Moscovici_2

18
Influence of a Consistent Minority on the Responses of a Majority in a Color Perception Task Author(s): S. Moscovici, E. Lage, M. Naffrechoux Source: Sociometry, Vol. 32, No. 4 (Dec., 1969), pp. 365-380 Published by: American Sociological Association Stable URL: http://www.jstor.org/stable/2786541  . Accessed: 05/04/2011 03:07 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at  . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at  . http://www.jstor.org/action/showPublisher?publisherCode=asa . . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. 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 [email protected].  American Sociological Association  is collaborating with JSTOR to digitize, preserve and extend access to Sociometry. http://www.jstor.org

Transcript of Moscovici_2

Page 1: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 1/17

Influence of a Consistent Minority on the Responses of a Majority in a Color Perception TaskAuthor(s): S. Moscovici, E. Lage, M. NaffrechouxSource: Sociometry, Vol. 32, No. 4 (Dec., 1969), pp. 365-380Published by: American Sociological AssociationStable URL: http://www.jstor.org/stable/2786541 .

Accessed: 05/04/2011 03:07

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at  .http://www.jstor.org/action/showPublisher?publisherCode=asa. .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

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 [email protected].

 American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to

Sociometry.

http://www.jstor.org

Page 2: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 2/17

Influenceof a ConsistentMinority

on the

Responses of a Majority in a Color

PerceptionTask

S.

MOSCOVICI,*

E.

LAGE

AND

M. NAFFRECHOUX

Ecole Pratique

des

Hautes Etudes,

Paris

Most

of

the

studies on social influence ave dealt with conformity,

ocial

pressure xercisedby majority roups, nd have used dependency

s

the

source of influence. his study concerns nnovation, ocial pressure xer-

cized by

a

minority,nd triesat the same time to prove that behavioral

style

s a

general ource of influence. n objectively lue stimulus

s used

which two subjects stooges) out of six call green n the experimental

groups. When the behaviorof the minoritys consistent, he

number f

green replies n the experimental roups s significantlyigher

han in

the control roup.This change n answer s not only verbal

greement ut

correspondso a change n theirperception ode, as shownby a

color dis-

criminationest. When the minority's ehavior s not consistent,ts impact

on

the majority

s

minimal.Thereforet is the consistent ehavioral tyle

of minoritieshat nsures he adoption f theirpoint of view.

THE

CONFORMITY BIAS

Specialised

iterature

ommonly

ssimilates

he

process

of influenceo

the

process

of

conformity

Allen,

1965).

On the

one hand, the tendency s to

assume that

any type

of

influence

eads

to conformity,nd

moreover hat

conformitys the solephenomenonchievedby means of influence.

n the

other

and,

when

xamining

he

ndividual,

t is

always ssumed hat

he

asks

himself hequestion Should follow he group r theminority? r in other

wordshe is faced

with

the alternative

f conformityr deviance.On the

contrary,

n

individual

requentlyoses

the

question

n

exactly

he inverse

manner: What

should

do so

that he

majority ill doptmypoint fview?

How

can

I

change

he

conception

f

others?

The

multiplicity

f

such

pos-

sible questions

ends o

contradict

he

afore

mentioned

ssimilation.

Without

going

nto the details stated

elsewhere

Moscovici

and

Faucheux,

1969)

we

can consider he nnovation s a

form f

social

influence.

n

order

o

study

theoreticallyndempirically

his

form,

he

nalysis

f

the ction f a

minority

upon

the

majority,

he

qualities

which

t

must

possess

n

order

o make

ts

*Fellow

(1968-1969)

at

the

Center for Advanced

Study

in

the

Behavioral

Sciences.

I also wish

to

acknowledge

the

assistance

extended to

me

by

the

James

Marshall

Fund.

365

Page 3: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 3/17

366

SOCIOMETRY

point of view accepted, onstitutes

sort of prolegomenon.

his research

proposes o showmore clearlyone

of these qualities and

to depart from

the customarymphasis

n attitudeswhich re linked o

conformity.

BEHAVIOR

STYLE AS

A

SOURCE OF

INFLUENCE

In

almost ll

of the research

one

to date

on social influence

nly

one

of

its possible sources

has been studiedtheoretically

nd experimentally:

e-

pendency.

Nonetheless,

or certain

easons,

we

cannot

make

use

of it in the

study

of

innovation.

irst

of

all,

it seemsclear

that

dependency

n

relation o

an

individual r a subgroupwhich

nnovates,

s

a

consequence

ather han

a

cause of an action imedat exertingn influence.he necessity o heedthe

advice

of electronics,omputer

r television

xperts ollows

he

adoption

of

electronic

quipment,

omputers,

r

television,

r

any

kind of

specific

ech-

nical invention.

minority hich

truly nnovates,

which

transforms

ocial

reality, nly rarelyhas powerat the

outset. n addition,

t is to be noted

that he ndividuals

r

subgroups

ho

change ules, alues,

orknowledge,

re

not judged as beingsuperior o others

nsofar s competence

s concerned.

In

short,dependency

n

relation o

the

phenomenon

hich nterests s

is neither decisivendependentariable, or a differentialactorwhich an

account for

nfluencewhich

s

exerted.

Thus,

we were

prompted

o seek

another

ource

of

influence

hich s

not

subject

to

the

imitations hich

we

have ustmentioned,

nd

which

omes loser o

expressing

he activeresolute

character

f a

minority.

e

believe

thatwe have found

t

in the behavioral

style

of the

individual r those

individualswho

proposea solutionto a

problem,

new

normfor

group.

