Opinions and Social Pressure
Author(s): Solomon E. Asch
Source: Scientific American , Vol. 193, No. 5 (November 1955), pp. 31-35
Published by: Scientific American, a division of Nature America, Inc.
Stable URL: https://www.jstor.org/stable/10.2307/24943779
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Scientific American, a division of Nature America, Inc. is collaborating with JSTOR to digitize,
preserve and extend access to Scientific American
This content downloaded from
147.251.60.130 on Fri, 22 Mar 2019 23:14:35 UTC
All use subject to https://about.jstor.org/terms
NOVEMBER, 1955
SCIENTIFIC
AMERICAN VOL. 193, NO. 5
Opinions and Social Pressure
Exactly what is the effect of the opinions of others on our own?
In other words, how strong is the urge toward social conformity?
The question is approached by means of some unusual experiments
T
hat social influences shape every
person's practices, judgments and
beliefs is a truism to which anyone
will readily assent. A child masters his
"native" dialect down to the finest
nuances; a member of a tribe of cannibals
accepts cannibalism as altogether
fitting and proper. All the social sciences
take their departure from the observation
of the profound effects that groups
exert on their members. For psychologists,
group pressure upon the minds of
individuals raises a host of questions
they would like to investigate in detail.
How, and to what extent, do social
forces constrain people's opinions and
attitudes? This question is especially
pertinent in our day. The same epoch
that has witnessed the unprecedented
technical extension of communication
by Solomon E. Aseh
has also brought into existence the deliberate
manipulation of opinion and the
"engineering of consent." There are
many good reasons why, as citizens and
as scientists, we should be concerned
with studying the ways in which human
beings form their opinions and the role
that social conditions play.
Studies of these questions began with
the interest in hypnosis aroused by the
French physician Jean Martin Charcot
(a teacher of Sigmund Freud) toward
the end of the 19th century. Charcot
believed that only hysterical patients
could be fully hypnotized, but this view
was soon challenged by two other physicians,
Hyppolyte Bernheim and A. A.
Li6bault, who demonstrated that they
could put most people under the hypnotic
spell. Bernheim proposed that hypnosis
was but an extreme form of a
normal psychological process which became
known as "suggestibility." It was
shown that monotonous reiteration of instructions
could induce in normal persons
in the waking state involuntary
bodily changes such as swaying or rigidity
of the arms, and sensations such as
warmth and odor.
It was not long before social thinkers
seized upon these discoveries as a basis
for explaining numerous social phenomena,
from the spread of opinion to
the formation of crowds and the following
of leaders. The sociologist Gabriel
Tarde summed it all up in the aphorism:
"Social man is a somnambulist."
When the new discipline of social psychology
was born at the beginning of
this century, its first experiments were
EXPERIMENT IS REPEATED in the Laboratory of Social Rela·
tions at Harvard University. Seven student subjects are asked by the
experimenter (right) to compare the length of lines (see diagram
on the next page). Six of the subjects have been coached beforehand
to give unanimously wrong answers. The seventh (sixth from
the left) has merely been told that it is an experiment in perception.
31
© 1955 SCIENTIFIC AMERICAN, INC
This content downloaded from
147.251.60.130 on Fri, ffff on Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
essentially adaptations of the suggestion
demonstration. The technique generally
followed a simple plan. The subjects,
usually college students, were asked to
give their opinions or preferences concerning
various matters; some time later
they were again asked to state their
choices, but now they were also informed
of the opinions held by authorities
or large groups of their peers on the
same matters. (Often the alleged consensus
was fictitious.) Most of these
studies had substantially the same result:
confronted with opinions contrary to
their own, many subjects apparently
shifted their judgments in the direction
of the views of the majorities or the experts.
The late psychologist Edward L.
Thorndike reported that he had succeeded
in modifying the esthetic preferences
of adults by this procedure. Other
psychologists reported that people's
evaluations of the merit of a literary
passage could be raised or lowered by
ascribing the passage to different authors.
