12.4 Writing a Scientific Paper

The last step of the analysis phase is to write a scientific paper. This may seem like a separate activity but you will soon realize that explaining your results to others is essential for your own understanding. The most important thing to make clear is that a scientific paper is not a laboratory report – it is a scientific argument. You are not writing a summary of activities carried out in a laboratory to demonstrate that you master a technique or a piece of theory. You are not expected to give a full account of everything that you did in the laboratory – mistakes, detours, troubleshooting, and so on – unless they have bearing on the conclusions and contribute to formulating a coherent scientific argument. The readers will not be interested in how difficult it was for you to arrive at your conclusions. They read to take part of the conclusions and want to see the evidence for them.

Writing is part of the analysis phase because it is a way to explain to yourself, as well as others, how your conclusions are connected with your data. If you cannot explain this, the analysis is not complete. Writing also involves putting your findings in relation to the knowledge in your field. As a result of your experiment this knowledge will have changed, if only a little, and your paper should make this change clear. Writing thereby also marks the beginning of the synthesis phase: that of adding a specific piece of research to the general scientific knowledge. These two phases are treated as one in this chapter as they are two sides of a coin: your research question addressed a limitation in the current knowledge, and the answer thereby makes a natural contribution to that knowledge.

The writing process begins with formulating the argument that you are going to make and does not end until the paper has been published. Your manuscript will go through a series of reviews and revisions along the way. Before submitting it to a journal or a conference you will probably discuss it with your colleagues and your academic supervisor. After submission, fellow scientists in your research community will review your text. These are appointed by the editor of the journal and will remain anonymous to you. Their comments will highlight issues with anything from the research question to the presentation of the results. The aim is to ensure that you have made an original contribution and that your paper is up to the accepted quality standards of your field.

We are not going to discuss all technical aspects of writing here, as there are whole books on this subject. There are also plenty of guides to technical writing on the Internet, as well as journal-specific author guidelines provided at journal home pages. Though there are minor differences between publishers, for example in reference style or in allowing authors to write in the first person, research papers generally adhere to a standardized format that is relatively fixed across disciplines. We will briefly discuss the parts of a scientific paper, as it is important to understand how results are communicated within the scientific community. A paper is generally built around the following rough headings:

• Background
• Method
• Results
• Analysis
• Conclusions.

The purpose of the background section is to present your research question. It explains the central problem of your study and justifies why it is worthy of closer investigation. It also describes the scope of your study. Which parts of the problem did you address and which parts did you leave out? The reader must be able to understand from your description what the research question is and why you have chosen to answer it. This naturally requires some discussion of the current status of knowledge with references to the relevant literature. It is good practice to start the background section with a general picture of the problem area and then, over a couple of paragraphs, gradually move closer to the specific problem that you have studied. This sets the scene for your research and guides your readers from the common ground that you share with them to the point that is of specific interest in your paper.

The background section also contains theory. Although scientific papers are written for specialists, it may be necessary to explain certain ideas and concepts that were used in the design or the analysis of your experiment. Any definitions that are needed to understand the line of reasoning should also be presented here. If the theory requires significant space, it may be appropriate to divide the background section into an introduction and a theory part.

The purpose of the method section is to explain your practical approach to the research problem. Here, you describe which data you collected, how you collected them and how you arrived at the results. It covers instruments and measurement techniques, experimental setups, materials, measurement procedures, as well as how you processed the data before the analysis. You must explain these things at a level of detail that allows the reader to replicate your study and arrive at similar results. The method section should also describe and explain the experimental design. This will include a discussion of how you handled background factors and noises.

The results section contains all the evidence that your conclusions are based on. It is not acceptable to present an arbitrary collection of tables and diagrams. The data must be presented in terms of the research problem. This will make it possible for the reader to follow your line of reasoning through the paper. When you use statistics, you should explain the use of the specific statistical techniques. It must be clear to the reader that they have been used correctly. Make sure that you include all the data that bear on your conclusions. If measurements are to be left out there must be very good reasons for it; these must be clearly stated. Recall the discussion about Robert Millikan in Example 6.5, who was accused of academic dishonesty for leaving out observations that he deemed unreliable. Your paper should give an honest and objective account of the support for your conclusions.

The analysis section interprets the data for the reader and explains how they answer the research question. The conclusions must be entirely supported by the data presented in the results section. If you want to discuss ideas that are not a direct consequence of the data, present them as new hypotheses and make clear that they require further investigation. This section should state whether the data support the initial hypothesis or not, explain the implications of the results, and point to further studies that are needed to provide a more complete answer to the research question. If the paper treats several subproblems, it may be perceived as repetitive to discuss the implications of the results in a separate section. For this reason, the two sections are sometimes merged into a combined results and analysis section.

The conclusions section simply contains a condensed summary of what has been learned from your study.

These sections are preceded by the paper title, a list of authors, and an abstract. The title should convey the essence of the paper in a single phrase. It may be a simple statement of the central conclusion or the research question. The title may also refer to methods or techniques used in the paper, if this is relevant to the reader. It is important to understand that researchers search the literature in bibliographic databases. This is where your paper is most likely to be accessed after publication. A literature search is based on keywords that are matched with specific parts of papers. You can search in the titles, the abstracts, the author lists, and so on. The search engine usually lists the titles of the papers that match the keywords. This is why it is important to choose a title that accurately targets a relevant part of the research community. They must understand from the title if your paper is worth a closer look.

The first step in taking a closer look is to read the abstract, which is usually a single paragraph summary of the paper. It describes the central research question, the method used to address it, and the central conclusions. If the abstract confirms the impression that the paper is relevant, the person making the literature search will download the actual paper from the publisher and read it in its entirety. Once you have made a literature search yourself, you will realize how important it is to convey the central message of the paper in the title and abstract.

