Recently I was bounced off the authorship of a letter to the editor. I had been one of four authors of a research paper published in a leading medical journal. Subsequently the journal received a critical letter from a reader, and I contributed to our joint response. After submitting our reply we learnt that the journal has an absolute rule that letters cannot have more than three authors, so (as the one who contributed least to the main paper) I had to be omitted. What is the point of such a limit? And how does it square with the ICMJE’s rules on authorship—“all who meet the four criteria should be identified as authors.” Well, it seems that that applies only if there aren’t too many of them.
Journals seem to love numerical restrictions. Limits I have seen relate to the number of authors, the number of words in the main text, the number of tables and figures, the number of references, the number of words in the abstract, and even the number of characters in the title (Snow White and the Seven Dwarfs has eight). Some of these limits seem to be introduced to save work at the journal, not to improve the quality of the articles they publish. Without such rules, decisions about the appropriateness of each aspect of every submission has to be subject to the judgment of peer reviewers and editors. With such rules things are much simpler—indeed some aspects can even be checked by a machine. But is this a good thing? No—such limits are anti-scientific.
The consequence of such rules is that authors will spend a lot of time manipulating their text to meet these limits. I suspect that often they will omit important material in order to meet arbitrary limits, such as tables and figures that would be of great value to readers, more detailed information about the methods used, or perhaps an explanation of why the study was done the way it was. With regard to a maximum number of references, Anthony Scialli (himself a journal editor) wrote in 1993: “Because of the journal’s rule, we were required to cut our reference list; we now simply state that other investigators using other techniques got different results, without being permitted to tell our readers where to go to read these different results.”
Clearly overly long manuscripts are undesirable, especially for paper based journals. But that judgment can surely only be made appropriately in relation to the content, and is better made after rather than before peer review. By all means encourage authors not to ramble on, and warn them that they may harm their publication prospects if they exceed certain limits. Some journals do say they will judge each case on its merits, and I commend the enlightened policy of Nature to “abolish space restrictions” on the methods section. (I’m not sure what to make of a journal with a minimum of 2000 words but no maximum!)
In addition, many journals add a restriction on time of submission for a letter to the editor (a minimal two weeks after publication for the Lancet) and—of course—the maximum number of words allowed (NEJM allows a measly 175). I wrote about this nonsense some years ago.
As Scialli wrote: “… scientific writers are likely to gain more from thoughtful editing than from arbitrary limits imposed by journals.” He observed that it was “unfortunate that some journals impose an additional layer of arbitrary nuisance on their contributors.”
Of course the journals hold all the aces, especially the “top” journals. If we authors don’t like it we can go elsewhere. The next journal will of course have a different layer of arbitrary nuisance.
Competing interests: None declared.
Doug Altman is the director and professor of statistics in medicine at the Centre for Statistics in Medicine, University of Oxford.