the effects of music in advertising on choice behavoiur
TRANSCRIPT
-
8/11/2019 The Effects of Music in Advertising on Choice Behavoiur
1/9
The Effects of Music in Advertising on Choice Behavior: A Classical Conditioning ApproachAuthor(s): Gerald J. GornReviewed work(s):Source: Journal of Marketing, Vol. 46, No. 1 (Winter, 1982), pp. 94-101Published by: American Marketing AssociationStable URL: http://www.jstor.org/stable/1251163.
Accessed: 11/03/2013 11:04
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].
.
American Marketing Associationis collaborating with JSTOR to digitize, preserve and extend access to
Journal of Marketing.
http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=amahttp://www.jstor.org/stable/1251163?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/1251163?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=ama -
8/11/2019 The Effects of Music in Advertising on Choice Behavoiur
2/9
Gerald
J. Gorn
h e Effects
o f
M usic
n
Advertising
n
hoice
Behavior
Classical
onditioning
Approach
OMMERCIALS
typically
contain both
product
specific
information and
background
features
such as
pleasant
music,
attractive
colors,
and humor.
This
paper
examines the
impact
of the
background
features on
product preferences.
One
experiment
was
conducted
to determine whether
background
features
of
a commercial
affected
product preferences
when
only
minimal
product
information was
presented.
A
second
experiment
examined the
relative
importance
of
background
features
and
product
information in
different
situations.
The
impact
of
product
information in a
commer-
cial on
beliefs and
attitudes
would
typically
be
inter-
preted
within an
information
processing
framework.
It is suggested here that a classical conditioning
framework
could
account for the
potential
impact
of
background
features on
product
attitudes.
Classical
Gerald
. Gorns a
professor
n
he
Faculty
f Commercend
Business
Administration,
niversity
f British
olumbia.he
author
would ike
to
thank he
Institut
'administration
es
Entreprises,
niversity
ix-
Marseille,
rance,
or ts
helpful
ssistance
n
formulating
nd
pretest-
ing
he
project.
Do
features like
humor, sex,
color,
and
music
in
a
commercial
merely
increase
our
attention
to
product
information in
a
message,
or can
they
directly
influence
our
attitudes?
The
results of
an
experiment
using
a
classical
conditioning ap-
proach
suggest
that
hearing
liked or
disliked mu-
sic
while
being exposed
to
a
product
can
directly
affect
product
preferences.
A
second
experiment
differentiated
communication
situations
where
a
classical
conditioning approach
or an
information
processing approach
might
be
appropriate
in
ex-
plaining product
preference.
conditioning
suggests
that
positive
attitudes
towards
an
advertised
product
(conditioned stimulus)
might
develop through
its association
in a
commercial
with
other
stimuli that
are reacted
to
positively
(uncondi-
tioned
stimuli).
Attractive
colors,
pleasant
music,
and
humor are
examples
of
potential
unconditioned
stim-
uli
in a
commercial.
Potentially,
classical
condition-
ing
could,
in
fact,
explain
the effect
of
many
variables
in
communication-attitude
change
situations.
For
ex-
ample,
the communicator
effect
may
to some
extent
be due to
the association
of the attitude
object
with
positive
affect
attached to
the
communicator.
In
marketing,
classical
conditioning
is
often men-
tioned
and
generally
accepted
as a
process
relevant
to
advertising (e.g., Engel, Blackwell, and Kollat 1978;
Schiffman
and
Kanuk
1978).
There
is,
however,
little
empirical
research
on
whether
preferences
for
objects
can
actually
be
classically
conditioned. In
psychol-
ogy,
where classical
conditioning
has
been
investi-
gated
more
extensively,
there is
little
evidence
that
attitudes
can
be
classically
conditioned
(Brewer
1974,
Fishbein
and
Ajzen
1975).
The
limited
popularity
of
classical
conditioning
may
be due to
several
difficul-
ties associated
with
typical
conditioning
experiments.