Goodreasons

xist

to

suppose

that n the

process

of

innovation,

he

way

in which

the

behavior

s

organized nd

presented

ould

suffice

o

provoke

he

acceptance

r the

rejection

f

a judg-

ment r a proposedmodelduring hecourseof social interaction. oreover

the consistency

f the

behavior

of

a

minority,

he

fact

that it

resolutely

maintains

well

defined

oint

of viewand

develops

t

in

a coherentmanner,

appears

s

if it

ought

o be a

powerful

ource

of

influence, hichunder he

circumstances

ould

not be a result

f

an

explicit ependency.

A series

of

experiments

ade

by

one

of the authors n

collaboration ith

(Faucheux

and

Moscovici,1967)

has

already

shown

the

impact

of

a con-

sistent

minority

pon

a

majority

when

preferenceudgments

oncerning

equiprobable timuli r the modificationf an implicit orm re involved.

In

thepresent

tudy,

which s

a

continuation

f the

previous ne, we should

like to prove

thatthisaction

s also

possible

whenthe

majority orm o be

changed

s

explicit

r

quasi-physical.

Why

are

we

expecting

uch

an

effect?

he

presence

f

a norm an be dis-

Page 4: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 4/17

INFLUENCE OF A CONSISTENT

MINORITY

367

tinguished

n

the

spontaneous nanimity

f

those

who

share

t,

and

in the

expectancy hat a high probability esponsewill

occur

in

the

face

of a

stimulus r a determinatebject. The validity

of

judgments

nd

opinions

(Kelley, 1967) and the stability f relationswiththeenvironmentre guar-

anteedowing o thisnorm nly

f

thesetwo criteria

re

expected.

Now,

let us

suppose

that a

subgroup iverges

rom

his

customary

mode

of

response nd that

he

provides n

alternative

mode

of

response

o the

same

object,the same stimulus.

he

diversity

hich

replacesuniformity

n

the

group

s

a creator

f

uncertainty

nd of

conflict;

doubt

is

cast

upon

thehierarchy

f

responses

f each

person

r

of the

group

nd the

variability

is increased.By insisting n his answer, minority

ill

not

only engender

a conflict,ut will ntensifyheconflict,ecause t posesits own udgments

and

opinions s having

he same

value,

as

being equivalent

o

those of the

majority Worell 1967). Moreover, his nsistence

roves

that

takingone's

stand

s

not

casual and that the

subgroup

as

no

intention

f

conceding

r

submitting

o

the group.

This exerts tremendous

ressure

owards

cceptance

of

the

new and

surprisingesponse.

We must

lso

add that these

conflict elations ssume

a

particular

haracter

n the

case where he stimulus

s

physical.

The

reality

to be

judged

n these

circumstances

s not

individual, rbitrary: t is com-

mon,

n

principle

niversal.

No

matter

who,

faced

with such a reality, ne

is

expected o

react

in

the same way, and each

one imagines hat he is

reacting s

he is

supposed o react.

In

an experiment

ited

by

Asch

(1962), Sperling

emonstratedhat the

influence xerted n an

individual

s

much

greater

when

he

believes

n

the

existence

f

an

objective response,

han when

he

does

not

believe in it.

Thus,

the

fact that

a

physical stimulus

s involved

does not necessarily

work

against

the exertion f

influence y

a

minority;

n

the contrary t

mayfacilitatet. The majority as one singlemeans to reducethe tension,

to

ignore

he

judgment

f

the

minority:

hat

is to

transformhe conflict

of

response

nto

a

conflict

f attribution. his

means

that t

must

be

able

to

explain

the

difference

ot as

being producedby

the

properties

f

the

stimulus,

ut

as

being producedby

those who

perceive

t: an

anomalyof

vision,

lesser

udgment apacity.

This

is

possible

when

minority

s

an

isolated

ndividual

Moscovici 1969).

In

the event

hat

nothing

n

the situation

ermits

uch an

attributionnd

thatmembersf theminority,onstituting dyad,cannotbe distinguished

from

members

f the

majority y

such

traits,

henthe

atter

re

even

more

obligated

ither o

adopt

the

response

f

the

minority

r

to

reject t, i.e.,

to

polarize.

No other

means s

left

to

them

to

restore he

invariability f

response

n their

elation

with

the

external

world.

Page 5: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 5/17

368

SOCIOMETRY

With

these

presuppositions

n

mind,

n

order o demonstrate

he

nfluence

of a

minority pon a majoritywithina group,

we

have conceived

an

experiment

n

which:

(a) Response onflicts increased y theconsistencyf theminoritynd

by the

consensus

mong

ts members.

(b) Objectivity

s an

implicit

xigency

f

judgments.

(c) The

responses

f

the

majority

nd

minority

re

exclusive,

onstituting

an alternative,

ithout

ither

ne

just negating

he

other, s,

for

example,

if one were to say that two unequal

amounts

fdots were said to be equal.

(d) The

difference

n

judgment

annot

be accounted for

by

individual

qualities. Thus

it was

necessary

or

the

minorityo

be

composed

f more

than one person.)Otherwise heconflictn response ould be transformed

into a

conflict

f

attribution,

ermitting ifferenceso be explained

by

personal

ccentricities,

or

example.

(e)

The

judgment

f

the

majority

n

the

aboratory

s

identicalwiththat

ofany

random

ample

outsidethe

laboratory,

o

that

the

udgment

f the

minority

an

be

expected

o

be

directly

ounter o the

normal

xpectations

in society.

EXPERIMENTAL PROCEDURE AND RESULTS

FIRST

EXPERIMENT.

The

subjects

were

iberal

rts,

aw

and

social science

students.

iven

henature

f the

experimental

aterial

emale

ubjects

were

preferred

ecause

of

their

reater

nvolvement

n

evaluating

he

color

of

an

object.

