This article contains some very useful advice, and I wish it were posted earlier (Stanford's application is due tomorrow, and many shortly thereafter).
I'm interested in graduate school because I'm considering a Ph.D. but I want to make sure research is actually rewarding. A Master's program seems like an excellent stepping stone, because a Master's degree looks great even if it turns out research is not for me.
Many American Ph.D. programs will let you pick up a Master's along the way just by completing the required coursework and submitting a form at the appropriate time - generally right before your comprehensive exams / selection of a thesis topic.
It's therefore possible to apply to Ph.D. programs, get accepted, spend the first couple of years figuring out if a Ph.D.'s for you, and then, if you decide negatively, leave with the "consolation Master's" - a degree completely indistinguishable from a terminal Master's, except:
- you were in a Ph.D. program, so you likely got a full fellowship instead of paying tuition yourself;
- you were in a Ph.D. program, so you had an easier time getting grants;
- you were in a Ph.D. program, so you got more access to and were treated more seriously by top-quality professors.
I got my consolation Master's in 2003 from a program I'd probably still be paying for today if I had applied for the Master's directly.
The third point you make about getting more access to top-quality professors is not exactly true. I have access to, quite possibly, one of the country's top researchers of high-frequency trading, and I am only working on a Master's. I also have access to two other professors who regularly hold seminars for major investment banking firms around the world.
Your access to those profs might not be as free as their PhD students' access, though. If it is, you might still have had more access as a PhD candidate in the same department.
I realized that, but I feel somewhat like a long shot to get into the programs I want in the first place (Stanford/MIT/Berkeley etc.), so going for a Ph.D. would only decrease my chances of getting in, as the author notes.
Also, my understanding is most schools provide some sort of TA or RA opportunity even for those in the Master's program. Is this not true?
FYI, Berkeley doesn't really have a Master's in CS, and the Master's program at Stanford is very different from the PhD program (you have to pay tuition and take a bunch of classes, whereas the Stanford PhD program has very few course requirements and you get much more departmental support).
I understand where you're coming from -- when I was initially thinking of applying to graduate school, starting with a Master's seemed like the safer and more sensible choice. But after talking to several people about it, the unanimous advice was to apply straight to the PhD program, for reasons similar to those enumerated above.
Support for Master's students really depends on the school. UCSD, where the author works, provides good support for Master's students. I have a friend who is attending UCSD for a Master's (albeit in Civil Engineering) who has received a TA position that covers his in-state tuition (but has to pay out-of-state fees), and pays a stipend.
Almost all the graduate schools I have researched have explicitly stated whether or not they offer any support to Master's students.
Afaik, finding support for CS students is not hard in many schools. Many other departments recruit RAs from the CS dept to help with their research. e.g. I had a RA with Arid lands research program 3 out of 4 semesters during my Master's. Though this will probably vary from uni to uni and this info can be found by getting in touch with some Master students in the dept.
Also, I had a TA in my last semester. So, its a YMMV for RAs/TAs depending on the university.
I know someone who wanted a Master's degree, but applied to Ph.D. programs so he could get it paid for. Things didn't go exactly as planned, and now he's a computer science professor.
He seems pretty happy with the outcome, though.
(I don't really recommend doing this, but if you're strongly leaning Ph.D. anyway and you have some research experience, you should probably just go for it.)
I'm in the same boat, applying to graduate schools right now for a M.S. in Artificial Intelligence, specifically machine learning and knowledge representation. Just submitted my application for the University of Toronto today. All others in the US I will be submitting around December 15.
I appreciate the insightful and helpful comments, and I agree with m0th87's reasoning for taking the Master's route. Personally, I decided to apply to the Master's because I want to work in the industry for 3-5 years before returning to school to complete a Ph.D. For some reason I feel like applying for a Ph.D to then take the consolation Master's is (in some way) cheating the system, and I can't bring myself to do that. Any thoughts?
If you want to work in industry for 3-5 years, I would do that either before or after graduate school, not in the middle. Once you've started graduate school (at least in a research program) and you're happy with your adviser/program, it will be a net loss of time if you leave school, work for 3-5 years, and then need to go back to school again.
Apologies, I misstated my intentions. What I meant to say is that I want to work in industry for 3-5 years, and then decide if I want to keep working, or pursue either a PhD or MBA.
One caution: depending on the school you go to, being in the Master's program might not afford you any opportunities to do any real research (outside of semester projects for your classes.)
I did a CS Master's at Michigan and there's a huge difference between the Master's student and PhD student experiences. Master's students are expected to do the coursework and get out. I was actively discouraged from even doing a Master's Thesis and I was told doing one was almost unheard of in the department. PhD students are in research labs right away and are guaranteed funding, which leaves essentially no room for Master's students to do research or teaching assistantships.
This arrangement worked well for me, since I wanted to do a lot of coursework and get into industry, but it would have been a horrible place to test my aptitude for and enjoyment of research.
On the other hand, a friend of mine got a Master's in HCI at Michigan and did a bunch of research and had funding for 3 of his 4 semesters. So it all depends on the department.
I'm interested in graduate school because I'm considering a Ph.D. but I want to make sure research is actually rewarding. A Master's program seems like an excellent stepping stone, because a Master's degree looks great even if it turns out research is not for me.