Why share the data?15 Jun 2015
Research data should be shared. This view is widespread in academia and common among research funders. Regardless, most research datasets remain locked away. Sometimes data are held back for less than admirable reasons, but they can also be held back with good intent. For example, access to data is sometimes restricted in order to maximize opportunities for early career researchers to publish papers using the data. I believe this is instead counter-productive for ECRs.
Open data turns ECRs into valuable collaborators
One concern is that open data will attract tribes of competing researchers to publish results that you could have just as easily produced. In my experience, this doesn’t happen. Sharing the data instead turns researchers on your team into valuable collaborators, since their intimate knowledge of the data cannot be replaced by an outsider.
Open data creates healthy incentives to get work done
Even if the fear that someone else will beat you to the scoop is largely unfounded, I do think it creates an incentive to complete work in a timely fashion. It’s otherwise too easy to put work on the back-burner, where it often stays indefinitely.
Data sharing is now a neccessary research skill
Funding applications now typically require a concrete plan for how you will share the data. Five years from now, you will not just need this data-sharing plan, you will also be competing with research teams with actual experience sharing data. It is thus important that you develop and demonstrate your data-sharing skills now.
Sharing data is ethical
The data aren’t yours. They belong to your participants, who have consented for you to use them for societal benefit. Further, you didn’t pay for the data to be collected. Yes, your blood, sweat and tears are in the project, but you were compensated for this. Thus any otherwise-reasonable action that limits the social value of the data is an unethical one. In my opinion, as soon as the pre-stated aims of a research project have been addressed, the data should be made as widely available as possible.