Scientific writing 3.0: A reader and writer's guide - Jean-Luc Lebrun, Justin Lebrun 2021

Introduction Q&A
Introduction Part II: Popular traps
Paper structure and purpose

Q: In their introduction, some writers present their main results, others just present their main goal. What is the best way?

A: Follow the journal guidelines on how to write the introduction. Some ask the writer to present the main results and some recommend not to state the results in order to keep the introduction short, to avoid repetition, and to focus on the goals. So there are indeed two ways to write an introduction, and people have strong preferences for or against each way. Each side has convincing supportive arguments. Here are those proposed by the side supporting repetition of results.

(1) The most frequent argument hinges around the famous quote: “Tell them what you’re going to tell them. Tell them. Then tell them what you told them.” This argument may apply to an easily distracted audience, but can we assume that readers are equally distracted?

(2) Another argument, this one more convincing, is that some journals no longer require a conclusion. The last section of any paper is the discussion. In that case, it may be worth repeating the main result in the introduction.

(3) Some say that many readers only read the introduction of the paper anyway, and therefore you had better mention your results there too… just in case. While this may be true, it is doubtful that readers with this behavior would skip the conclusions (if available).

Amongst those who suggest an alternative to the repetition of the results, Michael Alley advocates “mapping the document in the introduction”. He gives the example of a journal article where the author successfully manages to present in story form an overview of the methodology, thereby answering the “Why this way” question. The story reveals the problem and the method used to solve it. It remains silent on the results, but mentions the impact of being able to solve the problem.

My personal view is that the introduction should keep things moving and interesting by mentioning the expected results and their foreseen impact. If you do have to mention the actual results, do so without much detail, and make them part of the why story. In all cases, whether you mention the results specifically or not, end your introduction with the main expected outcome of your work.

Q: Can I cut and paste sentences from introduction to abstract?

A: The reader and the reviewer easily detect such shortcuts. Do not give readers the impression that you are in a hurry. The abstract is not written in the same way as the introduction, and the introduction is not written in the same way as the body of your paper. The verb tenses are different, the style is different, and the role of each part is different. Cutting and pasting extends beyond the mere transport of words. It carries writing styles, precision levels, and verb tenses that are fine in their original settings, but not necessarily fine in their new settings.

Q: Can I paraphrase sentences from the introduction of some earlier papers since the context of my new paper is the same?

A: Doing so contributes to boring papers, and you may be found guilty of self-plagiarism. The reason introductions feel repetitive is often because the writer rewrites a paper for different journals, or because there is not much of a knowledge gap filled between two successive papers from the same writer. To avoid such problems, think of each paper as a unique piece of communication to a unique reader. Do not rewrite, write afresh.

Q: When does one write the introduction of a paper?

A: In his excellent little book “A Ph.D is not enough”, Professor Feibelman gives sound advice:

“Virtually everyone finds that writing the introduction to a paper is the most difficult task. […] My solution to this problem is to start thinking about the first paragraph of an article when I begin a project rather than when I complete it.”⁶

When you write the introduction of a paper early in your project, you still have the excitement of the journey that lies ahead to energize your writing: the tantalizing hypothesis, the supportive preliminary data, and the fruitful methods. Yet, some argue that the introduction should be written at the end of the paper, once the contribution is clearer. So when does one write the introduction?

Let the content of your introduction dictate the timing. If you write a teaser introduction that states the goals, the context of your work, but only the expected results and impacts, you can indeed write your introduction early, assuming that your work keeps its original focus. Writing the introduction can also be done early if you prefer to write a succession of short papers, because your limited focus is unlikely to change.

But if your paper is the result of collaboration between many researchers spanning over a few years, it may be impossible to write the introduction early. You would then have to rely on your excellent writing skills to recapture the essence of your earlier goals and motivations to keep the reader interested along a good storyline.

Q: How long is the introduction of a paper?

A: A director of research I knew used to systematically reject any paper in which the introductory segments made less than 30% of the whole paper. To him, these introductory segments are necessary to bridge the reader’s knowledge gap. They include the introduction, but also the technical background section that immediately follows it. In the writing seminars we conduct, we rarely see papers with such a thorough background. The majority of papers we see have introductions that represent 10 to 15% of the paper. Occasionally, the sum of introductory segments is slightly above 20%, and for short letters, the introduction is only one paragraph.

So, how long is the introduction? To answer this question without thinking of the reader is unreasonable. The introduction is for the non-expert reader. The writer must have a precise idea of the type of non-expert reader likely to benefit from his or her paper. How much does that scientist know? How much background does that scientist require to use the contribution in whole or in parts? One can make assumptions of the non-expert reader based on journal and keyword choice. Does the title of the paper contain many highly specific terms? Or is your chosen journal very niche, publishing only incremental results in a small field? If so, the interested reader is more likely to be an expert and requires less introduction. If on the other hand you aim to be published in a journal like Science that has a very broad readership, much more introductory material is needed. Note that a paper with multiple contributions is likely to sacrifice the technical background to fit within a given number of pages. That is why it is better to write multiple papers, each with a single contribution and with an appropriate gap-reducing introduction.

