OIsevvalionaI Leavning as a Funclion oJ SnIoIizalion and Incenlive Sel

AulIov|s) AIIevl Banduva, Joan E. Ovusec and Fvances L. MenIove

Bevieved vovI|s)
Souvce CIiId BeveIopnenl, VoI. 37, No. 3 |Sep., 1966), pp. 499-506
FuIIisIed I Wiley-Blackwell on IeIaIJ oJ lIe Society for Research in Child Development
SlaIIe UBL http://www.jstor.org/stable/1126674 .
Accessed 02/10/2012 0846
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
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].
.
Wiley-Blackwell and Society for Research in Child Development are collaborating with JSTOR to digitize,
preserve and extend access to Child Development.
http://www.jstor.org
OBSERVATIONAL LEARNING AS A FUNCTION
OF SYMBOLIZATION AND INCENTIVE SET
ALBERT
BANDURA,
JOAN E.
GRUSEC,
and FRANCES L. MENLOVE
Stanford University
This
study investigated
the
effects of symbolization
on
delayed repro-
duction
of modeling
stimuli in a test
of
the
contiguity-mediational
theory
of
observational
learning. During exposure
to the behavior
of
a
film-medi-
ated
model,
1
group of
children
engaged
in concurrent
verbalization,
a
second
group
observed
passively,
while a third
group engaged
in com-
peting symbolization. Half of
the children in each
of
the treatment condi-
tions observed the model's behavior under a
positive
incentive
set;
the re-
maining
Ss were
provided
no incentive to learn the
model's
responses.
Ss
who
generated
verbal
equivalents of
the
modeling
stimuli
during exposure
subsequently reproduced
more
matching responses
than the
passive
view-
ers, who,
in
turn,
showed a
higher
level
of acquisition
than children in the
competing symbolization
treatment. Observational
learning,
however,
was
not
influenced
by
incentive set.
Most
conceptualizations
of imitative or observational
learning
have
been
developed
and tested
largely
on the basis of a limited
paradigm
re-
quiring observing
Ss to
perform matching responses
and to secure
positive
reinforcers as a
precondition
for their
acquisition (Baer
&
Sherman, 1964;
Miller &
Dollard, 1941; Skinner, 1953).
These
theories, however,
fail to ac-
count for the occurrence of
delayed reproduction
of
modeling
behavior
originally
learned
by
observers under
exposure
conditions
permitting
no
overt
performance
of
matching responses (Bandura, 1962; 1965a;
Bandura
&
Walters, 1963).
Since observers can
acquire only perceptual
and other
implicit responses resembling
the.
sequences
of
modeling
stimuli while
they
are
occurring, symbolic processes
which mediate
subsequent
behavioral re-
production
must
play
a
prominent
role in
observational
learning.
Recent theoretical
analyses
of observational
learning
(Bandura, 1965c;
This work was
supported by
Public Health Service research
grant
M-5162
from the National Institute of Mental Health. Author's address:
Department
of
Psychology,
Stanford
University,
Stanford,
Calif. 94305.
CHILD DEVELOPMENT
Sheffield, 1961)
emphasize
the role of stimulus
contiguity
and associated
cognitive
or
representational responses
in the
acquisition process. According
to
contiguity-mediational theory, during
the
period
of
exposure modeling
stimuli elicit in
observing
Ss
configurations
and
sequences of sensory
ex-
periences
which,
on the basis of
past
associations,
become
centrally
inte-
grated
and structured into
perceptual responses.
There is some evidence from research in
sensory conditioning
(Conant,
1964; Leuba, 1940;
Naruse &
Abonai, 1953)
to
suggest
that,
in the course
of
observation,
transitory sensory
and
perceptual phenomena
can be con-
verted to retrievable
images
of the
corresponding
stimulus events. In addi-
tion to
imaginal responses,
the observer
acquires,
once verbal labels have
become attached to
objective
stimuli,
verbal
equivalents
of the model's be-
havior
during
the
period
of
exposure
(Bandura, 1965b; Bandura, Ross,
&
Ross, 1963).
The latter
findings
thus
provide
some basis for
assuming
that
symbolic
or
representational responses
in the form of
images
and verbal as-
sociates of the model's behavior constitute the
enduring learning products
of
observational
experiences.
It is likewise assumed in the
contiguity-media-
tional
theory
that
symbolic matching responses possess cue-producing prop-
erties that are
capable
of
eliciting,
some time after
observation,
overt re-
sponses corresponding
to those that were
modeled,
provided
the
requisite
components
exist in the observer's behavioral
repertoire.
