Data loss

One of the biggest problems that can happen when you work with research data is data loss. If a data file without backup copies becomes corrupt, chances are that the data may be lost. But loss of data can also happen if no-one understands the data any longer, or if they cannot be read.


Even though current technology offers automatic backup and allows for thorough data rescue actions, data loss is still a common complication. Data which are stored in one single place are particularly at risk. If your recorded interviews are saved only on your digital recorder, if your field measurements or notes exist only on paper, if your dataset is stored only on your laptop, or the database is saved only on a flash drive, the project data can be lost. The disaster is only a mistake, a hardware error, or a burglary away.

Every project needs to have a plan for secure storage of data and other material throughout the project process. You also need to explore the conditions for backup in your selected storage solution. Most universities offer some form of local storage with regular backups, but that may not be the case for all storage solutions. You cannot guarantee that the Box solution offered by some universities is secure; the data could still be stored on a server that isn't backed up. And if that server crashes, the data could be gone for good. (Cloud services, in general, don't provide a backup guarantee.)

Regardless of which storage solution you use, you should consult with your local IT Services about their recommendations for backup and which requirements are placed on the data. (See Protect the data for more information.)


Our memory is not as good as we think it is. Details about data that were collected at the beginning of a project may be completely forgotten towards the end of it – after a few years you may have forgotten so much that the data are useless. If you document the data properly, you minimise the risk of losing data because you have forgotten what they mean and how they can be analysed.

It is equally important to document the changes you make between different file versions. One version of a data file might be destroyed, but if you have documented the differences between previous versions, you may be able to recreate a lost version from earlier versions of the file. Documentation takes a bit of time, but it's better to do that now, than to try to remember days or weeks of work later.

Secure formats and software

File formats and software may be fairly secure in the short term, but they are a growing risk factor over time. To avoid data loss due to having saved your files in poorly documented file formats, you should also make copies in archival formats. Doing so will also protect you from software that later can no longer be used, or that can no longer read previous file formats (not backward-compatible). (See Software for more information.)