Apparently the sheer weight of
numbers or authority sufficed to change
opinions, even when no arguments for
the opinions themselves were provided.
Now the very ease of success in these
experiments arouses suspicion. Did the
subjects actually change their opinions,
or were the experimental victories scored
only on paper? On grounds of common
sense, one must question whether
opinions are generally as watery as these
studies indicate. There is some reason to
wonder whether it was not the investigators
who, in their enthusiasm for a
theory, were suggestible, and whether
the ostensibly gullible subjects were not
providing answers which they thought
good subjects were expected to give.
The investigations were guided by certain
underlying assumptions, which today
are common currency and account
for much that is thought and said about
the operations of propaganda and public
opinion. The assumptions are that peopIe
submit uncritically and painlessly to
external manipulation by suggestion or
prestige, and that any given idea or value
can be "sold" or "unsold" without reference
to its merits. We should be skeptical,
however, of the supposition that the
power of social pressure necessarily implies
uncritical submission to it: independence
and the capacity to rise above
group passion are also open to human
beings. Further, one may question on
psychological grounds whether it is possible
as a rule to change a person's judgment
of a situation or an object without
first changing his knowledge or assumptions
about it.
I
n what follows I shall describe some
experiments in an investigation of the
effects of group pressure which was carried
out recently with the help of a number
of my associates. The tests not only
demonstrate the operations of group
pressure upon individuals but also illustrate
a new kind of attack on the problem
and some of the more subtle questions
that it raises.
A group of seven to nine young men,
all college students, are assembled in a
classroom for a "psychological experiment"
in visual judgment. The experimenter
informs them that they will be
comparing the lengths of lines. He shows
two large white cards. On one is a single
vertical black line-the standard whose
length is to be matched. On the other
card are three vertical lines of various
lengths. The subjects are to choose the
one that is of the same length as the line
on the other card. One of the three
actually is of the same length; the other
two are substantially different, the difference
ranging from three quarters of an
inch to an inch and three quarters.
The experiment opens uneventfully.
The subjects announce their answers in
the order in which they have been seated
in the room, and on the first round every
person chooses the same matching line.
2
SUBJECTS WERE SHOWN two cards. One bore a standard line. The other bore three lines,
one of which was the same length as the standard. The subjects were asked to choose this line.
32
Then a second set of cards is exposed;
again the group is unanimous. The members
appear ready to endure politely another
boring experiment. On the third
trial there is an unexpected disturbance.
One person near the end of the group
disagrees with all the others in his selection
of the matching line. He looks surprised,
indeed incredulous, about the
disagreement. On the following trial he
disagrees again, while the others remain
unanimous in their choice. The dissenter
becomes more and more worried and
hesitant as the disagreement continues in
succeeding trials; he may pause before
announcing his answer and speak in a
low voice, or he may smile in an embarrassed
way.
\Vhat the dissenter does not know is
that all the other members of the group
were instructed by the experimenter
beforehand to give incorrect answers in
unanimity at certain points. The single
individual who is not a party to this prearrangement
is the focal subject of our
experiment. He is placed in a position in
which, while he is actually giving the
correct answers, he finds himself unexpectedly
in a minority of one, opposed
by a unanimous and arbitrary majority
with respect to a clear and simple fact.
Upon him we have brought to bear two
opposed forces: the evidence of his
senses and the unanimous opinion of a
group of his peers. Also, he must declare
his judgments in public, before a majority
which has also stated its position
publicly.
The instructed majority occaSionally
reports correctly in order to reduce the
possibility that the naive subject will suspect
collusion against him. (In only a
few cases did the subject actually show
suspicion; when this happened, the experiment
was stopped and the results
were not counted.) There are 18 trials
in each series, and on 12 of these the
majority responds erroneously.
