Multi-attribute Utility Model Research Articles

Economic and non-economic motives for residential preferences and choices

Previous studies have repeatedly shown that consumers' residential preferences are more predictable from a modified multiattribute utility (MAU) model than are their residential choices. The aim of the present studies was to test the hypothesis that the differential predictability may be due to housing cost influencing choices more than preference ratings. In the first study, preference ratings were again found to be more predictable than choices regardless of whether cost did or did not discriminate between alternatives. Further analyses revealed that rather than different importance of cost, the use of simplifying heuristics due to different demands imposed by choices and preference ratings accounted for the observed differential predictability. In the second study, preference ratings, ratings of willingness to pay, and choices were directly compared under conditions when cost did not discriminate between the alternatives. In addition to replicating the results of the first study, the correlation between choices and preference ratings was found to be higher than between any of them and the willingness-to-pay ratings. Further support was thus obtained for the conclusion that the differential predictability is not due to different importance of costs.

A knowledge acquisition method based on a multi-attribute utility model

A knowledge acquisition method based on a multi-attribute utility model

Decision Modeling: Tools for Strategic Thinking

For 10 years, the Decision Techtronics Group (DTG) has helped public organizations use decision modeling to improve group decision making. When the New York State Insurance Department faced the task of making recommendations to resolve the medical-malpractice crisis, it developed, with the help of DTG, a system dynamics simulation model and a multiattribute utility (MAU) model to analyze policy options. Decision makers assessed the short- and long-term impact of proposed policy changes and their utility to various stakeholders. The department was able to make sound recommendations to the state legislature. This case demonstrates that the successful application of modeling to strategic decisions requires (1) an understanding of the various modeling techniques, (2) the ability to diagnose a real problem and to select appropriate techniques (3) the willingness to rethink the problem in terms of the modeling techniques, and (4) the construction of models with full participation of the decision makers.

Models of Tradeoffs in a Hierarchical Structure of Objectives

Multiattribute utility models and cost-benefit models often arrange the objectives into a hierarchical structure in order to identify appropriate subobjectives and to select appropriate attributes to measure the subobjectives. This paper discusses nonadditive models that use the hierarchical structure in order to examine issues of tradeoffs that models in present use are not able to include. The models are sufficiently structured so that they can be applied by procedures that are similar to those available for other prescriptive models of preferences.

Prognosis After AIDS: A Severity Index Based on Experts' Judgments

Survival after diagnosis of acquired immunodeficiency syndrome (AIDS) differs greatly depending upon the type of complications the patient experiences. A detailed understanding of this variation in survival is an essential component of health care planning so that demands for hospital beds and other health care resources can be more realistically anticipated. We developed two severity indices for predicting the prognosis of AIDS patients using additive and multiplicative multi-attribute utility models. A panel of physicians described the scoring of each index. On 97 randomly selected patient profiles, we compared both severity scores to the clinical assessments of the expert panel. The multiplicative index was more accurate than the additive index. We describe the multiplicative scoring system for AIDS severity.

Prediction of preferences for and choices between verbally and numerically described alternatives

In order to investigate whether the format (verbal or numerical) of the information presented about different options affects the predictability of preference judgments and choices, a number of hypothetical housing alternatives were described in terms of six attribute dimensions to 36 adult subjects. The descriptions were either purely verbal, purely numerical, or mixed verbal and numerical. Preferences and choices were predicted by means of a combined multi-attribute utility and expectancy-value model. It was found that the predictions of the preferences were more successful when the most important attributes were described numerically than when those attributes were described verbally. Regression analyses of aggregated data further showed that numerically described attributes tended to carry a greater weight for the prediction of both preferences and choices than did verbally described attributes in the mixed numerical and verbal descriptions. These findings suggest that the higher precision conveyed by numerical descriptions may have helped the subjects to evaluate different options in a consistent manner.

Comparison of Four Behavioral Theories

A major concern among occupational health nurses is the efficacy of their health education and health promotion programs. The success of programs may be improved when they are built on a sound and appropriate theoretical foundation. Four models that present a logical and reasonable approach to behavioral change include the Health Belief Model, the Theory of Self Efficacy, the Theory of Reasoned Action, and the Multiattribute Utility Model. Behavioral models provide a framework for the development of interventions and strategies which will facilitate behavioral change.

