I got VERY lucky. My advisor is absolutely amazing. He goes far out of his way to make sure I am taken care of. He finds conferences for me to attend, fellowships I can apply for, and research collaborations with other professors. He even takes the grad students in his group hiking.
But I realize this isn't typically the case. I feel bad for some other grad students in different departments that are essentially free labor, with advisors who aren't as great as mine. I'll reemphasize: I got lucky. But if you are considering getting your PhD, eliminate the dice-roll from the equation and find a kind, caring person to do research with, because it makes a world of a difference. I would argue it is the most important aspect of getting a PhD, and it can make or break your time in grad school.
Everyone told me grad school would be the most difficult, grueling time of my life. In fact, it's the exact opposite. I get to wake up every day, work wherever I like, and I no longer have classes so the only thing I have to do is research. I get to use the world's best supercomputers and neutron sources. I'm essentially getting paid to have fun. Plus, I have access to all my university's resources (health insurance, workout centers, library, a large student body to interact with) without any of the administrative overhead that professors have (grant writing, internal politics, teaching).
Well said.
Find out how many of your advisor's former students want to continue working with him or her. That might be a good measure of how much they want to deal with him or her once they don't need to.
A shitty advisor will make your life horrific - will pile a toxic combination of guilt, uncertainty, with just the right amount of compliments to make you keep going while hating your life. I've seen many brilliant colleagues completely burn themselves out because of this.
I've had an excellent advisor both for my PhD and later my postdoc - and while I've had strong disagreements with both, they are now excellent friends and people who I really cherish.
Before you accept a PhD in a lab, talk to your advisor's students privately, and ideally also to her ex students.