How do people respond to group pressure
in this situation? I shall report first
the statistical results of a series in which
a total of 123 subjects from three institutions
of higher learning (folOt including
my own, Swarthmore College) were
placed in the minority situation described
above.
Two alternatives were open to the
subject: he could act independently, repudiating
the majority, or he could go
along with the majority, repudiating the
evidence of his senses. Of the 123 put to
the test, a considerable percentage
yielded to the majority. Whereas in ordinary
circumstances individuals matching
the lines will make mistakes less than 1
© 1955 SCIENTIFIC AMERICAN, INC
This content downloaded from
147.251.60.130 on Fri, 22 Mar 2019 23:14:35 UTC
All use subject to https://about.jstor.org/terms
per cent of the time, under group pressure
the minority subjects swung to acceptance
of the misleading majority's
wrong judgments in 36.8 per cent of the
selections.
Of course individuals differed in response.
At one extreme, about one quarter
of the subjects were completely independent
and never agreed with the
erroneous judgments of the majority. At
the other extreme, some individuals went
with the majority nearly all the time. The
performances of individuals in this experiment
tend to be highly consistent.
Those who strike out on the path of independence
do not, as a rule, succumb
to the majority even over an extended
series of trials, while those who choose
the path of compliance are unable to free
themselves as the ordeal is prolonged.
The reasons for the startling individual
differences have not yet been investigated
in detail. At this point we can
only report some tentative generalizations
from talks with the subjects, each
of whom was interviewed at the end of
the experiment. Among the independent
individuals were many who held fast because
of staunch confidence in their own
judgment. The most significant fact
about them was not absence of responsiveness
to the majority but a capacity
to recover from doubt and to reestablish
their equilibrium. Others who
acted independently came to believe
that the majority was correct in its answers,
but they continued their dissent
on the simple ground that it was their
obligation to call the play as they saw it.
Among the extremely yielding persons
we found a group who quickly reached
the conclusion: "I am wrong, they are
right." Others yielded in order "not to
spoil your results." Many of the individuals
who went along suspected that
the majority were "sheep" following the
first responder, or that the majority were
victims of an optical illusion; nevertheless,
these suspicions failed to free them
at the moment of decision. More disquieting
were the reactions of subjects
who construed their difference from the
majority as a sign of some general
deficiency in themselves, which at all
costs they must hide. On this basis they
desperately tried to merge with the majority,
not realizing the longer-range
consequences to themselves. All the
yielding subjects underestimated the
frequency with which they conformed.
W
hiCh aspect of the influence of a
majority is more important-the
size of the majority or its unanimity? The
experiment was modified to examine this
EXPERIMENT PROCEEDS as follows. In the top picture the snbject (center) hears rules
of experiment for the first time. In the second picture he makes his first judgment of a pair of
cards, disagreeing with the unanimous judgment of the others. In the third he leans forward
to look at another pair of cards. In the fourth he shows the strain of repeatedly disagreeing
with the majority. In the fifth, after 12 pairs of cards have been shown, he explains that "he
has to call them as he sees them." This suhject disagreed with the majority on all 12 trials.
Seventy.five per cent of experimental subjects agree with the majority in varying degrees.
33
© 1955 SCIENTIFIC AMERICAN, INC
This content downloaded from
147.251.60.130 on Fri, 22 Mar 2019 23:14:35 UTC
All use subject to https://about.jstor.org/terms
question. In one series the size of the opposition
was varied from one to 15 persons.
The results showed a clear trend.
When a subject was confronted with
only a single individual who contradicted
his answers, he was swayed little:
he continued to answer independently
and correctly in nearly all trials. When
the opposition was increased to two, the
pressure became substantial: minority
subjects now accepted the wrong answer
13.6 per cent of the time. Under
the pressure of a majority of three, the
subjects' errors jumped to 31.8 per cent.
But further increases in the size of the
majority apparently did not increase the
weight of the pressure substantially.
Clearly the size of the opposition is important
only up to a point.
