20.109(F15):Guidelines for writing up your research
Instructions for writing specific sections of a scientific paper
Formal research articles and less formal lab reports are the principal ways scientific data is conveyed to the rest of the scientific community and preserved for future examination. Each scientific journal has its own idiosyncrasies regarding format particulars, but the most common elements of a scientific article, in order of presentation, are:
- Title
- List of Authors
- Abstract
- Introduction
- Materials and methods
- Results, including figures and tables
- Discussion
- References
The requirements for each section are outlined below. If you want more information, you can find parts of this text in an on-line collection of instructional materials used in the Purdue University Writing Lab (http://owl.english.purdue.edu). Other parts are inspired by Robert A. Day’s book, “How to Write and Publish a Scientific Paper” from Oryx Press, a copy of which can be borrowed from the teaching faculty. Still other content is based on the research and teaching experiences of former 20.109 instructors Natalie Kuldell and Agi Stachowiak.
Title
The title should be short (about 10 words), interesting, and it should describe what you found. Avoid very broad titles (e.g., 'DNA recombination'), and try to specify your own unique result within the limited space, while still maintaining readability of your title.
List of authors
Indicate that you are the primary author of the report by underlining your name. Others who substantially contributed to your work, such as your lab partner, should also be listed.
Abstract
The abstract serves as a condensed version (not >250 words) of your report, from background (both context and motivation) to key results (and how they were found) to implications for the future. By convention, it should be single-spaced and not include citations.
The importance of a good abstract cannot be overstated since computers generally index the words in a paper’s title and abstract, and thus these may be the only parts that many people read. AT MIT we are very lucky to have access to almost any journal we could want, but not everyone is in this position. The abstract may also be the way a journal’s editor decides whether to send your paper out for peer review or reject it as uninteresting and not generally relevant.
Introduction
The introduction provides a framework for the story you are about to tell, and thus serves two main purposes. For one, you must provide sufficient background information for a reader to understand the forthcoming results. Just as importantly, you must motivate the audience to keep reading! How? Reveal the significance of the work through connections to both prior scientific accomplishments and interesting future applications.
You are welcome to use your own creativity and judgement as to what a good introduction should look like; however, note that most introductions are "funnel" shaped in terms of content:
- Opening paragraph(s): most general, "big picture" paragraph(s). Here you should introduce the reader to the broader context of your experiment and motivate why your research is important. The best introductions tell a coherent story rather than present a dense but unconnected list of facts.
- Middle of introduction: "zooming in" somewhat. Once the reader has a frame for thinking about your research, you can present background information in more depth and motivation with more specificity.
- Wrapping up: most specific, a description of your particular investigation. Here you should make your overall methodology clear without getting into minute detail. In many journals, the introduction concludes with a brief preview of key findings and their implications (2-3 sentences total).
Materials and methods
The methods section should allow an independent investigator to repeat any of your experiments. Use sub-section headings to allow readers to quickly identify experiments of interest to them (e.g., "Protein conjugation to hydrogels" or "Cell culture and fluorescent labeling"). When commercially available kits were used, it is sufficient to cite the name of the kit and say that it was used according to the manufacturer’s protocol. The key to a good methods section is developing your judgement for what information is essential and what is extraneous.
Note that the methods section should be written in the past tense, since your experiments are already complete at the time you are writing your paper. This section should also be written in complete sentences and paragraphs, not in bullet point form.
Results
The purpose of the results section is to present your data in a relatively unbiased way, but with some guiding framework. Begin with a short description of the goal and strategy of your overall experiment, and then delve into specific sub-sections that describe each piece of the work.
Titled sub-sections help support your high-level narrative and make dense papers easier to read. Effective sub-sections are organized by functional content, not by what you did each day in lab. For example, two successive titles might be "Conjugating lymphoid tissue-associated adhesion proteins to hydrogels" and "T cell migration in protein-coated hydrogels." Less helpful titles for the same content would be "Hydrogel-protein reaction" and "Microscopy and velocity analysis."
