Best and Worst Advice for Academic Writing

Among the countless pearls of wisdom on academic writing, the one that I consider the best and worst advice was from my late PhD adviser. Having never drafted anything mildly academic before—but having spent a sufficient amount of time doing the underlying analytical work for one in the early stages of my PhD—he thought I ought to write up my research progress for him. “Oh, and it would be a good idea to cite 20 journal articles in this write-up”, he added as I was leaving his office.

This felt like the worst idea ever. At the time, I was but a simpleton graduate student—one who thought he had never made contact with a journal article before (I had!) and that all human knowledge was contained in textbooks (it isn’t!).

Such ignorance had as much to do with my arriving at UC Davis from a mediocre Indian university as it did with my being a mediocre undergraduate student; I was steeped in the culture of grade maximisation coupled with deep understanding minimisation. Its foundations were reading textbooks narrowly and shallowly over a short period of time—usually compressed into a short 24-hours span to memorise as much as I could (including step-wise solutions to math problems!)—for regurgitation in an exam. I would then promptly forget all of this within 24 seconds of walking out the exam hall. While it proved a surprisingly effective strategy for acing exams, it proved to be terrible in terms of being prepared for advanced research.

So, unsurprisingly, writing this first research report with 20 journal citations felt impossible because I didn’t know where to begin. This paralysis coupled with my general shyness to ask for help was compounded by my adviser’s generally very hands-off approach to mentorship. The sheer agony from merely thinking about this meant I barely made any progress on the introductory part of the write-up over a quarter. Eventually, when my adviser asked for the report, I hastily prepared one that only cited textbooks; I confidently suggested to him that we were in an under-explored field. Sure enough, he was aghast and pointed me to the references section of one of his papers that he had given me. I had been too focused on recreating its results (which is possible in mathematical and numerical simulation-heavy research with hyper-focused effort) that I completely failed to understand that it was, in fact, a journal paper; as a result, I also missed seeing that it referenced recent work by others. Had I paid more attention to how his article was laid out, not only would I have conjured those 20 citations with ease but would have a far better model of my subfield and what contributions remained to be made.

I felt tricked in that moment but, over the years, I have come to see this issue in other students, as well. This issue is also not limited to students from countries with weak research cultures. So far, I’d been working hard to produce results, but not smart to make a case for why this was unique; I was missing the forest for the trees.

Now, I feel differently about his advice of preparing an initial write-up with 20 citations; it is maybe the best high-level research advice ever given to me. It has been especially helpful in enabling me to expand my research interests. The most basic benefit of this rule of twenty citations (as I now call it) is that it has altered my method of reading papers. I start with the abstract and if it resonates, then I skim the introduction to identify references that could be most relevant to my understanding a field. I do this till I have a list of 20 or so papers¹ to skim; I summarise about 5-8 that are particularly interesting and consider what questions haven’t been answered but might need to be. Then I do another round of literature survey to determine if there’s a gap in my knowledge or a genuine gap in the field; if it’s the latter, then it becomes the kernel of insight for a new contribution or research proposal. This process is not exactly fast; as a student, it can easily take a couple months to build a confident map of the research topic.

Fifteen years later, I parrot the rule of twenty citations as advice to my undergraduate students; they typically grimace at the idea, much like I did at their age. When advising PhD students, I reframe the rule as a firm requirement for their first reports. It serves as a test to determine how voraciously they read and the conclusions they then make from the literature. PhD-level students are supposed to be able to surprise their readers with new world models. Of course, by the time they finish, they realise that a field is vaster than merely twenty papers. But this only serves to reiterate the value of beginning with twenty—it is a manageable number.

The more one reads and describes papers in their own words, the more one increases the surface area of being able to notice (or, even better, start to identify totally new) connections. If others have missed noticing this, then there’s a gap to make a novel contribution. This is essentially what you are awarded a PhD for—generating a novel question that you can solve. In the long-term, this approach can become the foundation for how one learns to diversify their interests, which is my favorite aspect of being a researcher.