First Look

Working Papers

The Judgment-Decision Paradox in Experience-Based Decisions and the Contingent Recency Effect

Abstract

The current paper explores a judgment-decision paradox in experience-based decisions: the finding that rare events are overweighted in probability judgments but are underweighted in repeated decisions under uncertainty. Two laboratory studies examine both decisions and probability assessments within the same paradigm. The results reveal overweighting and negative recency in probability assessments but underweighting and positive recency in choices. At the same time, there remains an overall consistency between choices and assessments. A third study validates the results in a field study. The results show that, after a negative rare-event (i.e., a suicide bombing) people believe the risk to have decreased (negative recency) but are more cautious (positive recency).

Paper not available

Learning and Equilibrium As Useful Approximations: Accuracy of Prediction on Randomly Selected Constant Sum Games

Abstract

There is a good deal of miscommunication between and among experimenters and theorists, and psychologists and economists, about how to evaluate a theory that can be rejected by sufficient data, but may nevertheless be a useful approximation. A standard experimental design reports whether a general theory can be rejected on an informative test case. This paper, in contrast, reports an experiment designed to meaningfully pose the question: "How good an approximation does a theory provide on average?" Even in the simplest class of two-person constant sum games there is mixed evidence on this question, and resulting controversy (dating from the 1950's in psychology and continuing through the 1990's in economics). We consider a random sample of 2x2 constant sum games to determine how close on average are the equilibrium predictions to observed play in repeated play of a game against a fixed opponent. We also compare the predictions of several models of learning and related models. We find that equilibrium becomes a better predictor of observed play as the players become more experienced, but that simple models of learning predict more accurately, over the 500 periods of repeated play that we observe. We also discuss how the predictive value of a theory can be measured in terms of how many pairs of experimental subjects you would have to observe playing a previously unexamined game before the mean of the experimental observations would provide a better prediction than the theory about the behavior of a new pair of subjects playing this game. We call this latter quantity the model's Equivalent Number of Observations (ENO), and explore its properties.

Paper not available

The Implicit Effect of Artifact-Driven Inferences on Perceived Procedural Justice

Abstract

This study considers organizational artifacts as triggers of procedural justice expectations, showing the effects of one artifact—an organizational wait queue—on perceived procedural justice. Queue designs that violated consistency or suggested bias were perceived as less procedurally just than those that implied consistency and lack of bias. Information about the bias implied by a queue improved perceived procedural justice. The analysis suggests organizational artifacts to be "secondary contract makers" and informs the study of procedural justice by suggesting a new source of inferences about and precursors to perceived justice. The study offers implications for the management of queues as well as other artifacts.