7AAVIN01 Chap.10 Metadata Quality, Ethics and Linked Data
Metadata Quality, Ethics and Linked Data
Metadata Quality, Ethics and Linked Data closes the metadata block and opens Week 4. It sets out five quality criteria, accuracy, completeness, provenance, timeliness and accessibility, and applies them to a record. It works through the ethics cases of Activity 3, from a replaced subject heading to location data that exposed a person, and offers a decision path for problematic terms.
It shows metadata at work beyond universities, describes the optional OpenRefine and Europeana activities, and introduces the Week 4 question of linking records across heterogeneous collections. These ideas give a critique its sharpest tools. Quality criteria turn a vague judgement into specific findings, and the ethics cases show that description involves choices about naming, exposure and bias that a record should document.
The optional activities extend both: cleaning a museum's metadata with OpenRefine shows that every fix is a change to be recorded, and exploring the Europeana Data Model shows what may be lost when very different collections share one model. The Week 4 overview then frames linked data as a set of open questions about connecting heterogeneous collections.
A worked example audits an invented collection record against two criteria, and a decision path shows one defensible way to handle a harmful term while keeping it findable.
What this chapter covers
- 01
Accuracy and completeness
- 02
Provenance of values
- 03
Timeliness and accessibility
- 04
Ethics of description cases
- 05
Rewrite, annotate or keep
- 06
Metadata beyond the academy
- 07
Cleaning data with OpenRefine
- 08
From records to linked data
Worked example · free
Apply the five quality criteria
- 1Machine tagging without a note of its origin is a provenance issue; record that a tool supplied the value.
- 1An unsourced date raises accuracy, since no one can check it against a reliable source.
- 1Missing alt text is an accessibility failure for people using screen readers.
Key terms
- Provenance
- The record of where a metadata value came from, such as a volunteer, an expert or a machine.
- Timeliness
- Whether a record's terms are still current and how quickly records are updated and shared.
- Linked data
- An approach that connects records across collections by using shared web identifiers.
- Europeana Data Model
- The shared model Europeana uses to aggregate records, separating objects from their digital representations.
- Accessibility
- The quality of a record being understandable and usable by people, including through alt text.
Metadata Quality, Ethics and Linked Data FAQ
What are the five criteria for metadata quality?
Accuracy asks whether values are correct and well sourced, completeness whether the essentials are described for the audience, provenance where values came from, timeliness whether terms are current, and accessibility whether people can understand the record.
How should a cataloguer handle a harmful historical term?
The lecture offers three options: rewriting it, annotating it, or retaining it with an explanation. A quoted title may need to stay as evidence with a note, while a cataloguer's own subject term can be replaced and the change recorded.
What is the Week 4 question about linked data?
Week 4 asks how effective the Semantic Web and Linked Data have been in connecting knowledge systems, how they work in practice, and what problems arise when content from very different collections is integrated.
Why does provenance matter for metadata quality?
Values can come from volunteers, experts or machines, and each source is acceptable if the record says which it was. Without that note, a reader cannot judge how far to trust a value, and later corrections cannot be traced.
What does Europeana's data model separate?
The Europeana Data Model separates the cultural heritage object itself from the digital resources, such as images or web pages, that represent it. That distinction answers the opening question of Week 3 about which object a record describes.
Why does metadata matter outside universities?
Industries depend on it. Music royalties rely on recording and song codes, bookshops on ONIX records, search results on schema.org markup, shared links on Open Graph tags, and streaming services on detailed descriptive tags, so missing or wrong fields cost people money and visibility.
Assessment move
Audit one real collection record from a museum or archive website against the five criteria and propose two improvements, as the optional collection audit suggests. Write the audit as a short paragraph with each criterion named. It doubles as a draft for the limitations section of a critique. Rehearse one ethics case aloud, explaining what you would change in the record and why, because examiners often test judgement this way.
Then look up an authority identifier for one person or place in your material and note its web address; that single step shows you understand how linked data begins.