Upload
vicentiu
View
217
Download
0
Embed Size (px)
Citation preview
8/10/2019 Blumer Public Opinion and PO Polling
1/9
Public Opinion and Public Opinion PollingAuthor(s): Herbert BlumerSource: American Sociological Review, Vol. 13, No. 5 (Oct., 1948), pp. 542-549Published by: American Sociological AssociationStable URL: http://www.jstor.org/stable/2087146
Accessed: 24/10/2010 16:46
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 Associationis collaborating with JSTOR to digitize, preserve and extend access to
American Sociological Review.
http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=asahttp://www.jstor.org/stable/2087146?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=asahttp://www.jstor.org/action/showPublisher?publisherCode=asahttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2087146?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=asa8/10/2019 Blumer Public Opinion and PO Polling
2/9
PUBLIC
OPINION
AND PUBLIC
OPINION
POLLING*
HERBERT
BLUMER
University
f
Chicago
T
HIS paper presents
ome observations
on
public
opinion nd on public
opin-
ion polling s currently
erformed.t
is hopedthat hese bservations
illprovoke
the discussion
orwhich, understand,
his
meeting asbeenarranged.
he observations
are not along the line of
what seems
to be
the chiefpreoccupationf students f
pub-
lic opinion
polling, o wit, the internal
m-
provement
f theirtechnique.nstead,
the
-observationsre designed o invite ttention
to whether
ublic opinionpolling actually
deals
with
ublic pinion.
The
first bservationswhich I wish
to
make are
in the natureof a prelude.
They
come
from mere ogical crutiny f
public
opinion olling
s
an
alleged
form f scien-
tific nvestigation.
What
I note
is the
in-
abilityof public opinion
polling
to
isolate
public opinion
as an abstract r
generic
conceptwhich ouldthereby ecome hefo-
cal
point
for
the
formation
f
a
system
of
propositions.
t
would seem needless to
point
out that
n an avowed cientific
nter-
prise seeking o study
a class of empirical
items
nd to
develop
seriesof generaliza-
tions
bout that lass
it is
necessary
o iden-
tify
he
class.
Such dentificationnables
dis-
crimination
etween he nstances
which all
within he class
and those
which
do
not.
n
this
manner,
he
generic
haracter
f
the ob-
ject of studybecomesdelineated.When the
generic bject
of
study
s distinguishable,
t
becomes
ossible
o focus
tudy
n that
ob-
ject
and
thus
to learn
progressively
ore
about that
object.
n
this
way
the
ground
s
prepared
for cumulative eneralizations
r
propositions
elative
o thegeneric bject of
investigation.
As far s
I
can
judge,
hecurrent
tudy f
public opinion
y polling gnores
he
simple
logicalpointwhichhas justbeen made. This
*Paper read before
the annual meeting
of the
American Sociological
Society
held in New
York
City,
ecember
8-30, I947.
can
be seen
throughhree bservations.
irst,
theres
no
effort,eemingly,
o
try o
dentify
or to
isolate
public
opinion
s an object;
we
arenot
given ny
criteria
which
haracterize
or
distinguish
ublic
opinion nd
thuswe
are
not able
to say
that a
given
empirical
n-
stancefalls
within he
class
of public
opinion
and some
other
mpirical
nstance
alls
out-
side
of the
class
ofpublic
opinion.
econd,
there s
an absence,
s far s I can
determine,
of using specific tudiesto test a general
proposition
bout
public
opinion;
this
sug-
gests
that
the students
re
not studying
generic
bject.
This suggestion
s
supported
by
thethird
bservation-a
paucity,
f not
a
complete
bsence,
of
generalizations
bout
public
opinion
despite
the
voluminous
amount
f polling
tudies
fpublic
opinion.
It
mustbe
concluded,
n
my udgment,
hat
current ublic
opinion
polling
has not
suc-
ceeded n solatingublic pinions a generic
object
of
study.
It
maybe
argued
that
the
isolation
f
a
generic
bject,
specially
n the realm
f
hu-
man
behavior,
s a goal
rather
han n
initial
point
of
departure-and
that
consequently
the-present
nability
o
identify ublic
opin-
ion
as a
generic
bject
s not damning
o
cur-
rentpublic
opinion
polling.
This should
be
admitted.
owever,
what mpresses
me s the
apparent
bsence
of
effort
r
sincere
nterest
on thepartofstudentsfpublic pinion oll-
ing
to move
in the
direction f
identifying
theobject
which
hey
re
supposedly
eek-
ingto
study,
o record,
ndto
measure.
be-
lieve
it
is
fair to
say that
those
trying
o
study public
opinion
by
polling
are
so
wedded
o their
echnique
nd
so preoccupied
with he mprovement
f
their
echnique
hat
they
hunt
side the
vital
question
f
whether
their echnique
s
suited
o the
tudy
f
what
theyare ostensiblyeeking o study.Their
work s
largely
merely
making
pplication
f
their echnique.
hey
are not
concerned ith
independent
nalysis
f the nature
f
public
opinion
n
order
to
judge
whether he ap-
542
8/10/2019 Blumer Public Opinion and PO Polling
3/9
PUBLIC OPINION
AND
PUBLIC
OPINION
POLLING
543
plication
f their echnique its hat
nature.