Disturbance of the majority's unanimity
had a striking effect. In this experiment
the subject was given the support
of a truthful partner-either another individual
who did not know of the prearranged
agreement among the rest of
the group, or a person who was instructed
to give correct answers throughout.
The presence of a supporting partner
depleted the majority of much of its
power. Its pressure on the dissenting individual
was reduced to one fourth: that
is, subjects answered incorrectly only
one fourth as often as under the pressure
of a unanimous majority [see chart at
lower left on the opposite page]. The
weakest persons did not yield as readily.
Most interesting were the reactions to
the partner. Generally the feeling
toward him was one of warmth and
closeness; he was credited with inspiring
confidence. However, the subjects repudiated
the suggestion that the partner
decided them to be independent.
Was the partner's effect a consequence
of his dissent, or was it related
to his accuracy? We now introduced into
the experimental group a person who
was instructed to dissent from the majority
but also to disagree with the subject.
In some experiments the majority was
always to choose the worst of the comparison
lines and the instructed dissenter
to pick the line that was closer to the
length of the standard one; in others the
majority was consistently intermediate
and the dissenter most in error. In this
manner we were able to study the relative
influence of "compromising" and
"extremist" dissenters.
Again the results are clear. When a
moderate dissenter is present, the effect
of the majority on the subject decreases
by approximately one third, and extremes
of yielding disappear. Moreover,
most of the errors the subjects do make
34
are moderate, rathei than flagrant. In
short, the dissenter largely controls the
choice of errors. To this extent the subjects
broke away from the majority even
while bending to it.
On the other hand, when the dissenter
always chose the line that was more flagrantly
different from the standard, the
results were of quite a different kind.
The extremist dissenter produced a remarkable
freeing of the subjects; their
errors dropped to only 9 per cent.
Furthermore, all the errors were of the
moderate variety. We were able to conclude
that dissent per se increased independence
and moderated the errors
that occurred, and that the direction of
dissent exerted consistent effects.
I
n all the foregoing experiments each
subject was observed only in a single
setting. We now turned to studying the
effects upon a given individual of a
change in the situation to which he was
exposed. The first experiment examined
the consequences of losing or gaining a
partner. The instructed partner began by
answering correctly on the first six trials.
With his support the subject usually resisted
pressure from the majority: 18 of
27 subjects were completely independent.
But after six trials the partner joined
the majority. As soon as he did so, there
was an abrupt rise in the subjects' errors.
Their submission to the majority was just
about as frequent as when the minority
subject was opposed by a unanimous
majority throughout.
It was surprising to find that the experience
of having had a partner and of
having braved the majority opposition
with him had failed to strengthen the individuals'
independence. Questioning at
the conclusion of the experiment suggested
that we had overlooked an important
circumstance; namely, the strong
specific effect of "desertion" by the partner
to the other side. We therefore
changed the conditions so that the partner
would simply leave the group at the
proper point. (To allay suspicion it was
announced in advance that he had an
appOintment with the dean.) In this
form of the experiment, the partner's effect
outlasted his presence. The errors
increased after his departure, but less
markedly than after a partner switched
to the majority.
In a variant of this procedure the trials
began with the majority unanimously
giving correct answers. Then they gradually
broke away until on the sixth trial
the naive subject was alone and the
group unanimously against him. As long
as the subject had anyone on his side, he
was almost invariably independent, but
as soon as he found himself alone, the
tendency to conform to the majority rose
abruptly.
As might be expected, an individual's
resistance to group pressure in these experiments
depends to a considerable degree
on how wrong the majority is. We
varied the discrepancy between the
standard line and the other lines systematicalJy,
with the hope of reaching a
point where the error of the majority
would be so glaring that every subject
would repudiate it and choose independently.
In this we regretfully did not
succeed. Even when the difference between
the lines was seven inches, there
were still some who yielded to the error
of the majority.
