My name is David Foster, I’m a scientific journal editor and I teach scientific writing at a UK university. I try to help researchers present their work efficiently and effectively for journal publication.
Sadly, there are many ways things can go wrong in preparing a research paper. Even outstanding research can be rejected or go unrecognized because it fails to get its message across.
Here's a list of 6 mistakes I see repeatedly in scientific writing.
1. Abstracts that don't hook and summaries that don’t summarize
Almost all research papers have an abstract, which gives you a chance to encourage people to read your paper. It’s free to access whether or not the paper itself is behind a paywall. You should make the most of this chance by ensuring the abstract is an informative and comprehensive summary.
Journals do, though, impose a word limit on abstracts, usually a few hundred words. It’s not necessary to use the maximum number available. Fewer is better. But too few can be much worse.
This abstract fails comprehensively (I’ve made some substitutions to avoid identification):
Singular and group effects are discussed for particle-particle scattering.
It has just ten words. It doesn’t give the context of the research, the reason for doing it, the nature of the experimental or theoretical approach, any research findings, or their implications. There’s little incentive for the reader to look at the rest of the paper.
In practice, abstracts are often written last and in a hurry. Yet they deserve the same attention as any other part of the paper.
(Editor's note, don't miss this Blogwig post: How to Write a Thesis Summary.)
2. Aimless introductions
Once someone has started reading your paper, you want them to keep reading it. The introduction provides motivation by setting the context of your research.
It should contain a review of relevant published work, and it should say why the topic is important or interesting, or what problem, question, or perceived need it addresses – all thoroughly supported by citations to authoritative references. Crucially, it should make clear the purpose, the aim or objectives, of the research.
One way to do this is with the formulaic “The aim of this work was ...”, but there are other, different ways of saying it. Ludwig gives lots of examples to take inspiration from.
Where should you place the statement about the purpose? It’s wise to postpone it until enough of the background has been covered that it can be understood. On the other hand, it shouldn’t be left so late that the reader loses the will to continue.
If the background material has to be set out at length, the purpose can be anticipated ("foreshadowed") at an earlier point, for example, by identifying the perceived need in general terms. In this example, it’s an energy management system:
To this end, a new energy management system is required.
And then, after the remainder of the background has been established, the purpose can finally be described in detail:
The objective of this paper is, therefore, to develop an energy management system that has multiple inputs, a central control, and multiple outputs. The multiple inputs are ...
Even if the reader doesn’t reach the end of the introduction, they will at least have an idea of your intentions.
3. Passive voice and top-heavy descriptions
The passive voice is common in scientific writing, with its emphasis on describing what was done rather than on who did it. It's therefore fine to use the passive voice, but you should do so thoughtfully. In fact, used carelessly, it can place unreasonable demands on the reader’s memory.
Here’s an example (I’ve again made some substitutions to avoid identification):
A hybrid device that operates in the frequency domain by converting the amplitude and phase information from the electronic crystal probe sensor and probe interference monitor (PIM) sensor, respectively, into frequency variation is presented.
This sentence has 32 words that the reader has to keep in mind before finally reaching the verb that defines the action.
You could make it more readable by moving “is presented” from the end to just after “A hybrid device” at the beginning. It then reads:
A hybrid device is presented that operates in the frequency domain by converting the amplitude and phase information from the electronic crystal probe sensor and probe interference monitor (PIM) sensor, respectively, into frequency variation.
The edit has, however, introduced a small interruption “is presented” between “device” and “that operates ...”. Interruptions of this kind are acceptable in English and are used to achieve what is sometimes called “end-weight”, where the heavier elements of a sentence – the details – are developed later to aid the reader’s comprehension, and avoid top-heaviness.
Still, this version remains fairly indigestible. It would be better to split it into two sentences, the one introducing the hybrid device, the other the details about it, like this:
A hybrid device is presented that operates in the frequency domain. It converts the amplitude and phase information from the electronic crystal probe sensor and probe interference monitor (PIM) sensor, respectively, into frequency variation.
