A
STUDY OF NORMATIVE AND INFORMATIONAL SOCIALINFLUENCES UPON INDIVIDUAL JUDGMENT
MORTON
DEUTSCH
AND
HAROLD
B.
GERARD
1
Research
Center
for
Human
Relations,
New
York University
B
y
NOW,
many experimental studies (e.g.,
1, 3, 6)
have demonstrated
that
indi-vidual psychological processes
are
sub-ject
to
social influences. Most investigators,however, have not distinguished among dif-ferent
kinds
of
social influences;
rather, they
have carelessly used
the
term
group
influence
to
characterize
the
impact
of
many
different
kinds
of
social factors.
In
fact,
a
review
of themajor
experiments
in
this area—e.g., those
by
Sherif
(6), Asch (1), Bovard
(3)—would
indi-cate
that
the
subjects (5s)
in
these experimentsas
they
made their judgments
were
not
func-
tioning
as
members
of a
group
in any
simple
or
obvious manner.
The
S,
in the
usual experi-ment
in
this
area,
made perceptual judgments
hi
the
physical presence
of
others
after
hearingtheir judgments.
Typically,
the
S
was
not
given experimental instructions which made
him
feel
that
he was a
member
of a
group
faced
with
a
common
task
requiring cooperative
effort
for its
most
effective
solution.
If
group
influences
were
at
work
in the
foregoing experi-ments,
they
were subtly
and
indirectly createdrather than purposefully created
by the
experi-menter.
HYPOTHESES
The purpose of this paper is to consider twotypes
of
social
influence,
normative
and
informational, which
we
believe
were
opera-
tive
in the
experiments mentioned above,
andto
report
the
results
of an
experiment bearing
upon
hypotheses
that
are
particularly relevantto the
former
influence. We shall
define
a
normative
social
influence
as an
influence
toconform
with
the
positive
expectations
2
of
another.
3
An
informational
social
influence
may
Dr.
Gerard
is now at the
University
of
Buffalo.
This research
was
conducted
under
a
grant
from
the
Office
of
Naval Research, Contract
No.
NONR
285(10).
By
positive expectations
we
mean
to
refer
to
those
expectations
whose
fulfillment
by
another leads
to or
reinforces
positive rather than negative
feelings,
andwhose
nonfulfillment
leads
to the
opposite,
to
alienationrather than solidarity;
conformity
to
negative expecta-
tions,
on the
other hand, leads
to or
reinforces
negativerather than positive
feelings.
'The
term
another
is
being
used inclusively
to
refer
be
defined
as an
influence
to accept informationobtained
from
another
as
evidence
about
reality.
Commonly these
two
types
of
influence
are
found
together. However,
it is
possible
to
con-
form
behaviorally with
the
expectations
of
others
and say things
which
one
disbelieves
but
which agree with
the
beliefs
of
others.
Also,
it is
possible
that
one
will
accept
an
opponent's
beliefs
as
evidence about
reality
even though one has no motivation to agreewith him,
per se.
Our
hypotheses
are
particularly relevant
to
normative social
influence
upon individualjudgment.
We
shall
not
elaborate
the
theoreti-
cal
rationales
for the
hypotheses, since they
are
for
the
most
part
obvious
and
they
follow
from
other theoretical writings (e.g.,
4, 5).
Hypothesis
I.
Normative social
influence
upon
individual judgments will be greateramong individuals
forming
a
group thanamong
an
aggregation
of
individuals
who do
not
compose
a
group.
4
That
is,
even
when
susceptibility
to
informa-tional social
influence
is
equated,
we
would
pre-
dict
that
the
greater susceptibility
to
norma-
tive
social
influence
among group members
would
be
reflected
in the
greater group
influ-
ence upon individual judgment.
This
is not tosay
that
individuals, even
when
they
are not
group
members, may not have some motiva-
to
another person, to a group, or to one's self. Thus,
a
normative
social
influence
can
result
from
theexpectations
of oneself, or of a
group,
or of
anotherperson.
