Nature editor: researchers should be forced to make data public

But funders still need to create standardised data repositories for all fields, says Magdalena Skipper

March 22, 2019
Man wearing smart headset
Source: Getty

The editor-in-chief of Nature has said that she would like to force researchers to make the data and code behind their discoveries openly available to improve transparency and make results more reproducible.

Magdalena Skipper’s comments came during a debate that heard warnings that it was still better for researchers to keep their data closed off from scrutiny for the sake of their careers.

“I would actually quite like to mandate, for example, data deposition, and code deposition” for Nature authors, she told a Berlin conference organised by the publisher.

Such a change could lead to researchers being judged not just on their findings, but their scientific practice, she argued.

Appointed in July last year, geneticist Dr Skipper has said that reproducible science is a priority for her tenure. But she stopped short of saying Nature – one of a handful of highly selective multidisciplinary journals perceived to give academic careers a major boost – would actually make data deposition mandatory, although she did say that it was piloting a platform where authors could share source code.

“The trick with mandating data of all kinds, right across the board, on a multidisciplinary journal like Nature, is that there are fields in which there are no structured repositories” for some types of data, she cautioned. Research funders needed to provide this infrastructure, she added.

“The funders have a very important role to play for motivating, rewarding and incentivising researchers” to “basically keep their house tidy in a way that makes their data reusable [and] makes the data interoperable,” Dr Skipper argued earlier in the debate.

Open access – the free availability of scientific papers online – has shot up the academic agenda over the past decade. But the Berlin debate focused on whether “open science”, the opening up of data, code and methodologies so that findings can be better scrutinised and built upon, had made similar strides.

Andrew Hufton, chief editor of Scientific Data, said that while academics were more aware of journals such as his that published underlying data, they were still seen as a “strange beast to a degree”.

“A frontier for us is that a lot of young researchers are quite familiar with the concept of open access, because that’s a very concrete thing. You hit a paywall,” he explained.

“But understanding that the concept of open science actually creates responsibilities for scientists themselves, is still something where there’s education [needed],” he said.

Dirk Ostwald, a computational cognitive neuroscientist at the Free University of Berlin, warned that researchers still had few incentives to make their data shareable.

“Currently in the academic system, given its competitiveness, it’s definitely better for any young researcher like myself to have a paper in Nature without making any of the data or the code available, than to work rigorously on the data, have it all nicely documented and be reusable,” he said.

Academics faced a choice, he said: “As a researcher, what do you want to maximise? Do you want to maximise the quality of your work, or maximise your career outcome?”

For example, researchers might choose to write code just to produce a publishable result, rather than make it accessible and reusable for colleagues later, he pointed out. “Journals can really enforce higher scientific standards, if they choose to do so. I see that happening more and more, but there is still a way to go,” he said.

Register to continue

Why register?

  • Registration is free and only takes a moment
  • Once registered, you can read 3 articles a month
  • Sign up for our newsletter
Please Login or Register to read this article.

Related articles

Reader's comments (2)

As a new recruit working in the pharmaceutical industry I was told that the only thing between another thalidomide and todays practises was the potential for bad publicity. In other words nothing had changed about the subterfuge in wording within a research paper except not to go as far as thalidomide. So the evidence that a new drug works for something is so far doctored as to mislead the buyer. If researchers come business men/women do that with drugs what else. Come on now stop misleading people show all parts of the research process. Everything.
Isn't it a little ironic for the editor-in-chief of Nature to be in favour of mandatory data and code deposition, whilst the journal itself does not offer an open access option, apart from making the accepted manuscript available six months after publication? I'm also curious about the pilot platform for sharing source code; will this be freely available at the point of deposition, or will it follow the same setup as Nature itself; behind a paywall or released after 6 months? You could be in the curious position of having access to the data, but not the published paper.