In order to test the
proposition
advanced in the above
theory
that
sym-
bolization enhances observational
learning,
a
modeling study
was conducted
in which an
attempt
was made to
manipulate
the relevant
symbolic
re-
sponses.
Viewers' verbalizations of
modeling
stimuli would be
expected
to
influence
mediating
or
representational response processes.
Therefore,
dur-
ing
the
response-acquisition phase
of the
experiment,
one
group
of children
engaged
in facilitative verbalization
designed
to enhance the
development
of
imaginal
and verbal associates of the model's behavior. A second
group
of
children
simply
observed
passively,
while the third
group
of children as-
signed
to a
competing symbolization
condition
produced interfering
verbal
responses
intended to counteract the establishment of
representational
re-
sponses.
The
degree
of observational
learning may
also be
partly governed by
incentive-related sets which exert selective control over the
direction,
in-
tensity,
and
frequency
of
observing responses.
It is also
entirely possible
that
different
symbolization
instructions could create in observers differential
anticipations
as to the
possibility
of their later
being
called
upon
to dem-
onstrate what
they
had learned from the modeled stimulus
presentation.
Such self-induced
sets,
if
operative, might
affect
attending
behavior and
thus confound the effects of
symbolization processes.
Hence,
in the
present
study
half of the Ss in each of the three observational treatments were of-
fered
positive
incentives to learn the model's
response patterns,
while the
remaining
children were
provided
no incentives.
500
BANDURA, GRUSEC, AND MENLOVE
It was
predicted
that the number of
matching responses acquired
ob-
servationally by
children in the facilitative
symbolization
condition
would
exceed the
corresponding
scores for the
passive
observers, who,
in
turn,
would show a
higher
level of
acquisition
than Ss in the
competing symboli-
zation treatment. It was also
hypothesized,
for reasons
given
above,
that
in-
centive set would enhance observational
learning.
METHOD
Subjects
The Ss were 36
boys
and 36
girls ranging
in
age
from 6 to 8
years.
They
were drawn from two
elementary
schools in a lower-middle-class com-
munity.
Design
and Procedure
The
investigation
utilized a 2
x
2 x 3 factorial
design.
Children of
each sex were
randomly assigned
to either the facilitative
symbolization,
the
passive observation,
or the
competing symbolization
conditions. Half of the
Ss in each of the latter
experimental
treatments
participated
in the in-
centive-set
condition,
while the
remaining
children were
assigned
to the
no-incentive-set
group.
Thus,
there were six
groups
of six Ss each.
The Ss in the facilitative
symbolization group
were instructed to ver-
balize
every
action of the model as it was
being performed
in the movie.
With children in the
competing symbolization group,
E
explained
that she
was interested in
determining
whether children could
pay
close attention to
a movie while
engaging
in another
activity simultaneously. They
were
therefore to count
"1
and a
2,
and a
3,
and a
4,
and a 5"
repeatedly,
at the
same time that
they
watched
closely
the filmed
presentation.
The Ss in the
passive
observation
group
were instructed
simply
to
pay
close attention to
the movie.
Children who were
assigned
to the incentive-set condition were also in-
formed that
following
the movie
they
would be asked to demonstrate what
they
had learned and that
they
would receive
candy
treats for each item
that
they reproduced correctly.
On the other
hand,
Ss in the no-incentive-
set
group
were told that
they
would return to their classroom
immediately
following
the movie.
Following
the instructional
procedures,
E turned on the movie
projec-
tor,
which
displayed
the
modeling
stimuli on a
glass
lens screen in the tele-
vision console.
501
CHILD DEVELOPMENT
The movie consisted of a 4-minute color film in which an adult male
model exhibited a series of
relatively
novel
patterns
of behavior and often
utilized
play
materials in unusual
ways.
Since this
study
was concerned with
issues of
response acquisition,
it was
necessary
to devise unusual
response
sequences
that have
virtually
a zero
probability
of
occurring spontaneously
in
preschool
children and hence meet the criterion of novel
responses.
In the
opening
scene the model entered the room with his
right
hand
cupped
over
his
eyes.
He then
clasped
his hands behind his back as he
surveyed
the
various
play
materials.
In the first action
sequence
the model built a tower with blocks ar-
ranged
in a
unique
manner and set a
plastic juice dispenser,
which was to
serve as a
target,
on
top
of the tower. After
picking up
a dart
gun
and a
baton,
the model
paced
off
a
specified
number of
steps
from the
target,
placed
the baton on the floor to mark the
range, got
down on one
knee,
and
fired the dart
gun
at the
plastic
container. The block tower was then dis-
assembled in a distinctive
way.
