Research Data Management: Quality, Organisation and Documentation

Quality control measures can improve the quality of the data collected. For example, by ensuring accuracy and consistency. Suitable measures should be in place before data gathering starts. The UK Data Service has a one page guide to ensuring data quality. 

Sensible file names and well-organised folder structures:

• Make it easier to find and keep track of data files.
  • It saves time... ("Where did I put it again?" or "Why o why did I call all these files data1, data2, data3... Which one has the data I need!")
  • Makes it easier for BU to find and identify data and documentation it is ultimately responsible for, long after projects have finished or individuals have left BU. 
• Improve the security of the data by:
  • Keeping sensitive data separate from anything else that can be shared more widely.
  • Making it easier to keep track of retention or disposal of data.

The UK Data Service has guidance on file naming and structure.

Data documentation is essential for data to be understood and re-used. Documentation is needed at a study level, and at data level. 

Documentation should be written during the course of the research project. Planning to capture this information (and sometimes assigning responsibility to do so), can prevent inaccuracies and wasted time trying to recall details much later on.

Study level documentation provides context about the research project. It is anything that's needed for someone else to understand exactly what the study is about and how it was conducted. This would be useful for other members of a research team, or for individual researchers coming back to the data after a long time. It is essential for anyone else with whom the data is shared so they could use the data in their own studies with confidence.

• Some of this will be recorded in the Data Management Plan. For example, details of the data collected, the methods, how the data is organised, quality control measures etc.
• Some of it though is likely to be documented separately. For example, a detailed record of interview procedures, instrument calibrations etc.

Data level documentation provides information about individual datasets. These might be embedded within a file (such as variable descriptions in an SPSS file) or in a separate READ ME file). Data level documentation will help others make sense of the data. For example, an explanation of what headings mean in a spreadsheet.