Skip the Ph.D., go straight to research

Satpalparmar got bad grades from an unknown university. He wants to become a researcher. And he asks me for advice.

I got good grades and a Ph.D. from one of the best universities in Canada, so who am I to give advice to people like Satpalparmar?

Nevertheless, here it is.

You do not need a Ph.D. to be a researcher. In fact, the requirement to have a Ph.D. is relatively recent. Go visit research labs in most governments and companies, and you will find smart people without a Ph.D. doing research. (I have famous people in mind, but I am afraid to point them out, for fear that they will take it the wrong way.) 

Here is a recipe that should turn you into a researcher, no matter who you are:

  • Read research papers.
  • Write and submit some research papers, books, booklets…
  • Prepare and give talks in the field, anywhere where they will have you.
  • Get involved in the field. Help organize events, conferences. Connect with the researchers.

Build up your resume. Get some recognition.

Published by

Daniel Lemire

A computer science professor at the Université du Québec (TELUQ).

9 thoughts on “Skip the Ph.D., go straight to research”

  1. I always send people to read phil agre’s networking on the network and how to be a leader in a field. Both are about getting a ph.d. but work just as well for b.a. and m.s. type people

  2. I agree with this. Just do it.

    Example: I can find no evidence anywhere that Doug Cutting has a grad degree of any kind.. and no one cares anyway.. the guy is research rainmaker.

    The above also assumes all researchers submit papers. See the many applied researches at Google, Yahoo!, Facebook, Microsoft etc who do not publish or only so so very infrequently.

    Groups like Lucene, Solr, Hadoop, Hive, Pig, Mahout et al are cutting new territory without stopping to publish much in detail either.

  3. Thanks for your prompt reply Daniel. What you suggested make lot of sense. I consider it as bottom up approach. In fact your suggestion are consistent with plans I have in my mind and thats the only reason I resign one year before my admission to any college. But as alway, bottom up approch have its own disadvantages:

    a) It will put huge pressure on me to make it work. Leaving a high paid job is not a trend in India. I am unmarried and going by trends and traditions in India I am expected to get married by next two or thre years. This means I need a steady income ( may be small but steady) for a decent life. Ther will be huge social pressure to justify what I am doing.

    b)Very few companies in India are working my web mining and web research. IITs (best of indian engineering education) are not very research oriented. This means there is huge competition for whatver research related work is happening in India. Plus you will not have best of field working with you. So it would be a huge compromise with work quality.

    c)This process is very long. I am 27 now. I believe it will take 2+ years to understand and conneting the dots and 2+ more years to come up with someting original and meaningful.

    I am looking for more pragmatic approach and I am willing to compromise on money/job and ready towork hard 24*7*365 for next couple of years if I could get opportunity to solve really good problems while working will best minds in web research.

    I doubt working alone or as non-Phd researcher will get me there.

    Thanks again for your time and prompt advice.

  4. Neal,

    There is no doubt that Doug is a great engineer. He’s builds useful systems, but I wouldn’t call him a researcher. Researchers build systems in order to run experiments to prove their theories about hard challenges. They then disseminate these findings internally or externally.

    I agree you don’t need a PhD to do research, but getting one teaches you the skills that should know to be a good researcher. I think it’s a good idea for your career, but I’m biased ;-).

    One way to get started is to be an assistant to a researcher and work for free. Even if you have bad grades, if you prove you can do research by publishing respectable papers and presenting at conferences, these will likely lead to an opportunity to do a PhD.

  5. I agree that you hardly need to complete a PhD to learn how to be a researcher. On fact, some of the strongest PhD candidates have already done great research as undergraduates. And I have two researchers in my group who don’t have PhDs, and they deliver great research!

    Exceptions notwithstanding, I think it’s quite difficult to get a job as a researcher without a PhD. So I have to disagree with your recommendation. If you want to pursue a career in research, understand that, by not having a PhD, you will often have to work harder to be taken seriously.

  6. @Satpalparmar

    Most likely, even with a Ph.D., you will not easily find a position as a researcher.

    Becoming a researcher is like becoming a novelist or an athlete. Some people have it easy, most don’t.

  7. The path to a Ph.D. is extremely risky. I know several people that spent 4-5 years and never finished. That is 4-5 years of lost salary and experience! On the other hand, taking a research job with B.S. is low risk and doesn’t stop you from getting a Ph.D. later. In fact, with even a handful of publications, it makes it much easier.

    I think that people with prior research experience finish faster. With the time to complete the Ph.D. being a major cost, i.e., 4-5 years of lost pay, this is a major advantage. I hope you consider this.

    Also, there are not just two options, but really three or even more:

    1. B.S. and straight to research
    2. B.S. & M.S. and then research
    3. B.S., M.S., & Ph.D. and then research

    In most programs, #2 leads into #3.

  8. Actually, the time spent in PhD is well spent. It’s sort of like boot camp.

    While there are extremely talented researchers who never wasted time getting a PhD and are doing great, they are exceptions, rather than examples.

    Undergraduate studies do not always expose a student to the rigorous research techniques a PhD students needs to finish her thesis. The time spent doing literature search, paper reviews, conference talks and the dreaded meetings with advisors is well worth it. You can’t simply replicate the skills, patience, and perseverance learned through this.

  9. In the same way you can be a novelist without being published by Farrar, Straus, and Giroux, and you can make a living playing the violin with an organization other than the Vienna Philharmonic, you can be a researcher without a tenure-track faculty position at Stanford.

    For instance, you can work with Dan Tunkelang at Endeca (though judging from the quality of their personnel, they’re pretty choosy in their hiring) or with me at Alias-i (we have a great non-Ph.D. research programmer).

    There are also tons of research jobs in universities in all departments. My wife Mitzi, with only a French Lit undergrad and Linguistics M.A., has worked, over the last 20 years, in research positions at Carnegie Mellon, Rockefeller, NYU, and Columbia.

    The main problem is that without the Ph.D., you can’t be the principal investigator on grants and can’t really get faculty positions at places like NYU. And it’ll be hard to get taken seriously at the “first rate” research labs like Microsoft, IBM, Xerox, Google, etc. But then it’s hard to get taken seriously as a researcher by the rest of the company in one of the Ph.D. heavy labs.

    Ph.D.s have a well-deserved reputation for non-practicality. I actually had a hard time convincing the management at SpeechWorks that my Ph.D., faculty position and stint at Bell Labs weren’t a drawback.

    If you want to do research and can get into a halfway decent grad school (meaning you can find an advisor), you’ll be able to carve a career for yourself in research somewhere. Or if you’re as smart as Mitzi, you might be able to do it without relevant grad school.

    And keep in mind that “second rate” research jobs in computer science pay much better than “second rate” novelist or violinist jobs.

Leave a Reply

Your email address will not be published. Required fields are marked *

To create code blocks or other preformatted text, indent by four spaces:

    This will be displayed in a monospaced font. The first four 
    spaces will be stripped off, but all other whitespace
    will be preserved.
    
    Markdown is turned off in code blocks:
     [This is not a link](http://example.com)

To create not a block, but an inline code span, use backticks:

Here is some inline `code`.

For more help see http://daringfireball.net/projects/markdown/syntax