In the next action
sequence,
which was di-
rected toward a
large
Bobo
doll,
the model sat on the
doll,
punched
it in
the
nose,
pummeled
it with a
mallet,
tossed rubber balls at
it,
lassoed it
with a hoola
hoop,
and
dragged
it with the
improvised
lariat to a far corner
of the room. The model then
picked up
two bean
bags
and
paced
off
back-
ward a
designated
number of
steps
from the
target,
which consisted of an
upright
board with three
large
holes. After
reaching
the
starting point,
the
model tossed the bean
bags
between his
legs
with his back to the
target.
He
then retrieved the
bags, paced
backward,
squatted
with his back to the
target,
and tossed a bean
bag
over each shoulder. The
closing
scene showed
the model
walking
across the room
whirling
the hoola
hoop
over his
right
arm.
Delayed
imitative
performance
is determined not
only by
observational
variables but also
by
rehearsal
processes
which
typically improve
retention.
An incentive set
may,
therefore,
influence the amount of behavioral
repro-
duction
by
both
augmenting
and
channeling observing responses during
acquisition,
and
actuating
deliberate
implicit
rehearsal of
matching
re-
sponses immediately
after
exposure.
Since the
present experiment
was
pri-
marily
concerned with issues of
response acquisition,
and the occurrence of
differential
anticipatory
rehearsal would obscure
results,
children in all
groups
were
assigned
the task of
counting
out loud
during
the brief
period
between the end of the movie and
reproduction. By
this
procedure,
an at-
tempt
was made to hold
interpolated
activities constant for all
groups
and
to
prevent
facilitative rehearsal of
responses. Immediately
after the
comple-
tion of the
movie,
E announced that
they
would now
proceed
to another
room in the mobile
laboratory,
and that she was interested in
seeing
how
high they
could count while
they
walked to the other end of the trailer.
Since the
interpolated
task was
highly
dissimilar to the modeled
activities,
it
was not
expected
to reduce retention to
any great
extent.
502
BANDURA, GRUSEC, AND MENLOVE
Test
for Acquisition
The test for
acquisition
was administered
by
a second female
experi-
menter in a room
containing
the stimulus items that were utilized
by
the
model in the filmed
performance.
In order to control for
any possible
E in-
fluences,
the
person
who conducted this
phase
of the
study
did not know to
which treatment conditions Ss had been
assigned.
The children were asked
by
E to demonstrate all of the model's re-
sponses they
could
recall,
and were
praised
and rewarded with
candy
for
each
matching response correctly reproduced.
A standardized
cuing pro-
cedure was
employed
to insure the same order of
reproduction
for all Ss.
They
were asked to show the
way
in which the model behaved in the
open-
ing
scene of the
movie;
to demonstrate what he did with the dart
gun,
the
Bobo
doll,
and the bean
bags;
and to
portray
the model's behavior in the
closing
scene of the film. The children were thus
given
a
fragmentary
cue
and asked to
reproduce
the entire
response sequence
in which the
par-
ticular stimulus
object
was
employed.
The E recorded the children's
matching responses
on a checklist con-
taining
the 38
responses
that had been exhibited
by
the model. In order to
provide
an estimate of interscorer
reliability,
the
performances
of ten chil-
dren were recorded
simultaneously
but
independently by
another rater who
observed the test sessions
through
a
one-way
mirror from an
adjoining
ob-
servation room. The two raters were in
perfect agreement
on 96
per
cent of
the
specific matching responses
that were scored.
RESULTS
Table
1
presents
the mean number of
matching responses
achieved
by
children in the various treatment conditions. In each incentive
condition,
TABLE 1
MEAN NUMBER OF MATCHING RESPONSES REPRODUCED AS A FUNCTION
OF SYMBOLIZATION AND INCENTIVE-SET CONDITIONS
OBSERVATIONAL CONDITIONS
Facilitative Passive
Competing
INCENTIVE
Symbolization
Observation
Symbolization
No Incentive Set:
Boys...........
16.8 14.5 11.5
Girls........... 17.5 13.2 6.0
Total......... 17.2 13.8 8.7
Incentive Set:
Boys...........
16.2 15.3 13.0
Girls.......... 14.8 11.7 9.8
Total......... 15.5 13.5 11.4
503
CHILD DEVELOPMENT
for both
boys
and
girls,
the mean
reproduction
scores attained
by
the ac-
tive
symbolizers
exceeds the
corresponding
means for the
passive
observers
who,
in
turn,
show a
higher
level of
acquisition
than Ss in the
competing
symbolization
treatment.
