To PhD or not to PhD

How do you know, when you are first starting out, which students you should have graduate with a MSc and which you should try and keep on for a PhD? That is a question I have been wrestling with for a little bit now and still don't have a clear answer. There are obviously cases where it is clear, but it is the borderline situations that make it a tough call. I realize that it depends heavily on the student being willing to stick around and interested in take the project to the next step, but even when that is the case there will be students that probably shouldn't take on a PhD or are doing it for the wrong reasons. I initially enrolled as a MSc student as a grad student, then switched into the PhD program after a year. My supervisor was willing to let me do that and we had discussed it at the onset. Now I am in the position of having students and considering whether to suggest they make the same switch... or not. I am perfectly happy with my students, but is there a clear difference in the first year of grad school between those who will be successful as PhDs and those who will just get some stuff done? That I don't know yet. It is particularly difficult to tell when the lab is going through the growing pains of starting from scratch. How much of a student's progress is hindered by having to battle with new or missing equipment and reagents? It is proposal writing time, so perhaps I will get a better feel for where my students are at once we work through that process and we can sit down next semester and talk about possibilities.