With the RDA10 Plenary Meeting in full swing, UdeM information scientist Vincent Larivière takes a look at what's happening in the global movement to share research data.
On the heels of the internet, a global movement of scientists has sprung up to share the research data they use. But while open access to data is increasingly discussed, there's no real consensus on how far it should go.
Data-sharing is widespread in certain fields, yet by no means common practice. “Across all fields as a whole, only a minority of scholars currently make their data available,” says Vincent Larivière, a professor at UdeM’s École de bibliothéconomie et des sciences de l’information.
Pressure is growing for scholars to share data, however. Some scientific publications, such as PLOS ONE and Nature, actually require that data be shared. “They're major journals, so they may cause others to follow suit,” sayd Larivière.
More research is put to use
By publishing data, scholars allow their research to be used by their colleagues. fostering further research . “Taken collectively, this leads to greater advances in human knowledge, which is good for everyone,” according to Larivière.
Why is the open-data movement not more widespread? “Scholars who spend years gathering data may prefer to keep it all for their own future research,” says Larivière. “Some also worry that data-sharing will create a generation of 'parasite' scholars who only use data gathered by others.” Crediting scholars whose data get reused may change that, however. “Some science journals have started establishing mechanisms to cite which data have been used, much like an index."
Validating data is key
Sharing also makes it possible for data to be analyzed anew, strengthing them through further validation. “Transparency can uncover weaknesses or errors in data-gathering,” says Larivière. “Scholars make mistakes too, but it is a lot harder to uncover mistakes if you only have access to the article and not the data.”
In some fields like biology and ecology, he adds, researchers who don't share their data lose a lot of credibility.
Shared data, re-used data?
But if more data get shared, will they necessarily be put to use? According to Larivière, scholars don't yet have the habit of poking around in the data of others. “There may be challenges associated with the way data are collected and with understanding what they say,” he points out. “If the person producing the data only provides a column of figures, it's easy to get lost. For the data to be able to be used again, to be comprehensible, some basic documentation has to be provided. Not all scholars want to spend time doing that.”
Reliability is a real issue as well. “Will scholars who are used to gathering data with their team in the field really trust someone else with them? Not necessarily. You have to think critically and not rely on any old data set.”
Funding bodies in the fray
Funding bodies, which play a key role in the research world, are starting to get involved. In the United States, the National Institutes of Health have made data-sharing mandatory, particularly in clinical and genome research.
“The United States produces three per cent of the world's research, so it exerts a hugh influence," says Larivière. Health Canada, for its part, has developed a clinical studies database, but has stopped short of encouraging scientists to use it.
One definite trend: more open access to scientific articles. “Approximately half of all articles are already published in open access,” Larivièere notes. “Federal funding agencies require it, but they also give scholarly journals a 12-month monopoly on sales after publication.” Quebec's main funding body, the Fonds de recherche du Québec-Santé, has followed Ottawa's lead, while other funding bodies are developing their own open-data policies.
In Canada, a lot of "research is funded by the public sector, which is to say by the taxpayer, who should have access to the results,” Larivière argues. “The idea is also to maximize return on the public capital that's been invested, because if you don't have access to past research, you can't produce new research.”
In Larivière's view, that's the first step in making science truly open. “After that, we can go deeper into the challenges of data sharing issues and establish frameworks that will benefit scholars and society as a whole.”
(Article by Martine Letarte)
Jean-Claude Guédon, open data’s tireless activist
The organizers of the RDA10 Plenary Meeting have invited Jean-Claude Guédon as guest of honour at the event’s banquet in order to salute his commitment to open data.
Guédon, a professor in UdeM's Department of Literature and Languages, is a founder of the Research Data Alliance, and has championed open access to research data and archives, as well as open-source-code software.
A supporter of the free circulation of scientific knowledge on the Web, Guédon is a signatory of the Budapest Open Access Initiative, where the open-archives movement began. He is a member of the Internet Society and of the Association francophone des utilisateurs de logiciels libres.
Université de Montréal
Tel: 514 343-7593