Sami Borg
Scientific Journals as Supporters of Open Access to Data
Research findings are usually published in scientific journals. Peer-review process is necessary to assess the veracity and quality of the findings. Reliability of published findings is a crucial element to all scientific journals. Therefore, scientific journals use the best experts in the field to review submitted articles.
At the moment, in many scientific fields the peer-review continues to concern mainly the information provided in the submitted manuscript but it is becoming increasingly common that journals require that authors make the data that forms the basis to the findings available for others. The peer-review can then be based both on the submitted manuscript and the associated data. Another essential point in open access to research data is that it will enable readers of journals to verify the published findings themselves. Open access is also a way to prevent dissemination of false information, as it enables post-publication critique of research outputs.
Data policies of publishers
The guidelines given by a scientific journal requesting that authors make their data available constitute the data availability policy of the publisher. Data policies have increased in many scientific fields over the past few years. This article presents a few of them.
Data policies tend to vary a lot because the structure, size and publication opportunities of data differ between scientific fields. There is also variation in the extent to which the findings can be reproduced. In their policies, publishers have to take a stand on several aspects of open access: when, how and where data should be published, and what standards should be used for data documentation.
Publishers, similarly to research funders, also have to consider whether to make their policies a recommendation or a mandatory condition of publication. Publishing data can be mandatory only when there are publication channels available for research data.
Publishers should also provide guidelines on how the data are to be cited and whether permanent identifiers such as DOI or URN should be used in citation information. Guidelines on citation promote appropriate citations and thus enable researchers to get merit as open data providers. Exact citations also facilitate data discovery, thus increasing the efficiency of research investments.
Life sciences as pioneers
While scientific journals in many disciplines have established data availability policies, major life science journals are forerunners in establishing specific and exact policies. The website of the journal Nature contains discipline-specific guidelines for both authors and peer-reviewers on data sharing. The guidelines list nearly 50 approved data banks or repositories where research data can be published. Data sharing is a condition of publication, and data must be made available to readers from the date of publication without undue qualifications as well as to editors and peer-reviewers at the submission of the article.
The journal Science also requires that authors make available all data necessary to assess the findings presented in the manuscript to any reader of the journal.
Data policies increasing in social sciences
In the field of social sciences, it is economic journals that most frequently have established data policies. Sven Vlaeminck (2013) analysed the policies of 141 eminent economic journals. One third of them had a data policy of some kind. The author found 29 journals that had a data availability policy requesting that authors submit their data prior to publication. Of these, ten used exactly the same or slightly modified version of the policy implemented by the American Economic Review. In addition, 11 journals had a “replication” policy which required authors to provide the data and other associated materials on request. Most data availability policies established were mandatory. Though data policies were increasing, over 70 per cent of the economic journals analysed had no published data policy.
Sergiu Gherghina and Alexia Katsanidou made a similar analysis of political science journals in 2011 (Data Availability in Political Science Journals). Their article lists three main elements that should be included in data availability policies of journals. The first is an explicit list of documents and data files that authors need to submit and are needed to make full replication possible. Not only data files but also a full set of supporting documentation is needed for replication. The second element describes what measures authors have to carry out to make their data available, and in what order. Thirdly, the journal needs to specify how binding these requirements are. Ghergina and Katsanidou emphasize that the elements require some technical and institutional support. Data archives are an example of such infrastructure.
About one fifth of the political science journals analysed had a data availability policy. The policies established were less often mandatory than in the case of economic journals. Data availability was typically requested only for quantitative data. Another analysis result was that the higher the impact factor and the longer the journal had been published, the more likely it was that the journal had published guidelines on data availability.
About one fifth of the political science journals analysed had a data availability policy. The policies established were less often mandatory than in the case of economic journals. It was also typical that data availability was only requested for quantitative data. Another analysis result was that the higher the impact factor and the longer the journal had been published, the more likely it was that the journal had published guidelines on data availability.
Sergiu Gherghina and Alexia Katsanidou emphasize, referring to the publications of Gary King (1995, 2006), that open access to research data should aim at making full replication possible. In their conclusions, the authors say that while political science journals have increasingly adopted data policies, there is still a lot to be done before data sharing becomes best practice in the field.
Joint responsibility for high quality
Data availability policies will become increasingly common if key shareholders of science start requesting open access. Research Councils UK as well as the National Science Foundation and National Institutes of Health in the US are all examples of major research funders requesting that grant applicants include a data management plan as part of their research plan. One of the key Finnish research funders, the Academy of Finland, has adopted a similar policy. International scientific associations of many different fields have supplied guidelines on open access to data.
Open access policies established by different shareholders will in their part affect and increase data policies of scientific journals. The responsibility for maintaining quality of science at a high level is shared by all shareholders of science. The biggest responsibility lies with individual researchers who produce new research outputs or peer-review them.
Sources:
- Availability of data and materials [web publication]. Nature. (Cited 2.10.2014.)
- Cech, T. R. (2003). Sharing Publication-Related Data and Materials: Responsibilities of Authorship in the Life Sciences. Washington: The National Academies Press.
- General Information for Authors [web publication]. Science. (Cited 2.10.2014.)
- Gherghina, Sergiu and Katsanidou, Alexia (2013). Data Availability in Political Science Journals. European Political Science 12/2013, 333–349. doi:10.1057/eps.2013.8. (Cited 2.10.2014.)
- King, Gary (1995). Replication, replication. PS: Political Science and Politics 28(3), 443–499
- King, Gary (2006). Publication, publication. PS: Political Science and Politics 39(1), 119–125
- Vlaemink, Sven (2013). Research Data Management in Economic Journals [web publication]. Open Economics. (Cited 2.10.2014.)
