>> people are more likely to harm when they are instructed to harm by a person wearing a white coat and saying 'the experiment requires that you must go on'.

> This is only true if people interpret a "white coat" to be a symbol of superiority and authority. Remove the white coat -- the symbol of superiority and authority -- and replace it with something else (a clown's nose? a sweatshirt? a fisherman's vest?) and people present with a different response.

yes, that is the idea. the notion is that what behaviour people exhibit is determined more by factors of the situation (whether the person is wearing a white coat or not) than by features of their personality. there is an account of the experiment here:

http://www.new-life.net/milgram.htm

> Ultimately 65% of all of the "teachers" punished the "learners" to the maximum 450 volts.

which is to say that 65% of people who participated in the experiment believed that they were delivering shocks that would kill the subject simply because a man in a white coat told them to. the results have been replicated. to the best of my knowledge no consistent personality traits were found to explain why some subjects didn't shock the person to that degree. the notion is that whatever test you have for how 'compassionate' a person is the test isn't as predictive as the situation is (where the situation tells us that the person will more likely shock until they believe the person has died). frightening, huh.

similarly if you want to predict whether the person will help the other person pick up their papers or not the situation (whether they find money or not) delivers a more reliable prediction than any personality traits do.