A few
words re in orderhere on an
ap-
proach
that consciously
xcuses tselffrom
any
considerationf sucha problem.
refer
to the narrowoperationalist ositionthat
public
opinion onsists fwhat public opin-
ion
polls poll. Here,curiously, he
findings
resulting romn
operation,
r
use of an in-
strument,
re
regarded s constitutinghe
object
of
study
nstead
of being some con-
tributoryddition o
knowledge f theobject
of
study.
he
operation
eases to be
a
guided
procedure n behalf f
an object of nquiry;
instead he
operation etermines
ntrinsically
its
own
objective. do
not care to consider
here theprofoundogicaland psychological
difficultieshat attend
he effort o develop
systematic
nowledge hrough procedure
which
s not a form f directed
nquiry.
All
that
wish
to note s that he
results f
nar-
row operationalism, s
above specified,
merely
eave or raise hequestion fwhat
he
resultsmean.Not having
conceptual oint
of
reference
he results remerely isparate
findings.
t
is
logically ossible,
f
course,
o
use suchfindingsodevelop conceptualiza-
tion.
fail
to
see anythingeingdone nthis
direction
y those
who
subscribe
o the nar-
rowoperationalist
osition n theuse ofpub-
lic
opinion olls.What s
logically npardon-
able
on
the
part
of
thosewho
takethe nar-
row operationalist
osition s for themto
holdeitherwittinglyr
unwittinglyhat heir
investigationsre a
studyof publicopinion
as this
erm
s conceived
n
our
ordinary
is-
course.Having rejected s unnecessary he
task of
characterizing
he
object
of
inquiry
for
he
purpose
f
eeing
whether
he
nquiry
is suited o the
object
of
nquiry,
t
s gratui-
tous and
unwarranted
o
presume
hat
after
all
the
nquiry
s
a
study
f
the
object
which
one refuses
o
characterize.
uch a
form f
trying
o
eat
one's
cake and have t
too
needs
no further
omment.
The
foregoingeries
f
ogical
bservations
has
been made
merely
o
stress
he absence
of considerationfa generic bjectbythose
engaged
n
public opinion
polling. Appar-
ently,
t
is
by
virtue
f this
absenceof con-
sideration
hat
they
re
obtuseto
the func-
tional
nature
f
publicopinion
n
our
society
and to questions
fwhether
heir echnique
is suited
to this
functional
ature.
n this
paper
intend
o udge
the
uitability
fpub-
lic opinion
polling
s a
means
of studying
publicopinion.This shallbe donefrom he
standpoint
f what
we know
ofpublic
opin-
ion
n
our ociety.
Admittedly,
e
do not
know
greatdeal
about
public opinion.
However,
we know
something.We
knowenough
about
public
opinion
frommpirical
bservations
o form
a
few
easonably
eliableudgments
bout
ts
nature nd
modeof functioning.
n addition,
we can make
somereasonably
ecure
nfer-
ences bout
thestructure
nd functioning
f
our society
and
about collectivebehavior
within ur
society.
This combined
ody
of
knowledge
erivedpartly
fromdirect
em-
pirical
observation
nd
partly
from
eason-
able inference
an serve
appropriately
s
means
of udging
nd assessing
urrent
ub-
lic
opinion
polling
s
a device
for tudying
public
opinion.
Indeed,the
features
hat
I
wish
to note
about
public
opinion
nd
its
setting
re so
obvious and commonplace hat I almost
blush
to
call themto
the attention
f
this
audience.
would
not do so
were
t
not
pain-
fully
lear
that he tudents
f
current
ublic
opinion
olling
gnore
hem ither
wittingly
or unwittingly
n
their
whole
research
ro-
cedure.
shall
indicate
by
number
he fea-
tures
o
be
noted.
i.) Public
opinion
mustobviously
e
rec-
ognized
s
having
ts
setting
n a
society
nd
as being function fthat ociety
n
opera-
tion.
This
means, atently,
hat
publicopin-
ion
gets
s
form
rom
he
ocial
framework
n
which
t
moves,
nd
from
he social
proc-
esses
n
play
n
thatframework;
lso
that
he
function
nd role
ofpublicopinion
s
deter-
minedby
thepart
t
plays
in the
operation
of
the
ociety.
f
public
opinion
s
to
be
stud-
ied
n
any
realistic
ense ts
depiction
must
e
faithful
o its
empirical
haracter.
do
not
wishto
be redundant
ut
I find tnecessary
to say thattheempiricalharacterfpublic
opinion
s
represented
y
ts
composition
nd
manner ffunctioning
s a
part
of
a
society
in
operation.
2.)
