Correspondence: Better writing and more space needed online
Martin Fenner
Thursday, 04 September 2008 09:49 UTC
Linda Cooper wrote a letter with the above title in this week’s Nature1. The title says it all. Papers in online-only journals are often unnecessary short. She also makes the point that the inflation of online supplementary material makes it more difficult to read a paper, especially when reading the print version.
My question would be whether longer papers are on average really easier to read. I think that the shorter format requires the author to be more careful with the language of every single sentence and can sometimes actually improve the writing.
1 Cooper L. Better writing and more space needed online. doi:10.1038/455026a
-
Replies
Jump to resultsResults
-
The British Medical Journal (BMJ) had an interesting article related to this topic a few years ago1. Some journals, including the BMJ, have started to publish two different versions of a paper: electronic-long; paper-short or ELPS. They argue that paper and online are two different media and require two different kinds of papers (online-only journals obviously would publish only the electronic-long version).
1 Delamothe T. Is that it? How online articles have changed over the past five years. doi:10.1136/bmj.325.7378.1475.
-
Maxine, I would suggest that there is a third type of reader, the Type 3 reader. These readers want and need BOTH a good, logical, and clear story as well as convincing data. They read papers critically, often at journal clubs and do not blindly accept the judgment of external referees; they argue about whether the data are persuasive or the story complete and want to know how the data were obtained and how they were analyzed. For Type 3 readers, a balanced paper that gives complete information in a well-constructed story really represents what most readers look for in all published manuscripts.
-
Maxine touches on an important point that I think remains unresolved for Nature. Can you really write an article that reaches both a general (though scientifically literate) and specialist audience; (I don’t imagine that the journal pretends to reach out to a truly lay readership.).
Linda, you seem to think that this is possible (and certainly it’s a worthy goal that we should strive to improve) but I wonder – is the degree of specialisation not too great an impediment? I flick though Nature weekly but rarely stop for long at any of the articles outside my field (molecular biology). Though I have a physics background, I’ve not encountered an astronomy paper that engaged me much.
Maybe this is a band-aid solution but what about insisting that any author whose paper is accepted has to write a short summary for a non-specialist? This could be published online.
-
Stephen, I really do believe that it’s possible to write for specialist and non-specialist audiences. It’s much more difficult to do this of course, and most researchers aren’t sure how to even try. A good place to start is with close and careful revising which certainly helps to eliminate useless words and phrases. When writers get rid of the clutter, they have more space to explain complex concepts. As well, over and over again in my classes I find that graduate students need help identifying the real focus of their papers. Once they can do that, they also figure how to tell a logical story about their important findings. And more importantly, readers – both those familiar with the field and those outside it – can more easily understand what the author is trying to communicate
-
Stephen asks: Even then of course, there will be the problem of generating a format of Supp. Info. that integrates well with the main article. But isn’t that just a bit of html? Yes, that is correct so far as text goes, for example at Nature we integrate Methods sections in this way. In the print paper there is a Methods summary, but in the online PDF and HTML (full-text), the longer Methods section is integrated.
Similarly, additional references can be incorporated this way, and they are picked up by indexing services (eg Thomson Scientific, ISI as was) if displayed in this HTML way.
But as we are saying, there is a limit to how useful this is, and to which “set” of readers. Do non-specialists want to download a massive PDF? Should there be two separate papers, one “extended print summary” and one “long online version”, so readers can download the version that suits them for their purpose?
-
The British Medical Journal (BMJ) now routinely publishes papers that are about 30-50% shorter in the print version compared to the online version. To find out how abriged versions of papers should be presented in the printed version, the Canadian Medical Association Journal (CMAJ) published research1 that compared three different shortened versions of the fulltext online paper:
- conventional (30-50% shorter than the online version)
- journalistic (shortened version written by a journalist)
- extended abstract
A sample paper in these three versions was made available in the supplementary material2. The fulltext online version is available on the BMJ website3. A survey of readers and authors found that 45% preferred the conventional version, 31% preferred the journalistic version and 25% preferred the extended abtract.
In the example given, I find the journalistic version is the easiest to read. And I prefer the online fulltext version to the print (conventional) version.
1 Müllner M et al. How should abridged scientific articles be presented in journals? A survey of readers and authors. doi:10.1503/cmaj.1040559
2 http://www.cmaj.ca/cgi/data/172/2/203/DC1/1
3 McGill N et al.. Mechanical methods of reducing blood transfusion in cardiac surgery: randomised controlled trial. doi:10.1136/bmj.324.7349.1299
-
Anonymous
One key issue for myself is the lack of explanation in the Methods section of current papers, including online-only journal publications. It is vary difficult to reproduce experiments when statistical methods, among other things, are not clearly explicated. It is therefore my view that online supplementary materials should give detailed explanations of assumptions, derivations and validation of methods used in the main paper. Saying one used “insert here” Software tool does not give specific enough details, especially since many current analysis tools are multi-purpose.
-
Yes, anonymous, agreed. Nature has some detailed statistical guidelines in its guide to authors just for this purpose.
Also, at this level of detail, it is possible for authors to upload many additional details into a protocols or other appropriate site, ideally in a 2.0 format so that those working in these precise areas can refine the details interactively. Ideally there would be bidirectional links between the protocol discussion and the original publication.
Results
-
