What to do and what not to do in scientific publishing

D A Hughes
Emeritus Professor
Institute for Water Research
Rhodes University

Introduction

Publishing the results of scientific research is an essential component of any career in the hydrological sciences, particularly for those who wish to get promoted in an academic institution. However, it is not an easy task, especially for young and emerging researchers who have limited experience of the publication process. Muñoz-Carpena et al. (2020) noted that in 2019 the editors of the Journal of Hydrology: Regional Studies rejected 70% of all submitted papers and 40% were desk rejections (i.e. rejected by the editors without being sent out for review). Getting published in any of the top international journals is therefore clearly a daunting task, even if the scientific content of the research is worthy of publication. Muñoz-Carpena et al. (2020) also identify some of the reasons for rejection and at least some of these can be avoided if more care is taken in the preparation of the submitted manuscript.

I personally believe that all editors and reviewers have a built-in ‘irritation meter’. The more mistakes (of whatever type) that are made in a submitted publication, the further up the ‘irritation scale’ goes the meter. Beyond a certain point, the reviewer (or editor) will almost certainly reject the paper, even if the topic is publishable. One of any author’s primary concerns should be to prepare their paper in a way that minimises the mistakes to keep the ‘irritation meter’ as low down the scale as possible. Part of this presentation is to highlight some of the common mistakes that are made (see also Walsh et al., 2009). Many of the points that are included in the following sections may appear to be obvious and very trivial. However, my experiences as the Africa region editor for Journal of Hydrology: Regional Studies since 2014 suggest that a significant number of submitting authors are either not aware of these issues, or choose to ignore them.

Format of the manuscript

All journals list the specific formatting requirements on their websites, but it is amazing how many potential authors seem to totally ignore these. Many of these requirements are designed to make it easier for a reviewer to read through the paper and contextualise any comments that they wish to make. A common mistake is not to include page and line numbers in the preparation of the manuscript. This makes it almost impossible for the reviewers to refer to specific parts of the paper. It is also standard practice to use double spacing. Other common mistakes are associated with citations and referencing, and it is important that potential authors check the specific guidelines for the target journal or, at the very least, look through previously published papers in the journal.

Plagiarism and duplicate submissions

Almost all international journals use editorial management software that includes some form of plagiarism (Koutsoyiannis and Kundzewicz, 2009) check that produces a report on which parts of the submitted text can be found in existing publications. If the percentage duplication is relatively high, it is quite likely that the editor will reject the paper. This duplication includes any material that has been previously published by the same author(s), referred to as self-plagiarism. Clearly, authors should avoid using ‘copy and paste’ approaches when writing their paper.

Authors are required to declare that they have not submitted the same (or similar) paper to another journal and yet I have experienced about five duplicate submission cases over that last 18 months in my role as an editor. Most journals collaborate in cross-checking and trying to identify cases where a paper has been submitted to more than one journal. Some of these have been picked up after the review process has started and this means that some of the reviewers have wasted valuable time on a paper that will be rejected (without ANY option for re-submission) by both journals.

It is also important to note that diagrams are typically subject to copyright and cannot be re-used in another publication without permission from the publisher. The clear message is therefore don’t ‘borrow’ a diagram from an existing publication. Even if you re-draw it in a slightly different form, you should cite the original source of the diagram.

Language and grammar  

Many of the issues associated with languages and grammar are quite sensitive given that the international literature is dominated by the use of the English language, while many authors come from countries where English is not the home language. While English is widely used in many parts of southern Africa, it is not the first language of many potential authors in the region (Hughes et al., 2014). There is, unfortunately, no simple solution to the problem and many adverse comments from reviewers are related to poor grammar. If the abstract contains poor language and grammar, it is also possible that potential reviewers will decline the editor’s invitation because they expect similar problems throughout the manuscript. This will inevitably cause delays in the whole review process.

At least some of the more obvious grammatical and spelling errors can be picked up by using the ‘Spelling & Grammar’ check in Microsoft Word (I assume other word processing software tools have a similar facility). Having your manuscript read through by a native English speaker will also help, while the expensive option is to make use of the many professional proofreading and editing services that are available.

Some guidelines on improving language and grammar are:

  • Avoid long and complex sentences.
  • Keep sentences relatively simple.
  • Don’t use very short paragraphs, but also ensure that paragraph breaks are appropriately placed, i.e. when there is a change in the key points being made.
  • Make sure that you don’t mix up indefinite articles (‘a’ or ‘an’) with definite articles (‘the’); a common mistake for authors from some language groups.

Always be concise and avoid repetition. One of the factors that drives up the ‘irritation meter’ of some reviewers is unnecessarily long-winded explanations and a lot of repetition. The best way to avoid these problems is by reading through the manuscript several times and asking the questions ‘can I get that point across in a simpler way?’ and ‘have I already said that previously in the paper?’. Once again, some of these language issues can be picked up by asking colleagues to look through the paper, asking them to specifically identify parts of the paper that they struggle to understand from a language (not scientific) perspective.

