Today PLOS announced a revised Data Policy that will come into effect on 1 March 2014, bringing in further measures to ensure that access to the data underlying research is an intrinsic part of the scientific publishing process. In that policy, the most strongly recommended data-sharing method is data deposition in an appropriate public repository. However, especially for smaller datasets and certain data types, the use of supporting information – those accompanying, peripheral files that are attached to a published article – may be appropriate. So, while we contemplated data policy, we at PLOS have also been thinking about supporting information: what we use it for, where it is helpful, where it isn’t, and how to approach it in an online, open data world. And we’d like to seek your thoughts too – so read on.
Supporting information can be useful
The accompaniment of research articles with supporting information is a relatively new development in scientific publishing, made possible with online formats. And the first thing we’d like to acknowledge is that it can be useful: it has opened new possibilities for research communication.
For example, researchers may:
- make the raw underlying data available for further work or replication;
- supply extra detail on methods;
- add figures or tables that are related to the study but not central;
- add multimedia or other file types that can’t yet be incorporated directly into the article;
- provide information to make their study more accessible (such as a translation into another language).
Journals also make specific uses of supporting files: for example, they may request that authors include documents showing adherence to field-specific guidelines such as ARRIVE or CONSORT, or provide the study protocol.
But it also creates challenges
As articles include more and more supporting information, we ask ourselves, and you:
- In practice, do peer reviewers assess supporting information sufficiently as part of their review? If not, some important information may not be reviewed; and perhaps some supporting information files require more review than others.
- Which information associated with the article should the journal host, and which information should be hosted elsewhere? It is worth considering: how immediately accessible the information will be from the article; the permanence of storage at either location; and how both human and machine readers would go about searching for, and within, the supporting information.
- If the use of supporting information has been shaped primarily by authors and journals, how has this affected readers? Do they miss information that could be important to the main study when it is in the supporting information? Is the distribution of information between the main article and supporting information determined by factors we can overcome in the online world, such as article length restrictions? Or would addition of some supporting information content to the main article impede comprehension?
We would like to make the case that while supporting information definitely has a role in research communication, there are ways of using it more responsibly – thereby improving the quality, coherence and completeness of published work. We recommend the following two approaches to responsible use of supporting information:
- Ask: does it need to be a supporting information file? Before making a supporting information file, consider incorporating the information into the main manuscript or (if it contains data) sending it to an appropriate public repository. The main manuscript is more immediately accessible and searchable by readers, editors, and reviewers than is supporting information; and established public repositories can optimize for the storage and retrieval of datasets.
- Consider machine-readability and reusability. Files can be made as usable as possible to reviewers, readers, and computational ‘readers’ by selecting a format from which data can be extracted: for example a spreadsheet rather than a PDF.
It doesn’t end here
We hope that by raising this topic with you – our authors, readers, reviewers, editors, and others – we can start to consider together the most helpful role for supporting information in research communication. Within PLOS, we’ll be continuing to develop our understanding of how supporting information is currently used, how we can best work with authors to optimize its use, and what technological solutions we could pursue to complement responsible supporting information file use. As we do this, we also want to know: what would your recommendations be? We invite you to comment here on the blog, via firstname.lastname@example.org or contacting the individual journals directly; we’ll also monitor Twitter and other sources.
Rosalind Attenborough, Publications Manager, PLOS Genetics
for the PLOS Supporting Information Files group