Kenneth R. MacCrimmon

Utility Theory: Axioms Versus ‘Paradoxes’

Few attempts have yet been made to interrelate the major axiom systems of expected utility theory such as those presented by von Neumann—Morgenstern, Marschak, Savage, and Arrow. This paper attempts to consolidate the theoretical and empirical research on these axioms by showing the correspondences among the major axioms of each system and highlighting their similarities and differences in dealing with concepts such as probability and utility.

Toward a Stochastic Model of Managerial Careers

This study describes how the career movements of managers and professionals within organizations may be described by a Markov chain model. This allows a formal description of the results of current career policies which can be examined for inconsistencies. Further, it allows predictions to be made of the effects of continuing present policies into the future. Thus, it provides a more rational basis for career policies of organizations and also for individual managers planning their own careers. From a sample of career movements of managers and professionals in a large industrial organization, data are presented, a simple model constructed, and inferences made. In addition to the normative uses of such models, it is important to note that they provide a means for examining in a dynamic way some basic behavioral science issues that have heretofore been approached in a static fashion.1 Victor H. Vroom is professor and Kenneth R. MacCrimmon is assistant professor in the Graduate School of Industrial Administration, Carnegie-Mellon University.

A portfolio of risk measures

A comprehensive set of sixteen measures of willingness to take risks has been developed. This set includes measures from three categories: measures from standardized risky situations having an underlying theory of risk, measures inferred from revealed choices in financial decisions, and measures derived from attitudes. A study of over 500 top-level business executives shows significant relationships within categories, but relatively little relationship across categories. Context differences, especially personal versus business situations and opportunities versus threats, underlie the responses.

Assessing Risk Propensity

The most accepted theory of risk is based on utility theory, as developed by von Neumann-Morgenstern (1947), Arrow (1971), and others. If the decision maker obeys the utility axioms, he acts as if he maximizes expected utility. The proof of the expected utility theorem suggests how to construct the utility function that reflects his risk attitudes.

The architecture of an information system for the support of alternative generation

Abstract:The pre-choice stages of the problem solving process are more difficult to support through information systems than choice itself. Systems that facilitate problem formulation and solution finding are typically either expert system programs with narrow application domains or programs dealing with easily quantifiable problems. In contrast, this paper introduces a decision support system, GENI, that operates independent of domain, aiding the user in general problem solving tasks. The system’s purpose is to support the problem solving process, rather than be a substitute for the human problem solver, by providing structure and by using different stimuli to prompt the user for data input. To deal with broad classes of complex qualitative problems, no single processing mode is sufficient. Hence the techniques in GENI differ mainly in the interface, that is, the types of stimuli displayed, and the types of data recorded. GENI’s user interface is built around a standard dialog function that can be modifi...

Multiple Criteria Dominance Models: An Empirical Study of Investment Preferences

When making choices from among alternative investment opportunities, investors may focus upon criteria such as expected rate of return, variance, probability of specific gains, probability of specific losses, and negative semivariance. We presented over 400 top-level business executives with nine personal investment opportunities and obtained their preference rankings for these ventures. This paper examines the extent to which their rankings can be explained by single or multiple criteria dominance models. In general, we find that with a single criterion model the mean is the criterion most focussed upon, but that more than eighty percent of the cases cannot be explained by any single criterion. Two criteria dominance models provide considerably better explanations especially when we include the probability of specific gains as one of the criteria. Models with more than two criteria do not provide better explanations.

Expert systems and creativity

In this paper we evaluate the requirements for an expert system designed to exhibit creativity. First we consider what it means to be an expert and what “expert system” means in an artificial intelligence context. Then we shift focus from “ordinary” intelligence to “creative” intelligence by examining the concept of creative behavior, followed by a discussion of the implications for the design of an expert system exhibiting creativity across various domains. In the final section of the paper we present a design overview of such a system.

Supporting problem formulation and alternative generation in managerial decision making

The pre-choice stages of the problem solving process are more difficult to support through information systems than choice itself. Systems that support problem formulation and solution finding are typically either expert system programs with narrow application domains, or programs dealing with easily quantifiable problems. In contrast, the authors introduce a decision support system (GENI) that operates independent of domain, aiding the user in general problem solving tasks. GENIs user interface is built around a standard dialog function which can be modified by means of parameter settings. The various settings result in different stimuli being displayed and different data items being recorded. The standard interface design was chosen to reduce the development and maintenance effort, as well as to simplify the users problem solving task.<<ETX>>

Second generation creativity support software

While first generation creativity support programs have had successes in improving problem solver performance in idea generation tasks, limitations of this type of software have also become apparent. The paper identifies these limitations and describes design principles for and components of second generation creativity support software.<<ETX>>

Decision Theory and Real Decisions

Decision theory has been extremely successful in various ways. As a theory per se, it has provided interesting research opportunities stemming from an elegant central core. It has served as a basis for identifying behavioral anomalies and then stimulated research into trying to incorporate explanations into a revised theory. On the application side, decision theory has been used as a basic building block in models throughout the behavioral and social sciences. The theory also has strong prescriptive implications and for several decades the key ideas from decision theory have been taught in basic courses in areas such as economics, business, psychology, and statistics.