Understanding Economic Behaviour.

to the Volume.- I: Economic Factors and Individual Satisfaction.- National Wealth and Individual Happiness.- Beliefs about Attainment of Life Satisfaction as Determinants of Preferences for Everyday Activities.- Need for Achievement and Entrepreneurial Activity in Small Firms.- The Kibbutz and Private Property.- II: Inflation, Unemployment, and Taxation.- Perceptions of Inflation.- Cross-Sectional Evidence on the Rationality of the Mean and Variance of Inflation Expectations.- The Perceived Causal Structure of Unemployment.- A Cross-National Comparison of Attitudes, Personality, Behaviour, and Social Comparison in Tax Evasion Experiments.- Perception and Judgment of Marginal Tax Rates After a Tax Reduction.- III: Utility and Decision Theory.- Beyond the Expected Utility Proposition in Rational Decision Making.- Levels of Aspiration, Promises, and the Possibility of Revaluation.- Another Attitude Towards Multi-Attribute Attitude Theories.- Application of the Valence-Instrumentality-Expectancy and Multiattribute Utility Models in the Prediction of Worker Effort.- Prediction of Preference Reversals by the Linearized Moments Model.- Economic Man or Social Man - Exploring Free Riding in the Production of Collective Goods.- IV: Consumer Behaviour.- From the Consumption of Necessities to Experience-Seeking Consumption.- Personality Traits as Elements in a Model of Eating Behaviour.- The Values and Consumer Behaviour - Some Empirical Findings From Austria.- Understanding Children's Economic Socialization.- A Model of Consumer Behaviour in the Situation of Shortages.- V: New Views on Economic Analysis.- The Quantitative Analysis of Economic Behavior With Laboratory Animals.- Subjectivism in Economics - A Suggested Reorientation.

Modelling the Process of Multiattribute Choice

In most of the multiattribute utility models it is assumed that the analytical forms of partial and overall utility functions are known, and hence the problem becomes one of estimating the values of model parameters. This assumption is rather restrictive since one may not possess prior knowledge as to the analytical forms of such functions. The present paper is an attempt to eliminate some of the shortcomings of the existing models by proposing two new formulations of the multiattribute choice problem without a priori limitation to any strict form of partial utility function.

Generic utility theory: Measurement foundations and applications in multiattribute utility theory

A utility representation is formulated that is generic in the sense that it is implied by many stronger utility theories, yet it does not make assumptions peculiar to particular utility theories. The generic utility representation postulates that the preference order on a subset of two-outcome gambles with fixed probability is represented by an additive combination of the utility of the outcomes. Although many utility theories imply that this representation should be satisfied, standard additive conjoint formalizations do not provide an axiomatization of the generic utility representation because the preference ordering is only defined on a proper subset of a Cartesian product. (The precise specification of the subset is stated in the paper.) An axiomatization of the generic utility representation is presented here, along with proofs of the existence and interval scale uniqueness of the representation. The representation and uniqueness theorems extend additive conjoint measurement to a structure in which the empirical ordering is only defined on a proper subset of a Cartesian product. The generic utility theory is important because formalizations and experimental tests carried out within its framework will be meaningful from the standpoint of any stronger theory that implies the generic utility theory. Expected utility theory, subjective expected utility theory, Kahneman and Tversky's prospect theory ((1979). Econometrica, 47, 263–291), and Luce and Narens' dual bilinear model ((1985). Journal of Mathematical Psychology, 29, 1–72) all imply that selected subsets of gambles satisfy generic utility representations. Thus, utility investigations carried out within the generic utility theory are interpretable from the standpoint of these stronger theories. Formalizations of two-factor additive and multiplicative utility models and of parametric utility models are presented within the generic utility framework. The formalizations illustrate how to develop multiattribute and parametric utility models in the generic utility framework and, hence, within the framework of stronger theories that imply it. In particular, the formalizations show how to develop multiattribute and parametric utility models within prospect theory and the dual bilinear model.

