Metadata Quality Assessment

Bruce and Hillmann's continuum identifies seven dimensions: completeness (are the needed elements present and populated?), accuracy (do values correctly describe the resource?), conformance to expectations (does the metadata meet community and user needs?), logical consistency and coherence (are values internally and externally consistent?), timeliness (is the metadata current with the resource?), accessibility (can people and machines read and use it?), and provenance (is its origin and processing history known?). In practice, accuracy, completeness, and consistency are the three most widely operationalized, as Park and Tosaka found, because they are central to discovery and relatively tractable to measure.

Completeness asks whether the metadata has the elements it should — whether required and recommended fields are populated at all. Accuracy asks whether the values that are present are correct, that is, whether they truly describe the resource. A record can be complete but inaccurate (every field filled, but with wrong values) or accurate but incomplete (the few values present are correct, but key fields are empty). Good assessment measures both, because high completeness can mask systematic errors, and high accuracy on a sparse record still leaves resources hard to find.

Bruce and Hillmann argue that there is no absolute, context-free standard of metadata quality: the same record can be excellent for one purpose and inadequate for another. A field that is optional for browsing may be essential for a specific service or for interoperability with a partner. So assessment fixes the schema, application profile, and the uses the metadata must support, and judges completeness, accuracy, conformance, and the rest against those expectations. This is why quality scores are tied to context and why the same collection can warrant different assessments when its intended use changes.