What is open science?
Knowledge beyond the ivory tower
Image: Julius Drost on Unsplash
What is Open Science?
Whenever I watch someone perform a magic trick, I wonder, “How did he or she do that?” In vain – as everyone knows, a magician will never reveal their tricks.
It’s a different matter with academics. Although you might also first wonder how one of your colleagues obtained the results they did when reading the first paragraphs of a journal article, you won’t have to for long. Unlike magicians, academics can and should disclose their “tricks”.
The “Open Science” movement has the goal of opening up academic findings and making them available to all people. Each step of the research process should be transparent and accessible, whether it’s the conception, methodology, study design, data collection, or publication. The aim is to not only improve the acceptance and credibility of scientific results among the general public but also to improve the quality of research itself. When scientific results can be easily checked, the public is more likely to trust them. The direct impact on society can also be quite large. For example, making traffic information available in real time can save people’s lives. However, there is a limitation to this openness: whenever personally identifiable information is used, it needs to be anonymized before it is published.
In today’s world making knowledge accessible is easier than ever before. Digitization has led to a shift in how scientists work and communicate.
Forms of Open Science
How exactly does open science work in practice? The different forms of open science provide us with an answer.
The concept of open access is the most well-known part of the open science movement and has already become relatively well-established and accepted. Open access refers to free and unrestricted access to academic information, primarily in the journal article format. Additional details and background on open access can be found in this blog article.
Similar to open access but for research data, open data aims to make data freely available without restrictions. The data that are shared can take many different forms, for example, questionnaire responses, images, lab results, statistical calculations, or video recordings. This data can accompany the open access article examining its results, and you’ll often find a section of “supplementary information” containing tables with data or links to data sets.
There are many advantages to making raw data accessible. First and foremost, other scientists can use the data for their own work, which can save time and money since the same data doesn’t need to be gathered twice. It’s also possible to use freely accessible data to support one’s own findings. Along with peer review, this verification is another means of checking the quality of scientific findings. If data are made available, it’s possible to examine it and (hopefully!) reproduce your colleagues’ results.
Despite the clear advantages, open data is still not as popular as open access. Data often remain saved on the scientist’s computer or external drive. Another issue is that the file sizes of some data collections can be quite large. To help resolve these issues, many institutes and organizations are now offering an internal server for research data that can then made available to the public. However, even when data is accessible, it’s not easy to use without the necessary subject knowledge, and even then it can still be difficult to search for.
To make it easier to locate and use research data internationally, several open data portals have been established. For example, the European Open Science Cloud provides a search portal for open research data gathered in the EU member states.
The aim of open methodology is to make the exact methods, study design, and tools used in a piece of research clear. Although it receives less attention than open access and open data, open methodology is just as important. In order to verify research results, it’s not enough to have access to data sets and the results of its analysis. Only if the steps taken to reach a conclusion are documented can these be reproduced and checked. One open methodology initiative in the Natural Sciences is Research Resource Identifiers (RRID) which has the aim of creating unique identifiers for antibodies, organisms, and tools.
Open Educational Resources
OER seeks to make education and educational materials accessible to a larger audience. Instructional materials are offered for free under a Creative Commons license and can be used and amended by anyone – pupils, students, teachers, teaching assistants, and professors. Materials on offer can vary a great deal and include lesson plans, worksheets, images, videos, and music. The portal OER Commons provides a search engine for Open Educational Resources.
Open Peer Review
Open peer review aims to improve the classic peer review process by removing the anonymity of the reviewer so that the process is no longer “blind”. The idea is to remove the possibility of a rival giving an unfair evaluation or delaying an author’s publication unnecessarily. In open peer review the discussion and critique between authors and reviewers is also sometimes made public so that readers can also see the discussion that took place. One platform for open peer review is PubPeer.
Open source software is software where the code is made freely available and can be developed further or used for other applications by anyone. The code is made available in portals such as GitHub or SourceForge. The software can be customized for one’s own use and improved. Some famous open source software programs include the Firefox browser, the LaTeX software package for text formatting or Zotero for reference management.
Roadblocks to Open Science
Perhaps you’re reading this blog post with the same astonishment that you might experience at a magic show, and you’re now asking yourself why all of these great ideas are not already well-established academic practice.
However, before many open science initiatives become common practice, there a few challenges that first need to be overcome:
- Publishers need to develop and get behind new publication models so that journal articles can primarily be published with open access.
- Research institutes, organizations, and universities first need to create the necessary infrastructure and working conditions so that research data and methods can easily be shared.
- The evaluation criteria for research impact need to be reconsidered. Researchers up for tenure should no longer be primarily judged according to the number of citations and publications in high-impact journals. New criteria would need to be established, such as how often a research data set was cited or the number of times a researcher’s work on social media (also known as Altmetrics).
- There will need to be a cultural and political shift in academia. For example, research grants could contain the requirement that results be published in open access journals and that data sets be made available in an accessible online portal.
- Researchers will need to free themselves of the fears that competitors could use their data faster for their own results or that their mistakes could be found. At the moment, the readiness with which researchers are willing to share their results depends strongly on the individual personality of the researcher. For many researchers, it’s important to know before publication how and by whom the data could be used. For others, formal recognition, such as citation counts and career advancement, could provide an incentive.
Although many challenges remain, we hope that the magic of Open Science will be an inspiration to you and your colleagues.
This week is Open Access Week (October 21 – 27). We want to know: do you plan to participate in any open science initiative this year? What do you think about open science? Let us know on our Facebook page!
The exhibit Open Up! in Hamburg, Germany provided the inspiration for this blog post. If you happen to be in Hamburg, we recommend a visit.