4
Generally
one
would
also
expect
that
group memberswould be
more
likely
to
take
the judgments of
other
group
members
as
trustworthy evidence
for
forming
judgments
about reality and, hence, they
would
be
more
susceptible
to
informational
social
influence
than
would
nongroup members. The
greater trustworthiness
usually
reflects
more experience
of the
reliability
of
the
judgments
of
other
members and
more
confidence
in the benevolence of
their motivations.
However,
when
group
members
have
had no
prior experience
togetherand
when
it is
apparent
in
both
the
group
and nongroup
situations
that
the
others
are
motivated
and in a
posi-
tion
to
report
correct
judgments,
there
is no
reason
to expect
differential
susceptibility
to
informational
social
influence
among
group
and
nongroup
members.
629
63
MORTON
DEUTSCH AND
HAROLD
B.
GERARD
tion
to
conform
to the expectations of others—e.g.,
so as to
ingratiate
themselves
or so as to
avoid ridicule.
Hypothesis
II.
Normative social
influence
upon individual judgment will be reduced when
the
individual perceives that
his
judgmentcannot be identified or, more generally, when
the individual perceives no
pressure
to
conformdirected
at him
from
others.
Hypothesis
III.
Normative social
influence
toconform
to
one's
own
judgment will reduce
the
impact of the normative social
influence
to con-
form
to the
judgment
of
others.
Hypothesis
IV.
Normative
social
influenceto conform to one's own judgment
from
anotheras well as
from
oneself will
be stronger
than
normative social
influence
from
oneself.
Normative social influence
from
oneself to
conform
to one's own judgment may bethought of as an internalized social process in
which
the
individual holds expectations withregard to his own behavior; conforming to
positive
self-expectations leads to feelings of
self-esteem
or
self-approval while noncon-
formity
leads to
feelings
of anxiety or guilt. Ingeneral, one would expect
that
the strength ofthese internalized self-expectations would re-
flect the
individual's
prior experiences withthem as sources of need satisfaction—e.g., by
conforming
to his own judgments or by self-reliance
he has won
approval
from
such signifi-
cant
others
as his
parents.
As Hypothesis IV
Indicates,
we
believe
that
contemporaneoussocial pressure
to
conform
to
one's
own
judg-ment
may
supplement,
and
perhaps
be
even
stronger
than,
the
individual's
internalized
pressure to
conform
to his own judgment.Two additional hypotheses, dealing with
the
effect
of
difficulty
of
judgment,
are
relevant
to
one
of the experimental variations. They
follow:
Hypothesis
V.
The
more
uncertain
the
indi-
vidual is about the correctness of his judgment,
the
more likely
he is to be
susceptible
to
bothnormative and informational social influencesin making his judgment,
Hypothesis
VI.
The more uncertain the indi-vidual is about the correctness of the judgment
of
others, the less likely he is to be susceptibleto informational social
influence
in making hisjudgment.
6
B
Although
we
have
no
data relevant
to
this hypothe-sis, we present it to
qualify
Hypothesis V and to
METHOD
Subjects.
One
hundred
and one
college students
from
psychology courses at New York University were
employed
as 5s. The
study
was
defined
for the 5s as
an experimental
study
of
perception.
Procedure.
We
employed
the
experimental situation
developed by Asch (1)
with certain modifications
and
variations which
are
specified
below.
For
detaileddescription of the procedures utilized by Asch andreplicated in this experiment,
Asch's
publication should
be
consulted.
The
basic features
of the
Asch situation
are:
(a) the 5s are
instructed
that
they
are
participatingin
a
perceptual experiment, wherein they have
to
matchaccurately
the
length
of a
given line with
one of
threelines;
(6)
correct judgments
are
easy
to
make;
c)
in
each
experimental
session
there
is
only
one
naive
S,
the
other
participants,
while ostensively
5s, are in
fact
stooges
who
carry
out the
experimenter's instruc-tions;
d)
each participant (i.e.,
the
naive
5 and the
stooges)
has to
indicate
his
judgments
publicly;
e)
on
12
of the 18
perceptual judgments
the
stooges announce
wrong and
unanimous judgments,
the
errors
of the
stooges
are
large
and
clearly
in
error;
(/) the
naive
5and the stooges are in a
face-to-face
relationship and
have been previously acquainted with one another.*counteract an assumption in some of the currentsocial psychological literature. Thus, Festinger (5)has written
that
where no physical reality basis exists
for
the
establishment
of the
validity
of
one's
belief,
one is
dependent upon
social reality
(i.e.,
upon
the
beliefs
of
others). Similarly, Asch
(2) has
indicated
that
group
influence
grows stronger as the judgmentalsituation diminishes
in
clarity.