Analysis
of variance of these data reveals that
symbolization
is a
highly significant
source of variance
(F = 13.01;
p
<
.001).
Further
comparisons
of
pairs
of means
by
the t test indicate that Ss
who
generated
verbal
equivalents
of the
modeling
stimuli
during presenta-
tion
subsequently reproduced significantly
more
matching responses
than
children who either observed
passively
(t
=
2.18;
p
<
.025)
or
engaged
in
competing symbolizations
(t
=
5.12;
p
<
.001). Moreover,
the
passive
viewers
acquired
more of the model's
repertoire
of behavior than children
who
engaged
in verbalizations
designed
to interfere with the
development
of
representational responses
(t
=
2.94;
p
<
.01).
It is of interest that
boys reproduced
a
significantly higher
number of
matching responses
than
girls
(F
=
5.70;
p
<
.05).
Contrary
to
prediction,
however,
observational
learning
was not influenced
by
incentive set nor
were there
any significant
interaction effects.
DISCUSSION
Although
the results of the
present study provide confirmatory
evi-
dence for the facilitative role of
symbolization
in observational
learning,
alternative
interpretations
of these
findings might
be examined. It is con-
ceivable that the method utilized for
preventing
the
acquisition
of
repre-
sentational
responses may
have interfered with observation of the
modeling
stimuli.
Considering,
however,
that the
modeling
stimuli were
projected
on
a
large
television screen
directly
in front of the S seated in a dark
room,
it is
improbable
that,
under such conditions of
highly
focused
attention,
concur-
rent
competing
verbalization could reduce
appreciably
the occurrence of
observing responses.
Indeed,
the
replies
of children in the incentive-set
condition to
questions
in the
postexperiment
interview indicate that
they
exerted
strong
efforts to observe and to retain the
responses
exhibited
by
the
model
("While
I was
counting,
I was
paying
attention to the
thing.
I
thought
back about it in
my
head
... I look at it all the time so I remember
it all the
time").
The marked external control of
observing responses through
the tele-
vised mode of stimulus
presentation
in all likelihood accounts to some ex-
tent for the absence of a
significant
incentive effect on the
acquisition
of
matching
behavior. As several of the children
put
it,
"I didn't look all over
the
place.
I
just
watched the movie and
kept my eyes
on it. .
...
I
kept
watching.
I couldn't take
my eyes
off it."
In situations where a
person
is
exposed
to
multiple
models
exhibiting
diverse
patterns
of
behavior,
knowledge
of the reinforcement
contingencies
504
BANDURA, GRUSEC, AND MENLOVE
associated with the
corresponding response patterns
and
anticipation
of
positive
or
negative
reinforcement for
subsequent reproduction may
exert
selective control over the nature and
frequency
of
attending responses.
The
effects of incentive set on observational
learning
would, therefore,
be most
clearly
elucidated
by
a
comparative study utilizing
stimulus situations
ranging
from
(a)
highly
focused observation of a
single sequence
of
model-
ing
stimuli or
(b)
controlled
exposure
to
multiple
models
requiring
selective
attentiveness to
competing
social cues
presented simultaneously
to
(c)
self-
selection of
frequency
and duration of
exposure
to
specific types
of models.
The latter
condition,
which
corresponds
most
closely
to observational learn-
ing
in naturalistic
situations,
would
probably
maximize the influence of re-
inforcement-oriented set.
The
hypothesis concerning
the influence of
anticipated positive
rein-
forcement for
matching responses
on observational
learning
was based on
presumed perceptual sensitizing
effects of an incentive set induced
prior
to
exposure.
The
fact, however,
that Ss who
expected
to be
subsequently
tested and rewarded for their imitative behavior
generally
achieved
slightly
lower
acquisition
scores than their uninformed
counterparts suggests
that
the benefits of incentive set
may
have been offset
by
detrimental effects.
There were indications in the interview
data,
elicited
by questions
about
the children's
thoughts during
the
film-viewing,
that the incentive-set in-
structions
generated
achievement anxieties in some of the children.
("I
was
thinking
what should I tell
you
and what should I do.... I was
just
think-
ing,
I
hope
I remember
everything.")
Observational
learning
could be ad-
versely
affected
by implicit
rehearsal of
preceding
events and
disruptive
thoughts concerning
an
impending
test if these
competing cognitive
ac-
tivities occur while the
modeling
stimuli are
being presented
at a
relatively
rapid
rate.
