Understanding impartial versus utility-driven quality assessments in large datasets

Establishing and sustaining very high data quality in complex data environments is expensive and often practically impossible. Quantitative assessments of quality can provide important inputs for prioritizing improvement efforts. This study explores a methodology that evaluates both impartial and utility-driven assessments of data quality. Impartial assessments evaluate and measure the extent to which data is defective. Utility-driven assessments measure the extent to which the presence of quality defects degrades utility of that data, within a specific context of usage. The quality assessment methodology is empirically assessed using real-life alumni data - a large data resource that supports managing alumni relations and initiating pledge campaigns. The results provide important inputs that can direct the implementation and management of quality improvement policies in this data repository.