IEEE Strengthens its Support for Open Science

Open science is the principle and practice of making research products and processes available to all while respecting diverse cultures; maintaining security and privacy; and fostering collaborations, reproducibility, and equity.

“IEEE strongly supports the sharing of research processes, methods, data, code, and findings as openly as possible to promote collaboration, accelerate innovation, and advance scientific knowledge,” says a statement recently issued by the IEEE Publication Services and Products board. The statement emphasizes that IEEE encourages its authors to share “other research outputs to facilitate verification and reproducibility of experiments and their conclusions.”

IEEE Fellow Gaurav Sharma, chair of the board’s strategic planning committee, toldThe Institute, “The idea [of open science] is to be transparent and collaborative in terms of the research process, instead of the traditional approach, where an individual or a team worked in isolation and kept their work under wraps until the time was ready for revealing only the final findings in a published article.

“Overall, the scientific community benefits from transparency,” Sharma says, “by having a window into the process by which science gets done and published. Open science also empowers individuals to participate in science and research much better than was previously possible.”

IEEE Life Fellow Sergio Benedetto, vice president, IEEE Publication Services and Products, told The Institute, “We have taken several steps to increase the availability of open access journals in our portfolio, and we definitely also want to enlarge our activities in the open science domain.”

Reducing fraud in academic research

Because open science aims to make things more transparent, it could help reduce fraud in academic research by making the data and source code used to generate the results available so that other researchers could reproduce the results.

“Research reproducibility is particularly important because it allows people to easily replicate and build on top of other people’s results,” Benedetto says. “It will be more difficult to cheat. If the results are not well replicated, then the authors will be requested to justify them. This will enhance the ethical conduct of researchers.”

IEEE DataPort and Code Ocean aid research reproducibility by enabling access to datasets and codes associated with research articles to which they are linked through the IEEE Xplore Digital Library.

“IEEE is one of the major players in ensuring ethics in publishing,” Benedetto says. “Open science is a tool that can help prevent unethical conduct.”

A more transparent peer-review process

Another method to boost research integrity is to remove the secrecy surrounding the current peer-review models known as single-blind review and double-blind review. A single-blind review lets reviewers know the names of the authors, but the authors don’t know the reviewers’ identities. In the double-blind model, neither authors nor reviewers know each other’s names.

Those models were established to ensure impartiality, reduce bias, and minimize potential conflicts of interest, but once the review process is completed and the article is published, maintaining anonymity and confidentiality isn’t always equitable, inclusive, or transparent, Sharma says.

The transparent peer-review model would, upon suitable opt-in by all parties, disclose the review reports and, possibly, the identities of the reviewers. The reasoning is that by disclosing the review reports there is greater transparency into the process by which a published article was acceptedSh, and reviewers can choose to claim credit for ideas or improvements they suggested in the reviews.

“We have taken several steps to increase the availability of open access journals in our portfolio, and we definitely also want to enlarge our activities in the open science domain.” —Sergio Benedetto.

Another benefit of the transparent review process is pedagogical: New researchers can learn to be better reviewers themselves from publicly available well-written reviews, which are not accessible for this purpose in the traditional approach. In addition, should the research stir debate or controversy during the review process, the concerns, questions, and responses also would become part of the final record of the published paper and available to the scientific community, Sharma says.

“The community at large benefits from having a window into the process by which science gets done and published,” he says. “You not only will see the end result but you also will see some of the debate.”

If datasets or the source codes are made available by article authors, IEEE is considering providing them to reviewers as well for assessing the article.

“Peer reviewers can check the results using the data and code, providing greater confidence in the published work,” Sharma says.

Sharma is conducting a pilot program on transparent peer review, in which—for the first time within IEEE—review reports and identities will be disclosed upon suitable opt-in.

Tools that remove barriers to reproducibility

IEEE has introduced many tools over the years that make it easier to reproduce research results.

“If you’re an individual researcher wanting to participate in citizen science, your resources do not necessarily allow you to build everything from scratch,” Sharma says. “But if there’s a large variety of tool kits that are available to you to build upon, then your task becomes much easier, and you can contribute more effectively.”

Here are some of the tools:

• Code Ocean is a cloud-based computational reproducibility platform that contains thousands of research results, which can be shared and reused. Once an author has added code, Sharma says, anybody can easily try it via a Web-based platform and validate it using their own data.
• IEEE DataPort is a Web-based platform that allows researchers to store, share, access, and manage their research datasets in a single trusted location. The portal accepts both standard and open-access datasets in many formats. The data is stored on the Amazon Web Services cloud and can be downloaded at any time.
• TechRxiv (pronounced “tech archive”) is a free, publicly accessible preprint server for unpublished research in electrical engineering, computer science, and related fields. By posting drafts, authors can receive early feedback from other researchers before they submit their work for formal peer review and publication. The research is time-stamped to show the ideas were originated by the author, so others can’t claim them.

“Transparency of research is the final goal of open science, something that has only pros and no cons at all,” Benedetto says. “Transparency improves the pace at which research evolves in every aspect. It makes everything freely available to researchers so they don’t have to struggle to replicate the results.

“In my opinion, transparency is everything, and IEEE is well equipped to make it happen in the near future.”