The stimuli

used consisted

of slides with

two

different

ypes

of

filters

mounted

n

them:

(1) photo

filters

ermitting

he

passage

of

a

beam

of

light

of

the

dominantwave

length (X=483.5)

in

the

blue scale;

(2)

neutral

filterswhich

reduced

ight ntensity

n

certain

proportion.

In a set of six slides,three lides weremore uminous han three thers.

These

variations

n

light

ntensity

ere studied

n

orderto make the task

more

ealistic

nd less

boring.

heir effectn

this

experiment as

controlled.

Each

experimentalroup

consisted

f four

naive

subjects

and

two con-

federates.

nce the

subjects

were seated in

a

row

before the

screen on

which

were

to

be

projected

he

slides, they

were

told

that

this

would

be

an

experiment

n color

perception.

t

the

same

time

they

were

informed

that

they

would

be asked

to

judge

the color

and variation n

light ntensity

of a seriesof slides (a brief xplanation f themeaning f light ntensity

was

furnished).

efore

passing

a

judgment,

he

whole

group

was

asked to

take

a

Polack

test

ollectively,

n order o

check

the

participants'chromatic

sense.

This

test

had

a

twofold

bjective:

first,

o

eliminate

hose

subjectswho

Page 6: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 6/17

INFLUENCE OF

A

CONSISTENT

MINORITY

369

perchance

mighthave visual

abnormalities; econd,

to

emphasize

he

fact

that

everyone n the

group had normalvision,

so

that

the

confederates'

response

will not be

attributed o a difference

n

vision, .e.,

to

a

personal

factor xternal o the experimentalituation.

After

he collective

orrection f the result o

the test,

nd

after

having

ascertained

hateveryone ees

normally,

he

subjects

were nstructed hat

responses

mightbe given and how

the experiment ould

be conducted,

to wit

replying

loud

and

naming

a

simple

color

as well

as

estimating

the

light

ntensity

n

numerical erms

ranging

rom

for the dimmest

o

5

for the

brightest).

ubjects were also told

that

the

preliminary

rial

would

be

just

for

practice

n which

each

subject

would

only

make

a

light

intensityudgment.

The

real purposeof these

preliminary

rials

was

to enable the

subjects

to get acquainted

with

the

color

of the

stimulus

nd to

immunize

hem

n

McGuire's

(1964)

sense of

word

against

the

future

nslaught

f

the in-

structed

minority

hichdoes not

share the

norm.

During

these

preliminary

trials

the

confederates

nsweredat

random.

Following

these

trials,

the

series

of

six

differentlides was presented

ix

times,

the order

of the

slides

varying ystematically

rom

ne

series

to

the next.

Thus

these

were

36

trials,

ach

one

lasting

15

seconds,

eparated y approximately

seconds

of darkness.

n

each

trial the

two confederatesxerted

nfluence

y calling

the

color

green.

n

this

manner,

he

confederates

ere both

internally

consistent

rom ne

trial

to the next

with each

other,

ince

they gave all

the

time

the

same

response.

At

the

end of

the

experiment

he

subject

filled ut a

questionnaire on-

cerning

he

stimuli nd

the

other

members f the

group.

As

usual, the real

objectives

f the

experiment

ere

explainedbefore eaving

the

room.

Two

variations

were

introduced

egarding

he

seating

of

the

two con-

federatesnd thepresentationf the stimuli.

(1)

Confederate

ariation:

n

12

groups

he

confederates ere eated side

by

side and

gave

the first

nd

second

responses,while

in

the

20

other

groups

hey

were

separated,

nd

occupied

the

first

nd

fourth

laces. The

variation

n

the

seating

of the second

confederate

as

aimed

at

modifying

the

interpretation

f

his

behavior,

hat is

to

say,

to

make

him appear

more

independent

f the first

onfederate.

(2)

The

stimulus

ariation:

n

order

to test the

impactof

the commit-

mentto the first esponse nd to permit possiblechange,we modified

the mode

of

presentation

f stimuli. n

13

groups

which

included those

in which

the confederates ere seated in

position

1

and 4, the

continuity

of

the

sequence

of the stimuli

was

interrupted

y introducingwo one-

minute

auses

after

sequence

of 12

slides.

Page 7: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 7/17

370

SOCIOMETRY

The orderof response

f

the subjects

remained he same

from

ne trial

to the

next for

the

duration

f

the

experiment.

SECOND

XPERIMENT.

e

wondered

whether

he

subjectsexperienced

n

influence hich, ven if it did not result n a change n verbal response

during

the experiment, id have a lasting

effect n theirperception.

We

expected

shift

n

the

blue-green esignation

hreshold hich

would

reveal

a reaction

hat was

repressed uring

he social interaction. ertain

ubjects

didrefuse oadopt openly heminorityesponse, eeling ompelled

o

remain

loyal to thegeneralnorm,

ven when

they

themselves

egan

to doubt

its

validity.

Here one

might xpect

latent

ttraction

manifesting

tself

by

an

extension f

the

designation green

to

stimuli

n

a

zone

which

a

control

groupwouldcall blue.The oppositereaction extension f thenotionblue

to

stimuli

n

the

green zone)

would

be

the

result of

polarization.

The

first tage

of this

experiment

s identical

o the

preceding xperiment,

that

s to

say

that

the

minority

xerts ts influence

n

the

majority.

t

the

end of this

phase

the

experimenter

hanked

subjects telling

them that

another

esearcher ho was also

interested

n

vision

phenomena,

ould ike

to

solicit

their

participation

n

another

esearch

roject, ndependent

f

the

one

in

which

they

had

just participated.

e left theroom and the second

experimenter

ntered

mmediately

nd

repeated

his

request.

The latter

hav-

ing obtained heagreement f thesubjects eatedthem round table and

said

to

them hat t was an

experiment

elated o the

effect f the

exercise

about

the vision

phenomena.