As every
sociologist
ught to
know
8/10/2019 Blumer Public Opinion and PO Polling
4/9
544
AMERICAN
SOCIOLOGICAL
REVIEW
and as every
ntelligent
ayman
doesknow,
society as an
organization.
t is
nota mere
aggregation
f disparate
ndividuals.
A hu-
man society
s composed
f diverse
indsof
functionalroups.n ourAmericanocietyl-
lustrativenstances
f functional
roups re
a corporation,
trade
association,
labor
union,
n
ethnic roup,
farmers'
rganiza-
tion.
To a
major
extent
ur total collective
life
s made
up of
the actions
and acts of
such
groups.
These groups re
oriented
n
differentirections
ecause
of special
in-
terests. hesegroups
iffer
n
terms
f their
strategic
osition
n
the ociety
nd in terms
of opportunities
o act. Accordingly,hey
differn termsof prestige nd power.As
functional
roups,
hat is
to say as
groups
acting ndividually
n
some
corporate
r
uni-
tary
sense,
such
groups
necessarily
ave
to
have some
organization-some
eadership,
somepolicy
makers,
ome individuals
who
speak
on behalf
f
the
group,
nd some
ndi-
viduals
who take
the nitiative
n
acting
on
behalf
f the
group.
3.)
Such functional roups,
when they
act,have to act throughhe channelswhich
are
available
n the
ociety.
f
the
fate f
the
proposed
ctsdepends
nthedecisions
f n-
dividuals r groups
who re
ocated t strate-
gicpoints
n the
channels f
action,
hen
n-
fluence
and pressure
s brought
to
bear
directly
r indirectly
n suchindividuals
r
groups
whomake
thedecisions.
take
t that
this
realistic
eature
f the
operation
f our
American
ociety
requires
ittle
explication.
If an actionembodyinghe
interests f
a
functional roup
uch as
a farmers' rgani-
ization
depends
for ts realization n
deci-
sions
of
Congressmen
r
a
bureau
or
a set of
administrators,
hen fforts
n behalf f
that
action
will
seek to
influence
uchCongress-
men,
bureau,
r administrators.
ince n ev-
ery
ociety
o
some
degree,
nd
n
our
Ameri-
can society
o a
large
degree,
here
re
indi-
viduals,
committees,boards,
legislators,
administrators,
nd executives
who
have
to
makethe decisions ffectingheoutcome f
the
actions
of
functional
roups,
uch
key
people
become
the
object
of direct nd
in-
direct
nfluence r
pressure.
4.
The key
individuals eferred
o
who
have to make thecrucialdecisions re
almost
inevitably onfronted
ith the necessity
f
assessing
he
various
nfluences,laims,
de-
mands, urgings,
and pressuresthat are
broughtobear on them.nsofar s they re
responsivend responsible
hey re boundto
make such an assessment
n
the process
of
arriving
t theirdecisions.
Here I want
to
make the trite emark hat
n
making
heir
assessments
hese
key
individuals ake
into
account
what
they udge
to be
worthy
f be-
ing taken nto ccount.
5.) The above
pointsgive a crudebut es-
sentially
ealistic
icture
fcertain
mportant
ways
n which
ur
society perates.
he
fifth
feature wishtonote s thatpublicopinion
is formed nd expressed
n
large
measure
through
hese
ways
of
societal
operation.
This point requires
little laboration.
he
formation
f
public
opinion
ccurs s a
func-
tion f
a
society
n
operation.
state
he
mat-
ter n
thatway
to
stress hat
the
formation
of
public opinion
oes
not
occur
through
n
interaction
f
disparate
individuals who
share
qually
n
the
process.
nstead
the for-
mationof publicopinionreflects he func-
tional
composition
nd
organization
f so-
ciety. he
formationf
public pinion
ccurs
in large measure hrough
he
interaction f
groups. mean nothing soteric y this
ast
remark. merely
efer o the commonccur-
rence fthe eaders
r officialsf
a
functional
group aking standon behalf f the
group
with eference
o
an issue
and
voicing
xplic-
itlyorimplicitly
his
tand
on behalf f the
group. Much of the interaction hrough
which ublic pinion
s
formed
s throughhe
clash
of these
group
views
nd
positions.
n
nosensedoes such group
iew
mply
hat t
is held
n
equal
manner
nd in
equal degree
by
all of themembers
f
the
group.
Many of
the members
f
the groupmay subscribe o
the
view
without
nderstandingt,manymay
be indifferentbout
t, manymay share
the
view
only
n
part,
and
many
may
actually
not
share he
view
but
still
not
rebel gainst
therepresentativesfthegroupwhoexpress
the
view.
Neverthelesshe
view,
s
indicated,
may be introduced
ntothe forum
f
discus-
sion
as the
view of
the
group
and
may
be
reacted o
as such.
To
bring
ut this
point
n
8/10/2019 Blumer Public Opinion and PO Polling
5/9
PUBLIC
OPINION AND
PUBLIC
OPINION
POLLING
545
another
way,
one need
merely
note that
n
the
more
outstanding
xpressions
f view
on
an
issue, he
ndividuals lmost
lways
peak
either
xplicitly
r
implicitlys
representa-
tivesofgroups. wouldrepeatthat n any
realistic
sense
the
diversified
nteraction
which
gives
rise
to
publicopinion s
in
large
measure
etween unctional
roups
nd not
merely
etween
isparatendividuals.
