Do you really have to reformat for each new submission?

There was a fascinating conversation at the blog of neuroscientist Dorothy Bishop recently. As a side point in a post on getting replication studies published, she says:

…if a journal commonly rejects papers without review, then it shouldn’t be fussy about the format in which a paper is submitted. It’s just silly for busy people to spend time getting the references correctly punctuated, or converting their figures to a specific format, if there’s a strong probability that their paper will be bounced. Let the formatting issues be addressed after the first round of review.

Palaeontologist Mike Taylor agreed in the comments:

…it seems unbelievably stupid that when preparing my research output for the world, such a huge proportion of the effort is dedicated to inserting and removing commas.

He blogged about this issue back in 2010, commenting that the value of this work by researchers is zero but its cost is enormous. In a recent article in Discover, he says:

Most journals have stringent formatting guidelines that authors must follow in submitted manuscripts. (A colleague of mine recently gave up attempts to submit his manuscript to a particular journal after it was three times rejected without review for trivial formatting and punctuation errors, such as using the wrong kind of dash. Seriously.)

I found it hard to believe that some journals are this strict. When I was an in-house editor (for Elsevier and BioMed Central), I always ignored the formatting of newly submitted papers. I was trained to focus on whether the topic of the manuscript fitted the scope of the journal and whether it would be interesting enough for the readers. I have never seen an article rejected because it was in the wrong format.

But it does seem that there are journals that reject because of formatting before they even decide whether to send a paper to referees. Mike mentions The Journal of Vertebrate Paleontology as being one, and Dorothy describes similar procedures at Journal of Neuroscience and Psychophysiology.

This seems to me to be a great waste of authors’ time. It is an insult to an author to ask them to reformat, only to then reject their paper on grounds other than the formatting – grounds that were presumably clear in the first version. If an editor is sure that they are not going to send a manuscript out to referees, why not just reject it straight away, so that the author can submit it quickly somewhere else?

I know how much time it can take to reformat from one journal style to another. In fact I quite enjoy doing this, though it helps that I am generally being paid to do it and that it is part of my job, not a distraction from important research. And I also have lots of practice in macros, wildcard searches and other tricks that make reformatting far quicker. If you’re a scientist whose main expertise is research, reformatting will be much harder and more frustrating. Scientists are good at research, and copyeditors are good at house style – why don’t journals get their (usually freelance) copyeditors to do this reformatting? The only reason I can think of is that we will charge for the work.

The best time for reformatting to journal style is after a paper is accepted. The second best is when revisions have been requested: given that there is now a good chance that the journal will publish the paper, it seems reasonable at this point to ask the author to get the formatting close to the journal style. Major things, in particular, such as whether the methods section is before or after the results, or keeping to overall length limits, are best done by the authors in revision.

I’d like to put together a list of journals that do or don’t require perfect formatting at first submission, so that authors can find out before submission what will be expected of them. This information isn’t easy to find. Of course, instructions for authors always imply that formatting should be done before submission, because (I suggest) the editors would love every author to do this, as it makes an editor’s job easier. But detailed formatting issues, like the kind of dash used (or other examples given by Mike Taylor)? I don’t believe any editor really needs this kind of thing to be sorted out by authors. If they do, they must frequently be disappointed!

If you feel that it would help your paper to get published if you reformat it precisely to the journal style before submission, I won’t stop you – after all, you will be making the job of the in-house editor easier, which might, you never know, tip the scales in your favour. And if you want to pay me to do the reformatting for you, I probably won’t say no. But at the risk of doing myself out of some work, I would recommend that authors don’t spend too much time or money on reformatting before the paper has been first submitted. And if a journal asks you to reformat when they haven’t promised to send the paper out to referees, I suggest asking why.

Your experience

Journal editors, do you expect papers to be formatted to your house style before you will consider sending them to reviewers? How strict are you about this? Why?

Researchers, which journals have asked you to reformat before they decide whether to send your paper to reviewers? Which journals don’t mind the formatting being incorrect?

The really dangerous formatting errors

No journal editor expects submitted papers to be perfectly formatted. But there are different types of error that cause different problems.

