Predicting the Strength of Preference for Labor Contracts Using Policy Modeling

Abstract

This study tests the predictive accuracy of policy models that describe the cognitive strategies used when subjects evaluate labor contracts. Subjects rated labor contracts that varied in wage offered, job security, and profit sharing plans. Profit sharing plans varied in two dimensions: the time when the profit was received and level of stability that was known about the company′s ability to maintain a profit. All the subjects appeared to average the subjective values of the contract features; however, subjects differed in the way they weighted the outcomes. One week later, subjects returned to choose between labor contracts and to indicate how much they preferred each choice. The preference orderings that were predicted by the policy model for the majority of the subjects were fairly consistent with the observed preference ratings. However, in the preference task, the observed preference ratings for the acceptable contracts were stronger than predicted by the policy model. Furthermore, contracts that were rated as least attractive were rated in the preferences task as more similar to each other than predicted. For the majority of subjects, these results indicated that policy modeling would be a useful method of representing information about an individual′s point of view to predict how they might compare contracts when asked to trade off the features. One small group of subjects rated the contracts based almost exclusively on wage during the rating task and then used wage, security, and profit sharing when comparing the contracts. The advantages of the policy model as a method of anticipating the value of labor contracts in the preference task are discussed as an approach to studying the resolution of conflict between two parties.