Page 473 -

15.3 Care and handling of research participants 465

administer an electric shock to the learner if incorrect answers were given and
that the voltage of the shock should be increased after each incorrect answer.
Shocks were described as being “extremely painful,” but incapable of causing
permanent damage (Milgram, 1963).
This description was an elaborate deception aimed at concealing the
true goal of the experiment: a study of the limits of obedience. As the
“learner” was in fact a colleague of the experimenter's, no actual shocks
were administered. However, the subject did receive a mild shock to provide
evidence of the authenticity of the equipment and the learner acted as if
shocks had been applied. The experimenter participated actively in the
deception, urging subjects to continue with the experiment even when they
expressed reluctance.
The results of the study were intriguing: of 40 participants, all continued
giving shocks until after the point where the “learner” kicked on the wall and
stopped responding to the test questions. Most (26 out of 40) of the participants
administered the maximum level of shock—two steps beyond “Danger:
Severe Shock.” Participation caused discomfort including nervous laughter,
embarrassment, and seizures for several subjects.
This experiment would not have worked without deception: had the
subjects known that they were not actually administering potentially painful
shocks, they presumably would have been even less reluctant to participate.
The deception created a scenario in which obedience had a real cost, in
terms of the distress associated with inflicting harm on a fellow human
being.
Milgram's experiment would not be considered appropriate human
subjects research in most current research environments. The extreme
nature of the psychological distress involved in these experiments and
the strong reactions experienced by some of the participants raise serious
questions as to whether such research can ever be conducted responsibly
(Milgram, 1963).
Virtual environments provide interesting possibilities for subsequent
investigations of similar phenomena without raising the ethical concerns
associated with Milgram's experiment as originally executed. In a “virtual
reprise” of those experiments, subjects were asked to administer shocks
to a female virtual human in an immersive environment. The use of a
computer-generated character eliminated the need for deceit, thus removing
some of the possible ethical objections. Although participants knew that
they were interacting with a computer-generated avatar, they responded
to the situation as if they were working with a real person, particularly
if they could see the avatar (as opposed to communicating via a text chat
interface) (Slater et al., 2006).