At the end of the paper there should be a list of research papers that you have referred to in the text. The specific format of this reference list varies between journals and is therefore described at each publisher's home page. There is usually an acknowledgement section too, where it is customary to thank relevant funding bodies for research grants. Such organizations often search the databases to evaluate the research that they have supported. For this reason, they may demand that researchers state the source of their funding when publishing their work.

Keep in mind that a research paper does not aspire to be a work of fine literature. It is supposed to be a concise report written in factual prose. Above all, it should be clear and brief. A long manuscript may even be reason for refusal. The best way to understand which style is appropriate for a given journal is to read a few of its articles. One of the main challenges is often to maintain a clear line of argument all the way from the background through to the conclusions. There are probably as many strategies to achieve this as there are authors, but you may find the following technique useful.

Before you start to write you need to make clear to yourself exactly what your scientific argument is. The central message of your paper is, of course, most clearly conveyed by your conclusions. For this reason it could be useful to start by writing down the conclusions and then finish the remaining sections in the reverse order to that in which they will appear in the paper. This is because each section provides the necessary support for the subsequent section: the analysis leads to the conclusions, the data is the basis of the analysis, the methods provide the data, and the research question determines which methods to use. This method of reverse writing allows you to follow the river back to its source, so to speak. If you begin by writing the background section, there is a greater risk that your line of argument becomes diluted in detours and digressions before you reach the end. There is also a risk that you become tired before you reach the sections that are most central to your argument. The analysis and conclusions sections are where you need to be the most sharp and focused in your writing. As these sections are so important, there are even people who habitually read research papers in the reverse order, starting with the conclusions and finishing with the background.

Practically, when writing a paper in this way, you begin by printing out the diagrams that demonstrate the validity of your conclusions. Pin them to the wall by your desk so that they are constantly available to you during the writing process. Create a text document and put all the section headings down in the right order. Write your conclusions under the final heading, briefly and to the point, in the form of a bulleted list. Copy this list and paste it under the analysis heading. Under each bullet, write down the main points of the analysis that lead to the conclusion. Do not worry about language and style at this point; just concentrate on putting down all the relevant information. Then move to the results section and describe the data and diagrams that underlie the analysis under each point. Continue backwards through the headings like this, describing the methods used to collect the data in the methods section and how the research problem led to these methods under the background heading. Whatever approach you take to writing, do not begin to write before you have all the conclusions in place. Writing a paper at the same time as you process and review the data is a safe method for obtaining a vague and confusing manuscript.

Another useful technique is to present your work orally to colleagues before starting to write. This presentation should include all the parts of a scientific paper in the right order. The feedback you receive will reveal whether you manage to convey your argument or not. Once you have worked out a clear and concise presentation, what remains is basically to transfer it into a text document and put on a finishing touch.

Authoring a research paper can be compared to being a tour guide. Needless to say, your readers cannot read your thoughts, so you must explain your thinking to them. Although the contents of the paper are summarized under standard headings, you must tie these together into a comprehensible chain of thoughts. After introducing the research idea, you must explain how it led to your approach, how the approach resulted in the data, and how your analysis led you to the conclusions. Reading a paper where the author has neglected to guide the reader is like walking through a museum where historical objects are on display without descriptions of what they are, where they came from, and what we might learn from them.

Finally, the peer review process is an integral and important part of writing up your research. It can be frustrating to obtain anonymous, critical comments on a text that is the result of great personal effort. Review comments are often brief and may sometimes be perceived as impolite, especially before you are used to the process. It is important not to take the criticism personally. Reviewing a paper is time consuming and, since referees are fellow researchers with many other tasks on their hands, it is understandable if they do not have time to formulate their comments as diplomatically as might be wished. We should rather see each comment as a valuable opportunity to improve the quality of our paper. They often indicate points where we have failed to explain our thoughts clearly, or where important information is missing. Intense work with a manuscript tends to make us blind to certain aspects of the text and it is valuable to get comments from people who have read it with fresh eyes.

A manuscript will typically be reviewed by several anonymous referees. After obtaining the review results you are expected to do two things. Firstly, you should write a rebuttal where you give detailed responses to each of the comments. Secondly, you should make appropriate updates to your manuscript. The rebuttal should indicate where these changes have been made, and it is good practice to highlight the changes in the manuscript so that the referees may find them easily. After submitting the revised manuscript and the rebuttal to the publisher, there is generally another period of waiting before you are notified of the journal's decision to publish or not to publish your paper. The publishing process usually takes additional time, so it may be a year between the original submission and the publication of your article in the journal.

In the mean time you will of course be fully occupied with your next research task. In all likelihood you entered the planning phase of this task when analyzing the data from the previous one, because analysis has a tendency to uncover new aspects of a problem and generate new ideas for experiments. As illustrated by Figure 12.1, the circle of research is never completed. In fact, it is impossible to say where it begins and ends.

Communicating one's results within the scientific community is one of the most important parts of doing research. Despite this, Ph.D. students in some places are not required to publish their work. Since writing is as difficult a skill to learn as anything else in research, it is doubtful whether a person can be considered to be an accomplished researcher before being published. Imagine calling yourself a journalist if you had never written a newspaper article. As research students are in training to become autonomous researchers, my recommendation is to publish as much as possible of your work prior to graduating. Writing papers is a central research skill and, as any other skill, it can only by learned by performing it. An additional advantage is that published papers confirm that your work meets the accepted quality standards in your field. When defending your thesis one day, your peer reviewed journal articles actually constitute objective proof of your research skills that can be used to meet the potentially more subjective opinions of the members of your grading committee.