. 53
( 115 .)


the learner in order to
administer the electric
Based on data from Milglram (1974).
Chapter 7 Conformity, Compliance, and Obedience 263

continue to do harm to others in a face-to-face, intimate-contact situation. For example,
there was no shortage of Nazis willing to shoot Jews at close range during the early
stages of the Holocaust.
Milgram also suggested that in the remote-victim condition a “narrowing of the
cognitive ¬eld,” or cognitive narrowing, occurs. That is, the teacher can put the learner
out of mind and focus on the learning task instead. As the victim becomes more observ-
able, such narrowing becomes more dif¬cult, and obedience is reduced. These results
suggest that it is more dif¬cult to in¬‚ict harm on someone you can see, hear, or touch.
This is why it is probably easier to drop bombs on a city of 500,000 from 30,000 feet
than to strangle one person with your bare hands.

Power of the Situation A second variable Milgram investigated was the nature of the
institution behind the authority. The original studies were conducted at Yale University.
To test the possibility that participants were intimidated by the schoolʼs power and
prestige, Milgram rented a loft in downtown Bridgeport, Connecticut, and conducted
the experiment under the name “Research Associates of Bridgeport.” He also had the
experimenter represent himself as a high school biology teacher. Under these conditions,
obedience fell to 47.5%, down from 65% in the original, baseline study. Although this
difference of 17.5% does not meet conventional levels of statistical signi¬cance, it does
suggest that removing some of the trappings of legitimacy from an authority source
reduces obedience somewhat.

Presence and Legitimacy of the Authority Figure What if the authority ¬gure was
physically removed from the obedience situation? In another variation on his original
experiment, Milgram had the experimenter give orders by telephone, which varied the
immediacy of the authority ¬gure, as opposed to varying the immediacy of the victim.
He found that when the experimenter is absent or tried to phone in his instructions to
give shock, obedience levels dropped sharply, to as little as 20%. The closer the authority
¬gure, the greater the obedience.
After Milgramʼs original research was publicized, other researchers became inter-
ested in the aspects of authority that might in¬‚uence obedience levels. One line of
research pursued the perceived legitimacy of the authority ¬gure. Two different studies
examined the effect of a uniform on obedience (Bickman, 1974; Geffner & Gross,
1984). In one study (Geffner & Gross, 1984), experimenters approached participants
who were about to cross a street and requested that they cross at another crosswalk.
Half the time the experimenter was uniformed as a public works employee, and half
the time the experimenter was not in uniform. The researchers found that participants
were more likely to obey uniformed than nonuniformed individuals.

Con¬‚icting Messages about Obedience Milgram also investigated the impact of
receiving con¬‚icting orders. In two variations, participants received such con¬‚icting
messages. In one, the con¬‚icting messages came from the learner and the experimenter.
The learner demanded that the teacher continue delivering shocks whereas the
experimenter advocated stopping the experiment. In the second variation, two authority
¬gures delivered the con¬‚icting messages. One urged the teacher to continue whereas
the other urged the teacher to stop.
Social Psychology

When such a con¬‚ict arose, participants chose the path that lead to a positive
outcome: termination of harm to the learner. When there was con¬‚ict between author-
ity sources, or between the learner and the authority source, not one participant went
all the way to 450 volts.

Group Effects A fourth variation involved groups of teachers, rather than a single
teacher. In this variation, a real participant was led to believe that two others would
act as co-teachers. (These other two were confederates of the experimenter.) When the
learner began to protest, at 150 volts, one confederate decided not to continue. Defying
the experimenterʼs instructions, he walked away and sat in a chair across the room. At
210 volts the second confederate followed. Milgramʼs results showed that having the
two confederates defy the experimenter reduced obedience markedly. Only 10% of the
participants obeyed to 450 volts (mean shock intensity 305 volts). Thirty-three percent
of the participants broke off after the ¬rst confederate de¬ed the experimenter but
before the second confederate. An additional 33% broke off at the 210-volt level after
the second confederate de¬ed the experimenter. Thus, two-thirds of the participants
who disobeyed the experimenter did so immediately after the confederates de¬ed the
Why does seeing two others disobey the experimenter signi¬cantly reduce the
participantʼs obedience? One explanation centers on a phenomenon called diffusion
of responsibility. Diffusion of responsibility occurs when an individual spreads
responsibility for his or her action to other individuals present. In the obedience situation
in which there were two other teachers delivering shocks, the participant could tell
himself that he was not solely responsible for in¬‚icting pain on the learner. However,
when the two confederates broke off, he was left holding the bag; he was now solely
responsible for delivering shocks. Generally, when people are in a position where they
can diffuse responsibility for harming another person, obedience is higher than if they
have to deliver the harm entirely on their own and cannot diffuse responsibility (Kilharn
& Mann, 1974). In short, having two people defy the experimenter placed the participant
in a position of con¬‚ict about who was responsible for harming the learner.
There is another explanation for the group effects Milgram observed. When the two
confederates broke off from the experiment, a new norm began to form: disobedience.
The old norm of obedience to the experimenter is placed into con¬‚ict with the new norm
of disobedience. The norm of disobedience is more “positive” than the norm of obedience
with respect to the harm to the learner. Remember that when participants were given the
choice between a positive and a negative command, most chose the positive. The lone
participants in the original studies, however, had no such opposing norms and so were
more inclined to respond to the norm of obedience. Evidently, having role models who
defy authority with impunity emboldens us against authority. Once new norms develop,
disobedience to oppressive authority becomes a more viable possibility.

