Unit 6.0 Summary
Unit overview
Unit study time
- 10 minutes
Why does metadata matter?
Metadata helps us answer important questions about data, including what was measured, how it was measured, who or what was measured, and whether the data is suitable for a articular purpose. We often use this information intuitively when deciding how to analyse or interpret data, but recording it as metadata makes those decisions explicit, structured and reusable.
By documenting research in a structured way, metadata can improve understanding, support comparison across studies, identify new research opportunities, improve quality and validity, support reuse by others and our future selves, and enable machine processing and automation.
Who should create metadata?
All research projects should create some level of metadata, regardless of their size or whether the data will be shared. What metadata is required, and who contributes to its creation, depends on the size, aims and context of the research. Metadata should be created by those with knowledge of the research and data, with the level of documentation reflecting its intended use.
What metadata should be created?
Key metadata types
Population metadata
Population, universe and unit type metadata provide essential context about who or what the data relates to and who was eligible to be included.
Data collection metadata
Data collection metadata describes how information was gathered, including collection methods, instruments and processes.
Question and measurement metadata
Question and measurement metadata capture exactly what was asked or measured, helping users understand the meaning and purpose of collected information.
Code and category metadata
Codes, categories and classifications provide meaning to responses and help ensure values are interpreted consistently.
Variable metadata
Variable metadata brings together key information about the data used for analysis.
Concept metadata
Concepts describe the underlying ideas being measured.
Study and dataset metadata
Study and dataset metadata provide the broader context needed to understand where data originated, how it was created and how it can be accessed and reused.
Together, these metadata components provide the context needed to understand what data represents, how it was created and how it should be interpreted.
There is no single set of metadata that is appropriate for every project. The metadata you create should be determined by the nature of your research, the characteristics of the data, how the data will be used, and who will need to understand or access it.
When deciding what metadata to create, you should consider:
Whether the data and metadata will be shared, and with whom.
- The size and complexity of the project.
- The scale of the research and the number of people involved.
- The time and resources available to create and maintain metadata.
- The significance of the research and the likelihood that the data will be reused, compared with other studies, or incorporated into future research.
- Whether the research is a one-off study or part of a repeated or longitudinal programme.
- Any relevant metadata standards, schemas, repository requirements or community practices.
The goal is not to document everything possible, but to create metadata that is appropriate for the project's aims, users and context.
How should metadata be created?
Creating metadata is not simply about documenting individual items. It also involves documenting the relationships between metadata elements and deciding what information is most valuable to record.
Questions, variables, concepts, studies and datasets form part of a connected metadata ecosystem. Understanding these relationships improves consistency, reduces duplication and supports reuse. It also helps people and systems navigate between different levels of description, creating a richer understanding of research data and the context in which it was produced.
Effective metadata creation involves balancing effort and value, while ensuring sufficient information is captured to support the intended use of the data. Understanding metadata therefore requires understanding both the metadata itself and the relationships that connect it.
Feedback please
Congratulations on completing the training course! If you can find 10 minutes to complete this survey, we'd love your feedback in order to improve the course.