To write the results section, use the figures and tables as a guide. One approach you might take is to start by outlining, in point form, what you found, going slowly through each part of the figures. Then take the points and group them into paragraphs, and finally order the points within each paragraph. Present the data as fully as possible, including stuff that does not quite make sense at first glance. Ultimately, each sub-section should begin with an overview sentence that motivates and introduces the present experiment and end with a sentence stating the primary conclusion reached from that experiment. (Sub-section headings and figure caption titles can also emphasize said conclusion.) The overview and/or concluding sentences should also provide a transition to the previous/next piece of data when possible. Within a sub-section, be sure to stick to one topic per paragraph; sub-sections will generally require a few paragraphs each.
Note that verbs in the results section are usually in the past tense. Only established scientific knowledge is written about in the present tense: “the world is round,” for example. You cannot presume that your own data are part of the body of established scientific knowledge. Thus, when you describe your own results, use the past tense: “a band of 1.3 kb was seen,” for example. There are exceptions to this general rule, such as “Table 3 shows the sizes of the DNA fragments in our preparation” or “In a 1991 paper, Ebright and coworkers used PCR to mutagenize DNA,” which are both acceptable statements.
Discussion
The purpose of the discussion section is to interpret and contextualize your data. You should begin by reiterating the purpose of your research and your major findings. Then you might do any or all of the following: connect your findings to other research (published or that of your peers); describe any ambiguities and sources of error in the data, and suggest future experiments to resolve uncertainties; explain where you expect your work may lead, and suggest specific experiments for extending your findings; describe any conceptual or technical limitations of the research. Finally, you should explain the significance of your findings to basic science research and/or to engineering applications. As with previous sections, the discussion should have a clear organization and narrative flow, whether or not you use sub-sections.
References
References are generally used in the Introduction and Discussion sections to support any claims that are not common knowledge. Include only those references that pertain to the question at hand. Journals vary considerably in their preferred format for the reference list. For this class, you should list the references alphabetically by the first author’s last name. Include all the authors, the paper’s title, the name of the journal in which it was published, its year of publication, the volume number, and page numbers. Please carefully follow the punctuation and format requirements. A typical reference should look like
Pavletich NP, Pabo CO. Zinc finger-DNA recognition: crystal structure of a Zif268-DNA complex at 2.1 A. Science 1991; 252:809-817
In the body of your report, this article would be cited as follows: “The crystal structure of the Zif268-DNA complex has been solved (Pavletich 1991).”
If two or more articles can be cited for this finding, then they are listed alphabetically, separated by a comma.
See also the MIT libraries information on citing sources, here.
Figures and tables
Some readers begin by scanning the figures, so these and their associated legends should provide a self-explanatory overview of your data. Decide what the data show, then create figures that highlight the most important points of your paper.
Tables are used to present repetitive data that is numerical. Graphs or illustrations, collectively called figures, are used to present numerical trends, raw data (like a picture of a gel), or a model that explains your work.
When you prepare your figures and tables, keep in mind that it is significantly more expensive for journals to publish figures and tables than text, so try to present the data in a way that is worthy of such added expense. The table below is an example of an ineffective table.
Temperature Repeats Cortical cells Ion flow 24°C 5 + + 24°C 5 - -
The information in Table 1 could be presented in one sentence, such as: “In ten experiments carried out at 24°C, ion flow was detected only in the presence of cortical cells.” This is a clearer and more concise way to present the information. In addition, all tables and figures must have numbers, titles, and legends (also called captions).
Figure and table legends
Legends to the figures and tables explain the elements that appear in the illustration. Conclusions about the data are usually NOT included in the legends (the choice is somewhat journal-dependent). As you write your first draft, you might state in a short simple sentence what the point of the figure or table is. In later drafts, make sure each element of the figure or table is explained. Your figure legends should be written in the present tense since you are explaining elements that still exist at the time that you are writing the paper.
Recommended order of assembly for a scientific paper
Your data (results and figures) is the heart of your paper. Begin by assembling and understanding it as clearly as possible. Then you can write the framing material – introduction and discussion, while compiling a few relevant references. (Actually, you will write a draft of the introduction about halfway through the module, once you have begun to develop some understanding of the field. The final revision is best written once you have all your results.) You should generally write the abstract last, once you have a big picture understanding of every element of your investigation. As for the methods section, you might write it in short pieces day-by-day, while the procedures are fresh in your mind.