The study provides clear answers to a
few relatively simple questions, and it
raises many others that await investigation.
We would like to know the degree
of consistency of persons in situations
which differ in content and structure. If
consistency of independence or conformity
in behavior is shown to be a fact, how
is it functionally related to qualities of
character and personality? In what ways
is independence related to sociological
or cultural conditions? Are leaders more
independent than other people, or are
they adept at following their followers?
These and many other questions may
perhaps be answerable by investigations
of the type described here.
E
e in society requires consensus as an
indispensable condition. But consensus,
to be productive, requires that each
individual contribute independently out
of his experience and insight. When consensus
comes under the dominance of
conformity, the social process is polluted
and the individual at the same time surrenders
the powers on which his functioning
as a feeling and thinking being
depends. That we have found the tendency
to conformity in our society so
strong that reasonably intelligent and
well-meaning young people are willing
to call white black is a matter of concern.
It raises questions about our ways of education
and about the values that guide
our conduct.
Yet anyone inclined to draw too pessimistic
conclusions from this report would
do well to remind himself that the capacities
for independence are not to be
underestimated. He may also draw some
consolation from a further observation:
those who participated in this challenging
experiment agreed nearly without
exception that independence was preferable
to conformity.
© 1955 SCIENTIFIC AMERICAN, INC
This content downloaded from
147.251.60.130 on Fri, 22 Mar 2019 23:14:35 UTC
All use subject to https://about.jstor.org/terms
100
i='
80
z
w
0
0<:
W
u.
6 0
Vl
w
I-
«
:::E
i=
Vl
w
40
I-
0
W
0<:
0<:
0
0
20
o
� � --
I�
I\J
YI\ ��r--� jf\
\l
....
V
•
,
2 3 4 5 6 7 8 9 1011 12
CRITICAL TRIALS
ERROR of 123 subjects, each of whom compared
lines in the presence of six to eight opponents, is
plotted in the colored curve. The accuracy of judg·
ments not under pressure is indicated in black.
100
80
i='
z
w
0
0<:
W
u.
60
Vl
w
I-
«
:::E
i=
Vl
w
40
I-
0
W
0<:
0<:
0
0
20
o
\ J� I'" Jl
'V V IV 1\
�J
\
V
... )� J\
\1
V
IV
•
,
123 4 5 6 7 891011 12
CRITICAL TRIALS
TWO SUBJECTS supporting each other against a
majority made fewer errors (colored curve) than
one subject did against a majority (black curve).
60.
50
i='
40
z
ill
0
0<:
W
30u.
Vl
0<:
0
0<:
0<:
W
20
10
o
........
V i""'"""'-oo
r---...�
j
I
I
I
17If
I
I
2 3 4 5 6 7 8 9 10 11 12 13 14 15
NUMBER OF OPPONENTS
SIZE OF MAJORITY which opposed them had an effect on the subjects. With
a single opponent the subject erred only 3.6 per cent of the time;'with two
opponents he erred 13.6 per cent; three, 31.8 per cent; four, 35.1 per cent;
six, 35.2 per cent; seven, 37.1 per cent; nine, 35.1 per cent; 15,31.2 per cent.
60
50
40
i='
z
w
0
0<:
W
30u.
Vl
0<:
0
0<:
0<:
W
10
II
� I� J
10
o
��I\III
I 2 3 4 5 6
J ....
I J \ �
I V , j
� �
-
J�
I
j LU �
L-V , I'
7 8 910 II 12 13 14 15 16 17 18
CRITICAL TRIALS
PARTNER LEFT SUBJECT after six trials in a single experiment. The
colored curve shows the error of the subject when the partner "deserted" to
the majority. Black curve shows error when partner merely left the room.
35
© 1955 SCIENTIFIC AMERICAN, INC
This content downloaded from
147.251.60.130 on Fri, 22 Mar 2019 23:14:35 UTC
All use subject to https://about.jstor.org/terms