Data Accuracy

From Open Risk Manual


Data Accuracy denotes the degree to which data correctly describes the "real world" object or event being described. Correctness may refer to the closeness of values in to the true values where closeness may numerical or conceptual in nature depending on the data type [1]

Accuracy implies agreement between the true value and the average of repeated measured observations or estimates of a variable. An accurate measurement or prediction lacks bias or, equivalently, systematic error. A measure of data accuracy would be the percentage of data entries that pass specified data accuracy rules


Plausibility checks aim to detect accuracy issues in the dataset. This is accomplished by reviewing the variables concerned with statistical approaches or domain knowledge

Issues and Challenges

While other Data Quality dimensions can be assessed by analysing the data itself, assessing accuracy of data can only be achieved by either

  • Assessing the data against the actual phenomenon it represents
  • Assessing the data against an authoritative reference data set (Golden Source)

See Also


  1. ECB, Supervisory Data Quality Framework, 2016