The readability could be further improved with some technical assistance from the author.
4. The "Mahalanobis effect" and late explanations
Perversely, your closeness to your research can be an impediment to explaining it. You may know an explanation or definition is needed, but, unlike the reader, you don’t automatically see where.
Here’s an example:
The nearest-neighbour measure using the Euclidean distance can be replaced by another measure, the Mahalanobis distance. This measure was shown to increase recognition accuracy ... The Mahalanobis distance is defined by ...
The author introduces the term “Mahalanobis distance” and goes on to consider its application with “This measure was shown to increase recognition accuracy ...” (I’ve omitted the rest of the text). Only later is it actually defined.
Unfortunately, a reader unfamiliar with “Mahalanobis distance” may not persevere, stopping instead at its first appearance.
They may assume the definition was given earlier and they missed it, so they work backwards trying to find it. Or they may decide that the material requires a more expert audience – and give up.
The solution, in most circumstances, isn’t to put the definition at the beginning of the paper, since the reader may well forget it or not see its relevance. Rather, it’s best placed where it can do most good, just before it’s used for the first time.
5. Discussions that merely recapitulate
The hardest part to write in a research paper is the discussion section. It’s hard because it has to explain what the research results or findings mean. Authors who are uncertain what to write may fall back on recapitulating the results. This is irritating for the reader and wastes an opportunity to add value to the paper and secure its eventual success.
What, then, should go into your discussion section? The typical elements are these:
- Reminder of the original problem
- Objective review of the results
- Relevance to other findings in the literature
- Implications
- Limitations
- Conclusion
Identifying the limitations might seem self-defeating, but they can equip the reader with the ability to use your results safely. They also show your objectivity and authority.
As for the conclusion, it may consist of a closing paragraph or stand as a section on its own. It can state what new knowledge or new thinking has been added, or what new problems have been identified for others to consider.
In any case, it’s useful to finish with a brief encapsulating statement – a take-home message. This example is a shortened version of one of the most famous, the ending of Watson and Crick’s genetics paper:
The specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material.
Tip: For consistency, your take-home message should also appear at the end of your abstract, in a similar form.
6. Random and partial referencing
References are included in research papers to give credit for previously published research, and to tell the reader who wants to know more where to look.
Together, they establish an environment for your research. Which areas it agrees with and which not. They show the reader your grasp of the field.
For all these reasons, references should be chosen carefully as you develop your ideas and prepare the text. They should never be chosen randomly, as an afterthought. It will be obvious to a knowledgeable reader.
How, then, to choose the references? If you have access, use one of the large online databases such as Clarivate Analytics’ Web of Science or Elsevier’s Scopus.
Unlike some freely accessible general databases, Web of Science and Scopus provide records with a degree of authenticity. They allow papers to be accurately ranked by impact, measured by the citations each receives from other recognized research works. This ranking will enable you to identify the most influential papers in the field.
Once you’ve chosen your references, ensure you give all their publication details. A partial reference may make the source untraceable by the reader, leading to frustration.
Try also not to include references that have only a URL (a web address). Their provenance may be uncertain, and the URL itself may later change – making the source inaccessible.
If a reference of this kind really is essential, then add as much other information as you can, for example, the title of the work, its author and publisher, its date of publication, and the date of your access.
Afterword
There’s more, of course, to scientific writing than avoiding these mistakes. The usual recommendation is to read as much as possible from the literature. But it takes time, and not all journals contain good writing.
To support researchers, I put together A Concise Guide to Communication in Science and Engineering. Drawing on almost 400 handbooks, style guides, and other reference works, it provides advice on research papers, literature reviews, and dissertations; scientific English and mathematics; statistics and graphs; publication, posters, and talks; and ethics, especially plagiarism.
Hopefully, it will help you find the right way to tackle most challenges in scientific writing.
It won’t, though, tell you how to find the right phrase. For that, I recommend Ludwig.