The
implication
of
Hypothesis VI is
that
if an individual perceives
that
a
situation
is
objectively
difficult
to judge—that
others
as
well
as he experience the situation in the same way (i.e.,as being
difficult
and as having uncertainty about theirjudgments)—he will not
trust
their judgments any
more
than
he
trusts
his
own.
It is
only
as his
confidence
in
their judgments increases (e.g., because he deems
that
agreement among three uncertain judges provides
more
reliable evidence
than
one
uncertain judge)
that
the
judgments
of
others will have informationalsocial
influence.
However
(at any
particular level
of
confidence
in the
judgment
of
others),
one can
predict
that
as his
confidence
in his own judgment decreases
he
will
be
more susceptible
to
normative social
influence.
With decreasing
self-confidence
there is likely to be
less
of a
commitment
to
one's
own
judgment
and,
hence,
less
influence
not to
conform
to the
judgments
of
others. Inspection
of the
Asch situation
would
suggest
that
informational
social
influence would
be
strongly
operative. As
Asch
has put it (2, p. 461):
The
subject knows
(a)
that
the
issue
is one of
fact;
(b)
that
a correct result is possible; (c)
that
onlyone
result
is
correct;
(d)
that
the
others
and he are
oriented
to and reporting about the same objectivelygiven relations;
(e)
that
the
group
is in
unanimous
opposition
at certain points with him.He further, perceives
that
the others are motivated
to
report
a
correct judgment.
In
such
a
situation,
the
subject's accumulated past experience would lead
him
NORMATIVE AND INFORMATIONAL SOCIAL INFLUENCES
631
To
test
the
hypotheses
set
forth
in the
foregoing sec-
tion,
the
following experimental variations upon
Asch's
situation were employed:
1.
The
face-to-face
situation.
This
was an exactreplication
of
Asch's situation except
for the
followingminor modifications: (a) Only three stooges, ratherthan eight, were employed;
7
b)
the
5
and the
stoogeswere unacquainted prior to the experiment; and
c)
two
series
of 18
judgments were employed.
In one series
(the visual series),
the
lines were physically presentwhen
the
S
and the
stooges announced their judgments;
in the
other series (the memory series),
the
lines were
removed
before any one announced his judgment. Inthe memory series, approximately three seconds afterthe lines were removed the first stooge was asked to
announce
his judgment. The sequences of visual and
memory
series were
alternated
so
that
approximately
half
the
5s
had the
memory series
first and
half
hadthe
visual series
first.
2.
The
anonymous
situation.
This
situation was
identical
with the face-to-face situation except for the
following
differences:
(a)
Instead
of sitting in the visualpresence
of
each other,
the 5s
were separated
by
par-
titions
which prevented them
from
talking to each other
or
seeing
one
another;
b)
Instead
of
announcing theirjudgments
by
voice,
the 5s
indicated their judgmentsby
pressing
a
button;
c)
No stooges were employed.
Each
5 was led to
believe
he was
Subject
No. 3, and
the others were No. 1, No. 2, and No. 4. He was told
that
when
the
experimenter called
out Subject No. 3
he was to indicate his judgment by pressing one ofthree buttons (A, B, or C) which corresponded towhat
he
thought
the
correct line was. When
an 5
pressed
a
given button,
a
corresponding bulb
lit onhis own
panel
and on a
hidden master panel.
Pre-
sumably the appropriate bulb also lit on the panels ofeach
of the
other
5s,
but,
in
fact,
the
bulbs
on any
5s
panel were not connected to the buttons of the other
5s.
When
the
experimenter called
for the
judgments
to
expect
that
he
could rely
on the
judgments
of
others,especially if they all agreed.
That
is, even if his eyes
were
closed he might
feel
that
he could safely
risk
his
life
on the assumption
that
the unanimous judgments
of
the
others were correct. This
is a
strong informa-tional social
influence
and one would expect it to beoverriding except for the fact
that
the subject has hiseyes open and receives information
from
a source whichhe also
feels
to be completely trustworthy—i.e.,
from
his own
perceptual apparatus.