The effect of varying value trees on multiattribute evaluations

This experiment examines the effects of varying the super- and substructure of value trees on the weights in multiattribute utility models. The value trees were variations of a tree used by the U.S. Department of Energy in its evaluation of alternative nuclear waste disposal sites. Using a questionnaire, 200 subjects assigned weights to the objectives and attributes in these trees according to four weighting procedures: Ratio, Swing, Trade-off and Pricing-out. There was a sizeable difference among weighting methods, in particular in that the pricing-out method assigned much more weight to the monetary attributes, while the other methods tended to weight health and safety aspects the highest. Objectives that were higher in the tree and that were described by more attribute detail, tended to receive a higher weight than those that were lower in the tree and were less detailed. However, these effects varried substantially among weighting methods. The results suggest a careful study of criteria for developing ‘appropriate’ value trees.

Predictors of Low Back Pain Disability

Low back pain has major socioeconomic implications; much of the costs relate to disability and compensation. Theoretically, the early identification of patients at risk to become disabled from a low back episode would lead to more aggressive intervention and reduction of subsequent disability. Low back disability is related to occupational, psychosocial, diagnostic, demographic, anthropometric, health behavior, and injury factors that have been reported in the literature. The multiattribute utility model is a new experimental approach to prediction of disability. The relative weights of the various factors that might be predictive of low back disability are determined by a panel of experts. Although this model is not yet scientifically proven, it offers a promising method of answering the question, "Can low back disability be predicted?"

Models for computer system evaluation and selection

Three main models for evaluating and selecting computer systems are presented and compared: (a) the additive-weight model, (b) the Eigenvector model, and (c) the multi-attribute utility model. A case study describes the application of these three models to the selection of a computer for an organization. In this study, the three models showed almost identical results in ranking the alternatives. Based on the data obtained from the case study and a comparison of the models' attributes, we recommend using the multi-attribute utility model in cases wherein the required assumptions of independence hold. To overcome the difficulties involved in understanding and applying the model, the development of an interactive decision-support system is recommended.

Evaluating landscape aesthetics: A multi-attribute utility approach

Regulatory functions require subjective judgments which are sometimes inconsistently and arbitrarily applied. Aesthetic assessments of land-use proposals is such a situation. The research presented here involved developing a multi-attribute utility model to measure the aesthetic impact of proposed land developments. It was hypothesized that such a model would help in the process of reviewing land-use permit applications by increasing the consistency and comprehensiveness of the evaluation; while at the same time increasing the efficiency of the process. A panel of 11 experts with backgrounds in architecture, historical preservation, law, art, planning and government was convened to construct the model. The panel met seven times for 2.5 hours each time over a 6-month period. The resulting instrument, called the Aesthetic Impact Model (AIM), consists of 34 attributes. AIM has been used on a variety of examples and has proved to be an excellent framework for discussing the aesthetic impact of proposed development projects. AIM has also been used very successfully in public gatherings of 20–30 people to discuss landscape aesthetics. Specifically, AIM has shown five strong points: it increases communication among interested parties by providing a common lexicon and clear and detailed feedback; it clarifies disagreements by narrowing them to specific attributes; it facilitates discussion of aesthetics by lay people; it rates projects on a systematic and consistent scale; it allows the performance of sensitivity (or “what-if”) analyses to determine the effect of changing the scores on specific attributes of a project. Some of the disadvantages of AIM are a false sense of objectivity and a verbalization of priorities by the regulating bodies that may have political ramifications and affect the chances of implementing the model.

