Mentoring choice: hands off or hands on?

In my PhD, I've had two mentors with very different mentoring practices. First one was very hands on, tracking even what time you were in the lab, and the second rarely visited lab. I realized things didn't work out with the first, hence my switch to the second. I am now very happy with what I did. I now know that my productivity goes down with increased pressure.

So, there must be two ends of the mentoring scale. One extreme is very hands-off, to the point that your advisor doesn't even know your project and the other would be you are not allowed to leave the lab or do any experiment without your advisor's approval. And you punch cards for your time in the lab. In my experience all mentors can be measured in this scale. One smart lad can even come up with a formula or index at some point. It's too late for me to quantify mentorship as I'm on the other side of the table.

The choice is there for both the student and the advisor.

If you are a (prospective) student, you must collect all the information you can to know the mentor's habits. Speak to the current students about life in the lab. Certainly none of them will speak badly about the advisor, especially if he is an intolerant slave driver. But you will notice the "red flags" during your conversation. The best people to find out about the advisor are the alumni. Take time to call them (not email, since they will not write in detail as any written evidence can be used against them) and ask direct questions. Of course the more data points the better since there could be personal grudges. Ideally, you should avoid the slave drivers (as they will make your most youthful times miserable) and stay away from totally hands-off people (as you will not be able to graduate in time). But I'd choose the latter anytime over the pressure cookers.

If you are a (prospective) mentor, you should think of how your mentoring style will be. If you don't, then you will do just like how you were trained. And that may not be the best. It is almost like setting up a company, with much less hierarchy. In my experience, the most important point is to decide how you will interact with the students and in what frequency. Because they are the ones that generate data (which will get you papers, recognition, funding, more students, etc.) you must know and decide how best you can keep them happy. And since each student is (very) different, your style must be flexible enough to cover all.