mental model is a mapping of the properties of the task to its representation in the mind of the worker. Traditionally, such models have been identified by the use of protocol analysis, but a method has been developed which finds the mapping through its effects on the coupling between human and machine. This method uses Conant's theory of system decomposition based on high-order Shannon information theory. An example of its use is given. Relevance to industry Operators make use of mental models of complex systems in process control and manufacturing. A method to identify such models is relevant to productivity, and to fault detection and management, since it gives insight into the way in which operators become coupled, by training and habit, to the dynamics of the processes they control.