The Role of Gender in Obedience
In Milgramʼs original research, only male participants were used. In a later replication,
Milgram also included female participants and found that males and females obeyed at
the same levels. However, later research showed that there is a gender difference in obedi-
ence. In an experiment conducted in Australia, Kilham and Mann (1974) found that males
obeyed more than females. In another study conducted in the United States, Geffner and
Gross (1984) found that males obeyed a uniformed authority more than females did.
Chapter 7 Conformity, Compliance, and Obedience 265

Another way to approach the issue of gender effects in obedience is to determine
whether male or female authority ¬gures are more effective in producing obedience.
In Geffner and Grossʼs (1984) experiment, the effects of experimenter gender, partici-
pant gender, and participant age on obedience were investigated. The results showed no
simple effect of experimenter gender on obedience. Instead, experimenter gender and
participant age interacted, as shown in Figure 7.8. Notice that there was no difference
between older and younger participants (“younger” participants being under age 30, and
“older” participants being over age 50) when the experimenter was female. However,
when the experimenter was male, younger participants obeyed the male experimenter
more than older participants did.

Obedience or Aggression?
Milgramʼs experiment used an aggressive response as the index of obedience. Could
it be that participants were displaying aggression toward the learner, which had little
to do with obedience? Such an interpretation appears unlikely. In situations where
participants were allowed to choose the level of shock to deliver to the learner, the
average shock delivered was 82.5 volts, with 2.5% obeying completely. This is quite
a drop from the 405 volts with 65% obeying completely in the baseline condition
(Milgram, 1974).
These results were supported by a replication of Milgramʼs experiment by other
researchers (Mantell, 1971). In one condition of this experiment, participants were
allowed to set the level of shock delivered to the learner. Compared to 85% of participants
who used the highest level of shock in a replication of Milgramʼs baseline experiment
(no feedback from the learner), only 7% of the participants in the “self-decision” con-
dition did so. These results and others (Kilham & Mann, 1974; Meeus & Raaijmakers,
1986; Shanab & Yahya, 1978) lead us to the conclusion that participants were display-
ing obedience to the experimenter rather than to their own aggressive impulses.

Figure 7.8 Obedience
as a function of the gender
of an authority ¬gure and
participant age. Younger
participants were more
likely to obey a male
authority ¬gure than older
participants. Younger and
older participants obeyed
a female authority ¬gure
Based on data from Geffner and Gross
Social Psychology

Obedience across Culture, Situation, and Time
Milgramʼs original experiments were conducted in the United States, using a particular
research technique. Would his results hold up across cultures and across experimental
situations? Some critics of Milgramʼs study, Dutch researchers Meeus and Raaijmakers
(1986), argued that the type of obedience required in Milgramʼs experiment”physically
hurting another person”was not realistic. Such behavior is rare in everyday life. They
argued that people are more often asked to hurt others in more subtle ways. For example,
your employer might ask you to do something that makes another employee look bad.
Would you obey?
Meeus and Raaijmakers (1986) studied a different form of obedience: administrative
obedience. Dutch participants were told that the psychology department of a university
was commissioned to screen applicants for various state and civic positions and that
the department was using this opportunity to test the effects of stress on test achieve-
ment. According to instructions, participants made a series of disparaging statements
about a person taking a test for a state job. Fifteen statements, each more disruptive than
the previous, were used. The mildest statement was, “Your answer to question 9 was
wrong”; a moderate statement was, “If you continue like this, you will fail the test”;
and the strongest statement was, “According to the test, it would be better for you to
apply for lower functions” (p. 323). Understandably, job applicants became increas-
ingly upset with each comment.
Most of the Dutch participants obeyed; 90% read all 15 statements. This resembles
the Milgram experiment in which participants had to increase shock in 15 stages as the
victim became more upset. In Milgramʼs terms, they gave the full 450 volts. When ques-
tioned about it, they attributed responsibility for the harassment to the experimenter.
In another variation on Milgramʼs experiment, Australian participants assumed
the role of either transmitter of the experimenterʼs instructions or executor (Kilham
& Mann, 1974). In the transmitter condition, participants relayed orders to continue
shocking a learner to a confederate of the experimenter who delivered the shocks. In
the executor condition, participants received orders indirectly from the experimenter
through a confederate of the experimenter. The hypothesis was that there would be
greater obedience when the participant was the transmitter rather than the executor
of orders, presumably because the participant is not directly responsible for in¬‚icting
harm on the victim. Results supported this hypothesis. Participants in the transmitter
role showed higher levels of obedience than those in the executor role.
Milgramʼs obedience effect has been supported by other cross-cultural research.
For example, obedience among Jordanian adults was found to be 62.5%”comparable
to the 65% rate found by Milgram among Americans”and among Jordanian children,
73% (Shanab & Yahya, 1977). The highest rates of obedience were reported among
participants in Germany. In a replication of Milgramʼs original baseline experiment,
85% of German men obeyed the experimenter (Mantell, 1971). Overall, it appears that
obedience is an integral part of human social behavior.
Finally, Milgramʼs ¬ndings have withstood the test of time. Blass (2000) evaluated
replications of Milgramʼs experiments conducted over a 22-year period (1963 to 1985)
and found that obedience rates varied from a low of 28% to a high of 91%. However,
there was no systematic relationship between the time that a study was conducted and
the rate of obedience. According to Blass, it does not appear that an enlightenment effect
has occurred. An enlightenment effect occurs when results of research are disseminated
and behavior is altered. If this happened there should have been reliably less obedience
in later studies of obedience than in earlier studies (Blass, 2000).
Chapter 7 Conformity, Compliance, and Obedience 267

