Much of the research on the optimal combination of multiple judgment implicitly assumes that the judgments are not affected by the combination rule. The current research demonstrates that in certain organizational settings in which the judgment (and their combination) are likely to affect the judges, this assumption is likely to be violated. A cognitive game theoretic analysis is proposed to analyze situations of this type. This framework is supported by experimental results that demonstrate some of the nontrivial predicted effects of the judgments. Practical implications are discussed.
Let Mj be the mean subjective probability estimate of J judges (or the mean of any monotone increasing transformations of the judgment), and let P(T/Mj) be the probability that an event is true given Mj. We report the results of analyses of real data sets in selected domains in order to explore the behavior of P(T/Mj) as J increases. The exploration is motivated by results of Wallsten, Budescu, Erev & Diderich (1997) and Wallsten & Diederich (1997), who proved that under reasonable symmetry assumptions and a conditional pairwise independence assumption, P(T/M)j approaches 1(0) as J increases, according to whether Mj is above (below) the center of the scale. Conditional pairwise independence is violated in these data sets to varying degrees, and there is no guarantee that the other assumptions hold. The theoretical result is quite remarkable and it is of practical interest to know the effects of averaging increasing numbers of judgment with real data.
Let Mj be the mean subjective probability estimate of J judges (or the mean of any monotone increasing transformations of the judgments) and let P(T/Mj) be the probability that an event is true given Mj. We report the results of simulations designed to explore the behavior of P(T/Mj) as J increases and the sensitivity of the process to violations of conditional pairwise-independence. The exploration is motivated by results of Wallsten, Budescu, Erev & Diederich (1997) and Wallsten & Diederich (1997), who proved that under reasonable symmetry assumptions and a conditional pairwise independence assumption, P(T/Mj) approaches 1(0) as j increases, according to whether Mj is above (below) the center of the scale. The theoretical result is quite remarkable and it is of practical interest to know rates of convergence under the ideal case and robustness of results to violations of the assumptions.
The concept of strategic surprise has been used by writers on military strategy to explain the successful amplification of resources during battle. We apply the same concept to the relationship between buyers and suppliers. First, we examine factors that produce vulnerability to strategic surprise in cooperative situations. We then examine the reasons why firms are caught by surprise in spite of their vigilance. We argue that firms use cognitive heuristics which produce misinterpretation of evidence and a consequent sense of false security. In particular, interaction norms combine with the representatives heuristic to increase the likelihood of misjudgment.
The negative outcome bias is characterizing the way managers evaluate the quality of past decisions and of the need to analyze the decision making process. It was found that managers tend to perceive a causal positive relationship between the valence of outcomes and the quality of preceding decisions. Similarly, managers are mobilized to initiate a learning process mainly after negative outcomes but not after positive ones. The negative outcome bias poses a severe threat to the optimality of organizational decision making and learning. Potential debiasing methods are discussed.