He then

described

he

material,

solated the

subjectsby

means of

cardboard creens

nd instructs

hem

to

writedown

the

responses ndividually

n a sheet

of paper.

The

material

onsisted f

16 disks

n

the blue-greenone

of

Farnsworth 00-hue et perception est.

Three

disks from ach

end of the blue

and green scale were absolutely

unambiguous,ut

the

other

0

stimuli

might ppear

ambiguous. fter aving

made sure that the subjectsunderstoodhe instructions ell, the experi-

menter nnounced he

beginning

f

the

test. Each disk

was presented n

a neutral

background

or

period asting pproximately

seconds; it was

placed

in

the center

f

the table so that

it

would be

visible to everyone.

The

series

of

16 disks was

presented

0

times

n

the continuousmethod.

The order

of

presentation

as randomized.

Afterthe discriminationest

the

first

xperimentereturned,

he

subjects

filled

n

the

postexperimental

questionnaire

nd

the

experiment

nded

n

the

same

manner

s the previous

one.

Ten

groups participated

n

this

experiment.

THaRD

EXPERIMENT.

In this

experiment

hich

was

identical

o the first

one, only

we

diversified

he

consistency egree

of

the confederates.

n

this

case they

answered

4

times

green

and

12

times

blue, the dispersion

Page 8: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 8/17

INFLUENCE OF A CONSISTENT

MINORITY 371

of blue answersbeing randomized. leven

groups participated o

this

experiment.

The control roupwas the same for

he three xperiments.or thisgroup

the presentation f the stimuluswas continuous. he control ubjectsalso

took,of course, he discriminationest after

heinitialexperimentalhase.

In all we had 22 control ubjects, r

four

groupsof

6

subjects,with the

elimination f two subjectswho failed to give

the discriminationesponse

according o the instructions.

RESULTS

THE PERCEPTUAL TASK.

Green responsesresponseswhich xpress he

influence f minorityn the experimentalroups)constituted .42 per cent

of

the answersof the 128 naive subjects

in the two first xperiments.

There is no significant ifferenceetween he two series

of

groups

on the

perception ests nor on the postexperimental

uestionnaire. mongthe

22

subjects f the control roup, nly one gave

two greenresponses, epresent-

ing 0.25 per cent of the responses f the uninfluenced

ubjects.That means

that the latterperceived he stimulus s really

blue and that this norm s

firmly stablished ocially.

The

differenceetween

ontrol nd

experimental

ubjects

on the basis

of Mann

Whitney's

U

test

(Z=2.10)

turns ut to be

significantp=.019,

one-tailed est).

Other

data show

this

influence

s

well.

Subjects changed

their

response giving

4 or more

green

responses)

n

43.75

per

cent

of

the

groups.

The

percentage

f

individualswho

yielded

was

32

per cent.

Thus we have two categories

f

groups,

hose

n

which

no

subjects were

influenced nd those in which subjects were

influenced.n the latter, t

can be seen that

57

per

cent

of the

subjects

or

two subjectsper groupon

the

average gave

the same

response

s the

confederates.

8.70 per cent

greenresponseswereobtained n these groups.

Thus, the quantity

of

green responses

which we obtained

was not so

much the resultof isolated

ndividuals

who followed

he

confederate,s

the result f a

modification

f

udgment

within

he

group.

The

confederates'

seatingposition,

nd

the

type

of

introduction-continuous,

r

discontinuous

-of the

stimuli

did

not

have

any

differentiationffect.

Moreover,

we

have

noticed that even

though

no color

contrast ffect

existed,

he

subjects

were more similar o the confederates hen

ight

n-

tensitieswereweak than whentheywerestrong Z=3.37, p<.003, Mann-

Whitney

U

test).

This

agrees

withthe Bezold-Briicke

henomenon

oncern-

ing perception

f color

with

differentuminosities.

et, irrespective

f

the

luminosity

he

proportion

f

greenresponse

was

significantlyigher

n

the

experimentalroups

han

n the control

roups.

Page 9: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 9/17

372

SOCIOMETRY

In the third xperiment,here ne or several esponses f the confederates

were nconsistent,

e obtained

nly 1.25 per cent green esponses.

similar

proposal

was

obtained

n

groups ompletelynconsistent50 per

centblue-

50 per cent green responses f the confederates).Althoughwe have to

explore

more

systematicallyhe variation f inter- nd intra-subject

ncon-

sistency,

he resultswe

have just mentioned re suggestive

f

a

marked

influence f

the

behavior tyleof a minority.

THE DISCRIMINATION TEST. The question here concernswhether

he

subjects who changed their social response under the influence f the

consistentminoritylso changed heir erceptive ode. In addition,

we also

wanted to verify he hypothesis hat the subjectswho

did not

change

their ocial response, venin thegroupwherethemajoritywas not at all

influencedt this evel by the minority,t least changed theirperceptual

code.

The measurement

f

the

thresholdmakes it

possible

to

verify

this

hypothesis.

ur

calculations ear

on the threshold

alues,

whichwere ob-

tained by

a

graphic

method

n

the smoothed ut

curve

of

individual

e-

sponses.

We

retained

hree

values:

(1)

the

50

per

cent threshold

ndicating

the point

in

the

ordered equence

of stimuliwhere the

subjectgives

as

many

blue

as

green udgments; 2)

the ower hreshold

alue indicates

the

point where

he

subject gives

75

per

cent

green nd

25

per

cent

blue

judgments; nd (3)

the

upper

threshold

alue,

where the

subject gives

25 per cent green nd

75

per cent blue judgments.

o

study the

influence

of

the consistentminority,

e

subsequently

liminated

he

results f three

subjects

in

the

experimental roups

who

polarized.