I
think
hat t is also
very lear
that
n
the
process
of
formingublic
opinion, ndivid-
uals arenot
alike
n
influence
orare
groups
that
are equal
numerically
n
membership
alike
n
nfluence.
his
s so
evident s
not to
require
laboration. t
is enough
merely o
point out thatdifferencesn
prestige,
osi-
tion, nd influencehat
characterize
roups
and
individuals
n
the
functional
rganiza-
tions
fa
societyre
brought
nto
play
n
the
formationf
publicopinion.
The
picture f a
series f
groups nd indi-
viduals
of
significantly
ifferent
nfluence
n-
teracting
n
the
formation
fpublic
opinion
holds
true
quallywell with
referenceo
the
expression
f
public
opinion.By
expression
ofpublicopinion meanbringinghepublic
opinion o bear on those
who
have to
act
in
response opublic
pinion.
his
expressions
not
in
the form f a
paradeor
arrayof the
views of
disparate
ndividuals,
n
an
open
forum.
Wherethe
views
are
voiced
n
open
forum
hey
re
likely,
s has been
ndicated,
to be in
one
way
or
another
he
xpression f
group
views.
But in
addition o
thevoicing
of
views
n
the
open
forum,he
expression f
public opinion s in the form f direct n-
fluence
n
those
who
are to
act
in
response
to
public pinion.
hrough
uch
means
s let-
ters,
elegrams,
etitions,
esolutions,obbies,
delegations,
nd
personal
meetings
nterested
groups nd
individuals ring
heir
iews nd
positions
o bear
on
the
key persons
who
have
to
make the
decisions. am
not con-
cerned
with
whether
uch
forms f
expressing
public
pinion hould
ccur;
merely
ish o
emphasize
hat
n
any
realistic
onsideration
ofpublicopiniontmustbe recognizedhat
such means
of
expressingublic
opinion
do
occur.
A
society
which
as
to act will
use the
channels
f
action
hat t
has in
ts
structure.
6.)
The
last
feature f
public
opinion hat
I wish to
note s that n
any
realistic ense
public
opinion
onsists f
the
pattern
f
the
diverse
iews
nd
positions
n
the ssue
that
come to
the ndividualswho
have to
act
in
response o thepublicopinion. ublicopin-
ion which
was
a
mere
display,
r which
was
terminal
n
its
very
expression,
r
which
never came
to
the
attention f
those
who
have
to act on
public
opinion
would
be
im-
potent
nd
meaningless
s
far s
affecting
he
action or
operation
f
society
s
concerned.
Insofar
s
public
opinion
s
effective
n
so-
cietal
action
t becomes o
onlyby
entering
into
thepurview
f
whoever,
ike
egislators,
executives,
dministrators,nd policymak-ers, have to act
on
public
opinion.
To me
this
proposition
s
self-evident.
f
it be
granted,
he
character
f
public
opinion
n
terms of
meaningful
peration
must be
sought
n
the
array
of
views
and
positions
which
nter nto
the
consideration
f
those
who have
to
take action
on
public
opinion.
It is
important
o
note
that
the
ndividual
who
has to act on
public
pinion
as
to
assess
the
public
opinion s it
comes
to his atten-
tion, ecauseofthevery act hat hispublic
opinion
omes
to him
n
the form
f
diverse
views
nd
usually
opposed
views.
nsofar s
he
is
responsive
o
publicopinion
he
has to
weigh he
respective
iews.
How
this
ssess-
ment
s made s
an
obscure
matter.
ut
one
generalization
ven
hough
rite,
an
be
made
safely,
o
wit,
that
the
ndividual
akes
nto
account
different
iews
only
to
the
extent o
which
such
views
count.
And
views
count
prettymuchon thebasis of how the indi-
vidual
judges the
backing of
the
views
and
the
implication
f
the
backing.
t
is
in
this
ense,
gain, that
the
organizationf
the
ociety
with
ts
differentiationf
prestige
and
power, nters
nto
the
character
f
pub-
lic
opinion.
As was
explained
bove,
the
key
person
who
has to
act on
public
opinion
s
usually
ubject o
a
variety f
presentations,
importunities,
emands,
riticisms,
nd
sug-
gestions hat
come
to him
through
he
var-
iouschannels nthecommunicativetructure
of
society.
Unless
one
wishes
to
conjure n
his
imagination
very
fanciful
ocietyhe
must
dmit
hat
he ervant
f
public
pinion
is
forced
o
make an
assessment
f
the ex-
8/10/2019 Blumer Public Opinion and PO Polling
6/9
546
AMERICAN
SOCIOLOGICAL
REVIEW
pressions
f public
opinion
hat
come
to
his
attention
nd
that n this ssessment
onsid-
eration
s
given
oexpressions
nly
othe
x-
tent
to
which
they
are
judged
to count.
