For a research article to be read, understood and remembered it must have a clear message. The structure and form of the article is part of this message.
Photo: Einar Nilsen
From time to time I hold lectures on how to write scientific articles, in which I draw on good books about this topic (1 – 3 ) as well as my own experience as a researcher and editor. I tend to start by asking the audience what characterises a good scientific article, and I receive a lot of answers. The recurring ones are: a relevant and important research problem, appropriate use of scientific methods, reliable data, balanced interpretations, a good discussion of the strengths and weaknesses of the study, good writing, illustrative graphs and well-ordered tables. Many emphasise that the results must be truthful, in the sense that they have not been produced by data manipulation, fabrication or other kinds of unethical behaviour. In the end we tend to agree on the most important aspect: The article must have a message.
Research articles are supposed to tell a story. Before they start writing, authors should therefore ask themselves: How can I persuade potential readers to read, follow, understand, believe in and preferably remember the story I want to tell? What impression should the reader be left with after having read the article? Such questions ought to be the starting point for the writing and the design of tables and graphs (4 , 5 ).
So what is a message? This will obviously depend on the study in question. An article’s message is linked to its key findings and their implications. A surprisingly large number of authors have trouble formulating this in an appropriate manner – «where’s the beef?», as the Americans often say. If you as an author fail to find a good and adequate formulation, ask yourself: How would I explain this to «the man in the street» or my own grandmother? This will set you off to a good start. Use simple words rather than complicated ones wherever possible.
Another good piece of advice is to put a lot of effort into the abstract. There are (at least) three reasons for this. First, those who read only the abstract should also be informed about the main findings and conclusions of the study. Second, the abstract should entice potential readers to read the entire article – they should be hooked. Third, the abstract is a good exercise in writing a short, precise and succinct text. The number of words in the abstract is limited, so each word must be carefully selected. The abstract should not be «thrown together» immediately prior to the submission of the manuscript. Try to write the abstract first, but re-read it and make amendments before submitting your manuscript.
Most medical journals require original articles to have a so-called IMRaD structure – four main sections about the background, material and method, results and discussion, respectively (minor variations occur) (1 – 3 ). This strict and conformist structure may easily be perceived as rigid and restrictive by an author who wishes to communicate his or her message to the widest possible audience. This is an erroneous idea. This structure is an aid in the efforts to achieve a logical and appropriate flow in your text and to emphasise the message. A recognisable, set structure signals that the article is scientific and serious, and that the authors are credible and trustworthy. Another advantage is that readers will know where in the article to find information that they are looking for.
Scientific articles in this journal should include a text box with the article’s main message stated in 3 – 4 phrases. Articles in some other journals – such as the BMJ , formerly the British Medical Journal – often have a text box with a brief statement of what is known about the topic in question and what the described study adds to it. Some journals prefer to include the main finding or conclusion of the study in the article’s title, but this is not the case in this journal nor in the New England Journal of Medicine (6 ). This may appear paradoxical, since we ask our authors to communicate their message as effectively as possible. The title also has functions other than to arouse interest among the target group: it should signal a scientific approach, seriousness and balanced reflection. One could put it like this: The form of the title is part of the message – in the same way as the structure and form of the article is part of the message. The form has a message: This is a scientific article, not a statement of opinion or a news report. Those who wish to voice their opinion can do so by writing commentaries, op-eds or other types of articles that have fewer restrictions with regard to the choice of format and formulations.
When estate agents are asked to state the three most important factors that determine the sales value of a property, they tend to reply: location, location, location. The Norwegian Labour Party’s general secretary claimed that there were three reasons for the party’s excellent results in the 2005 elections: the trade unions, the trade unions, the trade unions. In the same vein, it is tempting to emphasise three criteria for a good scientific article: message, message, message.
