An Open Mind on Open Data the Move to Make Scientific Findings Transparent Can Be a Major Boon to Research, but It Can Be Tricky to Embrace the Change

CAREERS INTERDISCIPLINARITY Biology and economics NETWORKING How junior faculty members NATUREJOBS For the latest career combined to save the world p.119 can recruit top staff go.nature.com/rygpmc listings and advice www.naturejobs.com

DATA SHARING An open mind on open data The move to make scientific findings transparent can be a major boon to research, but it can be tricky to embrace the change.

BY VIRGINIA GEWIN junior researchers, who are carving out their data-sharing requirements, and as a growing niche. On the one hand, the drive to share is number of enthusiasts call for more openness, t is a movement building steady momentum: gathering official steam. Since 2013, global junior researchers must find their place between a call to make research data, software code scientific bodies — including the European adopters and those who continue to hold out,

and experimental methods publicly avail- Commission, the US Office of Science and even as they strive to launch their own careers. BAKER/GETTY ANDY Iable and transparent. A spirit of openness is Technology Policy and the Global Research One key challenge facing young scientists is gaining traction in the science community, and Council — have begun to back policies that sup- how to be open without becoming scientifi- is the only way, say advocates, to address a ‘crisis’ port increased public access to research. cally vulnerable. They must determine the risk in science whereby too few findings are success- On the other hand, scientists disagree about of jeopardizing a job offer or a collaboration fully reproduced. Furthermore, they say, it is the how much and when they should share data, proposal from those who are wary of — or best way for researchers to gather the range of and they debate whether sharing it is more likely unfamiliar with — open science. And they observations that are necessary to speed up to accelerate science and make it more robust, must learn how to capitalize on the move- discoveries or to identify large-scale trends. or to introduce vulnerabilities and problems. ment’s benefits, such as opportunities for more The open-data shift poses a conundrum for As more journals and funders adopt citations and a way to build a reputation

