Don't even ask how I came open the Stanford Prison Experiment but I found it extremely interesting, particularly in the light of the abuses at Abu Ghraib a few years ago.
In a nutshell, this guy at Stanford took a bunch of students and stuck them in a basement to replicate a jail. Half played prisoners and the other half guards. The experiment was to run for two weeks. It was shut down after six days because the guards became too sadistic and the prisoners really started to suffer a lot of emotional damage. Six days. That's fricken nuts. Read the Results section to see some of the craizness.
So this is what I take away from that experiment and the Abu Ghraib ordeal. In normal prisons you have a pretty exhaustive training and screening process to weed out the 'bad apples'. I assume prisons are also regularly supervised from outside sources to prevent abuses, etc. As a result, while abuses probably still happen, they are as systemic or widespread as the Abu Ghraib and the Standford ones become.
These kids playing guards weren't throughouly screened nor trained so it makes sense to me that you end up with more people that have sadistic tendencies than you would in professional prisons. I think a similiar thing happened in Abu Ghraib and who knows where else where you have soldiers put into a position they weren't screened or training for enough and you end up with more people with sadistic tendencies in those positions of power. As a result, you end up with the abuses that came to light. I think the Stanford experiment supports this pretty well.
Anyway, just thought it was pretty interesting.