Journal of
Marketing
Vol. 46 (Winter1982), 94-101.
4 / Journalof Marketing,Winter1982
This content downloaded on Mon, 11 Mar 2013 11:04:59 AMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/11/2019 The Effects of Music in Advertising on Choice Behavoiur
3/9
Difficulties
With
Classical
Conditioning
Demand
Characteristics
Fishbein
and
Ajzen
(1975)
suggest
that
attitude
hange
in
conditioning experiments
may
be a
function
of
the
demand
characteristicsof the
experimental
situation.
They
believe
that
the
subject
must
be
consciously
aware of the presence of the unconditionedstimulus
when the
conditioned
stimulus is
present
in order
for
conditioning
to
take
place.
This
awareness
may
lead
subjects
to
believe that
the
experimenter
wants
them
to
respond
positively
to a
conditioned
stimulus if
it
is
paired
with a
positive
unconditioned
stimulus.
But
awareness of
the conditioned
stimulus/uncondi-
tioned
stimulus
contingency
should
not
automatically
mean that
attitude
change
is
more
likely
the result
of
demand
characteristics than
conditioning.
In
fact,
there is
evidence of
attitude
conditioning
even
where
demand
ype
responses
were
minimized
through
elab-
orate cover
stories
later verified
as
believed
by
sub-
jects (Zanna, Kiesler, and Pilkonis 1970).
Lack of
Awareness
The
consumer
may
not
always
be aware that
the
un-
conditioned
stimuli
in a
commercial
may
affect
his/
her
product
attitudes.
Even
more
generally,
the con-
sumer
may
not
be
aware of the
real forces
impacting
on
both
attitudes
and
behavior. For
example, many
in-store
features
may
influence
consumers
although
they
may
not
be
easily
evoked
in
post-purchase
ex-
planations
of
buying
(amount
of
shelf
space
devoted
to
a
product,
color
of
the
package,
and
so
on).
Nisbett
and
Wilson
(1977)
reviewed
empirical
evi-
dence
casting
doubt on
people's
ability
to
introspect
accurately
regarding
heir
cognitive
processes.'
They
suggest
that
people
often
speculate
about
the
potential
causes
of
their
own
behavior
and
select
stimuli
which
they
think
are
probable
reasons for
their
behavior.
They
may
conclude
that
what
they
think
should
be
the
cause
of
their
behavior
is in
fact
the
cause
of
their
behavior
(Nisbett
and
Wilson
1977).
In
a
typical
communication
situation,
people
may
not
realize or
accept
the
impact
of
unconditioned
tim-
uli
on
their
responses,
since
unconditioned
stimuli
should
not
logically
be
related to
behavior.
Possible
classical conditioning effects might, therefore, be
underestimatedand
underreported
n
self-reports.
In
analyzing
their
behavior,
people
may
attempt
o
con-
vince
both
themselves
and
the
researcher
hat
they
are
strictly
rational.
Cognitive
bias
can
result
from
an
at-
'The
Nisbett
and
Wilson
(1977)
article
is
controversial;
see
both
Ericson
and
Simon
(1980)
and
Weitz
and
Wright
(1979)
for
criti-
cisms.
tempt
to
think
well
of
oneself
(Greenwald
1980).
A
rational
analysis
of
behavior
in a
communication
sit-
uation
might
be
biased
in
favor
of
information
and,
therefore,
of an
information
processing
model.
For
example,
it
may
be more
logical
to
say
that
you
like
the
product
more
after
seeing
a
commercial,
because
you
now
believe it
has
X
characteristic,
rather
than
because
you
liked
the
music
in the
commercial.
Thus
evidence
supporting
classical
conditioning
s
unlikely
to emerge in
self-reports.
Arousal
In
testing
for
classical
conditioning, product
infor-
mation in
the
commercial
mustbe
kept
minimal,
oth-
erwise the
unconditioned
stimuli in
the
commercial
might
merely
be
arousing
interest
in
product
infor-
mation.