Their

50

per

cent

threshold as

lower than

that of all the

control

group

thresholds.t

was

their ower threshold

alue,

which ndicates

generalization

f

the

notion

of

blue

in

the green

one.

Then, by comparing

he

50

per cent,

75

per cent,

and 25 per cent thresholds f the experimental roups 37 subjects) and

the control

roups 22 subjects)

we obtained

Table 1)

the

expected hift.

All

of

the

data reflect he effect f interaction etween

minority nd

majority

n

the

modificationf the

perceptual ode.

This

modification

ffects

TABLE

1

Shift

n the

Threshold

orPerception f

the

Color

Green

Control

Group ExperimentalGroup

P

(one-tailed

Threshold Mean

SD

Mean SD

t

level)

50

47.39

1.21 48.03

1.38 1.78

.038

75

46.16

1.42

46.85

1.54 1.68 .047

25

48.41 1.14

49.19

1.28 2.33

.01

Page 10: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 10/17

INFLUENCE OF A CONSISTENT MINORITY 373

more subjects than the change of verbal responses.This proposition

s

supportedby other data. On the one hand, if withinthe experimental

groups distinctions apparentbetween ubjects who sometimes dopted

the minority esponse nd subjects who never adopted the minority e-

sponses,no such differencemerges n the discriminationest for

the three

thresholds nderconsideration. n the contrary,t must be observed

hat

shift s even morepronounced orgroupswhere he majority id not change

than t is for thosewhere t changed, nd the Student's of 1.50 is close

to the 1.68 value, while t would be significantt

.10.1

We

had

made the

assumption hat

in

the

groups

where

therewas

no

change

n

social response, r where he green response ad been in

some

way repressed ne would observe greater umber f green udgments

in the discriminationest. One can see that this is indeed the case. The

difference etween the groups where the majoritydid not change and

where

he majority id change s significant

X2=

14.94, p<.002). We can

conclude hat the consistentminority as an even greater nfluence n the

perceptive ode of the subjectsthan on theirverbal response o the slides.

Of

coursethe experimentalechnique mployedwas not without ts faults.2

But the resultsobtained should be mentioned nly for the new research

line

it gives us.

THE POSTEXPERIMENTAL QUESTIONNAIRES.

The postexperimentalues-

tionnaires e had devised howed s that: (a)

The

divergence f opinion

r

response f the consistentminority onstrains he subjects to a cognitive

activity earingupon the stimulus. he perceptive hange

s not

produced

by

a

pure attraction owards he minority.b) The relativecertainty f

the

majority

s

probablyweakenedas a result of the confrontation ith

the minority, nd its problemwas to explain not why it followed he

minority,ut why t did not follow t.

(a) The CognitiveActivity f The Experimental roup.To begin with

we can

put

forward

hat

occasionally eeing green slides,

or

seeing green

in blue slides is not due to a simple acquiescence to the response of

the

minority.

1

Thomas nd Bistey 1964) report study sing he same stimuluss our study nd

theyfound hat subjectswho called the

stimulus green r mostly reen howed

significantlyreater eneralizationoward

he ongerwave ength han hosewho called t

blue or mostly lue. Our results re n the opposite irection.

2

Using he sametest,Brown ndLenneberg1958) showed hat there s a relation-

ship between olor-naming

nd color

recognition

hich s a function

f

stimulus

x-

posure-time.hus

we should have varied

the

exposure ime.Neverthelessincewe

dealtwithhighly

odable

olors,

e should e able to recover hem

rom heir ame.But

in general ur study

s in

agreement

ith heirswhich

hows hat

nconsistency

ithin

the group orrespondso inconsistencynd

hesitationn the

ndividual.

Page 11: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 11/17

374

SOCIOMETRY

Having

raised the

question:

To

what

extent

s it

possible

for

these

lides

to be perceived

s green we ascertained

hat

subjects

n the

experimental

groups did not

accept this possibility

n

a more

significant egree

than

subjects n the controlgroups.On the otherhand, however, ubjects n

the

former

roups

did

prove

more inclined o

accept

the

green

response

than

subjects

n the latter

groups

t 2.64, p<.008). Thus,

we can

infer

that

the desire o reach

n

agreement

ith he

minority

ed to an

inclination

to

see what

the

latter

were

seeing,

to make an

effort o look for

green

in the blue stimuli.With this n

mind we asked the

subjects: how many

differentuances

of color did you

distinguish? ubjects

n the

experimental

groups erceivedmore hantwo

nuances,while ubjects

n

the control

roups

saw at mostone or two (Z 2.12, p<.0342). A differentiationan also be

made betweenubjectswithin he

experimentalroups.

ubjectswho yielded

to the

minority ay

more

nuances than

those who did

not

yield

to

the

minority.Z=2.79, p<.005).

Moreover,whether

heydid or

did

notyield

to the minorityubjects

n

groups

n

which

change

n

response

ccurred

perceived

more hades than those

n

groupswhere

he majoritymaintained

its position, nd

always responded lue (Z= 1.78,

p<.076). Using an

ap-

propriate uestion,

we then sked

subjects o specifyhese hadesby naming

the

colors

which

composed

them.

No

matter

what these

shades

were

or

how many werecited, for purposes

of this analysis

we retained nly the

highestpercentage

f

green

found

on

the

response heet, using it

as an

index of the extreme

imit of a subject's attempt

o find this color.

All

subjects

n the experimental roupsdistinguished

more green than

those

in

the control

groups Z=2.99, p<.003). Of course,

n the experimental

groups, ubjectswho yielded to

the minority aw

more than 30 per cent

(Z-

4.92,

p<.001).

