so they forgot to tell you how to write a paper

18 min readJul 7, 2021

For whatever reason, you may find yourself in a PhD program while somehow not actually knowing how to write a paper. It happens! You’re not a fraud (that I know of) and you’re not doomed to failure. Hopefully you’re reading this before you get to the big paper. If you’re reading this and trying to write the big paper, consider reading my other thing about the big one. You may also be a go-getter and reading this pre-grad school, in which case I congratulate you on being ahead of the pack here. Lots of people learn how to write papers in grad school, but certainly you will have a less stressful time if you can focus your energies instead on other stuff, like sleeping and eating.

The purpose of this post is to tell you how to organize and pace yourself while you write a paper. When I refer to “a paper” I mean an academic article for journal submission or qualifying papers and the like — this can apply to things like term papers for a class but you will find that instructors have specific stuff they are trying to get you to do. It’s a good idea if you’re writing a term paper for a class to investigate (by reading the syllabus or emailing your prof) what the learning goal is, and prioritize that. The point of a term paper usually isn’t to have a good finished product at the end, but to use the process of writing to cause you to absorb information… so it’s good to check.

The other important caveat about this post is that it’s not the One True Way to write a paper. It’s a way that helps me, and when I tell other people they say it helps them, so I’m sharing it with you. If you don’t find this workflow useful, you are not obligated to do it and it won’t necessarily mean your paper sucks.

Overview

This workflow is broken into three parts. Pre-writing is all the stuff that I need to do before I can write. Writing is, you know, the part where you make blank pieces of paper less blank. Post-writing is all the stuff that I need to do after the pieces of paper are quite full. I don’t necessarily complete the stages fully before moving on — sometimes I do Stages 1 and 2 on a section, then do stages 1 and 2 on another section, and so forth until the whole piece is done and I can do stage 3. In general it is hard to do stage 3 to something I haven’t done stage 2 for at all, though (because, like, it is hard to edit before I’ve written anything).

The idea of these workflows is to give you a framework for figuring out how to structure your time. One of the difficulties of writing as part of your job is that people don’t tell you what you need to be working on day to day, so it’s easy to show up to work on Monday and look at your big heap of things that need to get written somehow, and get totally overwhelmed and not be able to start. It’s really easy to let this snowball and result in last-minute panic and you stay up all night in a fugue state writing 20 pages, but that’s unsustainable and the fatigue is cumulative and even if that “technique” seems to work okay for you right now, at some point you’re going to hit your limit. Better to get out ahead of the problem and learn some task-wrangling skills and save your future self a lot of misery.

Stage One: Pre-writing

It would be helpful, before you start writing, if you knew what you were going to write.

The above sentence could mean a lot of different things. For example:

It would be helpful if you have collected and analyzed data (if that’s the sort of thing you’re into)
It would be helpful if you knew what length or form requirements you are subject to (to know the scale of the project if nothing else)
It would be helpful if you knew genre expectations for this particular piece (e.g. specific journal preferences, trends, flavors that aren’t explicitly part of their style guide but are just tendencies)
It would help to know what parts will go in what order.

Here are suggestions on how to determine each of those things. After that will be suggestions about how to organize and use that information

Pre-writing stage 1a: what is the finished piece for?

Hopefully you know what kind of thing you are writing. If you are writing a research paper for you graduate program requirements, or writing with the intent to submit to a particular journal, or writing with the intent to publish outside the academic literature, those all have format and genre expectations.

In general, you need the following information:

What’s the minimum/maximum length?
Do you need a lit review? How robust an introduction? Abstract?
Who is the audience for this? (Is it undergrads, or three specialists on the planet, or your parents, or your twitter followers, or what?)

For graduate/university program requirements, gather this information from the following sources:

your program/department web site may have requirements
your advisor might have opinions
you should definitely ask peers/graduates of your program if you can read theirs. (This assumes that you are not the first person to ever do a thesis/paper in your department, which for your sake I hope is a safe assumption).

Pre-writing stage 1b: what is the finished piece going to be about?

