The thing is, at least in terms of psychology (as far as I understand it, at least, and as I'm taking a degree in the thing, I should hope I understand it right!) the question isn't about causing harm, per-say, but causing
Unnecessary harm. If the harm caused is due to a key part of the experiment, something the experiment can not do without,
and the experiment could garner important test data,
And that test data possibility outweighs the harm it may cause, then that (just about) falls into the area of ethical. (Of course, what counts as useful data and unnecessary pain is a tricky subject at best, but that's why we have ethical committees to decide it.)
Here's what I think of it: The experiment, as it stands, was ethical. The pain it may or may not have caused was a necessity, because otherwise the experiment would pretty much have fallen on it's face. The general idea behind it was to test how far people could be pushed by authority figures into doing something you didn't actually want to do, and finding something that people didn't want to do besides accidentally killing a bloke would have been difficult at best. So any psychological harm that may have been caused was almost a necessity.
So yes, the experiment was ethical.
Running the experiment nowadays, however? Less so.
We've now got more questions about why this happens. People are raising objections to how the experiment works even here, and it was even run fairly recently with pretty much the same results, which at least implies that we haven't changed in a good 40-50 years. in essence, we've learned just about all we can from this experiment - especially because it's been run with nearly every outlaying parameter being changed. Running the experiment would be unethical, not necessarily because the experiment itself is unethical, but because the experiment is, by and large, now unnecessary - If you want people to find out the depths people will go because of authority figures, you don't really have to run the experiment, just show people previous experiments and explain it. Until we can come up with a very good reason to re-run it, it's unnecessary - which means the pain it causes is unnecessary.
At least, that's my best guess on why it's been deemed unethical. It could be that we're all pansies, but who's to say.
P.S.
thethird0611 said:
During the experiment, the recording even talked about 'Heart Problems'...
Not to be a pedantic bastard (Well, OK, I am, but still), but I believe that mentioning heart problems was only part of experiments after the first, to see if it would have any change in obedience. As in, "If we say this person has heart problems, will they still shock them to death?"
Indecently, for those curious - experimentation found that this is one of the few parameter changes that actually have no effect good or ill - as in, just as many people "shocked" a "person with heart conditions" as someone without.