According to the Organisation for Economic Co-operation and Development (OECD), "research data" are the factual records: numerical data, textual data, images and sounds. These documentation are used as primary sources for scientific research and are accepted by the research community as a means of validating conclusions. A research dataset is a systematic and partial representation of the subject or object studied.

Research data does not include laboratory notebooks, preliminary analyses, article drafts, future development plans, peer reviews, personal communications with colleagues, laboratory animals, specimens or strains.

It is important to make accessible open research data. This is one of the pillars of the Open Science movement, which aims to make the entire process and results freely available to society. Open data has undoubted benefits, such as:

• Promoting transparency and integrity in research.
• Encouraging new scientific discoveries through the re-use of data.
• Providing a return on public investment in research.
• Integration into international networks of data repositories.
• Transparency and accountability of public funding.
• Commitment to society and the advancement of science.
• Overcoming inequalities between countries and institutions that cannot afford to pay for accessing scientific information.
• Enabling research to be reproducible and credible, as it has the potential to expose fraud and misconduct.
• Increasing the visibility and the impact of the research.
• Encouraging collaboration.
• Avoidance of duplication of research efforts.
• Preservation of data for future use.

It is very common that three types of data will exist before the research ends:

• Raw data: data obtained and/or created during the research.
• Processed data: data extracted or derived from the raw data.
• Referenced data: a subset of the processed data that analyze and draw conclusions.

Choosing which data to keep depends mainly on how the research will be used:

• Verification of results: data are used to facilitate transparency and replication of the research. In this case, it is recommended to preserve both: processed and referenced data.
• Re-use of data: data can be used in future research, both by the same researcher and by others. In this case, it is recommended to preserve the raw data.

In both cases, the documentation accompanying the data must be sufficiently clear to meet the above criteria.