It would probably be helpful to figure out what you’re writing about before you start writing. If you’re writing about ~science~ (like I usually am), these are the things I would recommend knowing:

What is the core research question? It would probably be good to know this before trying to collect data… also, what is the hypothesis you’re testing? (If you’re into that sort of thing)
What data will you be talking about? Did you collect it yourself? Are you… going to collect it yourself? Go collect your data at this stage, please, and not any later stages.
What does ~the literature~ say about your thing? How to write a lit review is its whole own separate guide (and I will write one! Eventually!) but you should at least have a general way of organizing information you have gleaned from other authors on this topic.
What’s a rough idea of your conclusion? Yes, you should know this before you start writing. (Yes, you’re allowed to change it later.)

If you’re writing about (any other field that does not involve a hypothesis or data collection) my general advice still mostly holds, but I also urge you to figure out the following:

What is the core argument or insight you’re bringing to the piece?
What is the structure of your argumentation?

Those are smaller and vaguer because it’s been a long-ass time since I’ve written outside my field, sorry to report. But if you have figured out the above information, you’re ready to write… the outline!

Pre-writing stage 1c: what will the finished piece say?

This is the outline stage! You should make an outline. Here’s WHY you should make an outline:

Structuring an overarching argument is easier when you’re doing it in bullet-point format, rather than just sort of discovering it as you write.
If you are writing for the approval of an advisor or editor, you can make them look at the outline! This means they’ll tell you to cut Section 3 before you spend three human weeks writing Section 3. You will be so, so glad you didn’t spend three human weeks writing a section that you were just going to cut, anyways.
The outline makes it a lot easier to conform to genre conventions (such as Journal Article)
Writing an outline will force you to admit, on paper, in writing, that you totally have not finished collecting or analyzing your data yet, have you?
The outline functions as a to-do list for basically everything else that happens.

Okay, that sounds super important! I make an outline for anything longer than about 200 words. I made an outline for this blog post. Please make an outline! Nod your head at the computer or phone as you’re reading — will you make an outline? Yes? Now, here’s how to write an outline:

Start from the coarsest-grained divisions possible. Journal articles in my field usually have five-ish sections (Introduction, Lit Review, Methods, Results, Discussion/Conclusion). You can do this part of your process completely templatically by looking at papers in your field and literally just ganking their section structure.
For each of the top-level sections (except, sometimes, the Intro and Conclusion), break them down into reasonably-large chunks. For example, the Lit Review Section tends to be organized (in my field) by major threads. So you’ll get three-ish sub-sections in the lit-review: Approach A, Approach B, Mixed A+B Approaches.
Once you’ve broken down all the top-level sections into the first subsection division, you can break them down one more time if it’s helpful. For example, if I were breaking down the subsections of the Lit Review, I would break down the Mixed A+B Approaches section to major features. It’s hard to come up with subsubsection names (and I often don’t include them in the final document organization), because they’re named things like “This is why So-and-So hates the Approach A, but Such-and-Such’s (2018) reply argues that it’s fine actually” or like “Xyz realized that approaches A and B are in no way mutually incompatible so here’s what their mixed version looks like, but here’s Abc’s alternate version of a mixed approach.”

The subsubsection divisions (the last step above) are really helpful for me, because they let me see how the logic of the paper flows in the organizational structure I’ve built. Outlines also let me move stuff around without making the writing really difficult.

You can keep sub-dividing, but beyond three subdivisions (for an article) you start getting down to essentially the point of writing the topic sentence of each paragraph. This can feel extreme, but can be very useful in getting your juices going! It’s a nice transition from pre-writing to writing, so if it feels doable and you’re not using it as a procrastination method, I say go for it.

By the end of the pre-writing stage, you should have basically all of the thinking done.

By the end of the pre-writing stage, you should have basically all of the thinking done. Your data should be collected, all your stats and analyses figured out, your conclusion pretty much decided. Why? Because writing is hard enough, and it’s even harder to think as you write.

Stage Two: Write!

