Collect and describe your research data
Before you collect data, plan how to name and organise your files, how to describe your data (which metadata to collect), and which file formats to use.
Also, consider whether you need any permissions, for example from the Swedish Chemicals Agency or the Swedish Radiation Safety Authority. You may also need permission if you are studying endangered species, taking samples in protected areas, using laboratory animals, or doing studies that involves human beings as research subjects.
If you undertake contract research or a research collaboration with industrial or academic partners, it is important that all agreements and research contracts are in place before you collect data.
Shortcuts to the content
- File formats
- Organising data
- Collection of personal data
- Ethical reviews
- Animal experiments
- Contract research and research collaborations
Save your files in an open or standard preservation file format rather than a proprietary file format. It can be simpler to plan to use the correct format from the start, as it can be difficult to convert to preservation file formats later when you need to share or store your data.
It is not always possible to use a file format that meets these criteria. For example, some disciplines have standards or traditions that determine the choice of format. Certain types of scientific equipment, analytical tools, or self-written software can also determine the file format. When planning to reuse or share your research data, it is useful to include information in the form of metadata about how your data can be read (for example which software to use) and to include any self-written software necessary with your data.
The Swedish National Data Service (SND) can advise on accepted and recommended file formats.
Organise your research data using carefully considered folder and file structures. Give your files and folders names with an appropriate hierarchy. Names should provide cues to the time, place, content, and other relevant information. It is best to have a simple version control strategy in place to keep track of different versions of your files, for example by recording the date in the file name or by imposing other forms of version control.
Describe your data in a structured way and provide all necessary metadata (data about data). Scince it is difficult to generate metadata retrospectively for large quantities of data, make sure this is something you do from the very beginning.
Metadata falls into three groups:
- Descriptive metadata
- Structural metadata
- Administrative metadata
Descriptive metadata enables others to search for your data, and includes the title, author, abstract or summary, geographical area, and DOI (Digital Object Identifier).
Structural metadata enables others to use your data, and includes the number of variables, number of research subjects, file structure, or internal file structure.
Administrative metadata enables others to understand and use your data, and includes regulatory requirements, any licences or embargos, technical information (the programs needed to open the files), and file formats and versions.
It is important to distinguish between metadata and documentation. Documentation in the form of articles and reports can only be read by people. Metadata must be machine-readable, which places specific demands on the structure, format, and standardised vocabulary. Study-level metadata can be stored in a README file or similar, while data-level metadata should accompany individual folders or files.
Data made available in databases should be organised and described according to current standards, which are often determined by the choice of repository or database.
When the data you collect include personal data there are special regulations to follow, in particular ethical guidelines and the General Data Protection Regulation (GDPR). Before you collect data, determine what applies in your case.
All research projects that involve personal data must be registered with Lund University, because the university is required by law to maintain a register of all processing of personal data at the university. If Lund University fails to do so it may be fined. Research projects should be registered in PULU (Personal Data Lund University).
In research projects where data is collected directly from research subjects or informants, informed consent should be obtained. Where data collection may affect subjects physically or if the data contains sensitive personal data, the research must be subject to an ethical review by an ethics committee before the research begins.
Will your research involve human studies, biological material taken from a human being, or the processing of sensitive personal data? If so, you may need to apply for an ethical review. As of 2004, it is forbidden to embark on certain types of research without approval from an ethical review board.
Will your research involve the use of laboratory animals? If so, you must obtain in advance the necessary permits and approvals to perform animal experiments.
If you undertake contract research or a research collaboration with industrial or academic partners there must be a research contract or agreement in place to regulate, for example, the right to results, confidentiality, and publication. The research contract will also set out the partner or organisation responsible for the long-term safekeeping of research data and other research material when the project is completed.
Each individual researcher, doctoral student, student, or other participant in the contract research or research collaboration must sign to indicate their agreement to the terms before the full research contract for the contract research or research collaboration is signed.
Responsible for research data at the Biology Library and the Geolibrary.
anja [dot] zimmerman [at] science [dot] lu [dot] se
Phone: +46 76 139 17 01
Responsible for research data at the Library of Chemistry and Chemical Engineering and the Physics- and Astronomy Library
kurt [dot] mattsson [at] science [dot] lu [dot] se
Phone: +46 79 063 08 23