Intrinsic data quality dimensions: expanding on Wand and Wang’s data quality model

Purpose: Studies show that data quality (DQ) issues are extremely costly for companies. To address such issues, as a starting point, there is a need to understand what DQ is. In his context, the 1996 paper “Anchoring data quality dimensions ontological foundations” by Wand and Wang has been highly influential on the understanding of DQ. However, the present study demonstrates that some of the assumptions made in their paper can be challenged. On this basis, this study seeks to develop clearer definitions. Design/methodology/approach: The assumptions behind Wand and Wang’s DQ classification are discussed, on which basis three counter-propositions are formulated. These are investigated through a representation theoretical approach involving analyses of deficient mappings between real-world and information system states. On this basis, an intrinsic DQ classification is derived. A case study is conducted to investigate the value of the developed DQ classification. Findings: The representation theoretical analysis and the case study support the three propositions. These give rise to the development of a DQ classification that includes three primary intrinsic DQ dimensions (accuracy, completeness and conciseness), which are associated with six primary value-level DQ deficiencies (inaccuracy, incorrectness, meaninglessness, incompleteness, absence and redundancy). The case study supports the value of extending Wand and Wang’s DQ classification with the three additional data deficiencies. Research limitations/implications: By improving the conceptual clarity of DQ, this study provides future research with an improved basis for studies and discussions of DQ. Originality/value: The study advances the understanding of DQ by providing additional clarity.