Descriptive metadata in DDI format
DDI started as a collaboration of data repositories
One of the most important prerequisites for data archiving and reuse is sufficient documentation of data, i.e. metadata. An international documentation format for research data was agreed upon already in the 1970s, but due to differences in documentation needs and practices, many local “dialects” of data documentation formed.
In 1995, however, ICPSR (Inter-university Consortium for Political and Social Research) founded a committee to develop a standard for the documentation of data in the social sciences. The committee included members with broad expertise in social-scientific research and documentation. The committee’s proposal for the new standard was named Data Documentation Initiative (DDI).
The DDI standard is maintained and developed by the international DDI Alliance hosted by the University of Michigan. Members of the DDI Alliance include data repositories and universities from all around the world as well as organisations developing computer programs for statistical analysis and data collection. Each member has a representative in the Alliance’s Scientific Board, which convenes once per year. DDI is open for use by anyone, but only members of the Alliance can contribute to the development of the standard. Earlier documentation standards, such as MARC, ISO690-2, SDMX and Dublin Core, were used as reference in the development of DDI.
XML format used for DDI
Data that are described using the DDI standard are documented in XML format. XML (Extensible Markup Language) is a file format meant for saving and presenting structured information on the internet. XML offers a hardware- and software-independent method for managing information. In a single XML document, the allowed elements (sections of the document e.g. titles, lists etc.) and their relations, order and repeatability are defined in the Document Type Definition (DTD) or schema.
DDI-Codebook used at FSD
All data archived at FSD have been described using the DDI-Codebook format, which is suitable for the long-term preservation of the metadata of individual survey datasets. The elements of DDI-Codebook (approx. 300 in total) can be divided into five sections:
- Document Description
- includes e.g. bibliographical information of the metadata (i.e. the “codebook”)
- Study Description
- includes e.g. dataset creators, keywords, abstract, sampling procedures, data collection, units of observation, target population, terms of access
- Data Files Description
- includes e.g. data structure and format, number of variables, size of files, software
- Variable Description
- includes e.g. variable and value labels and question texts
- Other Study-Related Material
- includes information related to the data that is not described elsewhere
FSD uses about a third of the elements of DDI-Codebook. All information elements are available in the Tag Library, along with guidelines for use.
DDI enables efficient discovery
Structured XML data descriptions aligned with the DDI standard significantly improve the findability of data. In addition, the XML files can easily be processed into different types of web and print publications.
DDI-Lifecycle for managing the data lifecycle
DDI-Lifecycle offers solutions for the documentation and long-term preservation of comparative, panel and series data. In comparison with DDI-Codebook, DDI-Lifecycle has a much more extensive structure.