without the need for conventional metrics, and some require, a data-management plan of what’s happening in my lab.” such as publication in high-impact journals. that makes data available. The US National Researchers also point to the time sink that The nascent era of openness is best embodied Science Foundation is among these. “There is involved in preparing data for others to view. by the Transparency and Openness Promotion used to be no enforcement, but that’s chang- Once the data and associated materials appear (TOP) guidelines for journals, first published1 ing,” says Karthik Ram, a data scientist at the in a repository, answering questions and han- in Science by researchers at the Center for Open Berkeley Institute for Data Science in California dling complaints can take many hours. Science in Charlottesville, Virginia. Adoption of and co-founder of ROpenSci, which develops The time investment can present other the guidelines by a journal or organization signi- open-source software programmes. Some phil- problems. In some cases, Ram says, it may fies to the research community that it supports anthropic funders, including the Bill & Melinda be difficult for junior researchers to embrace transparency, openness and reproducibility Gates Foundation in Seattle, Washington, and openness when senior colleagues — many of (whether an experiment can be replicated by the Wellcome Trust in London, also mandate whom head tenure and promotion committees the original researcher or by someone else). open data from their grant recipients. — might scoff at what they may view as mis- Those tenets apply to all aspects of science, Others, such as the Gordon and Betty Moore placed energies. “I’ve heard this recently — that including experimental design, data sharing and Foundation in Palo Alto, California, encourage embracing the idea of open data and code makes the publication of null findings and replication sharing but do not require it. Still, the trend traditional academ- studies. As Nature went to press, 538 publishers is clear, says Carly Strasser, who oversees the “Open science, ics uncomfortable,” and journals — including Elsevier and Springer foundation’s Data-Driven Discovery Initiative. data sharing, says Ram. “The con- Nature — had signed up to the TOP guidelines, “Open science, data sharing, software sharing software cern seems to be that along with 57 organizations, among them the is the future of science,” she says. “It’s only sharing is open advocates don’t American Association for the Advancement of going to get more difficult to engage in science the future of spend their time Science, which publishes Science. without being open.” science.” being as productive But many young researchers, especially as possible.” A DRIVE TO REPRODUCE those who have not been mentored in open An open-science Some fields have embraced open data more science, are uncertain about whether to share stance can also add complexity to a collabora- than others. Researchers in psychology, a field or to stay private. Graduate students and post- tion. Kate Ratliff, who studies social attitudes at rocked by findings of irreproducibility in the docs, who often are working on their lab head’s the University of Florida in Gainesville, says that past few years, have been especially vocal pro- grant, may have no choice if their supervisor or it can seem as if there are two camps in a field ponents of the drive for more-open science. In another senior colleague opposes sharing. — those who care about open science and those one of the latest examples of irreproducibility Some fear that the potential repercussions who don’t. “There’s a new area to navigate — issues, investigators tried to replicate results of sharing are too high, especially at the early ‘Are you cool with the fact that I’ll want to make from 100 psychological studies but succeeded stages of a career. “Everybody has a scary story the data open?’ — when talking with somebody in fewer than half of them2. about someone getting scooped,” says New York about an interesting research idea,” she says. A few psychology journals have created University astronomer David Hogg. Those fears incentives to increase interest in reproduc- may be a factor in a lingering hesitation to share GLASS HALF FULL ible science — for example, by affixing an data even when publishing in journals that Despite complications and concerns, the ‘open-data’ badge to articles that clearly state mandate it (see Nature 515, 478; 2014). upsides of sharing can be significant. For exam- where data are available. According to social Researchers at small labs or at institutions ple, when information is uploaded to a reposi- psychologist Brian Nosek, executive director focused on teaching arguably have the most tory, a digital object identifier (DOI) is assigned. of the Center for Open Science, the average to lose when sharing hard-won data. “With my Scientists can use a DOI to publish each step of data-sharing rate for the journal Psychological institution and teaching load, I don’t have post- the research life cycle, not just the final paper. In Science, which uses the badges, increased ten- docs and grad students,” says Terry McGlynn, a so doing, they can potentially get three citations fold to 38% from 2013 to 2015. tropical biologist at California State University, — one each for the data and software, in addi- Funders, too, are increasingly adopting an Dominguez Hills. “The stakes are higher for tion to the paper itself. And although some say open-data policy. Several strongly encourage, me to share data because it’s a bigger fraction that citations for software or data have little cur- rency in academia, they can have other benefits. Many advocates think that transparent data LEARN TO SHARE procedures with a date and time stamp will pro- tect scientists from being scooped. “This is the Open-data pro tips sweet spot between sharing and getting credit for it, while dissuading plagiarism,” says Ivo Scientists who are cautious about open proprietary data formats, such as Microsoft Grigorov, a project coordinator at the National science can start small by sharing data for spreadsheets, or colour-coded cells that are Institute of Aquatic Resources Research Secre- a project that they have already completed. readable only by humans. tariat in Charlottenlund, Denmark. Hogg says Specialists in the field offer this advice: ●● Permanently archive data in reputable that scooping is less of a problem than many ●● Document a data-deposition plan while repositories such as FigShare or Zenodo, not think. “The two cases I’m familiar with didn’t working on publications, so that the data on a personal website. involve open data or code,” he says. and the paper will be ready for publication at ●● If you choose to share data from a Open science also offers junior researchers the same time. It is not necessary, however, new project, make sure to generate the the chance to level the playing field by gain- to release data alongside a paper, unless a relevant metadata as you go. It is very hard ing better access to crucial data. Ross Mounce, funder mandates it. to reconstruct important details after the a postdoc studying evolutionary biology at ●● Craft a very explicit statement about data fact. Tools such as those on Zenodo enable the University of Cambridge, UK, is a vocal reuse — including who can use the data, researchers to document such details champion of open science, partly because his how to use them and how to attribute them. throughout a project, so that all you have to fossil-based phylogenetic research depends on ●● Machine-readable data will be most do is flip a switch when you are ready access to others’ data. He says that more open- easily combined with other data sets. Avoid to share. V. G. ness in science could help to dissuade what some perceive as a common practice of shutting

out early-career scientists’ requests for data. There is some evidence to support that statement. A study in 2014 sought data from TURNING POINT 217 studies published between 2000 and 2013. But the team could secure only 40% of what they requested, and responses varied 3 Andrew Simons according to the requester’s seniority . McGlynn says that many of the obsta- cles — whether real or perceived — to open found a masters programme in international science can be sidestepped. He is on the development at the John F. Kennedy School of editorial board for the journal Biotropica, Government at Harvard University in Cam- which encourages — but does not require bridge, Massachusetts. I was able to tailor my