Everything

ends

to point to

the fact that members

of the

majority

made an

effort

o

take into account the viewpoint

f the

minority,o verifyhe objectivebasis of its judgment. t no timedid they

remain

passive,

nor were theycontent lindly o

accept or reject a

norm

opposed to

theirown.

The

effect

f

this

was probably

the modification,

as we

saw,

of their

own

perception

r their

definition f green and of

blue.

(b) Perception

f

the

Consistent

inority.

aive

subjects,who constituted

the

majority

n

the

experimental

roups

were more

nclined

to

see green

in

the blue slides

than the control

ubjects (and

actually did see more

green).The psychologicalroblemwhich heyhad to solve was the follow-

ing: why, althoughhaving agreed

that

the

minority's nswer was not

without

oundation,

id

they

not

yield

to

it,

since

a

physical timuluswas

involved?The

only possible explanation

or

such a

contradiction as

the

Page 12: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 12/17

INFLUENCE

OF

A

CONSISTENT MINORITY

375

assertion hat they were less certainthan the minority.

hus

while

they

were nterested

y what was proposed o them, hey

considered

hemselves

to be more competent han

the minority,ince they represented

ormal

perception-thereforeheyhad the right o yield or not to yield.Needless

to say, these trends an be

accountedfor n otherways.

In

spite

of the

resultsof

the

Polack Test,

subjects did not believe that a person

who

always perceived these

slides as green could

have a

very good

color

perception.

ven

if

he had

good vision,

his

competency

n the

area

of

colormustbe inferioro that

of the majority

f

people. On the other

hand

the consistent

atureof theminority esponse

n

the

face of

the

different

judgments

mitted

y the

majority, uppliedgreat self-assurance.

ithout

coming o any definiteonclusion,t can nonetheless e seen thatthefirst

interpretationpplies

to the two series of

predictions

onsidered

ogether,

while

the second concerns ach series

eparately.

Now let

us

examine

he

results

btained

more

n

detail.

n the first

wo

questions ubjects

were

sked to

judge

each

of

the

persons

who

participated

in the experiment,ncluding hemselves,n a 10-point cale (from

good

to

bad),

as

to

their

capacity

first

o discriminatentensities

nd

second

to

perceive olors.

A

comparison

f

the

grades

which

ubjects gave

to them-

selves,

onfederates

nd

other

ubjects

for

color perception

s

very

nstruc-

tive.On

the

whole, ubjects

onsidered

hat

the

confederates'

olorperception

was

not

as

good

as

theirs,

oth

in

the

groups

where

the

confederates

ere

seated nextto

each other t=9.98, p<.001), and

in

the

groups

where

hey

were

separated t 7.02,

p<.001). They also

considered

hat confederates

did

not

perceive olorsas well

the

othermembers f the group (t

10.83,

p<.001). Nevertheless,t

was felt hat the secondconfederate ad a better

color

perception han the first onfederateZz=2.04,

p

.04,

Mann-Whitney

U

test).

Thus

the

members f the majority udged themselves

more com-

petent han the minority,nd they xperiencedittle nxiety egarding heir

perceptive apacity.

What

about

certainty?

n

their

postexperimentaluestionnaireubjects

had to

classify the personswho participated

n

the experiment,

ccording

to

whether

heywere moreor less sure of

their esponses. ubjects

udged

confederates

o

be more sureof theirresponses han they were

(t-5.02,

p<.07)

and thanother

members f the

group t=4.42,

P<.07).

A

difference

revealed tself also in the perception f the two confederates.

he con-

federate eated in the first ositionwas judged as being more sureof his

response

han

the second

confederate,oth in the groupswheretheywere

seated next to each other

t=2.54, p<0.7) and in the groupswherethey

were

separated (t 3.22, p<.07). These evaluationswere shared

by all

Page 13: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 13/17

376 SOCIOMETRY

subjects,whether heywere amongthosewho responded ike the consistent

minority,r whether heywere n the groupswhere he majority esisted ll

influence.hreetrends learly merge rom hese esults: a) subjects udged

themselvesmore competent nd less certain han confederates; b) judg-

ments f competence nd of certitude f confederates ad an inverse ela-

tion; (c) the confederaten the second positionwas perceiveddifferently

from he one in the first osition nd as being closer to other subjects.

These trendscorroborate bservationsmade in other experiments. hus,

Brehm nd Lipsher 1958) provedthat perceived rustworthinessould be

greaterwhen the communicatorook an extreme osition n either ide of

the

issue, than when he took a moderate osition.More recently, isinger

and Mills (1968) studied he effect f the discrepancy f the communicator

positionupon

his

sincerity

nd

competence. hey proved

that a

communi-

cator on

the

opposite

ide

will

be

perceived

s

more

ncompetent

nd

more

sincere

n

comparison

ith

a

communicator ho

is

opposed but more

mod-

erate.

These

experimentsuggest hat the response

f

an individual

r an

extreme

ubgroup

as

moreweight.But what

interests s here s the

fact

that obtaining he same results s ours,theyofferndirect upport n favor

of the

view

that

consistency,specially

f a

minority

ith

a

norm

opposed

to thenormof themajority,s at the same timean indexof extremism.

Now,

this

extremism,

o

the

extentthat it shows itself

uncompromising,

engenders n anxiety

inked

to

the

disagreement,

nd

places

the others

n

a

situationwherethey

must

either

oncede

or

polarize

n

order to

reduce

this

disagreement

nd

diminish he

anxiety.

As

nothing ermits

hem to

polarize,

hen

n certain

groups, ubjectsyielded.

The

trendsdiscovered

lso

enlightened

s

about

the

role

of

the second

confederate.

n

a

sense,

he does

not

contribute

ny supplementary eight

to

the

response

f

the

innovator,

he first

onfederate.

