After the conference, I asked the IM&T RDS Support team if they would be happy to contribute some thoughts to this blog.
Q: CSIRO has been assigning Digital Object Identifiers (DOIs) for a while now. Can you describe how the minting of a DOI happens at CSIRO? Do you have any rules in place around what does and does not get a DOI?
A: We use the ANDS Cite My Data service and automate the assigning of DOIs. As the CSIRO Data Access Portal enables different access levels for deposited data collections only collections that are published to the Public have a DOI generated. All collections with Public data get a DOI.Q: You have done a lot of thinking about maintaining DOIs over time, and procedures for deciding when changes to a dataset represent a new version requiring a new DOI. Can you share some of your thinking about that?
A: Any changes to the data itself will automatically generate a new DOI (with the previous version still being accessible with a note letting people know that there is a later version). Any changes to the Attribution fields (Creator, Contributors, Title, Publication Year) in the metadata need to be checked by a Data Administrator for a decision about whether a DOI is to be retained or a new DOI generated (depending on whether typos are involved or complete changes to these fields).Q: At the conference, you said that researchers responded positively to the idea of getting a DOI and being cited, and that this was a 'carrot' for self-deposit of data. I wondered if you could elaborate on that, and say a bit more about why you think researchers have responded so well.
A: The DOI is familiar to researchers as they are issued for journal publications, and generally available from the “advanced online publication” stage. We think therefore that the possibility of getting a DOI for a data collection helps them see the data collection as a valid published citable scientific output akin to their or others' journal articles. We are also citing the as yet small body of evidence suggesting a link between sharing and citing data and citations to articles.Q: CSIRO researchers operate in a very different environment from university researchers. Do you think those differences will be significant in terms of how data citation feeds into things like rewards and promotions for researchers?
A: We are not sure that CSIRO researchers do operate in a very different environment to University researchers. Although our reporting metrics are directed by different authorities they are very similar to those that the University sector uses. Our researchers still operate within many of the same disciplines as university researchers, and there is significant collaboration between CSIRO and the university sector. Data citation and sharing practices seem to come from the discipline, rather more than from the institution, so until there is broad external practice and acceptance there may not be significant impact on rewards and promotions. This is why we promote the link between sharing data and enhancing citation rates.Q: Is there anything else you'd like to highlight about data citation at CSIRO?
A: Like other institutions, we are in the early days of promoting this practice. We are keen to share our ideas, as well as grab ideas from others.
Thanks to CSIRO for sharing their experiences so far - I look forward to carrying on this conversation over the coming months as Griffith starts to address some of these questions.