Past installments of Unsolicited Advice dealt with such mechanical topics as how to choose an undergraduate school or graduate school, or how to get into graduate school. (Hell if I know how to get into undergraduate schools.) Now we step fearlessly into somewhat more treacherous territory: how to be a good graduate student. As always, this is one idiosyncratic viewpoint, and others should be offered in the comments. It's treacherous, of course, because there is certainly no right way to go about being a good graduate student. Once upon a time, as part of my ongoing campaign to discredit the notion of make-or-break general exams, I had the physics department at Chicago do a survey of their faculty, asking them to give a subjective rating of all the Ph.D. students who had graduated in the last five years (and with whose work they were familiar). We then plotted the resulting scores against how well they did on the candidacy exam. Result: there was a small handful of students who completely dominated on the exam, and were pretty much recognized as excellent physicists, clustered in the corner. Other than that, a complete scatterplot -- there was no correlation between test scores and success in physics (among this highly-selected sample). But if you plotted candidacy-exam scores against incoming physics GRE scores, it was almost a perfect correlation. There are some students who are the kind who are really good at physics in an exam-type environment, and who have the ability to carry through that talent to actually doing research. But there are others who struggle with the tests, yet nevertheless are great physicists. And vice-versa: you can be a crappy physicist, whether or not you do well at the GRE's and general exams. The point being, there are many ways to be a successful physicist, and a corresponding number of ways to be a successful grad student. So the first piece of advice, possibly too vague to be useful, is: Look to maximize your talents. Typically, your first year or two in grad school you have some flexibility. You're taking classes (this is written from an American perspective, sorry), and possibly also doing research, but you haven't necessarily been tied down to a final choice of thesis advisor, or even research field, or even theory vs. experiment. This would be a good time to be honest with yourself -- what are you really good at? You might have had your heart set on building the next great particle accelerator ever since you deconstructed your parents' stereo when you were twelve, but when someone puts a soldering iron in your hand you just can't seem to stop breaking things. But you did get a perfect score on the GRE. Well, maybe it's time to face the music and switch to string theory. But I'm burying the lede here. If I had to concentrate on a single useful piece of advice for grad students, it would be: Take the initiative. The deep truth of grad school is that the transition from undergrad to grad is when you go from primarily being "a student" to primarily being "a scientist." As a student, your primary responsibility was to do what your professors told you to. As a scientist, your primary responsibility is to do good science. Many students struggle in grad school, especially in the early years, because they are implicitly waiting to be told what to do. Don't wait -- try to figure out what you should be doing, and do it. Check the arxiv in the morning to look for interesting papers. Go to colloquia and seminars, even if you don't understand them -- nobody really understands them, and it's the best way to get a feeling for what those things are that you should be working toward understanding. Talk to people! Knock on professors' doors (or, more politely, email them to make an appointment), and chat with them about what they are doing and what you might like to be doing. Even better, talk to senior grad students and -- best of all -- postdocs! They have more time than professors, and have a better understanding of the situation you are in right now. (When it comes time to apply for postdocs yourself, you're going to need three letters of recommendation from scientists who know you and your work very well. If you can only think of one or two people who might qualify, you've badly mismanaged your time in grad school.) Come up with ideas! A good advisor will set you on a productive path for your first research projects, but that's no reason why you shouldn't also be trying to come up with good ideas yourself; at some point that's going to be your job, after all. And when it comes to the nitty-gritty of actually doing research, whether it's theory or experiment, don't expect anyone to hold your hand at every step -- use your brain to try to figure out what should be done next. At some point you will sit back and realize that it's kind of fun. And then it will dawn on you that you've passed the threshold toward which you've been progressing for quite a number of years -- you're an honest-to-goodness scientist. We can't pretend, of course, that being a scientist is just a matter of willpower; you do have to learn some stuff. One of the eternal grad-school dilemmas is how many courses you should take, vs. how quickly you should just devote yourself to doing research. I'm going to have to be wishy-washy here, as there is no right answer, although it's certainly possible to go too far in either direction. If you dive into doing research without having a proper grounding in coursework, you can end up being an expert in the one particular hyper-specialized thing that you are researching, but be left with a rather fuzzy grasp of all the rest of physics. Not only does a situation like that doom you to a lifetime of sitting