What Are Research Data?

 

Definition

Research data are data that[…] are generated as part of scientific activities, including source historical research, experiments, measurements, surveys, et cetera, and that can be digitally and electronically stored.

DFG, 2010

The research data types, formats and structures vary considerably across disciplines. Research data are also called primary data or raw data.

 

Sources

  • Observations: realtime data, survey data
  • Experimental data: lab measurements, chromatograms
  • Simulations: simulation measurements, model measurements
  • (Parts of) software programs
  • Citations or references: collections of data sets already published
  • Other sources: research reports, project data files, correspondence, proposals and applications
 

Formats

There is a broad range of research data types and formats:

  • Models: statistical models, 3D models
  • Multimedia data: JPEG, TIFF, MPEG
  • Numerical data: Excel, SPSS
  • Software/programming languages: Java, C++
  • Text: Word, PDF, XML
 

Examples

  • Documents, tables, data files, lab documentation, videos, audio recordings, workflows, methods, presentrations, transcripts
 

The Domain Model

Das Curation Continuum (Domänenmodell) nach WissGrid WissGrid

Within a research project, there are typically several domains or work environments. They are differentiated, for example, by the means of data exchange, the groups or persons that may access the data, and the type of use.

  • The private domain is the work environment of an individual researcher.
  • The shared domain is the work environment of a research group.
  • The permanent domain is the work environment created for long-term archiving purposes.
  • The domain "Access and Follow-up Use" is the work environment open to researchers world-wide, across all disciplines and sectors.

Each research project requires the use of the first three domains over its lifetime.

It is important that the migration between domains functions as desired. This requires careful planning. The data management plan facilitates a secure migration between domains.

  • Already in the private domain of the individual researcher, the basis for a later data transfer to other domains should be created.
  • For the transition to the group domain, rules for the joint generation and use of research data are required.
  • In case that a long-term storage of data is desired and publication is planned, additional information to facilitate understanding and follow-up use of the data across disciplines should be provided.
  • Importantly, the data could be relevant for a number of different research contexts. Frequently, data generated in one discipline today provides the basis for tomorrrow's research in another discipline.