Databrary is a data library intended to foster sharing of data, specifically video, within developmental science. We are jointly-funded by NSF and NIH. Most behavioral scientists do not currently share data openly. Databrary seeks to reduce barriers to data sharing to make it the norm, not the exception.
Video enables use cases that extend beyond research
Unlike many forms of flat-file, text-based Data, video is, to a large degree, self-documenting. As a result, we anticipate that most Databrary Investigators and Affiliates will use Databrary for a variety of pre-research or non-research uses, including viewing videos for methodological details or procedural examples or to find excerpts for teaching or scientific presentations.
Since video data is inherently identifiable, all users must become authorized to access any of the shared data on Databrary. Databrary's access model is unique in that authorized investigators are granted user and contributor privileges to all shared data in the library.
The Databrary Investigator Agreement combines data use and data contribution.
Many data repositories or data sharing arrangements require that data use agreements be signed prior to accessing a specific dataset for a discrete research purpose. Most data repositories have a separate agreement associated with data contribution. The Databrary agreement is new and perhaps unique in enabling open access for both data use and contribution privileges. We believe that combining the agreements will help lower barriers to data sharing. This is also in keeping with Federal mandates to share research data and emerging practices related to openness in scientific research.
When is IRB approval necessary?
In signing the agreement, all Investigators promise to seek IRB approval whenever their use of Databrary requires it. But, prior IRB approval is not required to access Databrary's resources for non-research purposes. In short, Databrary does not require IRB approval in order to become an authorized investigator.
The kind of Databrary use cases that require a specific IRB protocol is a local decision. There are many "pre-research" and teaching use cases that may not require IRB approval.
Databrary contains Identifiable Data for which participants have given permission to share
Databrary contains metadata that are not sensitive, de-identified data, and identifiable data (e.g. images or videos with faces). All identifiable data in the library have been shared with the explicit permission of the people depicted.
In order to contribute and share identifiable data with Databrary, Investigators must get approval from the people from whom the Data were collected. Investigators promise to provide accurate information about the participants' desired release level. Importantly, people whose identities might be discovered must have given informed consent in order for their data to be shared with other researchers on Databrary.
Databrary has developed a Release Template that helps to clearly inform participants of the potential risks, should they agree to share their data. The template is one of the resources that Databrary has made available and may be adapted and added to any IRB protocol. However, a given Institution may deem other release language to provide equivalent protections.
Whether De-identified Data may be shared with Databrary without seeking explicit permission depends on the Investigator's IRB. Some IRBs allow De-identified Data to be shared without explicit permission of the person tested.
The Databrary agreement is intended to formalize the obligations by the scientists who will view, browse, download, analyze, or contribute data. Those obligations include proper use of the information, protection of participant identities, and seeking approval when a specific use case requires it.