This post is written with some tongue-in-cheek; I do not mean for you to take these points as unassailable. No doubt there are some omissions here, but I had to pick my top ten (plus). These are things that looking back at my M.Sc. (Imperial) and D.Phil. (University of Oxford) experiences have forged the way I do scientific research, and built my academic character. My B.Sc. was too long ago to recall anything accurately! I find the exercise of writing them down useful (see the second commandment…).
Without further ado, my scientific research (and post-graduate degree acquiring) commandments:
- Thou Shalt Read
The most important thing in academia is to be well-read. During my D.Phil. I averaged, at least, reading a paper a week (doesn’t sound like much, does it?). That’s a thorough read (not skimming over abstract, results, and conclusion). Superficially, you will read many more (a fleeting look at the abstract and results), quite possibly one-a-day. The papers I read were not uniformly distributed over the three and half years my doctorate took – you will read more in the beginning than towards the end. Some tutors will suggest to write a literature overview of the area in the first year. I haven’t done so myself, but that is a useful proposition to get acquainted with a scientific domain (and a potential, first publication too!).
Note that while you read, you should be taking copious notes, because you are bound to forget things years down the line. Which brings us to …
- Thou Shalt Write
Write down everything. Use physical notebooks, wikis, text files (markdown), WordPress, anything. There is a certain pleasure in jotting down physical notes with a beautiful pen, and satisfaction in crossing out tasks. Again, write down everything; keep lists; TODOs, papers to read, experiments to try, meetings notes. More importantly start writing from day one. I used to keep notes of everything, meetings with supervisors, conference reports, notes on important papers. I would write so often, I came up with a blog for our Oxford Protein Informatics Group, OPIG (aptly named blopig). I am chuffed this is still in use today.
Bonus: writing helps you keep organised.
- Thou Shalt Publish
The logical consequence of writing. Like it or not, the currency in research are publications. When applying for positions in academia, most interviewers will look at your publication list (keep your profile on google scholar updated). You can do without publications, if you have some other glaring skills they are looking for or come highly recommended by a trusted source, but having a number of publications in decent journals will greatly improve your chances of success. Also, your viva voce will be easier if you have published your research in peer-reviewed journals. There is a significant overlap between questions from examiners and comments from reviewers. I take this “publishing” a step further; on twitter; in blogs; in the newspaper. Wherever they will have me.
- Scientist, communicate!
This is perhaps one of the most underrated suggestions; communications with your supervisor(s), with the general public (science outreach), with your colleagues, at conferences, making posters, and through journals. You are going to spend so much time discussing your research, that you should pay attention to actively train yourself in this. Universities organize “soft” courses in communications, and you should make time to attend a few.
- Collaborate and take interest
Your colleagues are an invaluable source of ideas and support. In an introductory lecture at Imperial, Prof. Michael Stumpf told us “You will learn from your colleagues almost as much as you learn from us [lecturers/supervisors]”. That stuck with me, as true and surprising. When doing a Ph.D. you are really narrowing down in some (typically) previously obscure scientific area. There are few people who can understand the implications of your research, and most of that set are sitting in your lab. Take advantage of this. Involve yourself in other people’s research both in your same university and outside. This is critical for cross-pollination of ideas. You will be richer for it.
- Choose a Subject you are passionate about
When deciding to go for a Ph.D., there are two main choices you need to make, or to a lesser extent one. “What is your research topic?” (the other is your supervisor choice). It helps to have a clear idea of what you are spending every waking second for the next three/four/eight years on. But I think this is rare. Some super-focused wunderkind who will have done their bachelor’s and master’s on a single topic might have it all laid out and planned, but this isn’t the general case (at least in my experience). Students do occasionally come up with their projects, but they find it hard to gauge how difficult/trivial a problem really is. Also, people coming from industry might not have experience in what makes a good academic problem worthy of PhD (this is mostly true where I come from; software). Most students have a general idea of the topic, but not specific details of the project. If this is the case, best to speak with someone well-versed in the area (an academic perhaps). The important thing is to find a subject you are passionate about, and eager to learn about. This starting enthusiasm will rocket you throughout, and will make you loathe the subject less a few years down the line. This point, being passionate, isn’t different than any other course-of-action you may take in life.
- Find a Supervisor who shares your interests
Once you’ve settled on a topic (or area), you need to choose your supervisor. And not the other way around. Ideally, you get along well with your supervisor; otherwise it is going to be a miserable ride for both. Remember your supervisor is not your friend, but has a very specific role to fulfill. To guide you towards the (successful, one would hope) end of your research project.
- Own your research project; be independent
The project you are undertaking is not your supervisor’s or anyone else’s. It is yours, and success depends on owning it early-on. Never ask your supervisor “how do you want me to do this?”, or such dross. It is true what they say, that the most awkward phrase to hear in a viva is “because my supervisor told me so”. It is your duty to design your research. Through scientific arguments, feel free to disagree with your supervisor (in a polite and respectful manner, we are humans too!).
- Use the correct tools for the job
There is a wealth of tools (freely) available for your research. Using the right tool will save you time and effort. Knowing how to program in today’s interdisciplinary nature is a must, so take some coding courses. Tools will affect the quality of your science and its presentation. Powerpoint, Visio or Inkscape for poster preparation. LaTeX for documents. Learn about less popular, but incredibly useful tools; e.g. Recoll to search your documents.
- Make sure your research is reproducible
Make sure your experiments are reproducible. No doubt, at some point you are going to repeat an experiment you have done months or years ago, on another machine and you do not want your results to vary due to a different environment. Docker, Conda (if using Python), and Git are examples of tools which help here. Science benefits from reproducibility. Other people can test your hypothesis by repeating the experiments and build on your results. The things to version and tag are source code, data, and libraries (versions of each).
- Bonus: Bring something to the table — enthusiasm, passion, etc.
Doing postgraduate research is a tantalizing experience. It is up to you to make it memorable. Bring something extra to the table. Your colleagues and supervisors will respect you for it.
There are many more hints I can give (e.g. backups, validate your work, don’t put too much pressure on yourself, persevere when times are difficult – and they will be, don’t get too attached to your methods or results, network, work hard, etc.). But perhaps these will make up a future post.
Please do get in touch (via twitter or comments below) if you reckon there are any glaring omissions.