Additional scientific writing tips
In our experience, concentrating on the following topics tends to repay student effort.
You may also wish to read the article published in 2006 by the EMBO journal called Writing Readable Prose. It makes a strong case for the role that writing (good or bad) can play in your career development, and also makes explicit some elements of clear writing.
Finally, you'll learn a lot by working with the BE Communication Lab fellows!
Structure and repetition
The readability of your paper can be substantially improved by its organization. Well chosen sub-section titles provide the reader with a distillation of the high-level points you want to make. Take the opportunity early and often to summarize major points - if you do this right, you will feel like you are repeating yourself too much! Each major section of the report should be more or less stand-alone, that is, understandable without having read the rest of your paper. Be sure and work from big to small to big (from broad points to details and back) throughout your paper. Finally, limit paragraphs to one topic apiece, splitting up longer linked ideas as needed.
Results vs. discussion
One difficult task for budding scientific writers is determining what goes in a Results vs. a Discussion section. Even at the professional level, different scientific journals may have somewhat different expectations for how much interpretation is appropriate in the Results. A good rule of thumb in most contexts is to draw technical conclusions in the Results, but save scientific conclusions for the Discussion. An alternative framework that should yield similar writing choices is to focus on relatively certain conclusions in the Results, and save more speculative (but still supported!) conclusions for the Discussion.
Example of a technical conclusion: "Sequence alignment revealed that both mutants exhibited mutations in at least one of the expected sites and thus were consistent with the possible nucleotide changes that could occur with the library used."
Example of a scientific conclusion: "A previously reported model proposes that the OmpR-P phosphatase reaction occurs by hydrolysis rather than direct phosphotransfer from H557 (Hsing 1997). Our mutant is consistent with a change to the structure of EnvZ that facilitates this hydrolysis reaction, thus increasing phosphatase activity while concurrently reducing kinase activity."
Holistic view of data
What should you do when two independent pieces of data suggest two very different conclusions about your overall experiment? The one thing you should not do is state that both conclusions are correct, in different parts of the paper, and hope that the reader will ignore the discrepancy! As a scientist, you must take a holistic view of your analyses. Is there any way to reconcile the two results? Is one method more trustworthy than the other or is there otherwise an indication to favor one conclusion over the other? If you have no way of determining which result is correct with your present data, can you propose future experiments to conclusively decide?
Concision
Eliminate unnecessary words and phrases. Writing concisely takes more effort than writing inefficiently. It also results in more readable text that expresses your ideas more powerfully.
Precision and detail
Being concise is not an invitation to be vague. Be sure to support your claims with specific examples, and to follow through on your ideas with sufficient detail.
Writing is revision
The first draft is a record of your ideas; the second draft is a crafted presentation of them. Draft as much of your report as possible early on, so you have the opportunity to revise it before submission. Even better, outline the content and high-level organization of your paper before you even begin drafting. Students who plan ahead in this way tend to submit particularly successful reports.
Evaluation
The following table describes the features of research papers that are excellent, good, and that need substantial improvement, respectively from left to right. Your own manuscripts will be evaluated with these features in mind.
Content
Section | Goal | Evaluation | ||
---|---|---|---|---|
Title | To give content information to reader |
|
|
|
Abstract | To concisely summarize the experimental question, general methods, major findings, and implications of the experiments in relation to what is known or expected. |
|
|
|
Introduction | To identify central experimental questions, and appropriate background information. To present a plausible hypothesis and a means of testing it. |
|
|
|
Materials & methods | To describe procedures correctly, clearly, and succinctly. |
|
|
|
Results | To present your data using text AND figures/tables. |
|
|
|
Discussion | To evaluate meaning and importance of major findings |
|
|
|
References | To give credit to foundational work (on which your own study is based) |
|
|
|
Style
Writing style and mechanics | Evaluation | |
---|---|---|
Voice |
|
|
Word choice |
|
|
Fluency |
|
|
Scientific format |
|
|