What do we mean by research data? In the KAFFE project, SND’s survey of researchers’ needs and the gap between policy and practice, it became clear that the term is used in different ways by different actors. That prompted SND to develop a broader and thoroughly considered definition, together with an explanation of how it was formulated.
During the work on the KAFFE project, it became clear that friction can arise when different actors use the same word in different ways. One such term is research data – a central concept for SND. There are several definitions from major actors, but they tend to differ from one another. In some cases, these differences can be significant, for example depending on the context, discipline, or type of data material. SND’s previous understanding of research data also proved to be too vague and context-dependent to fully encompass research data as a whole.
A definition must not exclude any research
Lisa Isaksson, Language Coordinator at SND, explains why it was important to develop a broad definition that is not tied to a particular context, discipline or type of data.
“SND’s definition of research data must not unintentionally exclude any research. Data can take many different forms, and we needed to identify the lowest common denominator for what can be considered research data. We also wanted to make visible aspects that have been implicit in other definitions, such as what ‘data’ actually are.”
Stefan Ekman, Senior Advisor at SND, says that it was important to explain why the definition is formulated as it is.
“When we began considering what definition might be appropriate, we quickly realized that few actors, if any, explain why they have defined research data in the way they have. We wanted to make clear how we had reasoned when formulating the definition. That is why we added explanatory discussion and justifications, with references to the theory and research we drew on.”
What data are and when data become research data
The definition consists of two parts: what data are, and what makes certain data research data.
“We have taken as our starting point that data become research data when they are considered to have scientific value,” Stefan explains. “This is based on the reasoning that the key question is not primarily what research data are, but when they are research data.”
“In this definition, we also do not limit the concept of research data to the material that the organization works with, as some other definitions do. The fact that SND only works with digital research data does not affect how research data are defined,” Lisa points out. “This makes the definition more robust and less dependent on how the organization changes over time.”
“Our definition also does not include a long list of examples of research data,” she adds. “However, there are numerous examples in the accompanying explanatory discussion, including a few words about research that does not use data.”
The definition
“Research data are representations of phenomena that, on the basis of scientific principles, are deemed suitable for use in scientific or scholarly analysis, or for otherwise providing the basis for, or validating, research results. They have been compiled, collected, created, or otherwise generated as part of the research process.”
The definition and the accompanying explanatory discussion are published on Zenodo: What we mean by research data. DOI: 10.5281/zenodo.18801001.