The foregoingix featuresre, I believe,
trite
utfaithful oints
bout
public
opinion
as
it
functions
n our ociety.
hey
may
erve
as
a background
or heexamination
fpub-
lic
opinion
olling.
may
state
herethat
n
this
discussion
am
not concerning
yself
with
theproblem
fwhether
he
ndividual
opinions
ne gets
through
he polling
nter-
view
rereasonably
alid.
My discussion,
n-
stead,
s concerned
ith
the
question
fthe
value
ofpoll
findings
ven
f onemakes
the
dubious
ssumptionhat he ndividual pin-
ions
that re secured
re
valid.
In
my udgment
he nherent
eficiency
f
public
opinion
olling,
ertainly
s
currently
done,
s
contained
n its
sampling
rocedure.
Its
current
amplingrocedure
orces
treat-
mentof
society
s
if
society
were
onlyan
aggregation
f
disparate
ndividuals.
ublic
opinion,
n
turn,
s
regarded
s being
quan-
titative
distribution
f
individual
pinions.
This wayoftreatingociety nd thiswayof
viewing ublic
opinion
mustbe
regarded
s
markedly
nrealistic.
he
best way
I
can
bring
his out
is
by
making
ontinuous ef-
erence
to the
common ense empirical
b-
servations
f
public
opinion
hatwere
noted
previously.
We
do
notknow
at
all whether
individuals
n
the sample
represent
hat
por-
tion of
structuredociety
hat s participat-
ing
in the
formation
f
public
opinion
n a
given ssue. That the samplewill catch a
number
fthem,
r even
a larger
umber
f
them,
s
very
ikely.But,
as
far s
I
am able
to determine,
here
s no
way
n
current ub-
lic opinion
olling
o know
much bout
this.
Certainly
he mere
fact
that
the
nterviewee
either ives
or
does
not
give
an
opinion
oes
not
tellyou
whether
e
is
participating
n
the
formation
f
public
opinion
s
it is
being
built
up
functionally
n
the
society.
More
important,
ssuming
hat
he
ample
atches
the ndividualswhoare participatingn the
formation
f
the
given
public
opinion,
no
information
s
given
of
theirpart
in this
process.
ne cannot
dentify
rom
he ample
or
from the
replies
of those constituting
thesample
the socialnitch f the
ndividual
in
that
portionof the social
structure n
which the
public opinion s
being formed.
Such
informations not given n
the con-
ventional tems f age,sex,occupation, co-
nomic
status, educational
attainmentor
class status.These are
rarely hemarksof
significant
unctional
osition
n
the forma-
tion
of publicopinion
n a given ssue. We
do not know
from he
conventional ind of
sample or from he
responses f the inter-
viewee
what
nfluence,
f
any,he has in the
formation r
expression f
public opinion.
We
do not know
whether e has a following
or
whetherhe
doesn't.
We do not know
whether r nothe is speaking nbehalf f a
group r
groups rwhether e
even belongs
to functional
roups nterestedn
the ssue.
If
he does,perchance,
xpress he views of
some such functional
roup,
we
don't
know
whether rnot
that
group s busily t work
inthe
channels fsociety ogive
vigorous x-
pression o theirpoint
of view.We do not
even
know
whether
e,
as
an
individual,
s
translating
is
opinioninto
what
I
have
termed reviouslyeffectiveublicopinion.
In
short,we know
essentially
othing f
the
ndividualn
the
ample
with
eferenceo
the
ignificance
f
him
r ofhis
opinion
n
the
public pinion hat s
beingbuiltup or which
is
expressing
tself
unctionally
n
the opera-
tionof
society.
We
do not know
whether
he
individual
as
the
position
f
an
archbishop
or
an
itinerant
aborer;
whether e
belongs
to a
powerful
roup aking vigorous
tand
on the ssue or whether e
is a
detached e-
cluse
withno
membership
n
a functional
group;whether e is
bringing is
opinion o
bear
in
some
fashion t
strategic
oints
n
the
operation
f
society
r
whethert
is
iso-
lated
and
socially mpotent.
We do notknow
what
role,
f
any, ny
ndividual
n
the sam-
ple plays
n
theformationf the
public
opin-
ion
on
which
e is
questioned,
nd
we
do not
knowwhat
part,
f
any,
his
opinion
s
given
has
in
the
functional
ublic opinion
which
existswithreferenceo the ssue.
What
has
just
been said
with
referenceo
the
ndividual
omponent
f
the
publicopin-
ion
poll applies
collectively
o
the
total
find-
ings.
The
collective
indings
ave
no
assur-
8/10/2019 Blumer Public Opinion and PO Polling
7/9
PUBLIC
OPINION AND
PUBLIC
OPINION
POLLING
547
ance of depicting ublic opinion n
a
given
issuebecausethese indingsgnore heframe-
work and the
functional
perationof
the
publicopinion.
f this
s
not clear
fromwhat
has alreadybeensaid, would ike to point
out
theenormous
ifficulty
hatoccurswhen
one seeks
to
assess
the
findings
f a
public
opinion oll
in
terms f theorganizationf
society
withwhich n
administrator,egisla-
tor, xecutive,
r
similarly laced person
has
to contend.