Reevaluating Milgram™s Findings
Milgram sought to describe the dynamics of obedience by comparing obedience rates
across different experimental conditions. A wholly different picture of Milgramʼs ¬nd-
ings emerges when a careful analysis of the audiotapes made by Milgram of almost all
sessions of his experiment was done (Rochat, Maggioni, & Modigliana, 2000). Such
an analysis by Rochat et al. showed that obedience within an experimental session
tended to develop slowly and incrementally through a series of steps. Rochat and col-
leagues classi¬ed participantsʼ behavior as either acquiescence (going along with the
experimenterʼs demands without comment), checks (the participant seeks clari¬ca-
tion of a restricted part of the procedure), noti¬es (the participant provides informa-
tion to the experimenter that could lead to breaking off of the experiment), questions
(the participant overtly expresses doubt or requests additional information about the
experimenterʼs demands), objects (the participant overtly disagrees with the experi-
menter and brings up some personal reason why he/she should not continue), or refuses
(the participant overtly declines to continue the experiment, effectively disobeying the
Rochat and colleagues found that the participantsʼ acquiescence to the experi-
menter was relatively brief. At the 75-volt level (when the learner ¬rst indicates he is in
pain), 10% of participants exhibited a low-level de¬ant response (minimum checking).
As the experiment progressed, opposition in the form of checking increased. By 150
volts, 49.7% of participants were checking, and by 270 volts all participants checked.
Additionally, 30% of participants either questioned, objected to, or refused the experi-
menterʼs orders at or before 150 volts, with an additional 35% reaching this high level
of opposition between 150 and 330 volts (Rochat et al., 2000). Interestingly, 57% of
the participants who eventually refused to continue began to protest before 150 volts,
whereas none of the fully obedient participants did so.
Regardless of the path chosen by a participant, he or she experienced a great deal
of con¬‚ict as the experiment progressed. Participants dealt with the con¬‚ict aroused
by the demands of the experimenter and the learner by becoming confused and uncer-
tain, and by showing high levels of distress (Rochat et al., 2000). Some participants
dealt with the stress of the situation by rationalizing away the suffering of the learner,
whereas others rushed through the remaining shock levels. According to Rochat and
colleagues, participants resolved their con¬‚ict in one of two ways. Some participants
completed the task to the 450-volt level in a “resigned or mechanical fashion” (p. 170).
Others resolved the con¬‚ict by becoming oppositional toward the experimenter by ¬rst
questioning and/or objecting to the experimenter and then later refusing, despite the
pressure put on the participant by the experimenter to continue (Rochat et al., 2000).

Critiques of Milgram™s Research
There were aspects of Milgramʼs experiments and others like them that were never
precisely de¬ned but probably in¬‚uenced levels of obedience. Consider, for example,
the gradual, stepwise demands made on the participant. Each 15-volt increment may
have “hooked” the participants a little more. This is in keeping with the foot-in-the-door
technique. Obeying a small, harmless order (deliver 15 volts) made it likely that they
would more easily obey the next small step, and the next, and so on (Gilbert, 1981).
Each step made the next step seem not so bad. Imagine if the participant were asked to
give 450 volts at the very start. It is likely that many more people would have de¬ed
the experimenter.
Social Psychology

What about the protests made by many participants? Very few participants went
from beginning to end without asking if they should continue or voicing some concern
for the victim. But they were always told, “You must continue; you have no choice.”
Perhaps, as some observers suggest, the experiments are as much a study of ineffec-
tual and indecisive disobedience as of destructive obedience (Ross & Nisbett, 1991).
When participants saw others disobey, they suddenly knew how to disobey too, and
many of them did so.
There is another, even more subtle factor involved here. The experiments have a
kind of unreal, “Alice-in-Wonderland” quality (Ross & Nisbett, 1991). Events do not
add up. The participantʼs job is to give increasing levels of electric shock to a learner in
order to study the effects of punishment on learning. The shocks increase as the learner


. 53
( 115 .)