Experiments in the Heuristics of Human Information Processing

A major contribution of Bayesian decision theory to accounting research is that by embedding a subjective interpretation of probability in a general theory of rational decision making, it provides for explicit recognition of uncertainty.

Abstract

Research in financial accounting, managerial accounting, and auditing has reflected increased use of Bayesian decision theory. This theory has been used in such areas as cost-volume-profit analyses (Jaedicke and Robichek [1964]; Johnson and Simik [1966]; Hilliard and Leitch [1975]), the investigation of cost variances (Kaplan [1969; 1975]; Dyckman [1969]; Li [1970]; Ozan and Dyckman [1971]), auditing (Tracy [1969]; Sorensen [1969]; Scott [1973]; Kaplan [1973a; 1973b]), and information economics (Demski [1972]; Feltham [1972]; Demski and Feltham [1976]).' A major feature of Bayesian decision theory is its reliance on subjective probability, which it regards as the quantified opinion of an idealized person faced with uncertainty (Savage [1972]). The subjective probability of an event is defined by the set of bets about the event such a person is willing to accept, and an internally consistent (or coherent) subjective probability can be derived for the person if his choices among bets about this event satisfy the axioms of the theory. A major contribution of Bayesian decision theory to accounting research is that by embedding a subjective interpretation of probability in a general theory of rational decision making, it provides for explicit recognition of uncertainty. Instead of focusing on models in which decision variables have known values or are treated as certainty equivalents, accounting researchers have tried to focus on models that incorporate the random nature of decision variables.