A 27th birthday is a scary thing for someone risking becoming a pop star, but it’s a happy occasion to celebrate the fully formed brain. Now, while I’m satisfied to claim the latter – but moreso while I’m pleased to have seen two research articles through to journal submission and postage on arXiv this summer – I’m inclined to take account of what has (and has not) been working for me on the project management front. What parts of a daily workflow really save time? How many new ideas can be juggled at once without dropping any? Where should prototyping for new research code go, and how should preliminary results be organized? This is the side of doing research about which I might have spent the least time thinking when starting out, but it’s become the topic that makes or breaks steady progress. I have found that having some experience from repeated failures to manage my work well has led to some decent, workable answers to questions like the above.
I’ll proceed with a few themes.
Articulate
As a prototypical example for working through complex thoughts, journaling is a good way to organize a narrative structure. It’s not just a means of discourse for when you’re otherwise isolated with unclarity, but it’s a way of modeling your thought process in a way that can be constructively criticized. Writing a lot – or speaking a lot, while you’re at it – can guide you to clearer thought patterns by simply availing a concrete option for specific repetition.
I don’t think journaling accesses the fullness of a consolidated understanding of complex thoughts. Whether you’re processing a personal loss or mulling over the possibly unsatisfied regularity conditions for an unusual estimation task, I think that having one narrative for explaining all of the relevant content helps to define and direct behavior. Why one narrative? If you cannot pin down the perceived conclusion or ultimate consequence of a line of thinking, then the multi-verse of risky decisions that could be made can be paralyzing. If you can manage a goal-oriented narrative, then you can commit more easily to the necessary next steps.
Consolidate
As a vague case study, suppose that you’ve attended a compelling research talk about a subject that is new to you, and you’d like to think more on it. You have three obvious options for what to do immediately next:
- Find and read some papers that relate to the subject
- Write up the questions that you have from just having watched the talk
- Do something unrelated
On the one hand, a natural instinct is to look up some papers because then you can get a better sense of whether you’ve missed an important high-level aspect of the topic in question, learn more about notation and narrative conventions in the relevant sub-discipline, and maybe inform yourself of new, better questions. On the other hand, writing questions in a document somewhere is (I) quicker, (II) helpful for forcing yourself to articulate some thoughts clearly (which for research productivity is infinitely better than many thoughts vaguely), and (III) a product at the very least, demonstrating that you’ve done some thinking. The remaining option, moving on immediately, does not obviously recommend itself if the concern on your mind is to take advantage of inspiration while it hits.
My instinct at this moment is to stand on a simple desideratum concerning such choices:
Is your expectation clear for where this is going?
If your answer to this question is, “No, I have no idea where my train of thought is going!” then you are probably better off leaving the matter aside for the day. Sleep on it and let your subconscious work on it if it’s truly interesting to you. If your answer is, for example, “Yes, this subject may be useful to me because soon I will need to ask myself questions to which this talk may have given answers!” then by all means, record these thoughts immediately, ideally doing so in the Overleaf project for that paper to which your questions are relevant. This practice will help you immediately consolidate this new information with a narrative that pre-exists, and by my estimation, this is your best option for not losing these thoughts to the aether where other dead-ends go to die. Finally, if your answer instead is, “Yes, I have a very concrete follow-up question and I am prepared to make progress toward answering it a top priority!” then probably the best course of action is to verify that this new question has not yet been answered by consulting the literature. If indeed it is an important and open question, then by all means, begin setting out the narrative for a brand new research direction! But the risk of spending time on yet another dead-end or otherwise forgotten idea is real, and this final option should therefore be reserved for special and clearly understood opportunities.
Replicate
We know that repetition is good for forming memory, both in terms of quality and lifetime. How can it come in handy with project management?
I think that the idea of fast prototyping is very useful when it comes to structuring and developing a new idea. Not spending too much time trying to get details right the first time, instead favoring quick drafts that provide structure that can be constructively criticized.
How do you prototype quickly? One way is to use previous work as a template. I have found this to be especially useful for developing research code. For research that depends on training neural networks, one can save a lot of time from using previously tuned nets when possible, because the training process is often peculiar to the setting and will be painful to start again from scratch.
However, a risk of the fast prototyping is that stereotypes formed from previous work will sneak in by the nature of “thinking fast.” This point should not dissuade you from taking advantage of previous experience, or else every new thing you try will be very energy expensive. Instead, it’s helpful to be reminded concretely that developing some sensitivity to when you are thinking too fast will be important as you attempt to venture into unknown territory over and over again – drawing on experience is a good starting point, but only when you are thoroughly convinced that your experience is relevant. This might set in only after the first few attempts at rapid prototyping.
Conclusion
I’ve given a few themes on research project management as I see relevant to my own professional goals these days. As I think about my own professional development, I find that some advice found broadly in, say, the self-help genre, but really advice of any sort, is too broad to be immediately applicable. Psychologists who write books need to write them for a broader audience than STEM grad students, or else they won’t sell well – at least for the reason that STEM grad students do not constitute a very large market. Needless to say, this post is no attempt at a rigorous substitute for professional advice, and I couldn’t purport to offer any such thing at my current stage, but instead my progressive attempt at intermediating between my limited reading from psychology broadly and my own experience. To all of this, I’ve enjoyed some success with articulating, consolidating, and replicating my formal thinking so that I can be more reliable and more often successful in my fast thinking. Furthermore, this seems to me like a general recipe for research work.