A Multiattribute Utility Model for Incorporating Risk in Fire Management Planning

Abstract The potential variability of investment returns often perplexes the investment selection process. On national forests, the relative efficiency of alternative fire management strategies is investigated using a modification of traditional cost plus loss methods. Current procedures do not indicate potential variability in the performance of these alternatives, however. A multiattribute utility model is presented as the appropriate framework for evaluating alternatives when decision-makers are concerned with the risk and efficiency implications of selecting a particular strategy. Among fire management planners, the term risk usually relates a perception of the variability of potential damages. A risk measure is defined and an index is created to indicate the performance of each fire management strategy relative to the potential performance of a standard established as a benchmark. Utility functions for the risk attribute were assessed using hypothetical lotteries where the value (utility) of risk was a function of its deviation from a risk benchmark. Two-attribute utility functions indicated surveyed respondents' preferences for attaining risk and efficiency related objectives. Analysis yielded two conclusions: (1) that fire management planners are not risk-averse, and (2) including risk as an attribute of the decision problem can significantly influence investment selection. Forest Sci. 32:1032-1048.

Explicated models constructed under time pressure: Utility modeling versus process tracing

Two multiattribute utility models, the additive (AMAU) and the multiplicative (MMAU), were compared to the average process model (AP), a variation on the concepts used in the process tracing approach. The principal difference between the AP and the utility models is that the AP uses introspection as part of its modeling technique. The models were compared in the context of judging severity of heart problems. Seven experts, working under time pressure, served as subjects. All three models were constructed in 45-min interviews with the experts. Comparisons gave varied results, some clearer than others. The AP model was more accurate than the AMAU model. On one data set the AP and the MMAU had similar accuracy rates, whereas on another the AP was clearly more accurate. Despite variations, these results suggest that anticipatory introspection improved the predictive accuracy of utility models constructed under time pressure.

Risk Preferences for Gains and Losses in Multiple Objective Decision Making

Payne, Laughhunn, and Crum (Payne, J. W., D. J. Laughhunn, R. Crum. 1984. An experimental study of multiattribute risky choice. Management Sci. 30 1350–1361.) found that managers were multiattribute risk averse for gains, but multiattribute risk prone for losses, a pattern that is inconsistent with both the additive and the multiplicative multiattribute utility models. In this paper we develop the reference risk-value (RRV) model, which is simple in structure yet capable of representing the kinds of multiattribute reference effects observed by Payne et al. We also report the results of two experiments that compare the descriptive validity of the RRV model with that of the additive and multiplicative utility models. Experiment 1 involved choices between risky multiperiod cash flows; Experiment 2 choices between risky job alternatives described by change in salary and change in type of work. In Experiment 1, subjects were multiattribute risk averse for gains, but multiattribute risk neutral for losses. In Experiment 2, subjects were multiattribute risk averse for both gains and losses, but significantly more so for losses. Because both experiments produced significantly different multiattribute risk preferences for gains than losses, both favor the RRV model over the widely used additive and multiplicative models. However, because the patterns of multiattribute risk preferences for gains and losses were strikingly different in the two experiments, these results argue against any direct generalization of the “reflection effect” to a multiattribute context.

Multiattribute utility models and the tenure process

AbstractA formal decision model is used to develop a decision support system to assist faculties in making personnel decisions that are systematic, explicit, consistent, and retraceable.

A decision theoretic methodology for severity index development.

Severity indices are a key element in evaluation of health programs. But the methodologies used in development of such indices are rarely clearly defined. Moreover, there have been no tests of transportability of the index development methodologies that do exist, so there is no guarantee that panels of physicians convened by different facilitators would produce indices with similar performance characteristics. This paper describes a replicable index development strategy, a heart disease severity index developed by using that methodology, and a test of that methodology's transportability.

Influence of Fishermen's Preferences on the Success of Commercial Fishery Management Regimes

Abstract Fishermen may opt for goals of stock rebuilding and stable yield when fish stocks are depleted and the short term outlook for catches is poor but reject these goals when a good year class appears. The purpose of this study was to determine whether this apparent ambivalence among fishermen is a consequence of changing preference for short and long term returns or can be accounted for by the change in short term pay-off associated with good recruitment. I derived realistic short and long term pay-offs by exploiting a population model of the Gulf of Maine herring (Clupea hareugus) stock under two regimes. The goal of one regime was to obtain long term stable yields while the goal of the other was to take windfall yields from occasional good year classes. I used a multi-attribute utility model to determine if there was a set of preference weights for short and long term returns that would cause fishermen logically to choose the stable yield regime when the short term outlook was poor but to choose th...