I am a PhD student designing algorithms for mathematical problems. I have the very bad habit of loosing my track on research problems. Suppose I am working on problem X, then, in the next week, I will start looking at Y,Z and other problems. Due to this, I feel tired at the end of the day with very less productivity with respect to X. I have always fear in my mind of failing, and that's why I try to look at other problems simultaneously. I don't know if it is okay for a PhD student (after mid-stage) to do this kind of things.

Question: How to focus only on the problem at hand?


2 Answers 2


You are "guilty" of a particular type of procrastination (in turn a type of avoidance behavior), in which you avoid working on your (self-assigned) task by working on some other task. Ergo, the usual strategies to counter procrastination apply. There are several examples on this site and elsewhere online; my preferred method is to combine the pomodoro technique with a progess tracker.

  • Pomodoro technique : Work on your task during limited and manageable time periods by setting a timer, e.g. 25 minutes, then force yourself to take a break of at least five minutes even if it completely disrupts your flow. This limits fear and thus avoidance, and it creates motivation to get back to work after the break. Don't allow yourself to work on your side-gigs until four (or so) pomodori are completed, but then go ahead and indulge.

  • Progess tracker: Use a 'done' list, an online habit tracker, a spreadsheet or whatever works to track each completed pomodoro. This creates motivation, by seeing your success, in particular if you accomplish not to break a streak, and facilitates habituation.

If your problem to focus disappears, you may want to be more liberal with the breaks between pomodori, so they don't disturb your flow, use a single pomodoro just to get you started, or do away with them all together.

As a side remark, it's good to work on several projects, but you should prioritize and stick to priorities.

  • For progress trackers, I've been quite happy with [Nirvana](nirvanahq.com). It has a focus view, which only shows the next task or couple of tasks. This avoids the main problem I've had with to-do lists - they quickly end up too long, making it hard to focus on the important tasks.
    – Anyon
    Commented Jan 30, 2019 at 13:39
  • @Anyon I think progress trackers shouldn't ever double as task manager to avoid the problem you mention. They should be strictly "done" lists, not "to do" lists. Commented Jan 30, 2019 at 15:19

Actually, it is probably a mistake to focus too closely on only one thing. Perhaps, though is that you just wander a bit more than you should. But focusing on one thing doesn't work for most human activities or you get "blind sided" by other things. One needs "peripheral vision" in research as well as in driving a car. Other ideas relate to the main thread of your research and you need to be aware of these if you are to avoid a "too narrow" perspective and didactic work.

But, it is possible to defer work on those other ideas that occur to you as you go. My personal method, nowadays, is to keep a pack of index cards on my desk and when other ideas come to me that are not related to the task at hand, I write a quick note on the card and return to work. I keep these cards together and can return to them whenever I want a "working break". I can add a few thoughts to a card.

As a grad student, I wrote up "interesting ideas that might be pursued" on individual pages and kept these in a notebook. When I finished my degree I had a pile of potential work to be explored.

Note that the mind doesn't work especially efficiently if you try to focus it too closely for too long. You can get stuck in what seems to be (and may actually be) a dead end. It is good to give your mind a break. Sleep is useful for this, but so is changing gears. The trick is to organize your work in such a way that you let this happen without wandering down side alleys for too long.

Having the "interesting ideas" captured so that they can be returned to later may be enough to let you avoid following those threads too far at this instant so that you can return to the main thread.