Reinventing Discovery

The scientific community is eager to learn about the latest discoveries immediately. And new avenues for open science are promising. Yet, before researchers are willing to freely share their expertise, incentive systems need to change to make it worthwhile, said Michael Nielsen, a quantum computation pioneer and advocate for open science, before an audience at the opening session of the 2010 SPARC Digital Repository meeting in Baltimore.

When scientists share results in journal papers, they receive a reputational reward in return. But posting something on a science wiki does not generate a reward, says Nielsen. “It’s not a technical program, it’s a social program,” he said. “New media has great potential, but unless there’s a reward it’s not going to work.”

In discussing solutions to this dilemma, Nielsen offered examples of projects where scientists have worked together to freely build knowledge online – ones that have worked, and others that have failed.

One success: The Polymath Project. Last year, mathematician Tim Gowers experimented with posting a problem on his blog and inviting others to chime in with answers.  The rules of collaboration were simple: Be polite to one another and post a single idea in each comment.

So the conversation began. Over the course of 37 days, 27 contributors posted 800 comments producing a total of 170,000 words. Together, the group built on each others’ ideas and pushed the research forward fast to solve the problem.

With the findings, two papers were written under a pseudonym. There were too many contributors to list as co-authors, but participants could put the papers on their curriculum vitas to claim credit for their part in the project, noted Nielsen.

While an exciting new experiment, Nielsen said one problem is that the Polymath Project is not preserved since it was not sponsored by any one institution. Still, it’s an approach that is a harbinger of things to come in the exchange of scientific information, he said, challenging creative expertise to come forward to address the issue of preservation.

Another example Nielsen highlighted was the movement of citizen science and, in particular, Galaxy Zoo. Online volunteers were recruited to help classify galaxy images – something that humans can do better than computers. To the shock of the project organizers, 250,000 people volunteered to help, generating an amazing amount of data: 150 million classifications of 930,000 galaxies.

“Projects like Galaxy Zoo are broadening who can be a scientist,” says Nielsen. Citizen science, Open Access, science blogs and science news sites are also building bridges to fill the gap between the scientific community and society as a whole, he said. Optimists believe these approaches could also lead to an explosion in the rate of scientific progress, said Nielsen.

But the new media needs to be refined. Alongside examples of successful open science projects are promising ideas that failed.

Nielsen pointed to wikis that developers hoped would be like a super textbook online and scientists collaborating on projects together. Unfortunately, ones such as Qwiki, started five years ago at Cal Tech, never took off.

Science comment Web sites, started by Nature, among others, envisioned scientists writing reviews of papers to filter readings for each other. Everyone was eager to read reviews, as evidenced in the high traffic to the site, but no one was willing to write them so the concept failed.

Another idea, a scientific social network – a Facebook for scientists – have also failed, noted Nielsen.

What’s going on?

“There is little incentive for scientists to contribute,” said Nielsen. “People are not taking it seriously. Scientists are reluctant to share their knowledge openly when they feel their research isn’t good, releasing it might help their competitors or they don’t get any academic credit, said Nielsen.

While new media is emerging, scientists are still looking for a conventional end product for their work—a research paper. People are willing to jump in to help with open science projects, if it generates papers.

By putting this constraint on their involvement, Nielsen said scientists are being timid and missing an opportunity. “Where we should be aiming is for extreme openness, said Nielsen. “All the information in people’s head should be put out there.”

There are practical concerns over privacy, ethics, and political implications, yet Nielsen maintains that openness should be the default position of scientists.

How do we get there? It can seem daunting to act as an individual and a change will require mass cooperation, contends Nielsen. Back in the 1600s, scientists such as Hooke and Galileo were secretive with their discoveries writing their notes in code. But in time, advances began to be valued and rewarded by society. “The public benefit is strongest if discoveries are widely shared,” said Nielsen.

Now, scientists continue to have incentives share their discoveries in older media, but structures are not set up to make it worthwhile in the new avenues.

Nielsen suggested it may take a move similar to the development of the Bermuda Principles in 1996, which led to more concrete Open Access policies, such as the NIH Open Access policies. Grant agencies could mandate certain types of sharing and encourage others. Information not only needs to be disclosed, but available in a usable format accessible to others.

Concluding his remarks with a note of optimism, Nielsen said: “I think we can start to realize some of the potential of these tools and we will see a second revolution in open science.”

#