Practical skills for writing papers, though absolutely central to the profession, are rarely explicitly taught. I don't know why such an important skill is so often overlooked.

This section aims to address that inefficiency by teaching pragmatic skills.
Fuck hidden knowledge. Fuck soft-skills learning by lucky osmosis or innate talent.

Before Writing

During the experiment design phase, think about writing while designing by asking yourself "What does this research look like as a paper?"

Consider how you will explain you process to others, i.e. reviewers. This approach prompts you to consider design processed more deeply, clarifying or removing superfluous components.

Since a picture says a thousand words, also ask yourself "What figure would this paper contain?"

Ideally, part of the design phase is creating a pre-registration on the Open Science Framework, but that is a topic deserving of its own section.

Also consider the potential importance and impact of the work itself, which will guide you when considering the tier of journal to which this work probably belongs. Ask yourself whether the research findings will be important to you sub-area or relevant to a broader audience.

How people read papers

When academics read papers, they tend to read them in the following order:

1. Title
2. Figures
3. Abstract
4. Maybe some other part of the paper, if you're lucky

From that, learn these lessons:

• Title: Write an informative title. Be concise. Use the title to ground your narrative and communicate your "take home message" if possible.
• Figures: Always use figures if you can! Make informative figures that are aesthetically pleasing. Ideally, your "take home message" can be communicated through a figure. Try not to include pointless figures that distract from the main message.
• Abstract: Abstracts can end up forgotten until submission time. This is foolish since the abstract is often the only substantive text people read! Do not "bury the lead" by being coy: your abstract should be as informative as possible and convey the most interesting results of your paper.

Software and Tools

Ask your supervisor if they have specific software or tools or a process that they expect their lab to use.
Collaborating becomes easier when you all use the same tools. You don't want to send someone a Google Doc if they expect a Word doc or vice-versa. There is no single best tool. In practice, the best tool is the one everyone is using so make sure your collaborators are on the same page.

I prefer these tools:

• Zotero for reference management
• MS Word for writing
• R for analysis; ggplot2 for charts/plots
• MS PowerPoint for simple diagrams/flowcharts
• (optional) Adobe Illustrator for advanced diagrams/graphics
• JANE (and relevant experiential knowledge/ask supervisor) for picking a journal

Finally, before writing, watch this.
This video discusses how academics are conditioned to write poorly during their undergraduate degrees. They are taught to extend their writing to meet page-limits rather than to write concisely. Many undergrads "try to sound smart" by changing their vocabulary into overly erudite verbiage (e.g. "utilize" rather than "use"). Break these bad habits immediately.

When you are writing a paper for publication, you are writing to an audience of experts that want to learn information. Your audience doesn't want you to "sound smart" or write more than you need. They want you to be clear, concise, and relevant. They want you to convey information in a way that helps them understand and use it.

Writing

There are two steps

• Write your first paper
• Write your n+1th paper

Write your first paper

I wrote my very first paper long enough ago that I have a hard time remembering exactly what I did. Here is what I recommend based on my reasoning that I find it easier to start with something rather than starting from a blank page.

• If your work is related to anything you did in any course, use the method described for writing your n+1th paper (start with the closest paper, then change/edit/cut/add).
  • If you had to write something for an Ethics application, e.g. a rationale, paste that information into your document.
  • If you have nothing at all, find a relevant paper you have read, then pull the structure from that paper. Take the headings from that paper as a starting point just to get an outline going.
• Bring in your hypotheses from your pre-registration
  • If you didn't pre-register, write your hypotheses before you analyze data!
  • If you didn't write your hypotheses before you analyzed data, fuck... write your hypotheses as they honestly were, but plan to write them in advance next time!
• Write the methods section. Methods is the easiest section to write because methods are just facts about what you did. Follow conventions from papers you've read. Be concise, but complete.
• Write the results section. Results are often the next easiest to write because this section typically involves reporting statistical analyses without much elaboration. This is also where you try to add relevant figures (always have at least one figure if possible!)
• Write the discussion section. This often starts with a verbal summary of the results, then branches into interpretations, then connects to the wider literature context.
  • Write a COG Statement as part of writing your limitations section.
    See Simons, D. J., Shoda, Y., & Lindsay, D. S. (2017). Constraints on Generality (COG): A Proposed Addition to All Empirical Papers. Perspectives on Psychological Science, 12(6), 1123–1128. https://doi.org/10.1177/1745691617708630
• Write the introduction section. Start broad, then narrow quickly and get to the point. You don't need to review everything. The introduction is there to set up your narrative, not review the entire field. The introduction is meant to flag important issues in the reader's mind. Each paragraph should convince the reader that this research is relevant and that reading the rest of this paper will provide them with answers or context that could help them understand this area.
  • Remember that your main audience is expert peers that also do research in this area. You are not teaching them the basics. You are flagging ideas for context. Your literature review should be short and purposeful. It should make it clear that your study was a next reasonable step to take. When writing your introduction section, aim for concise, not exhaustive.
  • For more detail, see Writing an Introduction.
• Edit, edit, edit
• Send to collaborators for comments
  • When you get comments back, read them closely. Take them as a way to learn what you may have missed on your first attempt. In the future, learn to imagine what comments your collaborators might have before sending it to them, then address your imagined comments first.
  • You don't have to accept every collaborator's change or piece of advice. Consider their opinions, but make reasoned choices. Set a time to meet if necessary.
• Edit, edit, edit
  • If you are the first author, you are the champion. The paper will generally progress at your pace until it goes to collaborators.
  • Notice when you are "editing for content" versus when you are "editing for voice". You will develop your own authorial voice —which words you pick, how you structure sentences, how you use commas and other punctuation— and sometimes your collaborators may change the voice of a section rather than the content of a section. If that happens, review whether their edits help clarify or whether they merely make the paper sound like your collaborator's "voice". Don't waste time changing someone's voice and don't accept changes to your voice once you have developed one.
• Submit
  • Note that submitting doesn't just take a few minutes! It can take well over an hour the first few times you submit to a journal. Submitting involves various administrative tasks, like making accounts on journal websites and filling in forms with author emails, affiliations, etc. Each journal has different requirements, too, some more cumbersome than others. This process can take quite a while so mentally prepare yourself for the banal frustrations of administrative purgatory. Rather than rail against this unpleasant inefficiency, learn what tends to be needed, then learn how you can make this process smoother each time you do it. For example, before planning to submit, always collect author emails, official affiliations, and preferred name spellings (Does someone like to use a middle initial?).

