Does the measure of information quality influence survival bias?

Abstract

Examines certain issues related to application of total quality management to the information systems setting, in particular to the measurement of output quality. Data produced by a system are often transformed for decision making by end users. Although ensuring the quality of the data can be achieved, it is difficult to control for all the inferences made by end users. Using a financial reports database as an example, demonstrates how the choice of measure of information quality can sometimes bias the survival time of firms in the sample. The analysis uses the Cox proportional hazards model to demonstrate that when immaterial data errors are present after firms have incurred a loss, sometimes they increase their survival chances. Such consequences are valid in other settings as well and demonstrate that defining and measuring the quality of information systems needs more considerations than is generally assumed.