Q: What can one tell about a scientist from reading the introduction of their latest paper?

A: Amazingly, one learns a great deal. If the introduction is well developed, and yet easy and interesting to read, the writer displays good communication skills. If the introduction contains targeted references (as opposed to bulk references) and few or no imprecise words, the writer displays good scientific skills. If the introduction contains no judgmental words, the writer displays good social skills. Communication, scientific, and social skills are essential qualities. Besides reviewing the CV or publication record, managers would be well advised to read the introduction from the latest paper written by their potential hire.

Q: How can I give a hint to the reader that my results are not able to reproduce the results claimed in some of the related works?

A: Use the past tense when referring to possibly erroneous findings. Use the present tense to show that, as far as you are concerned, you are in no doubt that the information is correct as shown in this sentence: Tom et al. identified a catalyst that increases the yield at high temperatures [7].

What sentence comes next, (1) or (2)?

(1) Slinger et al. subsequently reported that the increased yield is not due to the catalyst [8].

(2) Slinger et al. subsequently reported that the increased yield was not due to the catalyst [8].

The correct answer is (2). It allows you to contradict Slinger’s findings with the following sentence:

We found evidence that the catalyst does increase the yield.

Let’s change the first sentence by expressing doubt on the findings of Tom et al.

Tom et al. identified a catalyst that increased the yield at high temperatures [7].

Slinger et al. subsequently showed that the increased yield is not due to the catalyst [8].

We also found evidence that the yield increase at high temperatures is not linked to the catalyst but to…

Q: In your list of four justifying questions (why this, why now, why this way, and why should the reader care), I noticed that the question “why you” was missing. Do I need to answer that question also?

A: The why you question is not answered directly. You have four ways to answer it. (1) Your track record — the trail of references from your previous papers. (2) The acknowledgements — if people fund your research, it is because they are confident that you can deliver something of value. (3) The most senior author mentioned in your list of authors — a well-known and well-cited author acts as a warrant. (4) The reputation of your research institution and its track record in well-cited research papers in your field.

But if you are new to the field, an outlier with no research sponsor, no academic heavyweight on your list of authors, and no glamorous university or research center to attach your name to, do not despair. Even if the odds seem against you, what matters in the end is the quality of your writing and the timeliness and impact of your research.

Do not forget that besides the four reader questions you mention, an additional two sets of questions on problems and solutions come from the reviewer: (1) Is the problem real and is it a useful problem to solve; and (2) Is the solution novel and better than other solutions.

Q: Should I always place the bracket containing a reference at the end of a sentence?

A: The reference must be unambiguous. For example in the following sentence, where should the reference be placed to avoid ambiguity when referring to the review paper, in * or in **?

Three speech recognition technologies * are prevalent today: Hidden Markov Models, Neural Networks, and statistical methods such as Template Matching or Nearest Neighbor **.

The correct answer is * because ** would be ambiguous as it may only refer to the nearest neighbor statistical method, not to the review paper that covers all three technologies. In short, the reference should be placed immediately after the information it references. If each technology has its own reference, then the following scheme would apply:

Three speech recognition technologies are prevalent today: Hidden Markov Models [1], Neural Networks [2], and statistical methods such as Template Matching [3], or Nearest Neighbor [4].

Q: What references should I put in the reference section?

A: Multiple types of references: (1) recent references because they show that you are keeping up-to-date with what is happening in the field; (2) references to papers published in the journal you are targeting⁷; (3) references to review papers when the information you use can only be found in them, but (4) references to the original paper and not the review paper when the review only points to the original paper without adding value; (5) references to papers YOU HAVE READ; (6) references to all the papers that have directly contributed results, data, or methods to your paper; and (7) references to the papers from the leader in your field.

Q: What references should I NOT put in the reference section?

A: References simply cut and pasted from one paper to the next, references to papers you have not read, references loosely connected to what you are doing, references to mildly relevant papers published by one of the following classes of people: your researcher friends, your manager, people from the same university or institute, or the unknown researchers who cited your paper — just to return the favor.

Q: How do you paraphrase? Is it good to use websites that paraphrase the text you submit?

A: Do not paraphrase while looking at what you want to paraphrase because you could easily introduce plagiarism that way. Instead, just read and get a good idea of what the other paper says. Then, with the paper out of sight, summarize the main points using your own words, and then you add the reference to that paper at the end. Unlike humans who benefit from having a deep semantic understanding, paraphrasing websites start from the words in the original text, not from a place of understanding. As a result, many times the paraphrasing introduces errors in your writing, and even sometimes distorts the original facts. Avoid!