Writing every day

July 09, 2022

The PhD is a professional degree and like all professional degrees, it prepares students for what are fundamentally jobs in communication. While the core of the work is technical — e.g., programming/coding, writing proofs, performing empirical analyses — that work is all for naught if you cannot communicate the fundamental insights and significance of your work to others.

This is why one of the things I always tells new graduate students is to write every day. Many new graduate students struggle under the lack of structure, inherent uncertainty, and deep questions about motivation in their pursuit of the PhD, so I believe these skills and coping strategies can be useful for most students.

There are many kinds of written communication that we should all practice every day. I still have a "write for an hour every day" in my daily to-do lists! A timer is ticking as I write this post right now! Below I detail recommendations for various forms of writing; you should eventually consider rotating through each of them as part of your routine:

I don't typically mean journaling, but that could be the right starting point for you

First I'd like to get a misconception out of the way: the kind of writing you do matters, but there is no one-size-fits all advice on how you should allocate your time. I've had some students think that writing every day means journaling.

There are only 24 hours in a day, so think of the type of writing you choose to do in terms of the opportunity cost of not doing some other type of writing.

Example. My undergraduate degree required a great deal of writing and I was developing a strict discipline of time-boxing via the pomodoro technique as part of a larger mindfulness practice at the start of my PhD. Thus when I started my daily writing practice, I did not need to journal, since I was more or less doing this already. Instead, I needed to focus on writing down and synthesizing my thoughts on the things I'd read, concepts I'd struggled with, and generally become more fluent in the language of scientific communication, as it was practiced in my research field.

My general advice is: if you already journal, or if journaling is already comfortable to do you, prioritize other writing tasks. If you struggle to get started writing at all, then by all means, journal!

Low-stakes writing: blog posts

I love blogging: I love doing it, I love the concept of it. What I love about it is that it is absolutely low-stakes: posts should be short (my typical length is probably 2-3 times too long!). They should be clear and accessible. You can freely admit to not knowing things. You can try out arguments and say things that are wrong. You can take the time and space you need to really explore an idea, and you can punt on the things you don't have time to write about or think about deeply.

Students often feel vulnerable when they first approach blogging. I usually assure them that no one is reading their blog posts! What I am really doing when I say this is making a "security by obscurity" argument. Of course someone will read your blog posts eventually, but given the sheer quantity of content on the internet, no one is reading them that deeply and no one has the bandwidth, nor inclination to judge you.

I'd also note that my most popular blog post has nothing to do with my research; it is a post with instructions on how to incorporate custom javascript in Jekyll blogs. It gets the most hits, has the most comments, and I have occassionally been tweeted at for it. One of the things I love about this is that I wrote something other people found useful, which isn't always the case with research! ;) In any case, this is all kind of cool, but goes to show that you cannot control what people will respond to.

So...if no one really reads your posts, what is the point? Here are what I consider the true benefits of blog posts:

Blogging provides a way to get "partial credit." Editing involves two distinct tasks: re-writing prose and removing content.

Re-writing prose. It can take time to find the right pitch for your abstract and introduction. With each re-submission, the world changes a little and so it stands to reason that the way we talk about timely problems changes a little. Sometimes we become better at explaining certain concepts. Blog posts can help students (and faculty and research scientists — anyone, really!) hash out the clearest prose for a particular audience.
Removing content. Most venues have a page limit. The page limit ought to map to a contribution of a particular size and there ought to be thematic cohesion in the document submitted. This means that sometimes technical content gets cut. There are sometimes strategic reasons for this (e.g., some advisors advocate reducing "surface attack area"), but it's generally best to think about it as a communication challenge: you need to save sufficient space to explain your work and why it matters. Your objective should not be to cram as much technical content into a draft as possible. It can be hard to murder your babies, but with blogging you don't have to!

Potential employers/collaborators/etc. know what you are working on. Publishing work in high-quality venues takes time. Many students in my field take over a year to submit their first paper. This is a problem when applying for internships, fellowships, etc., because it means you have no outward-facing discussions of your work. It can be quite difficult for new researchers to find venues where they can advertise and speak about their work, especially for those who started in the pandemic. Blogging is one way to make that information public.