BRUCE CHALLGREN — authors to contact the original researcher coursework to explore aspects of human health. when they use someone else’s archived data, which can be embargoed for up to three What brought you back to Ethiopia in 2008? years. “Not only will you get their valuable I had done short stints there and in Honduras, insights, but it’s inclusive and fair,” he says. and I returned as director of programmes Communication also helps for those who with a group that worked to alleviate chronic worry about jeopardizing a collaboration, he food insecurity in rural areas. We developed says. Concerns about open science should be an initiative that provided food and cash discussed at the outset of a study. “Whenever to 300,000 people. We also planted trees you start a project with someone, you have throughout the country. to establish a clear understanding of expec- tations for who owns the data, at what point Why did you decide to pursue a PhD in they go public and who can do what with economics? them,” he says. From 2008 to 2011, Andrew Simons led a I couldn’t help thinking, instead of helping It isn’t hugely difficult to share data (see programme in Ethiopia for a US-based non- 300,000 people, what if I had the ear of govern ‘Open-data pro tips’). Online reposito- profit relief organization. The former biologist ment and could suggest policies that could ries such as FigShare or Zenodo make it recently earned a PhD in applied economics help 7–8 million people? I was inspired by the increasingly easy to deposit scientific con- from Cornell University in Ithaca, New York, work of Chris Barrett, an applied economist at tent for widespread consumption. More than as a pathway to explore policies that could Cornell who works on global food security and 400 virtual communities have formed to help to improve global food security — reliable critiques food-aid projects worldwide. He has share data, software and documented work- access to affordable and nutritious food. a lot of influence on governments, which are flows so that a user can deploy them straight interested in his advice on how to make food- away, says Tim Smith, who oversees collabo- What sparked your interest in helping security efforts work better. My experience in ration and information services at Zenodo. developing nations? Ethiopia paved the way for me to work on a The repository launched in May 2013 at In 2000, as a biology undergraduate, I spent a handful of projects in East Africa for my PhD. CERN, Europe’s particle-physics laboratory semester in Latin America studying tropical near Geneva, Switzerland. biology. I lived with rural families in Guatemala Can you describe some of the projects that you And although there is a time cost asso- and Nicaragua, where I saw grinding poverty. worked on in Ethiopia? ciated with uploading and organizing One night, I saw a woman rummaging through I monitored the use of fuel-efficient stoves. For raw data, subsequent queries can often be the garbage to find clothing. It was heartbreak- 6 months, we tracked 1.7 million temperature averted by adding reader-friendly instruc- ing. I thought a lot about poverty and the ‘right’ data points from sensors in people’s homes to tions at the start. Hogg recommends that response from someone living a relatively understand when and how they used the stoves. researchers simultaneously upload tutori- wealthy life in the United States. In addition, I worked on a project to turn animal als and examples of how to use the content. bones into a soil fertilizer. These projects aim to In the end, sharing data, software and How did you shift away from biology? solve real problems — problems that will never materials with colleagues can help an early- I went straight to a summer internship at a be solved just by soil science or by applied eco- career researcher to garner recognition — a biophysics lab at Texas A&M University in nomics. We’ve got to combine insights from all crucial component of success. “The thing College Station. There, I saw a powerful con- these areas to find useful solutions. you are searching for is reputation,” says trast between the economically privileged, Titus Brown, a genomics researcher at the who had access to technology, and the poor, How have these experiences positioned you University of California, Davis. “To get who had no such access. I had always thought I for the job market? grants and jobs, you have to be relevant would go into molecular genetics and work on I have a wider tool kit than does someone who and achieve some level of public recogni- crops that could improve nutrition and food has studied just one discipline. I have an eco- tion. Anything you do that advances your security. But during my internship, I started nomics hammer, but I also have a few others presence — especially in a larger sphere, thinking more broadly about how technology to pick from. I want a job at a public-policy outside the communities you know — is a could be used to help the poor. school — I’m gearing up to apply for more than net win.” ■ 100 academic positions this year. I like working Did you pursue more opportunities overseas? with non-governmental organizations, but I feel Virginia Gewin is a freelance writer in Yes. I did a short internship in the Dominican that an academic route will give me the chance Portland, Oregon. Republic with a US-based, Christian interna- to design research with people who can provide 1. Nosek, B. et al. Science 348, 1422–1425 tional-relief organization that sent groups to meaningful input on policy discussions. ■ (2015). build a clinic in the slums of Santa Domingo. 2. Open Science Collaboration Science 349, 6251 (2015). As they got more money, they went on to INTERVIEW BY VIRGINIA GEWIN 3. Magee, A. F. et al. PLoS ONE 9, e110268 (2014). build homes. While there, I searched for and This interview has been edited for length and clarity.