As I have stated arlier uch
an
individualwho is presumably esponsive o
public pinion as
to assess
publicopinion s
it comes o his attentionnterms f thefunc-
tionalorganization
f
society
o
which
he
is
responsive. e has to view thatsociety n
termsof groups
of
divergentnfluence;
n
terms
f
organizations
ith
different
egrees
of
power;
n
terms f
ndividuals ith ollow-
ings;
in
terms
f indifferent
eople-all,
in
other
words,
n
terms
fwhat nd whocounts
in
his part
of
the social
world.This
typeof
assessment
hich
s
called for
n
the nstance
of
an
organized ociety
n
operation
s
well-
nigh mpossible
o
make
in the
case of
the
findingsfpublic opinion olls.We are un-
able
to
answer uch
questions
s
the
follow-
ing:
how
much
power
nd
influence
s pos-
sessed
by
thosewhohave
the
favorable
pin-
ion or theunfavorablepinion;who
re
these
people
whohave the
opinion;
whom
do
they
represent;
owwell
organized
re
they;
what
groupsdo they belong
to
that are
stirring
aroundon the
scene
and that
are
likely
to
continue
o do
so;
are those
eople
whohave
thegiven pinion erymuch oncernedbout
their
pinion;
re
they oing
o
getbusy
nd
do something
bout
t; are theygoing o get
vociferous,militant,
nd
troublesome; re
they
n the
position
to
influence owerful
groups nd individuals ho are known;does
theopinion epresent studied olicy f sig-
nificant
rganizations
hich
will
persist nd
who re
ikely
o
remember;
s the
opinion n
ephemeral
r
momentary
iew
whichpeople
will
quicklyforget?
hese
samplequestions
showhowmarkedlyifficultt s to assessthe
resultsof public opinionpolling from he
standpoint
f the
hings
hat
have to
be taken
into account
n
workingn an organized o-
ciety.
This
difficulty,
n
turn, ignifies hat
current
ublic
opinion
ollinggives n inac-
curate nd
unrealistic
icture
f
public
opin-
ion
becauseof
the
failure o catch
opinions s
they
re
organized
nd as
they
operate
n
a
functioningociety.
What have
said will
appear
to
many
s
distinctly
nvalidon
the
ground
hatpublic
opinion
olling
has
demonstrated
hat
t can
and
does
detect
ublicopinion
aithfully,y
virtue
of
its
marked
success
in
predicting
election
eturns.
his
contention
eeds
to
be
investigated
arefully,
articularly
ince in
most
circles
olling,
wherever
pplied,
s re-
garded as
intrinsicallyalid
because of
its
rather
pectacular
uccess n
predictinglec-tions.What think
eedsto
be
noted s
that
the
asting f
ballots
s
distinctlyn
action f
separate
ndividuals
herein
ballot cast
by
one
individual as
exactly he
same
weight
as
a ballot
cast
byanother
ndividual.
n
this
proper
ense, nd in
the
senseof
real
action,
voters
constitute
population f
disparate
individuals,
ach
of
whom
has
equal
weight
to the
others.
Consequently,
he
sampling
procedure
hich
s
based on
a
population f
disparate ndividuals s eminentlyuited to
securing
picture
f what
he
voting
s
likely
to
be.
However,
o
regard he
successful se
of
polling
n
this
area as
proof f its
auto-
matic
validitywhen
pplied
to an
area
where
people do
not
act
as
equally
weighted is-
parate
individuals
begs
the
very
question
under
onsideration.
would
repeat hatthe
formationnd
expression
f
public
opinion
giving
iseto
effective
ublic
opinion
s not
an actionof a population fdisparatendi-
viduals
having
qual
weight
ut
s
a
function
of
a
structured
ociety,
ifferentiated
nto a
network f
different
indsof
groups nd
in-
dividuals
having
differential
eight
nd
in-
fluence
nd
occupying
ifferent
trategic o-
sitions.
Accordingly,o
my
mind,
he
uccess
attending
olling
n
the
prediction
f
elec-
tions
gives
no
validity
o
the
method
s
a
means of
studying,
ecordingr
measuring
public
opinion
s it
forms
nd
functionsn
oursociety.
There is
a
very
mportant
ontention
n
this
connection
hich
has
to be
considered.
