If you are looking at online-only material and wondering how The American Naturalist might handle it, here is a quick guide:
- Is it essential to telling the article’s story clearly? Then it should be part of the article (or possibly be a separate print appendix). Papers can be as long as they need to be—but not any longer.
- Is it essential for replicating the paper? Is it something a reader will want to see 30 years from now? Then it should be in the appendix—in the source file, edited, and typeset.
- Is it nice but not necessary? Is it a format that can’t be printed on a piece of standard paper? Is it so extensive that it’s prohibitively expensive? Is it laid out in such a way that redoing it makes little sense? Then it should be in author-supplied file formats.
- Is it code? Then it’s in author-supplied files. Is it data? Then it’s in Dryad.
Here's why we do what we do...
We handle online-only material differently from many other journals, so it seemed like a good idea to explain why we do what we do. First though, it’s useful to look at the 20 years I’ve been managing editor and see the problems I've already witnessed. We had supplementary files in Excel back in 1998. Microsoft no longer supports Excel 1993 so proprietary software is a liability. We have moved every file (thousands of them) to new platforms three times. Files that aren’t integrated can be misplaced. We have systems to track them down and repost them, but over time, inevitably, this will be harder and harder. Data archiving policies have been developed, and alternative workflows for scientists, particularly with commercial sites, have developed, so the landscape is more complicated. Commercial sites (e.g., GitHub or FigShare), even if they aren’t particularly oriented toward profit, are financed by venture capital and are at the mercy of market forces and may not have the same kind of institutional commitment to long-term access and preservation, so they aren't a clear alternative.
So here is a breakdown of how we handle online-only material.
Back in 2003, we had a terrible backlog of papers. To try to squeeze in more papers into our print page budget, we decided to move the appendixes for papers from print, where they had always been, to be online-only appendixes. However, often these appendixes are essential for replication or for understanding the breakdown of the math. The University of Chicago Press was one of the first presses to have an electronic edition, and it had worked up its definition of the best practices for preservation. The Press determined that online-only appendixes had to be in the same source file as the main text, they should be edited and typeset for consistency with the article and for clarity, and they would be preserved for as long as the article itself would be preserved. The journal hopes to still be publishing for the next 150 years. The questions we ask about supplementary material, therefore, are, Does this material need to be preserved? What is the consequence if this material became unusable or lost 30 years from now? Will editing and typesetting this material make it less clear? Because figures are processed for the best quality possible, the tables are edited, and the math typeset, appendixes use up expensive and limited resources (people) and so there are page charges for appendixes (which can be waived with the other page charges). An appendix, if it is short or very useful, may still be printed in the journal. We may also print a short essential appendix (e.g., a table defining the variables) and put other appendix material online only.
Other Supplementary Materials
Supplementary materials generally are nice but not necessary. Separate files supplied by the authors, particularly separate files in proprietary formats, are detachable and “losable,” may not be usable as the owner of the proprietary software upgrades, and are not always well edited and well presented. However, in the time since 2003, authors have gotten used to the idea of putting all kinds of materials related to a paper online at no cost, so it’s sometimes not clear what needs very long-term preservation (i.e., appendix style) or what is informative but not as necessary. In addition, there are items such as massive excel tables, code that is better in its original file format, or files that are just not typesettable, like video or sound files. Material may also be at other locations (e.g., GitHub or FigShare) but the best practice is to have information, especially code essential for replicating the paper, together in one place with the article. Our goal is preservation that is as long-term as possible and that is not dependent on the vagaries of the venture capital marketplace. We post any author-supplied files (posted in the original format supplied) at no charge.
In 2010, journals in ecology and evolutionary biology agreed to enforce the Data Archiving policy, originally proposed by the NSF and made possible by the development of Dryad. See the announcement of the policy here: http://www.journals.uchicago.edu/doi/full/10.1086/650340. I happened to attend the initial meeting launching the project in 2005—so I remember that the idea was two-fold: providing the necessary information so that a study could be replicated and preserving data into the future. The image they kept using in 2005 was the researcher who retired, turned off her computer, and lost a life-time of irreplaceable information. As a result, the journal requires that data be archived in Dryad or a noncommercial data repository that is open (or accessible) and that is designed for long-term preservation. Dryad can also host other materials related to the article. The journal article has an active link to Dryad and Dryad has an active link to the article so the connection is clear. Data files may also be stored with the article at UCP but they must be archived at Dryad. The American Society of Naturalists covers the costs (via a change in the page charge structure).
References in Appendixes
As of 2016, we are requiring that references cited in the appendixes be printed in the main body of the article. In pursuing a question about how the field’s reliance on the impact factor was creating its own impact on research, it became very clear that original research was not getting its just due. Lists of references in author-supplied files are not counted as citations by Google Scholar or by Web of Science. Google Scholar counts cites in our form of appendix (i.e., mining the html tagging) but Web of Science counts only the Literature Cited in the main PDF. We wish to support original research so we commit pages to printing the references. Occasionally a meta-analysis paper presents a reference list that becomes prohibitively expensive to print, but we’ve hit only one of those since this policy began.