Background
features
such
as
sex, music,
color,
and
humor
have
typically
been
treated as
arousal
stimuli
or
stimuli
that
reinforce
the
informa-
tion in the
commercial
rather han
unconditioned
tim-
uli (see Engel, Blackwell, andKollat 1978, for a dis-
cussion
of
this
material).
n
the
print
medium,
Kroeber-
Riel
(1979)
found
that
ads
that
were
arousing
pro-
duced
better
recall
of
the
information in
the
ad.
A
greater
knowledge
of the
product
information
could
presumably
create
more
positive
product
attitudes.
However,
by
minimizing
product
information
in a
commercial
one
can
investigate
the
potential
ability
of
unconditioned
timulito
change
consumer
attitudes
directly.
Mere
Exposure
Versus
Classical
Conditioning
Mere exposure effects can be confoundedwith clas-
sical
conditioning
effects.
While
people
may
develop
favorable
attitudes
owards
products
advertised n
the
context
of
unconditioned
stimuli,
these
attitude
shifts
may
be
simply
a
function
of
mere
exposure
to
the
advertised
product.
Zajonc
(1968)
has
emphasized
that
exposure
leads
to
liking.
Krugman
(1965)
sug-
gested
that
in low
involvement
situations,
people
may
purchase
a
product
simply
because
they
have
been
exposed
to
it
before
through
a
commercial.
It is in-
teresting
to
note that
with
the
Krugman
model,
ex-
posure
leads
to
behavior,
with
attitude
not
concep-
tualized as an intervening variable. With classical
conditioning,
however,
the
person
is
not
choosing
an
object
simply
because
it's
been
seen
before.
His/her
affect
is also
involved,
so
that
if
the
product
(condi-
tioned
stimulus)
is
paired
with
a
negative
uncondi-
tioned
stimulus t
would
be
avoided
despite
exposure.
The
following
experiment
nvestigating
he
impact
of the
background
eatures
of
a
commercial
on
prod-
uct
preferences
allowed
for
a
test
of
mere
exposure
versus
classical
conditioning
effects.
Effects
of
Music
n
Advertising 95
This content downloaded on Mon, 11 Mar 2013 11:04:59 AMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/11/2019 The Effects of Music in Advertising on Choice Behavoiur
4/9
Experiment
One
Basic
Structure
and
Hypothesis
In
the
present
study,
subjects
were
exposed
to a neu-
tral
product
(conditionedstimulus)
in
the context of
a
background
eature
(unconditionedstimulus),
often
found in
commercials.
To minimize the difficulties
mentioned earlier
in
employing
a
conditioning para-
digm, a numberof steps were taken. To lessen de-
mand
characteristics,
an
unobtrusive,
behavioralmea-
sure of
productpreference
was used instead
of
a
paper
and
pencil
measure.
Information
on
the conditioned
stimulus was
kept
as close
to zero as
possible
to dem-
onstrate hat the
unconditioned timulus
could
directly
affect
product preference
even
where
product
infor-
mation
was minimal.
A
negative
unconditioned
stim-
ulus condition
was structured o test
for mere
expo-
sure versus
classical
conditioning.
Consistent
with
a
conditioning
interpretation,
t was
hypothesized
that
subjects
would
prefer
an
unexposed
versus
exposed
product
if
the
exposed
product
were
paired
with a
negative unconditionedstimulus. In contrast, where
the
conditioned
stimulus was
paired
with
a
positive
unconditioned
stimulus,
subjects
would choose the
exposed
rather han
unexposed
product.
Sample
244
undergraduates
andomly
assigned
during regis-
tration to two sections of
a first
year management
course at McGill
University
served as
subjects.
Pilot
In a
pilot,
a
group
of 10
subjects
evaluated
10
dif-
ferentpieces of music on a scale rangingfrom dislike
very
much
(1)
to like
very
much
(5).
