Simulations of the Relationship Between an Information System's Input Accuracy and Its Output Accuracy

The economic consequences of data quality are partly determined by the relationship between the quality of the input data and the quality of the information that an information system outputs. However, the relationship between an information system’s data accuracy and its output information accuracy is hard to assess. A popular belief on this issue is reflected by the saying “garbage in garbage out”, namely, the accuracy of the output of an information system is positively and tightly linked to the accuracy of its input. Yet, this belief has not been validated. The relationship between an information system’s input accuracy and its output accuracy is the subject of this paper. A traditional assumption in research of the relationship between input accuracy and output accuracy is that errors are independent. Motivated by a belief that dependence between errors is, in fact, common, this research examines the potential effect of the dependence factor on the relationship between an information system’s input accuracy and its output accuracy. The main research questions are: (1) How does dependence between errors affect the relationship between input accuracy and output accuracy? (2) Is the relationship between input accuracy and output accuracy positive? An earlier paper [1] analyzed these questions applying the information structure model, rooted in statistical decision theory. The inquiry in [1] was limited, however, to an information system that uses a single input for calculating the output. This paper extends the scope of the investigation to situations in which an information system produces output using multiple rather than a single input. The inquiry applies simulations for that purpose.