You do have to write the paper, in order for it to become written.

Luckily, if your pre-writing was very thorough, the writing step is totally divorced from the thinking step. The writing step is often difficult for people because, like, communicating ideas is just hard?

I’m going to list styles of “make the wordcount go up” that have worked well for me. There’s lots of advice out there on the “make the wordcount go up” part, especially about things like how to pace your work sessions and develop good work habits, so I won’t go into those too much here. I will say that many people find writing to be very cognitively intensive, so don’t beat yourself up if you find this part more taxing than other kinds of work.

That being said, here are the styles!

Option A: Mad-Libs style

This is a way to even more intensively separate the act of writing from the act of thinking. Basically, there are two phases of this style:

Phase 1: write as fast as possible, for as long as possible. Do NOT stop for anything — if you can’t think of a word, just write [NOUN] and get back to it later. Same goes for citations, figures, numbers, everything.

Phase 2: okay, you did that! You did that until you got from the top to the bottom! Now, uh, go play mad-libs. All those spots you just said [NOUN] or [CITATION] or [FINISH THOUGHT HERE IDKfjdkasl;] you now get to go fill in with nouns and citations and finished thoughts.

The advantage of this approach is that you can separate out two different types of tasks — looking up a citation and formatting it and putting it in your bibliography is a different type of task than writing an introductory paragraph! This lets you budget your brain hours more strategically, so you can do the “fast sloppy whatever” writing when your ideas are flowing and you need to just get them down without interruptions, and you can do the “actually fill in the Mad-Libs” when your brain is all done with ideas but is totally ready to do some small, slightly annoying tasks.

Option B: Bedtime Story style

This is a very popular method, even in fields like mine where there’s lots of data and diagrams and LaTeX typesetting. I genuinely do not know why, but it definitely works for some people. Basically, just, like… tell the story. Use your outline to guide the general direction, but write it as you would tell it to a friendly colleague over drinks.

I recommend only tackling one section (or subsection) at a time with this method, and I also recommend writing the Introduction last (after the Conclusion, even!). This also seems to work slightly better if you take a slightly more informal tone in the prose, and go back and edit it to sound more formal later.

Option C: Style of painting a wall neon yellow

This is the style I recommend if you’re really struggling to keep in mind the full organization of your paper as you write.

Here’s the metaphor: you know if you’re trying to paint a wall neon yellow, the paint is going to be kind of patchy and thin and look like shit on the first coat? Rather than start at the side and put three coats on a single square foot of wall, it makes sense to paint the entire wall (even though the paint is thin and patchy and shows through) and let it dry completely, then come back and do another complete coat on the entire wall.

For writing, what this means is that you go to your outline, and just keep adding subdivisions from top to bottom. After the fourth sub-division, you essentially are going to have the main idea of each paragraph; now you can go and add two or three bullet points per paragraph. Do this from the top of your paper to the bottom (paint the entire wall!).

Then you go back to the top, and start turning those bullet points into sentences. Do it all the way, from top to bottom. Add sub-bullet points iff needed, but don’t stop and futz with them too much. Keep adding that recursive fractal detail until it basically looks like a very badly-formatted paper, and then take all that bullet-point indenting away and it will instead look like a paper. Hot damn! Good job, the wall is now yellow.

Option D: Style of brain-dumping and then translating

This is as much a style of getting “unstuck” as it is of actually writing. I’ve never written an entire paper this way, but I have written huge sections this way. I do this when I feel like I know what I want to say, but the act of turning the ideas into sentences is really hard. Here is the strategy: open a new document (a textpad or other very low-distraction program, or, literally, a text message to yourself or a friend) and just explain what you’re trying to write. Use text-message writing, emojis, keysmashes, don’t fix misspellings, just spew all your ideas in there! This is especially fun if you do have a patient friend who is willing to endure you sending them 6900 texts in a row about, like, plant sex or whatever you’re writing about.