  1. Common formatting errors that are easy to spot and easy to fix, like double spaces at the end of sentences
  2. Things that aren’t really errors, which will be ignored when the paper is typeset, such as spacing between paragraphs
  3. ‘Errors’ that are really matters of house style, such as italicisation of the ‘p’ in ‘p-value’
  4. Things that are hard to spot and that could cause the reader to misunderstand something crucial.

It is this last category that I would like to focus on here. These are the kinds of errors that every scientific copyeditor should know to look out for, but not all do. If you do one of these things, you will cause extra work for whoever copyedits your paper. Not a major problem, you might think. But if the copyeditor is not experienced in this kind of problem, or is rushed or underpaid, or if the journal doesn’t copyedit it at all, the error might not be spotted. As you will see from the examples below, this could have serious consequences.

Misuse of underlining

OK, so you want to quote a mean and standard deviation. But that symbol you want to add between the two, that ‘plus or minus’ sign (±), where is that in your wordprocessor? It’s not always obvious where to find the symbol you want. So some clever authors think ‘I know, I’ll put in a plus sign and underline it! It looks the same, doesn’t it?’

Well it looks the same as long as the underlining never gets lost. But the journey from the author’s wordprocessed file to the final published article is a long one, and the underlining could be stripped out at various points: preprocessing, copyediting, typesetting or conversion to XML or HTML. Some typesetters, I’m told, remove all formatting from documents and add it back manually by looking at a pdf of the original. Sure, proper professional scientific typesetters used by journals are unlikely to do this, but can you be sure? Underlining is generally not used in published work, so some copyeditors also remove it – again, they should always check first for any underlining that means something, but can you be sure that they will see your clever underlined plus sign?

Similar logic also leads to the underlined less than sign, or greater than sign, to give a version of ≤ or ≥.

So I wonder, now, how many papers there are out there that quote the mean + standard deviation not the mean ± standard deviation because of this error. I haven’t actually seen any myself – have you?

Greek letters

A similar difficulty presumably leads some authors to fudge Greek letters. Take microliters, an extremely commonly used unit in molecular biology papers. The correct symbol is ‘µl’, but I’ve lost count of the times I’ve seen ‘ul’. Maybe this wouldn’t cause too much misunderstanding if left in, but the fact remains that it is incorrect.

More dangerous is to put in an ‘m’ and change the font to Symbol. This can lead to the same kind of problem as underlining – the Symbol font can all too easily be stripped out by mistake during processing. And do you really want the amounts to be given in milliliters rather than microliters? In some rare cases this could actually be dangerous, if someone took a drug at 1000 times the intended dose because of this kind of error.

Another surprisingly common error is to use the German ß instead of β. ß means double-s and is completely unrelated to β, despite its surface similarity.


This one I find even more surprising than those mentioned above: using the wrong symbol for the degrees sign (°). This symbol isn’t hard to find in the Insert Symbol dialog in Microsoft Word, or in the LibreOffice equivalent. Superscript lower-case letter o or superscript zero aren’t good substitutes, again because the formatting might be stripped out. There is a very similar looking symbol called ‘Masculine ordinal indicator’, which is used in languages such as Italian to abbreviate ‘primo’ (‘first’). If you use this it probably won’t cause any harm, but it will still be incorrect.

Finding the right symbol

The best way to insert any symbol in Microsoft Word is to find out its character code in Unicode and then type the code and then press ALT+X to convert it into a character. For example, type 2265 and then, with the cursor just to the right of the 5, press ALT+X – this gives ≥. This site has a convenient search function – type in ‘greater than or equal to’ and you get a list of possible characters, of which U+2265 (ie Unicode 2265) is at the top. In Open Office or Libre Office, the process is similar but slightly more complex and is explained here.

Thanks to @nfanget, @EndoMetabPub, Katherine Timberlake and Kathleen Lyle (the latter two of SfEP) for suggestions for this post.

Your experience

Editors, have you seen any other formatting errors that are as dangerous as those mentioned here? Researchers, have you seen any examples that got through to the published version of a paper?