We make

the

hy-

pothesis hathis behavior ervesas an exampleto the other ubjects; he

demonstrateshat

omeone s

capable

of

choosing

he

minorityesponse hat

there

s

a

choice

possible

betweenthe

two

alternatives nd

to a certain

extent, ustifies

hem.

n

short,

f

the

effect f

the first

onfederates an

influence ffect,

he

effect

f the second

would

be what economists all

a

demonstrationffect.

n

any

case the

minority's

nfluence

annotbe at-

tributed

o a

possible eadership ecognized y

the

group. Questioned s

to

which

persons

n

the

group they

would

like

to

find themselves n a

similar ituationwith, ubjectsdid not choose confederates orefrequently

than

any

other

member f

the

group.Likewise,

when

asked:

Who would

you

like to see lead the discussion

about

the

experiment)

n

the

group?

a

slight,nonsignificant

rendcan be observed o

choose

confederatesess

than

other

naive

subjects.

Page 14: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 14/17

INFLUENCE OF

A

CONSISTENT

MINORITY 377

DISCUSSION AND CONCLUSION

The

experiment hichwe have just described

hows,

t least

as far

as

female subjects are concerned, hat by being consistent minority

s

capable of influencing majority t the level of

verbal and perceptual

responses. ut this factmustbe examinedmoreclosely.

GENERALITY OF THE BEHAVIORAL STYLE AS A SOURCE

OF INFLUENCE.

We

have at the beginning f this articleput forward

he

idea

that the

con-

sistency f the behavior s a source of

influence

hen

a minority

s con-

cerned nd

when an innovation rocess

s

involved.

And

it

clearly ppears

that conformitys an effect f consistencynd not

of dependence

owards

the

majority f the group. To substantiate his

conclusion,

we

will

limit

ourselves o Asch's experiments. e know hat n these xperiments group-

majority an induce a single ndividual o give answersgoing counter o

perceptual vidence.The conditions equiredfor this

effect

o occur

are

the

usage

of

a nonambiguoustimulus, he need to

respondpublicly,

nd

the

presence

f a

unanimous

majority.

This

majority, ccording o Asch

(1962:497) gives

rise

to a propensity o adopt the

erroneous conformist

responses

f

the group.Our interpretations, of

course,different,ut first

let

us look

to

the data and theirmeaning.We can

consider

hat unanimity

in a group orrespondso inter-individualonsistency,o consistency hich

results

from oincidence nd

identity

f

response

f

several

subjects

to

a

given stimulus.

At

the

same

time,

the

sequence

of erroneous

esponses,

the identity

f

responses

f

each confederate

hrough

series of

stimuli,

expresses nternal,

ntra-individual

onsistency.

What

do

we see

when

we

examine

Asch's

results?

We

see that

a

unanimous

majority

from

wo to

sixteen

onfederates

rovoked

he

acceptence

f

erroneous

esponses or

one

third

32 per cent)

of

the

responses

f the

naive

subjects.

The

increase

in thenumber f confederateso more than threehas thereforeo effect

on

the

frequency

f

these

esponses. hus,

there

s no direct

elation

etween

the

magnitude

f

this

social

pressure

nd

conformity.ow, onlyone single

confederate

n

a

group

made

up

of seven or

eight

personshas to break

the

unanimity y giving

orrect

nswers

for

the

number

f

conformiste-

sponses

to

drop

to

10.4

per

cent

or 5.5

per cent.

Thus, a group of three

unanimous

ersons

s more

nfluentialhan

group

f

eightnon-ununanimous

persons.

his

is tantamounto

saying hat t

is

the

nter-personalonsistency

of,

ratherthan

the

strength

f

social pressurewhich s more important,

and

comes closest

to

accounting

or the

variation

n

the

rate of influence.

Asch's

(1955)

and Allen and Levine's

(1968)

experiments ive much

weight

o this nnovation.

hey thought

hat f social

support

was

important

in

order

to reduce

conformist

onstraint,

he

dissenter

ught

to

give the

Page 15: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 15/17

378

SOCIOMETRY

response

which the

subjects privately onsidered

o

be correct.On the

contrary,

n thecase of unanimity

heregroup

consistency as the

critical

variable,

a

dissenter's

isagreement ith the

group,

whether r not his

responses erecorrect nd inagreement ith he subject'sprivate udgment,

was

sufficient

o

decrease conformity.

he resultsof the two experiments

show

that t is lack

of

unanimous onsensuswhich s the

decisivefactor.

What

s

the ffect

f ntra-individual

onsistency

vertime-of the dentical

repetition

f subjects

responses o a series

of stimuli?

As we know,Asch

used

two

types f trials:

neutral rials n

which he confederates

esponded

in a correct

manner,

nd critical trials

n which

the confederatese-

sponded

n

an erroneousmanner.

iachronistically,

group ppeared

all

themoreconsistent ith tselfwhen therewere more critical trialsthan

neutral ones. Asch

(1956) varied

the proportionf the

neutral

rials n

relation o the critical

rials (1/6,

1/2, 1/1, 4/1) and

although

he dif-

ferenceswere not significant,

decrease

n

the

percentage

f conformist

responses

was observed 50 per

cent,

36.8

per

cent,38.6 per cent,

26.2 per

cent)

as

the majority

became less

coherent

n

time.

scoe and Williams

(1963)

obtained imilar

esults.On the whole,

onsidering

he information

we

have

at hand

today,

we can

say

that it is the behavioral tyle

of a

majority

r a minoritynd not the pure amount

of social

pressurewhich

is revealed o be at the origin f influencexerted.

CHANGE

OF

VERBAL

AND PERCEPTUAL RESPONSES.

We

have seenthatthe

alteration

f the

answer,

while not

negligible

t the conscious

ocial

level,

is

more

marked,

t

the atent ndividual evel.