Then, put that textpad side by side with your nice clean word processing doc (or LaTeX editor or whatever) and translate! Translating “jdfksal;fdsa the nomial domnin is EXACTLY LIKE THe clausal domain im gonna shit” into “the parallels between the nominal and clausal domains are well-attested in the literature; the data presented here support theories that formalize that similarity.” Or whatever!

The point of all these methods is, basically, to make your wordcount go up. Ideally at the end of Writing Stage Two: The Writing Part you have something very vaguely paper-shaped and sized. It is often not very good at this stage, but having words on the page feels really good. You’re allowed to ride that high! Just don’t get cocky and forget to do the equally-important post-writing stage!

(Do, also, budget your time to leave roughly equal amounts of time for the writing and post-writing stage. Trust me!)

Stage Two point five: decorations

I waffle about whether this stuff properly counts as “writing” or “post-writing.” It does mess up my neat three-stage thing, but so goes it. Here are decorations you might need to consider — either during your writing, or after:

Figures: are they good? Are they in there? Are they captioned? Did you, like me, spend a week relearning R (every single time) so you could make really nice plots? Are they colorblind-friendly? Do they fit on the page? Do they have the right axis labels? Make your figures good!
Formatting: are your margins and spacing and fonts and page numbers how they should be? If you’re a linguist like me, you also need to make sure your examples are numbered and typeset correctly, and referred to correctly in-text. Make the page look good!
Metatext: do you need a title, abstract, keywords? Author order and appropriate related stuff? Table of contents? Section headers?
Bibliography: DO NOT DISCOUNT how long it might take, even if you’re wisely using bibliography management software. You will need to make sure all citations are in there, and correct.

This stage is not to be ignored, and scales up to become quite massive. If you’re writing something big, like a dissertation or a book, budget a good bit of time for this stuff. Like, seriously, a month or more if you can spare it. You’ll want to do this stuff before you go on to the “proper” post-writing stage of editing or showing it to someone. I do some of this stuff as I write (mostly figures and bibliography) and some after the paragraphs are mostly paragraph-shaped (metatext, formatting).

This stage should be guided by, like, guidelines. If you’re writing a journal article or a dissertation, there will be extremely specific guidelines from the graduate school or the journal you’re writing for. If you’re writing a pre-dissertation paper (sometimes called qualifying papers) or term paper, your instructor or advisor should have at least some guidelines for you. They may not specify stuff like font or margin size — use your best judgment, keep it as normal-looking as you can — but they’ll almost certainly have a preferred citation style, for example.

(For QPs, I strongly recommend asking another student of your advisor who’s farther along than you if you can read their QP that’s been approved. Do read as many of these as you can. For dissertations — ditto, but they’re a larger reading commitment, so maybe skim ‘em.)

Stage Three: Post-writing

OKAY, good job, you wrote it! It’s now a large document with words and maybe even a few pictures! Please give yourself a pat on the back, you’ve done an enormous feat, and you’re two-thirds done now!

Wait, seriously, only two-thirds? you ask me, because you are thinking, surely Kirby is lying about Stage Three actually taking up a THIRD of my time. No, my friend, I am not. I am basing this on extensive experience, both mine and others’.

If you are writing a dissertation, people are going to read it and make you do stuff to it. Ditto journal articles. If you are writing a term paper, you could just turn your first draft in for a grade, but I strongly strongly recommend against this. It will not be your best work, for one thing, and for another thing it means you’re depriving yourself of incredibly important ‘practice runs’ for Stage Three when it becomes non-optional in your future. I’m still assuming that you’re a PhD student reading this, meaning that if you’re writing term papers now, you’ll be writing QPs and/or a dissertation and possibly journal articles in the not-too-distant future. If that does apply to you, then term papers provide you an excellent opportunity to rehearse the process of writing while it’s still nice and small-scale, before you have to do a much larger version of this whole thing. Practice Stage Three while it’s still little and cute, before it grows up to have big old teeth and claws!

Stage Three is basically “someone has opinions on what you made, and then you change stuff.” There are two ways to go about this: either you get someone else to have opinions, or you have the opinions yourself.

