What is the typical amount of vacation time, per year, during a PhD program in the United States. I am particularly interested in 1. PhD programs in the physical sciences, and 2. PhD programs at competitive, research-focused universities.
I'm not aware of explicit policies regarding vacation time during a Ph.D in the US. This is likely because in the US, the Ph.D is viewed as educational, and not as a job.
As a student, and now as a professor, vacation time was always an informal discussion between advisor and student. This is of course ripe for abuse. A job in the US usually grants 15 days of vacation (not counting weekends) per year, but in my experience that's MUCH less than what you might effectively get as a Ph.D student (in computer science) but might be comparable to what you get as a Ph.D student in the physical sciences.
Unfortunately, I don't have personal experience with physical science programs, but via my wife and other friends I've seen that physical science programs are fairly rough on vacation. This is partly because you need to be around to tend to long-running experiments ("the flies died!") and partly because of the nature of lab work and the much more intense style of lab-based science.
My recommendation would be, once you have some options, to ask students in the programs you're applying to, working with advisors you're interested in targeting.
Well, this is dependent on a lot of factors:
Your advisor. He/she controls a good part of your life (and probably pays you out of a grant), so vacation time goes through him/her. I've seen advisors who begrudgingly let students take a couple of weeks off in the summer, to those who don't keep track and their students seem to be on vacation all the time. I've also seen advisors who mandate some vacation time to avoid burnout. (and all of the rest of the answers should be caveated with "and if your advisor allows")
Your workload and motivation. Get your work done (e.g., paper submitted, TA duties done, dissertation chapter written), and you can safely take a week or two off.
School schedule. While graduate students don't really stick by a strict semester/trimester/quarter/summer/etc. schedule, it more or less dictates when you can take vacation. I never took more than a day or two off during Spring Break, and my summers were strictly time to get more work done (or to teach a summer class).
School / department policy. Departments sometimes regulate time off, but I'd say that is rare, as vacation is, again, left up to the advisor.
Despite the low pay, you are an employee, and you need to abide by the rules set by your employer (either the school, your advisor, or, possibly, your fellowship rules). Sometimes students don't get paid during the summer, and it is easier to take vacation.
The bottom line is that if you are taking too much vacation, you are delaying getting your research done, and that can have ramifications on when you graduate, how your research progresses (and if you get scooped!), and what your advisor thinks of your work ethic. Should you take time off? Yes. Should you be greedy and think it will be like undergraduate school where you get three or four months off a year? No.
Having done a PhD in the United States and then come to Denmark to do a post-doc, I can say that this is going to depend on a huge number of factors. In the US, during my PhD, many of my colleagues seemed to be on vacation all of the time (many of them did not finish). But, my advisor never took vacation and I, accordingly, worked on his schedule. I never took vacation and was always "on email," etc. when I was away for whatever reason. We had no guarantee of vacation other, I guess, than national and university holidays, but most people work from home (or in the lab) during those times any way.
In Denmark, however, this is completely different. Everyone in Denmark is guaranteed five weeks of paid vacation every year and even in academia, people use it seriously. Because it is July, all faculty and graduate students in my department are gone. Literally all of them and they are not on email and claim to not work, most for a three week holiday abroad.
So, in short, this is going to depend on a lot on country, department norms, and your advisor's style/expectations.
Short answer: around 4 weeks (20 days).
Okay, given the number of anecdotal answers given, let's try to do something different and find some hard data… It is rather easy, because most institutions have explicit vacation policies posted online.
Princeton graduate school (which I think fits your stated criteria) has Guidelines on Student Vacation Time, which say:
graduate student degree candidates may take up to (but no more than) four weeks of vacation, including any days taken during regular University holidays and scheduled recesses
Caltech Graduate Studies Office states:
The Institute policy is that graduate students are "entitled to two weeks' annual vacation (in addition to Institute holidays)." […] There are 11 Institute holidays this calendar year […] In total, graduate students are entitled to 21 vacation days per calendar year. These days do not accrue from year to year.
MIT's policy for Graduate Students is the following:
[…] observe normal Institute holidays and are entitled to two weeks of vacation with pay if their appointments are for the full calendar year. Their vacation schedule must be approved by their supervisors
Two weeks vacation and all official Georgia Tech holidays are allowed during each calendar year. Advisors must be notified of all vacation time and absences. Mid-term and intermission breaks are not vacation days unless scheduled as such.
In summary, vacation/holiday time ranges from 4 to 5 weeks at the institutions listed. It should be noted that the above are the actual vacation policies, so real-life situation might be different: less strict, if your advisor is understanding and it doesn't impede your work; more strict, if there is a negative culture in your workplace.
To add some personal experience from two different systems to the other answers (Chris Gregg and Suresh; which I fully agree with).
