Based on your personal experience, how may someone enter a completely new research area as a 3rd-year Ph.D. candidate? Is it better to start with reading journal papers? Or maybe taking some courses? Or watching YouTube videos? Or what?! I am basically looking for the most efficient/fastest way of doing this!

And if you want to know my story, I am a Ph.D. candidate in structural engineering with a focus on deterministic earthquake modeling, but for some reasons now my advisor wants me to start working on probabilistic modeling of hurricanes and floods using machine learning methods, and I am a little paranoid because I have no idea about any of these! So what are your suggestions?!

  • Which country are you a student in? i.e. are you in your nominally final year, or half way through?
    – awjlogan
    Commented Mar 7, 2018 at 10:46
  • @awjlogan I'm in the U.S. I have passed courses and published some papers but my dissertation is going to be on a topic that I have no experience in!
    – RezAm
    Commented Mar 7, 2018 at 14:48
  • I'm wondering if this question is too subjective. Anyone has different methods that suits them best for learning and honestly I'm surprised you, as a 3rd year PhD, don't know what suits your best. I guess that means you were lucky in learning things until know. Maybe a better approach would be to ask if it is okay for your superior to ask you that and how to deal with it, given that it seems to be a very time-consuming task.
    – SK19
    Commented Mar 8, 2018 at 12:50
  • @SK19 Don't wonder!! I definitely have my own strategy; the reason I'm asking this is to see how other people deal with such situations, not that I don't know how to do research!! I think that would be appropriate if you do not judge people or condescend them by saying illogical sentences like "that means you were lucky in learning things until know" .. If you are willing to share your way of doing research (if you even have one for real), then go ahead! Else you seriously do not have to post unrelated comments my friend!
    – RezAm
    Commented Mar 8, 2018 at 17:26
  • You are right, I misinterpreted your question insofar that I assumed you had no technique yourself, since you didn't write anything about it. I did not intend to judge or condescend, I apologize for the misunderstanding. And I'm sorry if my grammar is sometimes confusing. As a non-native speaker, that happens to me occasionally. Assuming a lack of need to learn research techniques, I merely intended to congratulate you on the rare, fortunate occasion.
    – SK19
    Commented Mar 8, 2018 at 17:36

1 Answer 1


I hope you find the following beneficial.

  • Firstly, you need to gain some familiarity with the new area and gain confidence on the general problems in the area. Survey papers, classic materials (e.g textbook chapters), youtube videos etc at this stage. Say 1 week. E.g, say a hypothetic paper is titled: "Statistical Modeling of Natural Disasters: A survey". However don't be too distracted by reading too much general papers as PhD is highly focused. Also, yotube videos should not distract and take too much time!
  • Secondly, get a little deeper e.g say a hypothetic paper is titled "Survey of probabilistic modeling of hurricanes and floods", "Birth and Death Processes for natural disaster modeling" etc. Say this takes another week: Here you look for something interesting to have an idea of potential technical papers.
  • Thirdly, you should already have a wide and narrow scope now, so you could pick say 2 journal papers. One classic reference (relevant, rich, clearly expressed, highly influential and perhaps well cited) say 2005 and one recent paper, say 2017-2018.
  • Fourthly, briefly look and get an idea of the basic concepts in the first reference (2005) to get familiar with the technical basics, basic relationships and professional technical terms. Then, I suggest you re-write the second (2017-2018) journal paper with your own hands etc. to keep yourself "on it" systematically. During this process, you could get some things and not get others. It is still fine at this stage. Also, at this stage it has been implicitly assumed that you have basic ideas of the statistical knowledge and tools to understand these materials.
  • Fifthly, you could summarize your handwritten note into a personal tutorial, note the open areas, assumptions, problem solved, intuitive questions etc. for future extension etc.. (It has also been assumed here that you are comfortable with the research tool e.g simulation approach etc.) Then begin to think about how to implement these (practicable and realistic) ideas, formulate a problem and hypothesis etc.

I hope others may continue from here. However, I hope this will get one on his feet, then everything coming after that may be more flexible. Note that I have not included survey paper writing as it may take more time.

Disclaimer: Thus is a suggestion and may not work explicitly for everyone at everytime

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .