Test Subjects Aren't Stupid

Jim Manzi makes a good point at the Corner about the well-known Milgram experiment:

... in which subjects were led into a mock experimental laboratory and told to press a button that supposedly delivered increasingly severe electrical shocks to other “subjects” (who were actually other researchers) at the instruction of the scientists running the experiment. It was all a set-up: no real shocks were delivered, but the subjects believed that they were delivering up to lethal voltages to other humans. Milgram was shocked (ha ha) to discover that a majority of subjects were willing to deliver all the shocks.

Manzi points out, rightly, that subjects of this experiment likely took into account the fact that the experiment was conducted within an environment that wouldn't allow any real harm to people supposedly being shocked. He pictures them thinking to themselves

I don’t understand exactly what’s going on, but all my assumptions about how the world works would be violated if Yale University could really run a torture chamber operated by random people picked off the street

Indeed, I think it's natural for subjects to be curious about what the researchers are testing. I don't think it's always difficult for subjects to figure out what's going on an act accordingly, to conform to what they believe researchers expect.

I remember having a good laugh with a friend who participated in a psychological study in college. He was given a series of questions which gauged aspects of his attitude toward women. Then he was told that there would be a second phase to the test, but that it would take some time to set up. He was handed some magazines by a researcher, "to pass the time while he waited." Those magazines turned out to be pornographic.

Of course, it was instantly obvious to my friend that the whole point of the test was to see if looking at those magazines changed his attitude towards women. It would be obvious to anyone, no? How often does someone nonchalantly hand you some porn? That sort of realization would, I think, affect the outcome.

The Milgram experiment may not seem poorly designed, but the fact is the participants realize they are participating in something. How hard would it have been to guess? How did affect it the outcome?