The results
of
the
pilot
led
to
the choice
of a
one-minute extract of
music from the movie "Grease" as the
positive
un-
conditioned stimulus
(x=4.3).
One
minute
of
classi-
cal Indian music
(x=
1.5)
served as the
negative
un-
conditionedstimulus.2
Two
different colors of a
pen
were
used as con-
ditioned
stimuli,
light
blue and
beige,
neutral
enough
so
that
associating
them
with liked
or
disliked
music
could
change
color
preferences.
A
pilot
on
23
subjects
had
revealedthat 80%
felt neutralabout
ight
blue and
beige pens,
whereas,
for
example, yellow pens
were
generally disliked, with black pens generally liked.
A
second
sample
(N=41)
of
subjects
were asked di-
rectly
whether
hey
preferred
beige
or
light
blue
pens.
It
was
expected
that if
subjects
were
neutralabout
the
two
colors,
then
half would
choose each.
The
results
2Although
musical
preferences
are
affected
by
prior learning,
music
was
treated as
the
unconditioned
stimulus since it was
not
being
con-
ditioned
in
the
present
experiment.
supported
his
notion,
as 22 out of 41
picked
the
beige
pen,
19 the
light
blue
pen.
Design
and Procedure
The
experiment
was carriedout
during
class time. The
experimenter explained
that an
advertising agency
was
trying
to select
music
(unconditionedstimulus)
to use
in
a commercial
for a
pen
(conditioned
stim-
ulus)
producedby
one of its clients.
They
would
hear
some music that was
being
considered while
they
watched the
slide of the
pen
which the
agency
was
planning
to advertise.
The
pen
was
inexpensive
look-
ing
and cost
forty-nine
cents.
Very
little information
regarding
he
pen's
attributeswas visible in the slide.
The
following
conditions
were structured:
)
con-
dition
LI-Liked
music,
light
blue
pen;
2)
condition
L2-Liked music,
beige pen;
3)
condition
Dl-Disliked
music,
light
blue
pen;
4)
condition
D--Disliked
mu-
sic,
beige
pen.3
While
subjects
had been
randomly
assigned
to the
two sections
of the
course,
a
counter-
balancing procedure
was used to control further
for
any differences that might have existed between the
two sections.
To counterbalance he music and the
color of the
pen
withineach
class,
one-halfof the
first
class was
randomlyassigned
to the liked
music-light
blue
pen
condition
(LI),
while the other
half was as-
signed
to the
disliked
music-beige
pen
condition
(D2).
In
the second
class,
one-half
of the students
were
randomly
assigned
to the
liked
music-beige
pen
condition
(L2),
with the other half
assigned
to the
disliked
music-light
blue
pen
condition
(DI).
The
two conditions
in
each class were
run
sequentially
with one-half
of
the class
taking
a break while the
other half
participated
n
the
project.
The
subjects
in each condition heard the music
while
they
watched
the
slide.
They
then evaluated he
music
on
a
scale
ranging
from dislike
very
much
(1)
to
like
very
much
(5).
They
were told that
they
would
receive either a
light
blue or
beige pen
for their
help,
donated
by
the
company
that
manufactures he
pen.
The
experimenter
held
up
each
pen
briefly
and com-
mentedthat
if
they
wanted
a
light
blue
one,
to
go
over
to the left side of the room to
pick
one
up
and
drop
off
their
question
sheet in a box
provided.
If
they
wanted a
beige
one,
they
were told to
go
to
the
right
side of
the room to
pick
up
the
pen
and
drop
off
their
question sheet. Using two different locations for the
boxes
(150
pens per
box)
of
beige
and blue
pens,
with
question
sheet
drop-off
boxes
next to
them,
allowed
3No condition
containing
both
positive
and
negative
music was
structured,
since it was felt
the
demand
characteristics
would be too
great
under
such
circumstances;
subjects
would
be more
likely
to
think
that
we
expected
them to
pick
the
color of the
pen
associated
with the
positive
music if
both
positive
and
negative
music were
used
in the
same
condition.
