Common mistakes when writing a scientific paper

After having reviewed a range of colleagues’ papers, and having written several manuscripts myself, I realized that  most of us tend to make similar mistakes. Those are generally the same mistakes that our students struggle with most.

Here I would like to highlight the most common mistakes – at least in my experience – that you should keep in mind whilst writing your manuscript.

 

The background literature was not checked thoroughly

Before you even start designing your study, make sure that you have read all papers focusing on your research topic, and that you are up to date. You need to present the full research background in the introduction to be able to argument why your study is important, which knowledge gap it is closing, or to back up your findings. Keep in mind that the papers’ reviewers will most likely be fellow researchers that are very familiar with your research topic, and will immediately notice if important parts of earlier research are missing.

 

The research aim, hypotheses, and predictions are unclear

Make sure that you have clearly stated your research aim, hypotheses, and predictions for your research – all based on the literature review you present in the introduction where you highlight what remains unknown. If the reader is unable to follow why you actually conducted the study, she/he will not be able to follow the paper.

 

The structure of the manuscript is confusing

It happens quickly that a reader is not able to follow the sequence of your hypotheses, methods, results and discussion. So, always make sure to keep the same order throughout the whole manuscript.

Example: You have hypothesis 1, 2 and 3, which are mentioned in the hypothesis. Now, in the methods, as well as results and discussion, you keep that order. First, how did you test for hypothesis 1 (methods), what did you find as a result for hypothesis 1 (results), how do your results support your hypothesis 1, and how does it relate to other studies (discussion). Then you do the same for your other hypothesis. It is easier for the reader if you shortly refer to your hypothesis again in every part of the manuscript before explaining the methods/results/discuss it. This will also help you to keep a good structure and flow!

 

The methods are not explained in enough detail

The methods need to provide every step you have conducted during your study in such detail that everybody would be able to reproduce your study 1:1. Be sure that you have mentioned everything you did, and everything you used for your research – details regarding equipment you used, the involved participants/animals or similar, time scales, data collection and data analysis – really everything. This part cannot be detailed enough.

 

The wrong statistics are used

Usually (at least in my experience), researchers are no trained statisticians, and often did not receive a good statistical training during their career, but rather learned the procedures learning-by-doing. Thus, depending on the research area, similar statistical tests are used as kind of a standard, and often papers replicate the tests used in earlier studies – but this does not always make sense. Be sure that you choose the correct test for your analysis; otherwise, your analysis is more or less useless. For example, make sure that you know if it is appropriate to use non-parametrical or parametrical tests for your dataset, and check for normal distribution before conducting further tests. Many universities have departments that can help you with the statistics, so make sure you make use of this option if you are not completely sure what to do. There are also many very helpful online platforms and videos that might prove useful.

 

The sections are mixed up

Do not mix up results and discussion – do not interpret or discuss the results when you present them in the result section, and only summarize the results shortly before discussing them. This does not apply for journals that combine results and discussion, so be sure to check the authors’ guidelines before you write your paper!

 

Example for a mix up: Only 3 of 10 bees flew at a higher speed than flies, suggesting that flies are commonly faster than bees. Only the first part of the sentence should be part of the result section, the second half is interpretation and belongs in the discussion.

 

The conclusions do not match the presented results

Statistically, results are either significant (p<0.05) or they are not. If there are no significant results, do not write that there is a “tendency” that e.g. the speed of bees is significantly different to the speed of flies. That is just not what your data supports. If you have results that are close to being significant (e.g. p=0.059) you can mention and discuss this in the discussion, and suggest that for example with a higher N the results probably would have been significant, and that future studies should consider using higher Ns. Also, be sure that in the discussion, the conclusions drawn from the data really are supported by those. Do not over-interpret your data to let them match your hypothesis.

I know, sometimes you really want your data to match your hypotheses. Especially when the values are not too far off the magical p<0.05, it is really hard to accept that you cannot prove your hypotheses at this point. But statistics are straightforward, so believe in it (as long as you chose the right test).

 

The writing is inaccurate

Scientific writing must be accurate. Although commonly in appealing writing, it is not advised to use the same word twice in a sentence, this does not apply to scientific writing, which must be on point. Check your text for fill words and make sure you only have words in the text that are needed to understand your sentence. Do not try to impress your reader by using exotic terms or complicated structured sentences. Make sure to avoid unspecific expressions (e.g. low frequency, high temperature, highly significant), but use quantitative descriptions instead (twice/h, 35ºC, p<0.001, respectively). Get used to always use the active voice, as it is more concise and clearer than the passive voice.

 

The citations and/or references are incomplete

Be sure that you credit all the studies whose results you mention in your manuscript. Always keep good track of your references. If you give a citation in the text (only author and year), always make sure that you give the full reference (author, year, title, journal, pages, DOI) in your reference list. It takes a lot of time to have to collect them all at once and maybe even have to look for the complete reference after you have finished your paper. The style you have to present the citations and references depend on the journal you aim to publish your manuscript in – so make sure that you aware of the style format beforehand to avoid extra work.

 

You have not picked a journal beforehand

Pick a journal you want to publish your manuscript in before you start writing your paper – make sure that this journal covered similar topics to your study in the past. To avoid having to change style and format of your manuscript, check the Instructions for authors as well as papers that have been published in the chosen journal. Writing your manuscript in the correct style of the chosen journal saves you a lot of formatting time – however, often you choose the journal after your manuscript was already written, or your paper was rejected by your chosen journal and you need to reformat it to hand it in for another journal. In these cases, you need to reformat your paper after it is already written in another style – this is time and nerve consuming. So, if you can, try to avoid it.

 

I hope this wrap-up provides some tips to support you in the process of writing your research article! Go for it!

 

Here you can find some basic tips for writing your research paper!