A Model of Error Propagation in Satisficing Decisions and its Application to Database Quality Management

This study centers on the accuracy dimension of information quality and models the relationship between input accuracy and output accuracy in a popular class of applications. Such applications consist of dichotomous decisions or judgments that are implemented through conjunction of selected criteria. Initially, this paper introduces a model that designates a single decision rule which employs a single binary conjunction operation. This model is extended to handle multiple, related decision rules that consist of any number of binary conjunction operations. Finally, application of the extended model is illustrated through the example of an online hotel reservation database. This example demonstrates how the new model can be utilized for ranking and quantifying the damage that errors in different database attributes inflict. Numerical estimates of the model can be integrated into cost-benefit analyses that assess alternative data accuracy enhancements or process or system designs.