Representations of Perceived Relations among the Properties and Variables of a Complex System

Three different techniques for representing human understanding of complex systems were compared. Novice veterinary students and cardiovascular research experts made judgments of the relations among the properties and variables of a complex system, the mechanical heart / blood vessel system. They also described the variables and properties by a series of bipolar ratings. A variety of analyses showed that the novices tended to conceptualize the system in static anatomic terms. Experts showed a more integrative conceptualization and distinguished more clearly than students between relations involving only system properties and those involving system variables. The methods of multidimensional scaling, agglomerative hierarchical clustering, and elementary digraphs were used to represent perceived relations among system variables and properties. It was concluded that the simplest form of representation, a digraph, has several advantages over the other representations.