Our

present

tateof

knowledge

does

not

enable

us to

ascertainwhether

t is of

a

perceptive

r

of a verbal

nature Goldiamond,

958).

However,given

thatmostof

the

experiments

in this

field

Tajfel,

1969)

with the notable

exception

f

Flament

1958)

report

nfluence t

the verbal level

and not at the

level of

perception,

theresultswe have obtained re all the more remarkable. hey obligeus

to

distinguish

etween

change

n

response

nd a

change

n

code,

between

influence

t

the

response

evel and

influence

t the

code level. n

this

sense,

we

have

the

right

o

say

that the consistent

minority,

n one

experiment,

provoked

real

modification

n

the normof the

majority, nd

not only

in its

response.

If

this

phenomenon

s rare

in the

laboratory,

t is not in

political

ife.

Thus,

a

politicalparty

often

dopts

the

deas or the

vocabulary

f

another

party r social movement.et citizens ontinue o vote for hissameparty,

to

respond

to this

party's slogans.

For

example,

n France the

Gaullist

government

n

framing

ts

own education

program,

dopted part

of

the

rhetoric

nd

the

program

roposedby

students

nd workers

n

May

1968.

Nevertheless,

hena Frenchman

otes for

the Gaullist

party

he

believes

Page 16: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 16/17

INFLUENCE OF A CONSISTENT

MINORITY 379

that he

is

responding

o

the

same

political

body

and in the

same

manner s he did in the past, althoughboth it and its representatives

have changedtheiropinions n very specific uestions.

ndeed, it is con-

ceivable that minoritiesre more capable of changing hemajority's ode

than ts social response,while the majority

would have more nfluence

n

the

ndividual's erbal

response

han on his intellectual

r

perceptive

ode.

This is an historical eality.Great nnovators ave

succeeded

n

imposing

their deas, theirdiscoveries, ithout ecessarily eceiving irect ecognition

for heir nfluence.orexample,manypsychologistsave assimilated otions

elaborated y psychoanalysis,ll the while refusingo recognize he value

of

psychoanalysis.

Thus, ifwe reallywantto understand he processof social influence,t

is not enough o studymore arefully he role ofminoritiesnd of innova-

tion. We

must

begin

to

explore

more subtle

mechanisms f influence han

those which

are at

work

in direct

and visible

acceptance

of

normsand

judgments roposed.

REFERENCES

Allen, V. L.

1965 Situational factors n conformity. dvances in ExperimentalSocial Psy-

chology 2:133-175.

Allen,V. L.

and

J.

M. Levine

1968

Social

support,

dissent

and

conformity. ociometry

31(June):138-149.

Asch, S. E.

1955 Opinions and social pressure. ScientificAmerican 193

(November) 31-35.

1962 Social Psychology.

Englewood Cliffs: Prentice-Hall.

Brehm, J. W.

and

D. Lipsher

1959

Communicator-communicatee

iscrepancy

and

perceived

communicator

'trustworthiness'.

ournal

of

Personality

27

June)

352-361.

Brown, R. W. and E. H. Lenneberg

1958

Studies

in

linguisticrelativity. Pp. 9-18 in Maccoby,

Newcomb, and

Hartly (eds.), Readings

in

Social

Psychology,

New York:

Holt, Rinehart

and

Winston.

Eisinger,R. and J.

Mills

1968

Perception

of the

sincerity

nd

competence

f a communicator s a

function

of

the extremity

f his

position. Journal

of

Experimental

ocial

Psychology

4(April):224-232.

Faucheux,

C.

and

S.

Moscovid

1967

Le

style

de

comportement

'une minorit6

t

son

influence

ur

les reponses

d'une

majorit6.

Bulletin

du

Centre

d'Etudes et Recherches

Psychologiques

16(Octobre-Decembre):337-360.

Flament, C.

1958

Influence Sociale

et

Perception.

Annie

Psychologique58(Fasdcule 2):378-

400.

Page 17: Moscovici_2

7/23/2019 Moscovici_2

http://slidepdf.com/reader/full/moscovici2 17/17

380

SOCIOMETRY

GoIldimond,

. and

L.

F.

Malpasm

1958 Locus of hypnoticallynduced hangesn

colorresponses. ournal f the

Optical ociety

f America 1

October)

1117-1121.

Iscoe, . and M. S. Williams

1963 Experimental ariables ffectinghe conformity

ehaviorof children.

Journal fPersonality1 June): 34-246.

Kelley,H. H.

1967 Attributionheoryn socialpsychology.

p. 192-241 n D. Levine ed.),

Nebraska ymposium

n

Motivation. incoln:

UniversityfNebraska ress.

McGuire,W.

1964 Inducing esistanceo persuasion. dvances

n Experimentalocial Psy-

chology : 191-229.

Moscovici, .

1969 Behavioral tyle as a Source of Social Influence.

ymposium

n

Social

Influence,

Xth

International

ongres

f Psychology,

ondon.

Moscovici, . and C. Faucheux

1969 Social Influence, onformityias and the Study of Active Minorities.

Center

f

Advanced tudy

n

Behavioral ciences traffordMimeo).

Tajfel,H.

1969

Social

and cultural actors

n

perception.

p. 315-394 n Lindzey nd

Aronsoneds.)

The Handbook

of

social

psychology,

ol. III (2nd ed.),

Reading:

Addison

Wesley.

Thomas, . R.

and

G. Bistey

1964

Stimulus

eneralization

s a

function f

the

number

nd

range

of

gen-

eralization

est

timuli. ournal

f

Experimental

sychology8(December):

599-602.

Worell, .

1967 Some

ramificationsf

exposure

o conflict.

rogress

n

Experimental

Personality

esearch

:91-125.