Tables and diagrams

A frequent problem is that authors include a table of values and then repeat many of those values in the part of the text that presents and discusses the table. A further point is that all tables and diagrams should be comprehensible on their own. This means that all the table headings and diagram keys should be clear so that readers can understand their content without having to search through the text. Good practice for tables is always to use the same number of significant digits for each table cell that refers to the same information, and preferably use right justification so that it is easier to see differences between cells (see the simple examples given in Table 1).

Table 1  Example of different table structures

Not well structured

Catchment Area (km2) Mean Annual Rainfall (mm)
A 500 950.567
B 320.55 800.7
C 20.5 1 100.87
D 1 000.0 300.6
E 22 000 870.987

 

Better structured

Catchment Area (km2) Mean Annual Rainfall (mm)
A 500.0 950.6
B 320.6 800.7
C 20.5 1 100.9
D 1 000.0 300.6
E 22 000.0 871.0

 

I have seen many examples of diagrams and figures that are almost incomprehensible, either because the text in the keys is too small, the colours of lines or areas are not different enough, or other problems of clarity. Always remember that when you are looking at a diagram in whatever software package is being used, it might look very clear on the computer screen, but may be a lot less clear when it is downloaded as a graphics image file. This is particularly true for comparative time series plots of hydrological variables that we often use in hydrological publications (e.g. observed versus simulated streamflow). Make sure that the individual lines are distinguishable and if not, then perhaps choose to illustrate the results with a shorter time period. Also make sure that the axes on graphs are clearly labelled with large enough text and including the units of measurement.

Equations and units

It is not usually necessary to include the equations of widely used statistical measures (such as the Nash Sutcliffe efficiency statistic) as long as a reference is provided to the source of the equation. Including these will make a ‘methods’ section longer than necessary. When equations are included make sure that all the variables and measurement units are properly described. Units should be defined using the standard SI notation such as m3 s-1, or mm yr-1, rather other forms like m3/s or mm/a.

Research topics

Clearly it is important to choose a journal that is suited to your specific research topic, based largely on the scope of the journal given on the website. Some papers are desk-rejected because they are simply not topics that are part of the scope of the journal. Most high quality international journals will only accept papers that have some degree of innovation and are not just local case studies using common hydrological analysis methods. If your research topic is a local case study, then you should emphasise the broader scientific relevance of the results, or discuss the results in a broader regional context.

There are some types of study that appear to be quite frequently submitted to hydrology journals that have little chance of success.

One of these is climate or streamflow trend analyses. While it may be OK to include trend analyses as part of a broader study, on their own they are unlikely to be considered innovative and they are often based on such relatively short periods of data that they are inconclusive. In contrast a recent paper by Ekolu et al. (2022) using 65 years of daily data covering Africa indicated that there are “significant decadal variations in all flood and drought characteristics, which explain aperiodic increasing and decreasing trends”. They further conclude that “This stresses the importance of considering multiple time-periods when analyzing recent trends, as previous assessments may have been unrepresentative of long-term changes.” Another problem with some trend studies is that report ‘non-significant’ trends. If you use a statistical test to check for trends and it fails, there is simply no trend – there is no such thing as a ‘non-significant’ trend.

Another topic that is popular, but also frequently problematic, is related to climate change predictions. The main problem with some of the submitted papers is that they only use the outputs from one climate model and fail to account for the well-known variations in future rainfall projections across a range of different climate models. Alternatively, they may use ensemble means, which will also fail to account for the range of uncertainty in future projections.

The previous paragraph raises a more general problem with some hydrological modelling studies, and that is the failure to acknowledge the types of uncertainty that exist within the application of any model, particularly in data scarce regions (such as southern Africa). It is often beyond the scope of a study to fully analyse all the uncertainties, or it may simply not be possible because of the lack of available data to quantify the uncertainties. However, they should still be acknowledged. Otherwise, the authors are essentially suggesting that they have more confidence in the accuracy of their simulations than is actually the case. There is no real excuse for ignoring uncertainty issues in hydrological modelling papers because the literature if full of very good examples of how to deal with uncertainty.

The review process

Most papers typically go through at least two reviews. This is true for experienced authors as well as those who are relatively new to scientific publishing. One of the things that can seriously irritate reviewers is when the authors do not adequately address their comments during the re-submission process. That is not to say that authors have to blindly accept every single review comment, but if they do not agree with the reviewer’s suggestions then they still need to provide an adequate explanation. The implication is that the authors need to provide a detailed response to the comments and indicate what they have changed in the revised manuscript. The simplest way to do this is to submit a response document that addresses each comment and to submit a ‘track-changes’ version of the previous version of the paper, as well as the ‘clean’ version of the revised paper. If you make it hard for a reviewer to check what has been changed, their ‘irritation meter’ is going to shoot up, possibly resulting in another ‘major revision’ (or ‘reject’) decision.

