Article Preview
TopDecision Making With Data: A Process Of Inference
The commonly accepted model in much of the BI&A research appears to be that humans, despite their limited information processing capabilities and the socially constructed power struggles that accompany life in organizations, consume the output of information systems to make informed and therefore more effective decisions. We have known for some time about the limits of human information processing capacity (Simon, 1955), and so the assumption of a purely rational decision-making process that underlies much of BI&A research is likely untenable.
Decision making with data is based on inferences drawn from cues in the data set (Chater, Tenenbaum, & Yuille, 2006). The drawing of inferences relies on inductive reasoning, which has been classified into either automatic (heuristic-based) or the more familiar rational choice model that uses rules-based processes (Ferreira, Garcia-Marques, Sherman, & Sherman, 2006). The heuristic-based approach involves pattern matching: the decision maker recognizes cues and the associated responses reflective of similar previous situations. The cues more or less trigger an automatic response selected from the range of recalled responses. In fact, the proponents of Naturalistic Decision Making (NDM) have argued that the while rules-based model might apply in some situations (Klein, 2008), in the real world, most complex decisions are actually made through this process of pattern matching.
The components to be matched include organizational or individual goals, critical cues, expectancies and typical actions (Klein & Klinger, 1991). For relatively simple decisions, this pattern-matching process is very quick. In this case, the situation is immediately recognized and a match to previous decisions is quickly found. The decision maker simply implements actions similar to those applied in the past.