Option A: Someone else gets to have opinions

Let’s say what you wrote was a qualifying paper. Good job! You’ll probably turn that in to your advisor, and they’ll have some opinions. This might be a cyclic process, actually — you might have been sending them one finished section at a time, and revising it as you go. I did this with chapters of my dissertation.

Anyways, when you send it to someone else to have opinions, it’s a good idea to tell them if you want specific feedback, such as:

I’m not sure if the organization of the sections flows nicely, could you comment on that?
How accessible is my prose? Are there parts of my writing that get really convoluted or confusing?
Does the framing of this argument make sense? I have a few other ideas about how to go about it, but this is my first attempt.

This is probably not the time when you should be asking for feedback on, say, your experiment design — that’s feedback you should have asked for during Stage One! A good advisor will have been helping you out at all of the previous stages, by the way. Here’s a blog post where I talk a little bit about what that other support looked like for me, and why it was necessary.

Anyways, your advisor might give you comments in a few different formats — comments on a DOC or PDF, or an email, or notes, or a meeting. One thing I strongly recommend doing at this point is making your own spreadsheet to organize the comments. I go into detail about how to do this in yet another blogpost, check that out!

Expect this stage to be some time waiting for the person(/people) to read it, some time for you to feel weird and freaked out, and then some time for you to actually make the edits. The cycle often repeats a few times.

If this wasn’t a QP, but instead a term paper or a journal article or something, it’s still a really great idea to get someone else to read it over. Ask someone who you’d be willing to return the favor to. This could be your advisor, or a classmate, or colleague or friend. Ideally it’s someone in your field who knows enough that you don’t have to explain the basics, and who knows the basic expectations of writing in your field.

Option B: It’s all you, baby

If this is a term paper you’ve written, or you want to give it a good read-over before showing it to someone else, you can do this step for yourself! This is a little tricky, because our eyes tend to glaze over when reading something we’ve just written. Here are some strategies for giving yourself meaningful edits and critique:

Wait a bit before doing this! At least 24 hours, ideally give your brain a few days to rest before attempting this.
If possible, print it out and mark it up in pen or sticky-notes.
If you can’t print it, make a duplicate of the document and comment it (in DOC or PDF form, whatever) as if it was someone else’s paper you were commenting.
Read it aloud to yourself!!!!!! CANNOT recommend this enough — I always find my weirdest typos and issues this way.
If your computer has this ability, get the computer to read it aloud to you. (My version of Word does this, and I love it.)

Once you’ve done any/several of the above, treat your own edits the same way you’d treat a colleague’s — note them down, and use them as a to-do list and work through them methodically. Again, my post on revisions is helpful here, even if the scary reviewer is myself.

Other parts of Stage Three

Besides the “getting comments and then addressing them” part of writing, other important post-writing work might include:

Figuring out how the manuscript management thing for a journal works so that you, like, actually submit the thing
Posting pre-prints somewhere, if you’re in a field that does this
Writing acknowledgements
Dealing with any copyright or publishing stuff, if it’s a journal article or dissertation or book/chapter
Filing the appropriate paperwork if it’s a qualifying paper or dissertation required for your program
Tweeting triumphantly about it?

Okay, now you get to call it done!

Good job! You wrote a thing! This might have taken a month or several years, depending on the size of thing you wrote and what else you’ve been up to — either way, please give yourself appropriate accolades, this is a wonderful accomplishment! Stage four is to party down and/or clean one million sticky-notes off your desk or close a bunch of tabs. Feels good, doesn’t it?

TL;DR

Writing something like a qualifying paper or dissertation has (at least) three stages:

Pre-writing, where you figure out what you’re going to write. Do your experiments and stuff, figure out what the paper’s about, make an outline, etc.
Writing, where you make there be words on the page.
Post-writing, where you do stuff to the words on the page.

Try to budget time for all three equally-important stages! Be thoughtful and intentional about your workflow to avoid last-minute panic and fugue states! Be safe out there, I love you!