The
subject
is
placed
in
strong
conflict
because the evidences
from
two sources
of
trustworthy information are in opposition.
In the
Asch situation,
it is
apparent
that,
in
additionto informational social
influence,
normative social
influence
is
likely
to be
operating.
The
naive
5 is in a
face-to-face
situation with acquaintances and he may bemotivated to
conform
to their judgments in order toavoid being ridiculed, or being negatively evaluated, oreven possibly out of a sense of obligation. While it maybe impossible to remove completely the impact ofnormative social
influence
upon any socialized being,
it is
evident
that
the
Asch
situation
allows much oppor-tunity for
this
type of
influence
to operate.
7
Asch
found
that three stooges were about as
effec-
tive
in
influencing
the 5s as
eight.
of
Subject
No. 1, of
Subject
No. 2, and of
Subject
No. 4, a
concealed accomplice manipulated masterswitches which
lit
bulbs
on
each
of the 5's
panels
that
corresponded to judgments presumably being made by
these
respective
5s.
Subjects
No. 1, No. 2, and No. 4
were, in
effect,
electrical stooges whose judgments
were
indicated on the panels of the
four
naive 5s (all
of
whom were Subject
No. 3) by an
accomplice
of the
experimenter
who
manipulated
master
switches
con-trolling
the
lights
on the
panels
of the
naive
5s. The
pattern
of judgments
followed
by the electrical
stooges
was the
same
as
that
followed
by the
livestooges
in the
face-to-face
situation,
d)
In providinga rationale for being labeled Subject No. 3 for each of
the
naive
5s, we
explained
that
due to the
complicated
wiring
setup,
the 5's
number
had no
relation
to his
seating position. Implicitly,
we
assumed
that
each
5
would realize
that
it
would
be
impossible
for the
others
to
identify
that
a
judgment
was
being madeby him
rather than
by any of two others. However,it is apparent
from
postexperiment questionnaires
that
many
of the 5s did not
realize
this.
It
seems likely
that
if
we had
made
the
anonymous character
of the
judg-ments clear and explicit to the 5s, the
effects
of
this
experimental variation
would
have been even moremarked.3.
The
group
situation.
This situation was identical
to the
anonymous situation except
that
the
subjectswere instructed
as
follows:
This
group
is one of
twenty
similar groups
who are
participating
in this experiment. We want to see howaccurately you can make judgments. We are going
to
give
a
reward
to the five
best groups—the
five
groups
that
make the
fewest
errors on the series ofjudgments
that
you are given. The reward
will
bea pair of tickets to a Broadway play of your ownchoosing
for
each
member
of the
winning group.
An
error will
be
counted
any
time
one of you
makesan incorrect judgment.
That
is, on any given card
the
group
can
make
as
many
as
four
errors
if you
each judge
incorrectly
or you can
make
no
errors
if
you each judge correctly. The five groups thatmake the best scores
will
be rewarded.
4.
The
self-commitment
variation.
This
variation
was
employed in both the
face-to-face
and anonymous
situations.
In it,
each
5 was
given
a
sheet
of
paper
on
which to write down his judgment
before
he was
exposed
to the judgments of the others. He was told
not to
sign
the
sheet
of
paper
and
that
it
would
not be
collected
at the end of the
experiment.
After
the first
series
of 18
judgments,
the 5s
threw away their sheets.
The 5s did not
erase their recorded judgments
after
each
trial
as
they
did
hi
the
Magic
Pad
self-commitment
variation.
4A.
The
Magic
Pad
self-commitment
variation.
Thisvariation was employed in the anonymous situation.
In it,
each
5 was
given
a
Magic Writing
Pad on
whichto write down his judgment
before
he was exposed to
the
judgments
of the
others.
After
each
5 had
beenexposed to the judgment of the others and had indicatedhis own judgment, he erased his judgment on the MagicWriting
Pad by
lifting
up the
plastic
covering.
It was
made
convincingly clear
to the 5 that
only
he
would
ever
know what he had written
down
on the pad.
