Evolutional Data Quality: A Theory-Specific View.

Abstract

A data evolution life cycle typically consists of four stages of data activities: collection, organization, presentation, and application. During the cycle, data evolve as per the needs and specifications of a theory. In this paper we propose a concept of theory-specific data quality (DQ) that stipulates DQ is defined and measured as the extent to which data meet the needs and specifications of a theory. Depending on when a theory is applied during data evolution, we define DQ respectively at collection, organization, presentation, and application levels. We derive measurement attributes based on a fishbone cause-effect diagram. We compare our models with existing ones and point our future research directions to further validate and apply the proposed measurement models.