The way this is handled clearly depends on where you are (I realize you are asking about US). In Sweden PhD students have a certain number of days (weeks). This is regulated by laws. However, a problem with trying to regulate vacation is that not all students or projects are alike and in the end you, as a PhD student, is responsible to complete your work in time (advisors obviously also have a role to facilitatet this). So in "my" system the stipulated vacation may be a blessing in disguise if taken very literally. The important point is that one must plan ones own time and that includes taking time off to reload batteries and rest. A difficulty is to balance these issues.
My own experience when I was a forreign graduate student in the US was that I really got a lot of work done during holidays and spring breaks when most people were away. I do not remember having any major vacation time but I always had days off for doing anything that could take my mind off graduate school (not that I was bored, I just felt rejuvenated by it). So managing time is more imprtant than having a long vacation. As I saw it grad school was my chance to get somewhere so it was worth working for.
Disclaimer: *n=1* at a research university
I can give you a reasonable estimate of the vacation hours as I am a current PhD student in the US, not in the physical sciences but in the computing sciences. At my institution (Cornell University), I can officially take 14 days off in the year when I am fully funded. This is what is typed on my funding letter which I receive at the beginning of every semester.
In practice, this depends upon your adviser. I have been fortunate enough to work with wonderful people who do not care how many days I take off as long as the work has been done according to their expectations. For instance, I took 2 months off in Dec-Jan last year but then I had finished my personal involvement in all the current projects and had submitted relevant papers for publication.
They are two ways to answer/interpret the question.
The first is: how much time are you "legally" entitled/required to take
The answer to this question will vary across country, university, and even how you are "employed" within a university. Even if this could be answered in general or for you specific situation, I am not sure it is helpful given the second interpretation.
The second is: how much time do PhD students actually take
This seems to me a more relevant question and is similar, if not identical, to: How hard do early-career academics in the United States work, really?. I believe that N-of-1 type answers are meaningless since I know people who takes zero vacation days and I know people who take in excess of 60 vacation days. The Sigma Xi society surveyed post docs and found that they take 12 vacation days a year on average. I provided details about this study in this answer of mine. I am not sure to the extent to which the vacation habits of post docs and PhD students are similar, but this is the only population study I am aware of.
As a Ph.D. you really have to rely on the 'mood' of your PI, or other circumstances: For example, I know a person who had to go back to China for a month to get his visa renewal.
But I think at least 1 week off for traveling during the summer is something that you could expect if you don't have an extremely strict PI.
Besides that, there is the Xmas week where the university is officially closed and you don't have to show up.
In addition to the other informative answers, I can't help but comment that (truly) one might ask oneself how/why this becomes a question at all. Of course it is a reasonable question, but it potentially hints at some aversion to one's "work".
I know others might disagree strenuously, but my own quite sincere and considered opinion is that academe is not a good career choice if the "job" is substantially oppressive or in any way "something to get away from". The reason is that quite a few people, your competitors, really love spending time at "the thing", and although of course "more hours" does not mean "more production", that kind of sustained engagement... and affection... does seem to enhance productivity and efficiency hugely.
That is, an academic job should be the kind of thing that one _does_not_need_ a vacation from, any more than one needs a vacation from eating or sleeping. Of course this is just one "ideal", but it is certainly the ground for my own life as a mathematician. That is (at the other end of one's career) when people ask me when I'll retire, regardless of exactly what I say, my thoughts are that it would be silly to stop accepting good pay for what I'd be doing anyway. :) :) (And, I add, for probably 12+ hours a day 6/7 days a week, and at least several hours every day of the year. I feel ill if I can't find a way to think about mathematics at least a few hours every day. If necessary, it seems that insomnia provides an opportunity...)
One may view this as a silly ideal, and my own experience as a bit of a caricature, but I think it is worthwhile for a potential academic to juxtapose such ranting with their own inclinations. E.g., if one can't feel an irrational affection for one's projects... it's time to consider other options.
So: vacation? I don't like the conventional notion of "vacation" (where you stop doing what you do ordinarily), any more than I'd like a vacation from eating or sleeping. Seriously. Yes, this creates some degree of conflict with family.
The worse conflict would be if one really does want to get away from one's (academic?) work. If so, then all the people who aren't necessarily as able as you, but who love it, will be zooming past you while you are on vacation. This is probably not just a "scare story", considering my personal observations over 40+ years.
Thus, conceivably, if you really think in terms of "vacation" from tasks that are not ... enchanting... then the real conclusion is that you should think about other possibilities.
(I thought that this ... arguably idealistic/extreme... viewpoint needed some representation.)
I'm working at a very competitive research university in physical sciences this summer as an undergraduate research intern. I asked a few of the PhD students i'm working with, and they told me that they get a maximum of two weeks off per year, not including one week that everyone gets off for Christmas. Thus a lot of them take their two weeks in addition to Christmas (a lot of internationals here don't care about Christmas much) to get an effective 3 week vacation.
Like others have said, it will depend entirely on your advisor. The PhD students I've talked to have told horror stories about one advisor in particular who wouldn't grant a PhD student an extra two days off to get married, because it would have put him over two weeks for the year. Others generally don't care as much as long as you get your work done.