Publication and archiving of research data

Which data should be published?

For publication and long-term archiving, it is advisable to select data that is necessary for replicating the research project and that will retain its value after the completion of the study. Particularly valuable is data that would be costly, difficult or impossible to collect again, such as data from long-term cohort studies or interviews with participants who may no longer be accessible to researchers in the future.

If the data contains personal information, this does not automatically mean that it cannot be published in any form. In many cases, data can be anonymised or pseudonymised so that it can be shared under certain conditions. It is worth considering whether this is feasible and whether the value of the data would be preserved in such a form. Properly anonymised data can be published with open access, which is preferable to leaving the data unpublished.

Anonymisation does not always mean simply removing names or identifiers. Data may contain several types of identifying information.

• Direct and indirect identifiers. Direct identifiers (name, personal identification code, email address, phone number) must be removed. With indirect identifiers, identification may be possible through a combination of characteristics, for example, a rare occupation combined with advanced age. In such cases, the relevant characteristics should be grouped into broader categories: age can be replaced with an age group, and occupation with a field of activity.
• Responses to open questions. Free-text responses often contain unintentionally identifiable information. Respondents may mention their workplace, place of residence, or a life event that makes them recognisable. Such responses should be reviewed and, where necessary, edited or removed.

Where to publish data?

For researchers of the University of Tartu Faculty of Social Sciences, DataDOI is the recommended repository. It is a dedicated research data repository managed by the University of Tartu Library and is available free of charge to researchers in Estonia. DataDOI is based on the widely recognised open-source Dataverse platform, assigns datasets a persistent identifier (DOI) through the DataCite Estonia consortium, and makes them discoverable through a range of international data catalogues.

DataDOI supports the DDI (Data Documentation Initiative) metadata standard for the social sciences. After uploading a dataset to DataDOI, researchers should additionally complete the social sciences and humanities metadata section. These fields comply with the international DDI standard for describing and documenting research data and help to improve the findability and reusability of datasets. Detailed guidance is available in the DataDOI User Guide.

Image

International subject-specific and general-purpose data repositories are suitable for storing data from international research projects, or where a funder or publisher requires the use of a preferred repository. Useful resources for selecting an appropriate repository include the Re3data.org registry of research data repositories and the General Repository Comparison platform.