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.
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.