I'm planning to get double honours in CS and Pure Math. I'm confused what's the biggest benefit of Ph.D. will be for me? People say you learn to do research, but I can do research(as I have the appropriate background, CS and Math) sitting at home or in industry. Then what's the biggest benefit of Ph.D.? Why should I do Ph.D.?

EDIT1: The only thing that is stopping me from getting a PhD is time.

EDIT2: Link to my new post.

  • 5
    Also see Doing research outside of academia, without a PhD?
    – ff524
    Nov 21, 2016 at 18:34
  • 17
    The biggest reason to get a Ph.D. is either 1) To teach, or 2) because some jobs require it. If you have no interest in either you certainly don't need a Ph.D. to do research.
    – Raydot
    Nov 21, 2016 at 19:35
  • 4
    It depends very much on what you mean by “research”. Of course anybody can do any research they want at any point, at any time. If, instead, you want your research to be published and reviewed, you most likely need Academia to recognise you in some way, that can be entering or possessing an academic title.
    – gented
    Nov 22, 2016 at 13:12
  • 25
    "The only thing that is stopping me from getting a PhD is time." A little humility wouldn't hurt, either. Nov 22, 2016 at 17:19
  • 4
    If you don't have time for a PhD, you don't have time for research.
    – henning
    Nov 23, 2016 at 14:05

5 Answers 5


I can do research sitting at home

Read a few research papers published in highly regarded venues in your field of interest. Are you capable of producing that kind of work at this point in your training?

  • Do you know how to identify important and original research questions?
  • Do you know how to select an appropriate methodology for answering a particular research question?
  • Do you have a "toolbox" of methods you can apply to different research questions as appropriate?
  • Are you able to critically evaluate your own and others' work as experts in the field would?

If not: a PhD is an apprenticeship in which you learn how to do the things mentioned above under the guidance of an expert advisor, in an environment that is designed for learning those skills, hopefully with a salary that allows you to focus on improving your research abilities.

  • 38
    @GaryGrey It sounds like your real question is "Will a PhD help me as a startup founder?" I suggest you ask that as a new question since it is substantially different from what you've asked here. (P.S. perfect grades are not really an indicator of research ability; a good understanding of the basics is necessary, but not sufficient, for doing research.)
    – ff524
    Nov 21, 2016 at 19:35
  • 2
    "Will I be able to do research and solve that problem with my background?" - @GaryGrey Only you can answer that. I will say that it is extremely difficult to get hired for machine learning jobs without at least a MS. Everyone has good grades, CS/Math background. You need to find a way to stand out and graduate work is one way path to do that.
    – Hobbes
    Nov 21, 2016 at 19:38
  • 8
    @Gary: I am a little puzzled by your last comment. Working at a startup and doing mathematical research seem very different. Is there some connection between the two that you envision? Nov 21, 2016 at 21:51
  • 6
    As far as the capability and talent for doing research is concerned good grades mean next to nothing. I have seen plenty of graduate students, some of them my own, who have come in with perfect grades, but ultimately were incapable of the kind of independent, original thought that is a sine qua non for a researcher, and a PhD student. So, long story short, there is no way for us to tell whether you would be able to do research work, nor is there a way to do so for a potential employer. If the job in question requires the proven ability to do such work, you will not be considered as a candidate
    – Pirx
    Nov 21, 2016 at 21:56
  • 7
    @GaryGrey Problems presented to you as an undergraduate and graded are almost uniformly problems that the instructor knows can be solved, and how to solve. Usually they are problems where they have already taught you the techniques to solve them, or use techniques you are presumed to know from prior classes. A PhD teaches you how to solve problems nobody knows how to solve, and (in some cases) nobody even knows the techniques required to solve them because they haven't been invented yet. Undergraduate "problems" are mostly toys meant to teach, not real problems.
    – Yakk
    Nov 22, 2016 at 15:28

Benefits of PhD:

  • an expert advisor training you to do research.
  • a workgroup and peer students to collaborate with.
  • a salary/stipend (hopefully) while you do it.
  • shiny title on your business card.
  • being a student, you can obtain discounts in conferences. It is often assumed that all "early-career participants" are students, because it's the most likely path.
  • in the industry, having a phd might unlock higher salaries.
  • if you want a career in academia, it's a sort of certification that you had some research training. Not having one is unusual. Of course, once you have a few journal papers published it starts to matter a lot less.
  • Thanks for the advice! The only thing that is stopping me from getting a PhD is time. I have excellent grades(almost perfect GPA). I'm very motivated to launch startup after graduating(maybe after working for a year or so). The only thing that's worrying me is that if I get stuck on a problem while working at my startup and the problem involves some sort of research, Will I be able to do research and solve that problem with my background(CS and Pure Math)? Say that problem is related to machine learning and involves creating a more intelligent machine.
    – Gary Grey
    Nov 21, 2016 at 19:19
  • 3
    Research, by definition, is supposed to be challenging. If I can guarantee that a graduate student can solve it quickly, then it's probably not a research problem. :) That said, with your background it's likely that you can read and understand (with proper effort) most of the research books and papers that you are going to need in your startup job, and come up with something now if you need it. Nov 21, 2016 at 19:40
  • Sorry, typo: last sentence -> something *new. Nov 21, 2016 at 20:04
  • Thanks Sir. I've posted it as a new question here - academia.stackexchange.com/questions/80227/…
    – Gary Grey
    Nov 21, 2016 at 22:44
  • @GaryGrey and while I have no experience in that, not all research is successful in the sense that it yields a useful working result. Think of all the times Edison messed up a lightbulb. It's research sure, but it just doesn't work because it's a matter of figuring it out. Granted, if research refers to reading and doing citations and things from previous publications with some original thought sprinkled in to glue it together, then I would say that it's a matter of how dedicated you are to those sorts of things.
    – user64742
    Nov 23, 2016 at 6:10

In general:

Research is a social process.

Hypothetically, you can prove theorems at home and submit them to journals and review journal papers from home. The industry is unlikely to support pure maths during your work time. But there are not so many people worldwide who are able to do research without verbal communication. The majority of researchers need a social working environment to some extent, at least every while and then. And get financed as well. (Submitting a paper without an institution name on it looks strange, btw.)

PhD is sometimes related to science (always in pure maths). PhD is something you cannot do without being associated to a university. What you get from a PhD degree has been mentioned in other answers.


In pretty much any field you need a terminal degree (usually but not always a doctorate) to get a teaching position.

But you can indeed do research without a terminal degree --or even no credentials at all-- though you probably won't be able to get paid for it, or published, unless you have a Master's.

  • 4
    A PhD is not just a credential that you need in order to get certain jobs (or to get published - you certainly don't need a PhD to publish your work, if it's high quality work.)
    – ff524
    Nov 21, 2016 at 18:17
  • 2
    There are journals that won't allow someone to publish unless they meet an arbitrary standard. As an example, Richard Bernstein MD could not get his paper published because he was an engineer at the time, not an MD. His paper was on the role of carbohydrates in the progress of diabetic debility, and his research was rock-solid.
    – MMacD
    Nov 21, 2016 at 18:26
  • 6
    While there may be some journals with unusual policies, the statement "you probably won't be able to get... published, unless you have a Master's." is just not accurate as a general rule. See Do journals in general have any kind of policy regarding papers submitted by someone without a research affiliation?
    – ff524
    Nov 21, 2016 at 18:33
  • We seem to be talking at crossed purposes. I'm talking about credentials, which aren't the same as affiliations.
    – MMacD
    Nov 21, 2016 at 19:40
  • 3
    @MMacD: I also strenuously disagree that one cannot get published for research without a master's degree. Every year there are math papers published in reputable journals authored (only) by current undergraduate students, i.e., by people with no academic credentials whatsoever. In fact I know of no reputable mathematics journal in which the "credentials" of an author play an explicit role in the evaluation of their work. Nov 21, 2016 at 21:55

Within a startup as you suggest you may hit a problem that requires research. Without research training you may not have enough breadth or depth in your field to estimate the scope of the problem i.e whether the solution is well known in the field, the solution is at the edge of current knowledge or whether the solution is currently intractable. The issue that bedevils all computer science / software engineering problems both in academia and from an entrepreunurial perspective is estimating how long it will take to solve a given problem and how much it will cost. Typically the issue here is the less knowledge one has about a given problem domain the lower the time estimate given and the higher the potential for getting the estimate wrong. Development time estimation is an almost intractable problem in it's own right for various well known reasons, that the agile methodology has grown up to address. So what are the benefits of a PhD: wider and deeper domain knowledge obtained from the literature review, a set of research tools, ability to formally present solutions, a measure of mentoring and having been able to tackle a problem in a relative "soft environment" of academia, together with the enhanced credibility of the qualification. If of course the problem you want to solve is not at at or beyond the current research wavefront for your field, has a tractable solution you may be able to achieve a quick implementation and get to market in short order. Clearly only you can asses that in first instance. However if you have no commercial experience and no higher academic qualification where will you get your venture funding from?

You must log in to answer this question.

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