Steps to a formal analysis of the cognitive-energetic model of stress and human performance

Abstract

Sanders' (1983) cognitive-energetic model of stress and human performance attempts to bridge linear stage and capacity models of information processing. It contains three different levels: a computational or cognitive level, an energetic level, and an evaluational level. The model involves a complex arrangement of component processes. The detailed effects of variations of some subset of component processes on the models' output cannot be read off directly from the corresponding flow diagram. It is argued that the identifiability of such effects can only be properly evaluated through a combined approach involving formal analysis and numerical simulation. The current paper outlines a formal analysis of the model conceived of as a stochastic PERT network. Additionally, Sanders' model is regarded as a special instance of Grossberg's (1982) functional-differential network. Results of illustrative applications of both representations confirmed the validity of the additive factor method which is the basic method with which Sanders built his model. More specifically, comparative simulation of a Sternbergian linear stage model and McClelland's cascade model suggested a qualitative similarity. It was concluded that the PERT, and, in particular, the neural network representations are appealing approaches in the assessment of the cognitive-energetic model of stress and human performance.