Sometimes, I find it helpful to establish the narrative by making a presentation before writing.
There are cases where a presentation can be needed before the paper gets written, too. For example, when giving updates during a lab meeting or when presenting preliminary results at a conference. Use these opportunities to figure out what structure helps convey the narrative arc of the research, which you can then use when you write the manuscript.

Write your n+1th paper

I find it easier to edit than to write on a blank page. With that in mind, here's how I write papers:

• Setting up the document; not writing. I find the last paper I wrote, or the most similar paper I most recently wrote, then make a copy of the Word Doc and rename it.
• Setting up the document; not writing. I change all the text to a different colour so I know that it hasn't been replaced yet.
• Write Hypotheses at the end of the Introduction; writing. Lucky me! I pre-registered my study because that's how to do good science. Thanks, past-me!
• Methods section; writing. I start making changes. I add or delete as-needed. I restructure every sentence so I'm not self-plagiarizing. After I change a sentence, I set its colour to black. I write as much of the Methods section as I can as soon as I start running the study; this section doesn't need to wait for data to be collected (and details to be forgotten).
• Run analyses; not writing. I run the analyses I pre-registered, then run anything extra as exploratory.
• Conceptualize results; not writing. I think about the results and the story they tell, then think about how to communicate that in a paper. I want to be concise and build a story the reader can follow. I come up with figure(s) to impart results faster.
• Results section; writing. I outline the results section according to the story I'm going to tell. This often has to be done from scratch because each paper is different. I paste in the results from my R analyses using packages that take linear models and return APA formatted results. I put in pretty plots made with ggplot2.
• Discussion section; writing. I start by writing the summary paragraphs. The style in my area is to essentially re-write results as text rather than numbers, then interpret, so I do that. I write the summaries first since that's easiest, then let the interpretation simmer in my mind.
• Revisit Introduction; writing. A lot of introduction text can be "refactored" from prior introductions if I've published in the area before. As with Methods, I make sure to change every sentence, but a lot of the same ideas are usually present. If I need new ideas for the story, I introduce them. If an idea is no longer relevant, I ruthlessly cut it. I want my introduction section to be concise, not exhaustive.
• Editing; writing. I edit a lot. I edit much more than I write. I'll bounce around to different sections, making sure wording/naming/framing/ordering is consistent. For example, I'll make sure the order of hypotheses is 1,2,3 in the Intro, in the Results, and in the Discussion. I do not do "synonym hunting"; I use consistent language to help the reader understand quickly. A paper isn't a novel that gets points for flowery language; a paper is for communicating information clearly and concisely.
• Preempting comments and reviews; writing. I think about comments my collaborators might give (based on working with them in the past). I think about issues reviewers might raise. I address them as much as I can in another editing pass before sending to anyone. Over the years, this has meant I essentially skip the first round of comments because I already know what gets flagged by people with whom I work or by reviewers.
• Send to collaborators; not writing. I send the manuscript to my collaborators, whether that is a peer or a supervisor. Then, I work on something else while I wait for their comments. I put this paper out of my mind completely.
• Revising based on comments; writing. I read all comments with charity and "benefit of the doubt". If it seems needed, I might schedule a meeting to discuss. Ultimately, when I am the lead, I only implement comments with which I agree. I try to take them all to heart, though, and address any shortcomings. One thing I do not accept is what I call "edits of voice". People I work with learn not to bother rewording sentences into their preferred wording. It took a while, but my PhD supervisor finally learned to stop adding commas all over the place since I rejected those changes.
• The cycle: Resend to collaborators, Revising based on comments, Resend to collaborators, etc. The better I get, the fewer revisions it takes. My Master's work took many revisions; now a paper can take as few as two.

Index

Return to Start Here

Jump to Writing an Introduction