The contentionan
be
stated s
follows:
An
election
y
public
ballot
s
in
itself
n
8/10/2019 Blumer Public Opinion and PO Polling
8/9
548
AMERICAN
SOCIOLOGICAL
REVIEW
expressionf
public
pinion-and,
urthermore,
it
is
effective
nddecisive
xpression
f
public
opinion.t
is,
n
fact,
heultimate
xpression
f
public
pinion
nd thus t
represents
he
proper
normfthe xpressionfpublic pinion.n theelectionyballot ach
voter,
naccordance
ith
the
basic
principles
f
democracy,
as
his
say
as
a
citizen nd
has
equal
worth
o
every
ther
citizen
n
casting
is
ballot. f
election
y
ballot
be
recognized
s
the
genuine
eferendum
n
which
true
public
opinion
omesto
expres-
sion, then
the
preeminence
f
current
ublic
opinion
olling
s the
device
or
recording
nd
measuring
ublic
pinion
s
established.
or, ub-
lic
opinion
olling
ith
ts
currentorm
f
sam-
pling
as
demonstrated
hat t can
predict
elia-
blyandeffectivelyheresults f theelection.
Accordingly,ublic
pinion
olling,
n
tself,
an
be
used
s
a
type
f
referendumo
record
nd
measure
he
rue
pinion
f
the
public
n
ssues
in
the
nstancesf
which
he
public
oes
not
go
to
the
election
olls.
Thus,
public
pinion
ol-
ling
yields
more
eliable
nd
accurate
icture
of
public
opinion
han
s
represented
y
the
confused,
ndefinite,
lanted,
nd
favor-ridden
expressions
f
opinion
hat
come
ordinarily
o
the
egislator,
dministrator,r
executive
ho
hastoactonpublic pinion.hepublic pinionpoll
tells
us
where
eople
tand.
t
gives
s
the
vox
popUl.
My
remarks
ith
eference
o
this
onten-
tion
will
be
brief.
t
should
be
evident
n
analysis
hat
he
ontentions
actually
nor-
mative
plea
and
not
a
defense
f
polling
s
a
method
f
study
f
public
opinion
s
such
public
opinion
unctions
n
our
society.
he
contention
roposes
that
public
opinion
be
construedn a particularway,to wit,that
public
opinion
ught
o
be
an
aggregation
f
the
opinions f
a
cross
ection
f
the
popula-
tion
rather
han
what t
s
in
the
ctual
func-
tioning f
society.
To
my
mind t
is
highly
questionable
whether
n
the
day
by
day
op-
eration
f
our
society
ublic
pinion
ught
o
be of
the
nature
osited
y
the
public
pinion
poll.
Many
appropriate
uestions
ould
be
raised
bouthow
and
to
what
extent
ublic
opinion
s
expressed
t
the
lection
olls,
nd,
moremportant,hethertwouldbepossible
or
even
advisable
for
public
opinion,
n
the
form
f
an
aggregation
f
equally
weighted
individual
pinions,o
function
eaningfully
in
a
society
with
diversified
rganization.
However,
uch
questions
need
not
be
raised
here. t is sufficiento
notethat f one seeks
to
justify ollingas a
method f
studying
public pinion ntheground
hat he ompo-
sition f
publicopinion ught
o
be
different
than what it is, he is not establishinghe
validity
f
the method
orthe studyof
the
empirical
world
s
it
is.
Instead,he is hang-
ing
on the
oat-tails f a dubious roposal or
social reform.*
In this aper have
presented riticismsf
publicopinion olling s a
method or he
recording
nd
measurement
f public opin-
ion.
These criticisms ave
centered round
the distortion
hat stemsfrom he use of
a
sample
n
the
form
fan aggregationfdis-
parate individuals
having equal weight.
These criticisms
houldnotbemisinterpreted
to mean hat uch
sampling
rocedures
in-
valid
wherever
pplied
or
thatwhereveroll-
ingmakesuse of
such a sampling rocedure
such
polling
s
intrinsicallynvalid.Clearly,
the criticism
pplieswhen
such
a
sampling
procedure s used
to
study matterwhose
composition
s an
organizationf
nteracting
parts
nstead f
beingmerely
n
aggregation
of ndividuals.Where hematterwhich ne s
studyings
an
aggregation
f
ndividual nits
then the
application f the
samplingpro-
cedure
poken
of is
clearly
n
order.
make
thisbanal statement
nly
to call
attention
o
thefact hat there
re
obviouslymany
mat-
ters about
human
beings
nd
their onduct
that
have ust
this
character f
being
n
ag-
gregation
f ndividuals
r
a congeries
f n-
dividual
ctions.
Many
demographic
atters
are ofthisnature.Also,many ctions fhu-
man
beings
n a
society
re
of this nature
-such
as
castingballots,
purchasing
ooth
*
I refer o such a program s
dubious because
I believe the much needed
improvement f public
opinion in our society should be in
the process by
which
public
opinion organicallyfunctions, .e., by
arousing,
organizing, nd effectively irectingthe
opinion of
people who appreciate that they have
an interest n a given issue. A reliance,
nstead, on
a
mere
referendum
y
an
undifferentiated ass,
havinggreat
egments f ndifferencend non-partici-
pation, s unlikely o offer desirable
public opinion.
At the best, n my judgment, uch a
referendum
could operate
as a corrective upplement nd not as
a substitute.
