When do you graduate with a Ph.D.?
October 2017 (perspective of an assistant professor)
This article is my attempt to summarize what it takes to earn a Ph.D. degree in computer science and closely-related fields in the U.S. around the time of writing.
Five years ago (time flies!) I wrote: Overview of U.S. Science and Engineering Ph.D. Program Requirements
This article is a more concise version of that one. It tries to answer a very specific question: When do you graduate with a Ph.D. (in computer science and closely-related fields in the U.S.)?
The simplest answer is: You graduate with a Ph.D. when a committee of 3–5 professors says that you can graduate. This might sound like a facetious answer, but it's technically the right answer. Only people with Ph.D.s can grant new Ph.D. degrees, so the only way to earn a Ph.D. is to have a committee of people with Ph.D.s give one to you.
The even simpler answer is: You graduate with a Ph.D. when your primary research advisor says that you can graduate. When they give you the “thumbs up” to graduate, that means that they feel confident enough in their ability to convince the other 2–4 professors on your committee to also approve you to graduate.
So what does it take to convince your advisor to let you graduate? Obviously the details vary per advisor, but here is what I've empirically observed from the past ~15 years that I've been in academia in computer science and closely-related fields:
This process usually takes 5–7 years from the time you start the M.S./Ph.D. program. Note that prior publications from before when you started your current program don't count, since they were from research done at a different university.
If you're publishing at less competitive venues, maybe bump up those numbers by 1 or 2 papers to get the “equivalents” at top-tier venues. But it's all fuzzy; there's no hard-and-fast rule. Also, in other fields, journals are what matter most, not conferences.
Second-author papers don't count for you since they count toward the first author's Ph.D., not yours. Papers you write at internships usually don't count either unless they are very closely related to what you're doing with your advisor back at school. (I published 3 papers out of my Microsoft Research internships, but none made it into my Ph.D. dissertation.)
Finally, everything here is a rough approximation based on my own empirical observations. After all, you could graduate in four years and instantly get a faculty job at MIT or a million-dollar industry research gig with just a single paper if it happens to totally blow everyone's minds and utterly revolutionize your field.
Standard disclaimers apply: The above is all based on my own observations and is not meant as a set of hard-and-fast rules.
For more details, read this longer article: Overview of U.S. Science and Engineering Ph.D. Program Requirements