Using the Conant Method to Discover and Model Human-Machine Structures

Abstract

The term Mental Model has been used to account for human performance in a variety of domains. Cross disciplinary interpretations of the term have appeared incompatible with each other, threatening to render it vacuous. Moray (1988, 1990, 1991, in press) proposes lattice theory as a comprehensive formalism of the structure of Mental Models, assimilating the apparently disparate interpretations. To date, there has been little objective, quantitative support for mental models of complex human-machine systems. The Conant (1972, 1976, 1988) Method of Extended Dependency Analysis (EDA) suggests a viable quantitative means of discovering human-machine structures, a crucial step toward an empirical validation of the Mental Models concept. Results of a process control experiment suggest that a complex system can be reduced to a manageable number of key variables. Further, successful operators identify these variables and demonstrate patterns of information flow which exploit crucial system relationships.