The important question concerning
the directions n which
public opinionmight ecure
its much needed
improvements, of course, outside
of
the scope of this paper.
8/10/2019 Blumer Public Opinion and PO Polling
9/9
PUBLIC OPINION
AND
PUBLIC
OPINION
POLLING
549
paste,
going to
motion
picture
hows,
and
reading
newspapers.
uch
actions,
which
liketo
think f
as
mass actions f
ndividuals
in
contrast o
organized
ctions of
groups,
lendthemselveseadily o thetypeofsam-
pling
hat
wehave in current
ublic
opinion
polling.
n
fact,
t is the
existence f
such
mass
actions f
ndividuals hich
xplains,n
my
udgment,
he uccessful se in consumer
research
f
sampling
uch as is
employed n
public
opinion
olling.
What
find
uestion-
able, and
what his
paper
criticizes,
s the
use
of such
sampling
with
ts
implicit
magery
and
logic
n
the
tudy
f a matter
which,
ike
theprocess fpublicopinion, unctionss a
moving
rganization
f
nterconnected
arts.
The
last
item
wish to
consider
riefly
refers
o
the
interesting
nd
seemingly
af-
fling
uestion
f
how
one should
r can
sam-
ple
an
object
matter
which
s a
complicated
system
f
interacting
arts,
having
differen-
tial
influence
n
the
total
operation.
erhaps
the
question
n
itself
s
absurd. At
various
times I
have
asked
different
xperts in
sampling
ow
one
would
samplean
organic
structure. ith single xceptionhese ndi-
viduals
ooked
t
me
askance
as if
the
ques-
tion
were
diotic.
But
the
problem,
think,
remains
ven
though
find
t
difficult
o
state.
n
human
ociety,
articularlyn
mod-
ern
society,
we
are
confronted
ith
ntricate
complexesof
moving
relations
which
are
roughly
ecognizables
systems,
ven
though
loose
systems.
uch
a
loose
system s
too
complicated,
oo
encumbered
n
detail
and
toofastmovingobe describednanyoneof
its
given
cycles
of
operation
dequately
and
faithfully.
et
unlesswe
merely
want
to
speculate
bout
it
we
have
to
dip
into
t
in
some
manner
n
order o
understand
hat
s
happening
n
the
given
ycleof
operation
n
which
we
are
nterested.
hus,
using
he
pub-
lic
opinion
rocess n
our
society
s an
illus-
tration
we are
able
to
makea
rough
harac-
terization
s
to
how
t
functionsn
the
case,
let
us say,
of a
national
ssue.
However,
f
we
want oknowhow tfunctionsnthecase of
a
given
national
ssue,we
are
at
a
loss
to
make n
adequate
description
ecause
of
the
complexity
nd
quick
movement
fthe
cycle
of
its
operation.
o,
to
know
what s
going
on,
particularlyo
know
what
s
likely
o
go
on
in the
atter tages,
we have to dip n here
and there. he problems
f where o dip in,
how to dip in, and how far to dip in are
what have n
mind n
speaking f sampling
an organic tructure.
I
suppose,
s
one
ofmyfriends as pointed
out,that
the answer o the
problem equires
theformulationf
a
model.We have
no
such
model
n the
nstance
f publicopinion s
it
operates
n
our
ociety.My
ownhunch
s
that
such
a model houldbe
constructed,
f t can
be
at
all, by workingackwardsnstead f
by
working orward.
hat is, we ought o begin
with hosewhohaveto act on publicopinion
and movebackwards long the lines of the
variousexpressions f public opinion that
cometo
their ttention,racing hese xpres-
sions backward through
heir own various
channels
nd in
doingso, noting he chief
channels, he key points
of importance,nd
the
way
in which
ny
givenexpression as
cometo developand pick up an organized
backing ut
of
what nitiallymusthave been
a
relatively morphous
ondition. erhaps,
such a
model,
f
it could be workedout,
would allow the developmentf a realistic
method f samplingn place of what seems
to
me to
be
the
highly
rtificialmethod
f
sampling
sed
n
current ublicopinion oll-
ing.
DISCUSSION
THEODOREM. NEWCOMB
University f Michigan
ProfessorBlumer has long been known s a
formidable ritic, nd
I am
sure there
re many
otherswho shared with me the anticipation f
seeing
him turn
his
battery
of
high-powered
guns upon the practitionersf public
opinion.
In my udgment, is guns n this nstance
have
misfired. his is not to say thatthosewho
study
attitudes y samplingmethods re beyond
riti-
cism; even
a lessercritic han
Professor
lumer
could pointto many hortcomingsn their art.
It
may truly
be said of
all
of
them,probably,
that n one
way
or
another
heyhave
done what
they ught
not to
havedone and thattheyhave
not done what theyought o have done,and in
some of
them there s littlehealth ndeed. His
targetwas
thus
an
easy one,
and I want
to raise
the question
of
why
t
is
that
he has neverthe-
less, n my udgement,missed t.
The first eason, think,
s that
he was not
quite selective nough
n
his
aim. I wish
he had