Just in case you thought we were overreacting in our preceding post, about electronic records being an imperfect solution for long-term data and document storage, consider the following current example.
Dr. Tim Vines, a visiting scholar in zoology at the University of British Columbia, set out to test the availability of biology research data over time. With a group of colleagues, he tried to track down the original data from 516 articles published between 1991 and 2011. The data sets were still available for the 2011 papers. But thanks to the types of changes in data storage technology that we pointed out, as well as outdated contact information for researchers, the availability of data sets dropped off rapidly year by year. By the time the researchers got back to 1991, only 20 percent of the original data was still available.
(The work is in the current issue of the journal Current Biology. The abstract is available, but the full article is behind a paywall.)
In many cases, Dr. Vines told to the Canadian Press, individual researchers are the only ones maintaining the original data. But over time people move, retire and die, in a kind of grim ladder of increasing unavailability. And even when the original researchers or data can be found, Dr. Vines told the CP, because of the effort it takes to retrieve data from outmoded storage devices, researchers “are not going to go through the length to get you the data, so the data’s effectively unavailable – it’s effectively lost.”
In a press release from UBC, Dr. Vines points out that “The current system of leaving data with authors means that almost all of it is lost over time, unavailable for validation of the original results or to use for entirely new purposes.”
He’s advocating that the data be archived online, by the journals that publish it. That’s fine, and doubtless a step up from keeping potentially crucial data on old floppy disks in the bottom of a drawer. But many journals are published on a shoestring by scientific societies, or by large international conglomerates that are in the process of having their once-profitable business models undercut by open access journals. They may yet outlast the length of a single academic career. But permanent? Let’s hope someone thinks to run off a couple of printouts, too, and store them in a library or archive. If one can still be found standing.
– Michel Hayden
The Decline and Fall of Practically Everything 