http://home.swbell.net/revscat/perilsOfObedience.html

I believe the conclusions are somewhat right but approach them with some skepticism. The variations had dramatically different results. Although they still indicted the conclusion could still be right.

I’m not sure how much of the results were obeying authority or some other impulse like cooperative, not disappointing someone, or sadistic tendency brought forth because the subject perceive they didn’t have to take responsibility for their action because they were taking orders.

I would like to see them set it up where there are two people (subject, fake-subject) on equal footing. One would administer the test and flip the switches while the other would determine what level of shock to administer, determine by guidelines. The experimenters can switch the jobs and establishing different situations to determine if there are other possible influences.

There are other ways to set it up to isolate psychology traits.
Nothing is proved until someone has thoroughly tried to disprove it.

That said, anything that brings more attention to the Milgram experiment (or to its cousin, the Stanford prison experiment) is welcome.

(A recent book, A Question of Torture, listed circumstantial evidence that Milgram's experiment was carried out at CIA behest. Considering the other stuff the CIA was up to, it would not be surprising.)