in on talks that you don't understand, but it might prevent you from making crucial connections that would actually be useful in your own work. But contrariwise, it's certainly possible to spend too much of your time taking classes. Classwork is what you have trained to be good at, and in some ways it's a comforting environment. But it's ultimately not the point of why you are in grad school. Likewise, sometimes you will really want to learn some particular subject, but your department doesn't offer a course in it. Here's where you should figure out that it's your responsibility to teach it to yourself. Especially these days, when there are not only five good textbooks but countless reviews on every subject available online, there's no excuse for waiting for a teacher to come along -- see the previous paragraph. Even once you get past courses and are unambiguously doing research, a similar dilemma presents itself -- calculating vs. contemplating. (That would be the theorist's version of the dilemma, anyway; experimenters are invited to suggest alliterative formulations of "tinkering vs. collecting data.") Being a scientist is a back-and-forth process, between on the one hand looking at the big picture, learning the basics, thinking deeply, coming up with new ideas, and on the other hand digging into the details, getting your hands dirty, and actually coming up with some tangible results. Science depends on both, although many people are happier on one side than the other. Despite what was said earlier about finding your strengths, here's a situation where you should make an extra effort to compensate for your weaknesses. You might be someone who loves doing calculations, producing page after page of equations, or file after file of simulation output. But if they don't add up to an interesting result, people aren't going to care that much. Or you might have deep and creative ideas about the nature of space and time or high-temperature superconductivity. But if you can't wrestle those ideas down to some specific calculations, your colleagues aren't going to be all that impressed. Sometimes, remember, the best ideas actually come about because you are simply fooling around with some calculations for their own sake. All of this has been necessarily vague, in accordance with the fact that there are many good ways to be a successful grad student. But at the end, the goal (for most people) is pretty concrete: to land a good postdoc. Do keep that in mind. So, no matter what your individual approach to success is, here is the eyes-on-the-prize advice: Be the kind of grad student that people would like to hire as a postdoc. What kind of student is that? Well, just ask yourself what you would be looking for, if you had a pile of promising postdoc applications in front of you. Some people are lucky enough to get general-purpose fellowships that are based simply on their genius; so if the genius thing is working for you, great. More postdocs are hired by some particular person or group, to perform some fairly well-defined kind of research. What those people are looking for is a postdoc who will contribute to their group, whether by being an awesome individual researcher, or by being a useful collaborator. So, be that person. While you're in grad school, establish a track record of productivity by writing papers. Even better, write good papers -- write about things that other people are interested in. What is it about your research or skill set that makes you useful to people hiring postdocs? Become the world's expert in some hot topic, or the master of some novel technique, along with establishing your broad-based competence. A good postdoc is expected to enliven a research group by being plugged into all the latest good stuff going on in the field, bubbling with new ideas and the energy and know-how to turn those ideas into tangible results. That should be you. (Certainly, not everyone will become a postdoc, nor should they. One of my best students didn't even apply for postdocs, after he determined it just wasn't for him. There are many other directions in which to steer your career after a successful time in grad school, and it pays to keep those possibilities in the back of your mind all along. But I'm not really the one to ask about them.) To be more concrete yet: Be a finisher. After several years of grad school, what do you have to show for it? Write papers, do analyses, build equipment, finish experiments. Demonstrate beyond any doubt that you can take the project from beginning to end, not just sit around the coffee room and lob probing questions. Give talks! Have something to say, and be confident that other people want to hear it. I've actually heard some students say that they love science, but don't like writing papers or giving talks. That's like saying you love being a butcher, just aren't very fond of cutting up animals. (Suggestions for more illuminating similes are welcome.) Writing papers and giving talks is the entire point of what you are doing. Be enthusiastic about it, and while you're at it, be good at it. There are so many smart people out there who write impenetrable papers or give incomprehensible talks, one good way to distinguish yourself from the herd is to learn to communicate effectively. But it won't help unless you have something tangible to communicate. September has long been my favorite month of the year, as campuses come to life with the incoming students, many of them starting off on a new adventure of one sort or another. Go get 'em, tiger. Update: Many other people, of course, have offered advice on how to be a good grad student. If you know of any, mention them in the comments and I'll link from here.