Any
deleterious
consequences
should be
greatest,
as was the
case,
among
the active
symbolizers
who were in the
optimal film-viewing
condition
and, therefore,
would most
likely
be hindered
by
the additional
motivational effects of
expectation
of a
performance
test
following
the ex-
posure
session.
The small but nevertheless
significant
sex differences in
acquisition
scores is
probably
due,
in
part,
to the fact that the modeled
responses
in-
volved
physical masculine-typed
activities. The
girls'
more adverse reaction
to the incentive set
may
also have been a
contributing
factor.
A
single
brief
exposure
to a continuous series of
modeling
stimuli is
likely
to
produce
some interference in the
acquisition
and correct
sequenc-
ing
of novel
matching responses.
Dramatic intrusions from the various se-
quential patterns
were
occasionally
evident in Ss' behavioral
reproductions,
as in the case of the child who fired the dart
gun
at the Bobo
doll, when,
in
fact,
the
plastic
container served as the modeled
target.
Massed
exposure
conditions
may
thus result in the
development
of some erroneous
modeling
responses
and the
loss,
through
interference,
of
previously acquired
ones.
505
CHILD DEVELOPMENT
Simultaneous
competing
verbalizations
during
observation of
modeling
displays
would not be
expected
to interfere too
extensively
with the de-
velopment
of visual
imagery, particularly
when the
modeling
stimuli are
highly
salient,
as in the
present experiment.
It
would, therefore,
be of con-
siderable theoretical
significance
to determine whether
any matching
re-
sponses
could be
reproduced
if,
in addition to
preventing
the
development
of verbal
associates,
visual
imaginal responses
were likewise
precluded,
masked,
or obliterated. Such
imagery-interference procedures
would
pro-
vide the most decisive evidence as to whether
representational
mediators are
necessary
for
long-term
retention and
delayed
behavioral
reproduction
of
modeling
stimuli.
REFERENCES
Baer,
D.
M.,
&
Sherman,
J.
A. Reinforcement control of
generalized
imitation in
young
children.
J.
exp.
child
Psychol.,
1964, 1, 37-49.
Bandura,
A. Social
learning through
imitation. In M. R.
Jones
(Ed.),
Nebraska
symposium
on motivation: 1962. Lincoln: Univer. of Nebraska
Pr.,
1962.
Pp.
211-269.
Bandura,
A. Behavioral modifications
through modeling procedures.
In L. Krasner
& L. P. Ullmann
(Eds.),
Research in behavior
modification.
New York:
Holt,
Rinehart &
Winston,
1965.
Pp.
310-340.
(a)
Bandura,
A. Influence of models' reinforcement
contingencies
on the
acquisition
of imitative
responses.
J. Pers.
soc.
Psychol.,
1965, 1,
589-595.
(b)
Bandura,
A. Vicarious
processes:
a case of no-trial
learning.
In L. Berkowitz
(Ed.),
Advances in
experimental
social
psychology.
Vol. 2. New York: Academic
Press,
1965.
Pp.
1-55.
(c)
Bandura, A., Ross, Dorothea,
&
Ross,
Sheila A. Vicarious reinforcement and imi-
tative
learning.
J.
abnorm. soc.
Psychol.,
1963, 67,
601-607.
Bandura, A.,
&
Walters,
R. H. Social
learning
and
personality development.
New
York:
Holt,
Rinehart &
Winston,
1963.
Conant,
M. B. Conditioned visual hallucinations.
Unpublished manuscript,
Stan-
ford
Univer.,
1964.
Leuba,
C.
Images
as conditioned sensations.
J.
exp. Psychol., 1940, 26,
345-351.
Miller,
N.
E.,
&
Dollard,
J.
Social
learning
and imitation. New
Haven,
Conn.:
Yale Univer.
Pr.,
1941.
Naruse, G.,
&
Abonai,
T.
Decomposition
and fusion of mental
images
in the
drowsy
and
posthypnotic hallucinatory
state.
J.
clin.
exp. Hypnosis, 1953,
1,
23-41.
Sheffield,
F. D. Theoretical considerations in the
learning
of
complex sequential
tasks from demonstration and
practice.
In A. A. Lumsdaine
(Ed.),
Student
responses
in
programmed
instructions: a
symposium. Washington,
D.C.: Na-
tional
Academy
of
Sciences,
National Research
Council,
1961.
Pp.
13-52.
Skinner,
B. F. Science and human behavior. New York:
Macmillan,
1953.
506