Decision Aiding Dangers: The Law of the Hammer and Other Maxims

Abstract

The potential dangers of decision aiding in the realm of computer-based command and control (C2) problem-solving systems are discussed. The emphasis is on the early steps in the system design and development process with a focus on requirements analysis, system modeling, and analytical methods assessment/selection. The first danger discussed is "methidolatry," the generalized assumption that computerization is always beneficial. The need for a framework for decision aid design and development is specified, and the importance of two variable clusters to aidability determination is highlighted: the decision situation (closed or routine, open, and crisis) and a typology of decisionmaking functions or tasks. The decision types/functions yield a matrix of C2 decisionmaking "occasions," which constitutes the universe for decision-aiding opportunities or perceived needs. The second danger concerns the lure of technology push where methodologists favor one methodological family or metafamily; artificial intelligence is the current favorite. Algoritunic versus heuristic metafamilies are contrasted, and four metafamilies and nine specific families are identified; in addition, 55 submethods are delineated. The concept of a decision aid methodology selector is introduced and briefly illustrated. A matrix comprising the methods inventory and the decision situation features 60 cells; each metafamily (or family) is rated as being of low, medium, or high current or potential relevance to each of the cells. A reference to the increasingly popular strategy of methodological synergy and cautions about other dangers and the regrettable lack of attention to a coherent theory of decision aid design and development are included.