I have noticed that some new authors expect a very rapid turn-around for processing their submission. In some cases, young scientists need a relatively quick decision because the award of their post-graduate degree is based on submitted papers. However, all authors need to be aware that it is increasingly difficult to find good reviewers who have the time or inclination to undertake reviews. As editors we often find that we have to invite more than 10 reviewers before we get 2 to agree. This inevitably causes delays in the initial processing of the paper. Reviewers are also less inclined to accept if they note that the title and abstract are not written clearly and with correct grammar and punctuation.

Structure and content

It is not straightforward to offer definitive advice on the structure and content of a scientific paper because the ‘best’ structure often depends on the type of paper. The following paragraph therefore offers some very generic advice on the different sections of a paper.

Introduction: This should be designed to set the context of the study and identify why it is novel and innovative. While it is necessary to provide some literature citations to support the issues discussed in the introduction, ensure that they are directly relevant and not simply ‘padding’.

Study area: Stick to the details of the study area that are relevant to your study.

Methods and data:  Ensure that you present your methods (and data) clearly and succinctly. If your methods (or model) have been widely used before then a simple summary, plus relevant citations, are usually sufficient. A flow diagram illustrating your methodology can often be very useful and informative, and can be used in place of several paragraphs of text.

Results and discussion: Some authors choose to just present their results in one section and then discuss them in a separate section. Others tend to present and discuss the results in the same section. The choice really depends on the type of study, but if the former approach is used then beware of a lot of repetition.

Conclusions: This is not meant to be a summary of the results and discussion section, a common error with relatively inexperienced authors. While there may be some inevitable repletion of issues raised in the results and discussion sections, the conclusions should focus on what the results mean in a broader context (possibly including the need for further research) and what scientific lessons can be learned from the study. Beware of reaching false conclusions and over-stating the reliability and significance of your results. Most editors and reviewers are well aware of the uncertainties involved in the study of natural systems and will not penalise an author if they clearly state the limitations of their methods, available data and results. In fact, the opposite is often the case, and they will appreciate authors who are honest and realistic enough to be aware of the limitations of the study.

References: Make sure that all the citations in the text are included in the reference list and make sure that they are complete (i.e. all author surnames and initials, article/chapter title, journal/book title, volume and pages, and doi). Also make sure that all those listed in the reference list are referred to at least once in the manuscript text.

Finally

One of the best ways to develop an understanding of what is required to write a paper is to read lots of papers by other authors. But don’t just scan through them to see if the content is relevant to your study, read some of them thoroughly for language and style, particularly those from well-known authors. I realise that this could be time-consuming and you might think that it occupies vital hours that could be spent on your own research project. But it is worth it. A senior professor from the Netherlands used to tell me that he kept a lot of English fiction novels in his office (covering different genera, including crime stories, sci-fi and romance, as well as many different authors). When he had a student with relatively poor writing skills, he would encourage them to borrow some of these books and go away and read them. The idea is that published novels are almost always written in good English (otherwise they wouldn’t sell) and that reading them would help the students with their own language and grammar skills. The students also get a rest from doing science.

 

Bibliography

Ekolu, J., Dieppois, B., Sidibe, M., Eden, J.M., Tramblay, Y., Villarini, G., Peña-Angulo, D., Mahé, G., Paturel, J-E., Onyutha, C. and van de Wiel, M. (2022) Long-term variability in hydrological droughts and floods in sub-Saharan Africa: New perspectives from a 65-year daily streamflow dataset. Journal of Hydrology, 613, Part A. https://doi.org/10.1016/j.jhydrol.2022.128359.

Hughes, D.A., Heal, K.V. and Leduc, C. (2014) Improving the visibility of hydrological sciences from developing countries. Hydrological Sciences Journal, 59(9), 1627-1635. http://dx.doi.org/10.1080/02626667.2014.938653.

Koutsoyiannis, D. and Kundzewicz, Z.W. (2009) Editorial—recycling paper vs recycling papers. Hydrological Sciences Journal, 54 (1), 3–4. https://doi.org/10.1623/hysj.54.1.3.

Muñoz-Carpena, R., Batelaan, O., Willems, P., Hughes, D.A. (2020) Editorial – Why it is a blessing to be rejected: improving science with quality publications. Journal of Hydrology: Regional Studies, 100717. https://doi.org/10.1016/j.ejrh.2020.100717.

Walsh, P.J., Mommsen, T.P., and Nilsson, G.E., 2009. Editorial: The do’s and don’t’s of submitting scientific papers. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology, 152, 291–292. https://doi.org/10.1016/j.cbpa.2008.12.007.