Multi-user test of the data quality matrix for product life cycle inventory data

Abstract

The data quality matrix for product life cycle inventory data proposed inWhdkma &Wlsnaus (J. Cleaner Prod. (1996), 4: 167-174) was subjected to a multi-user test, in which 7 persons scored the same 10 datasets representing 10 different processes. Deviations among scores were listed, and the causes for deviations were determined and grouped into a limited number of well-defined classes. For the majority of the scores, the different test persons arrived at the same score. Deviations occur most often among neighbouring scores. Only a smaller number of the deviations (less than 10% of all scores) affect the overall assessment of the data quality and/or uncertainty of the corresponding dataset. Based on the analysis of the causes of the deviations, improvements to the matrix and its accompanying explanations were suggested and implemented (reported in the appendix to this paper). The average time consumption for the scoring by the different test persons was less than 10 minutes per data set. It is concluded that the time consumption and the number of deviating scores can be kept at an acceptable level for the pedigree matrix to be recommended for internal data quality management and for comprehensive communication of quality assessments of large amounts of data.