Writing papers gets easier. Many students struggle to write their first paper. Scientific writing uses specialized language and different communities expect papers to have different structure. In addition to the mechanics of what goes into a paper, each research community as has its own dialect. It takes time and practice to build fluency in that dialect. Just as you cannot learn a language passively, you cannot learn to write for a community passively. Lab presentations, invited talks, workshops, and more informal forms of rapid communication (e.g., talking!) are a great ways to build this fluency, which will make writing papers easier!

Blogging as research notebook

One of the primary benefits of blogging your research is that it can function as prior art! Sometimes students fear being scooped. I'm not going to discuss recourse for stolen work here, but junior students should feel confident about sharing their work and establishing a record and timeline via blogging.

Higher-stakes writing: tech memos/reports

Good papers go through many iterations of writing, often via resubmissions. Each venue target may have a different target audience. Time matters too: tastes and attitudes about novelty and worthiness may change dramatically between submissions.

What is a tech memo? A tech memo is usually longer than a blog post, and often a bit more formal. While blog posts should be fairly self-contained, tech memos need not be. A short tech memo might eventually correspond to an expanded section of a paper or an appendix. Writing the technical content of an upcoming paper submission as a tech memo can be very helpful for focusing your time and resources on the pitch closer to the deadline. Tech memos can be written as tutorials, proposals, or other forms not quite appropriate for a research paper, but helpful for explaining work done.

Why tech memos are useful to you and others.

How I found them useful. I wrote quite a few tech memos at the end of my PhD. They were helpful for me to work through ideas during the pandemic, when I didn't have many people to talk to. They allowed me to pull on threads that might not go anywhere and be more effective and efficient at creative thinking.
Ease into paper-writing. First papers aren't just hard because of the actual writing of sentences; it's the whole package of balancing low-level details with high-level vision, of hyping up your work while imagining how Reviewer #2 might interpret what you did, of balancing the theoretical contributions with the empirical contributions. Tech memos don't have to be all things; they can be just one (plus LaTeX practice!).
Protect your work/contributions. Sometimes students leave academia; having a record of tech memos (whether internally or externally) can help the next student who picks up the work, while also clearly establishing you as a co-author. Lab-based tech memo repositories can help with institutional memory vis a vis prior work, preventing churn.

Finally, elevator pitches for your work are incredibly important. I hate these and am very bad at them. I don't get anything out of them when other people pitch them and hate hate hate hate doing this myself.

Somewhere (perhaps in an old Twitter account), I have a bookmarked Twitter thread about how to pitch your research on Twitter. Since I cannot find that original thread now, here is a link to an article that looks quite reasonable.

I clearly prefer longer-form writing, which probably means I should work on my Twitter game! In any case, tweeting about work (especially in a way non-conversational way) definitely counts as writing!

Addendum: There is no one "right" way to write!

I've seen many students have anxiety about writing in graduate school. Sometimes this is because they are not native speakers and are self-conscious about their writing. Sometimes this is because their native dialect, while English, is not considered "proper" academic writing. I have also seen faculty handle this anxiety poorly and too often contribute to it.

My take is that there are two fundamental goals of writing:

Clearly communicate your ideas to an audience.
Signal your membership in the in-group to that audience.

Most criticism of writing is a mix of these two. I tend to deal with #1 by asking increasingly detailed questions and expressing that I am the one who is struggling to understand, but that I need their help to explain it to me. I often struggle to communicate new ideas and connections myself and try to model equal partnership in communication.

#2 is thorny. I try to separate criticism of voice and style from community norms regarding technical content. The latter includes things the specific section titles, use of \paragraph{<Topic>.} in papers, community-specific vocabulary, etc. The former is any writing that sounds like an outsider in my internal voice, but where the voice is no impediment to my comprehension. For me this includes overly pedantic requirements regarding punctuation, grammar, and diction that do not change the clarity nor meaning of the text. I could write another blog post on the merits of being explcit about the costs and benefits of adapating one's voice, but this post is too long already!