Brannon Denning has some good advice for first year law students. In that vein, i'll summerize and extend for grad school. I just finished a 8 year program getting my MS and PhD.

First, why does it take so long. Well, alot of people end up switching labs after a few years, and get nothing to show for it. I got lucky and got my MS out of it, otherwise I would have been really stuck. Further, alot of people end up with projects that don't go anywhere. This may be due to bad fit of the student and the project, or badly thought of projects by the PI.

Statistics by the National Research Council (I think!) show that time to graduation in most hard sciences is about 7 or 8 years. In 'soft' fields like philosophy, history, english, it's even longer, because those programs are (grossly generalizing) unstructured. At least our school is structured: First two years, coursework and tutorials. Second two years, start project and do candidacy exam. 4+ years, finish PhD project.

So, for incoming students, prepare thyself for the long journey. You won't be out in 3-4 years like law/med students (unless you are really good or come in with a MS). On the other hand, you won't incure a ton of debt (they pay you a small living stipend).

Part of this preparation is what Brannon mentioned; exercise I found to be critical. Outside interests also important, but incredibly frowned upon in science. (what, you are not breathing science 100% of the time?). Some people just can't understand this notion, and they have miserable lives b/c all they do is science. Of course there is a work/life balance. Alot of students and researchers treat it like a job, a low paying hard job. They come in at 9am, and leave at 5pm, five days a week. That won't work in science. You have to be able to put in the hours required; I often come in at night to do time courses (and hate it) due to the necessity of the system. But if you do that, you need to set aside more downtime, ie really take off Sat and Sun, not just part days in the lab. If you are cruising during the week and only putting in 6 hours a day or less, you should probably come in on sat or sun to get stuff done. I feel like I've done my job if I put in at least 40 hours/week here, and that's experiment time, time when experiments are incubating.

So I guess time management is important, but also important is experiment management, asking the right questions, using the right controls, etc. This takes a long long time to develop, but the sooner you start the better you'll be. I've wasted years on bad research tracks that never led to anything useful; that happens alot and you need to be prepared for it. But you should be asking important questions, such as: has this question been addressed in the literature? How will answering this question help extend the field of knowledge? Which gap in knowledge will it fill? What do I need in the experiment to know that it works? Is it possible for the experiment to yield a meaningful, interpretable answer? If so, what would be the next step (or is the expt a dead end?). These are all critical questions, and the sooner you can define your PhD project, and formulate ways to fill a certain, small gap in the field, the sooner you can graduate.

Anyway that's my advice for entering students. How to think like a scientist is the thing you get out of grad school. Yes, you'll know techniques but that's not relavent. Finding good questions to ask, and being able to answer them, is the essence of getting a PhD.

If you are not interested in this, then either get a masters (which basically will make you a well-paid technician), or don't bother to waste 6-8 years of your life, do something you'll enjoy instead. If you don't enjoy science, there is no way you will be able to survive.