96 / Journalf Marketing, inter 982
This content downloaded on Mon, 11 Mar 2013 11:04:59 AMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp -
8/11/2019 The Effects of Music in Advertising on Choice Behavoiur
5/9
for a
relatively
unobtrusive
recording
of choices
and
linking
of these choices
to the evaluation
of
the
mu-
sic.4 As
a result of the music evaluation
ratings, ap-
proximately
10
subjects
were
eliminatedfrom
each
of
the four conditions.
They
were eliminated
if in
the
liked music
conditions,
they
did
not evaluate
the
mu-
sic from
"Grease"
as
either
like
very
much
(5)
or
like
somewhat
(4),
or in
the
disliked music
conditions,
if
they
did not evaluate the Indian
music
as either
dislike
very much (1) or dislike somewhat (2). For the re-
maining subjects,
it was
predicted
that
they
would
pick
the color
of
the
pen they
saw when
they
heard
the liked
music and the
alternative
color
when
they
heard
the disliked
music.
Once out
of
class,
subjects
were
handed the
fol-
lowing
free
response question: "Why
did
you
pick
the
color
of the
pen you
picked
rather than
the
other
color? If
you
have
any
reasons
please
list
them
be-
low."
There
was
room for
thiee
answers.
Finally,
a
discussion
about the natureof the
proj-
ect
was held with
10 students at
the end of class
one
and 10 studentsat the end of class two. Not one men-
tioned that our real
purpose
was to influence
their
particular
olor
preferences.
When told that this
was
in
fact the
purpose,
some
wonderedhow
we were
able
to detect their
preferences,
since
they
did
not
put
the
color of the
pen
they picked
or their names
down
on
the
question
sheet.
They
also did not
see
anyone
re-
cording
their
choices
when
they
picked
the
pens
from
the boxes.
Results
Comparing
he
effect of
liked
versus disliked
music
(Li
and
L2
collapsed
versus
DI
and
D2
collapsed),
there was a very clearcutimpactof the music in the
expected
direction.
As
can be
seen in
Table
1,
74
out
of
94
subjects
(79%)
picked
the
color
of the
pen
as-
sociated with
the liked
music,
while
only
30
out
of
101
subjects
(30%)
picked
the
color of
the
pen
as-
sociated with
the disliked
music.5
When
asked for
possible
reasons
for
their
choice,
126
out of
205
subjects
(62%)
indicated
that
they
did
have
a
particular
reason.
Of
these
126
people,
114
(91%)
mentioned
color
preference
as their
reason.
4A
behavioral rather
than
verbal
measure of
preference
was
used,
in
part,
because it could be recorded
unobtrusively.
In
addition,
be-
havioral
measures of
attitudes
are
generally
preferred
to
verbal
mea-
sures,
although
less
prevalent
in
the
literature
(Carlsmith,
Ellsworth,
and
Aronson
1976).
'The color of
pen
did
not
seem to
matter.
36
out
of
43
people
(84%)
picked
the
light
blue
pen
when it
was
associated
with
the
liked
music
(L,);
38 out
of 51
(74%)
picked
the
beige
rather than
the
light
blue
pen
when
the
beige pen
was
associated
with the
liked
music
(L2).
Similarly,
17
out of 52
(33%)
chose
the
light
blue
pen
when it
was
associated
with
disliked
music
(D,)
while
13
out
of
49
(26%)
picked
the
beige
pen
when it
was
associated
with
the
disliked music
(D2).
These
differences
were
not
statistically
significant.
TABLE1
Liked
Versus
Disliked
Music
and Pen
Choices
Pen Choice
"Advertised"
"Non-advertised"
pen
pen
Liked
music
74
20
94
Disliked
music 30
71
101
104
91
195
x2
=
47.01
(p