5.
The
public commitment
variation.
This
variation
632
MORTON
DEUTSCH AND
HAROLD
B.
GERARD
was
employed
in
both
the
face-to-face
situation
and in
the
anonymous
situation.
In it, the 5s
followed
the
same
procedure
as in the
self-commitment
variation
except
that
they
wrote down their initial judgments
on
sheets
of
paper
which
they
signed and which they knew wereto be handed to the experimenter after each series of
18
judgments.
RESULTS
The
primary
data
used
in the
analysis
of the
results
are the
errors made
by the
5s,
which
were
in the direction of the errors made by
the
stooges.
We
shall present
first the
data
which
are relevant to our
hypotheses;
later
weshall present other information.
Hypothesis
I,
The
data relevant
to the first
hypothesis
are
presented
in
Table
1.
The
tablepresents
a
comparison
of the
anonymous
situa-
tion
in
which
the
individuals
were
motivated
to
act as a
group with
the
anonymous situation
in
which
there
was no
direct attempt
to
inducemembership
motivation;
in both situations, no
self
or public commitment was made. The
data
provide strong support for the prediction
that
the normative social
influence
upon individualjudgments
will
be greater among individuals
forming
a group than among individuals who
TABLE
1
MEAN
NUMBER OF SOCIALLY INFLUENCED ERRORSIN INDIVIDUAL JUDGMENT AMONG GROUP
MEMBERS
AND
AMONG NONMEMBERS
Experimental
TreatmentGroup, anonymous, no
commitmentNongroup, anonymous,no commitment
N
1513
Mem-
ory
Series
6.87
3.
ISVisualSeries
5.60
2.77
Total
12.475.92
p
values*
.01
.05
.001
Based
on a
(
test, using
one
tail
of the
distribution.
do
not compose a group. The average member
of
the group made more than twice as manyerrors
as the
comparable individual
who didnot
participate
in the
task
as a
member
of a
group.Qualitative
data
from
a
postexperimental
questionnaire,
in
which
we
asked
the 5 to
describe any
feelings
he had about himself orabout
the
others during
the
experiment, alsosupport Hypothesis
I.
Seven
out of the fifteen
5s
in the
group
condition spontaneouslymentioned a
felt
obligation to the other groupmembers; none
of the
individuals
in the
non-group condition mentioned
any
feeling
of
obli-gation to go along with the others.
Hypothesis
II.
To
test
the
second hypothesis,
it is
necessary
to
compare
the
data
from
the
face-to-face
and anonymous situations amongthe individuals who
were
otherwise exposed tosimilar experimental treatments. Tables
2 and
3
present
the
relevant
data. It is
apparent
that
there was less social
influence
upon individual
judgment
in the anonymous as compared with
the
face-to-face
situation. This lessening
ofsocial
influence
is at the .001 level of
statistical
confidence
even when
the
comparisons includethe commitment
variations
as
well
as
both
the
visual
and the
memory series
of
judgments.
The
interaction between
the
commitmentvariations and the anonymous,
face-to-face
variations,
which
is
statistically
significant, issuch as to reduce the over-all
differences
between the anonymous and
face-to-face
situa-tion; the
differences
between the face-to-face
and the
anonymous situations
are
most
strongly
brought
out
when
there
is no
commit-
ment. Similarly,
if we
compare
the
anonymousand
face-to-face
situations, employing thememory rather than the visual series, the ef-
fect
of the
normative
influence
upon judgmentsin the
face-to-face
situation is increased some-what,
but not
significantly.
That
is, as we
eliminate counter-normative
influences
(i.e.,
TABLE 2
MEAN NUMBER
OP
SOCIALLY INFLUENCED ERRORS
IN
INDIVIDUAL JUDGMENT
IN THE
ANONYMOUS
AND IN THE
FACE-TO-FACE SITUATIONS
SituationFace-to-faceAnonymous
Self-Commitment
Public Commitmento CommitmentVisual Memory
Total
N
Visual Memory
Total
N
Visual Memory
Total
N
3
2 77
4.083.157.085.9213
13
.92
.64
.75
.73
1.67
1.37
12
11
1.13
.92
1.39
.46
2 52
1 38
13
