A NEW PERSPECTIVE OF SUBLIMINAL PERCEPTION

Ronnie Cuperfain
T.K. Clarke
Ronnie Cuperfain graduated from Dalhousie University in
1982 with Honours in Marketing. After graduation he joined
Reni's Needlecraft Canada Corporation Ltd. He was recently
promoted to Vice-president Finance.
Keith Clarke is Assistant Professor of Marketing. School of
Business Administration, Dalhousie University. He received
his Ph.D. from the University of Illinois. His research interests
include the impact of situational factors both in consumer
behavior and in research methods. His work has been published
in the Journal of the Academy of Marketing Science, the
Canadian Journal of Administrative Sciences, and several
professional conference proceedings.
ABSTRACT
Studies investigating subliminal stimulation have generally
held that such procedures are ineffective in an advertising
context. This article provides a new outlook based on right
brain processing. Evidence is presented that indicates that
subliminal stimulation has impact when thestimulus is selected
based upon the model presented.
Received July 11, 1984. Accepted for publication September
4. 1984.
There has been much written about subliminal
perception but little development of any relevant
theory concerning its manner of operation in an
advertising context. Key (12) claims to find sub-
liminal stimuli in all areas of advertising and feels
that it is an effective means of circumventing con-
sumers' ability to defend themselves against the adver-
tiser's message. Key's work has not found favor
with the research community, the major reason
probably being as stated by Moore (21), "Key appears
to invent whatever features of perception and mem-
ory would be necessary to achieve the results imputed
to embedded stimuli." His work has been more in
keeping with a Freudian approach to explaining the
impact of subliminal stimulation rather than the
more scientifically rigorous physiologically-based ap-
proach reviewed and taken in this paper.
Nevertheless, Key has helped to popularize the
notion of subliminal perception and reports of its
use continue to appear. For instance, the use of
subliminal techniques in offices and retail outlets
have been reported (4). None of these reports have
explained how subliminal perception operates,
only that it does. Though people may not know how
it works, a recent study indicates that a large propor-
tion of the population is now aware of the concept
and believe it to be in widespread, successful use (30).
While there has been some theoretical development
in the area of psychology, little of this theory has
carried over to advertising studies. The majority
of results have indicated that subliminal perception
has no impact in an advertising situation.
In support of this position are many studies show-
ing no impact of subliminal presentation on responses
(6, 3, 24, 10). From this has come the conclusion
that ". . . subliminal directives have not heen shown
to have the power ascribed to them by advocates of
subliminal advertising" (21). As quoted in Weiner
(26), "you can demonstrate the basic phenomenon
of subliminal perception. But you can't use it to
make people buy things." If this is indeed the case,
there is no cause for alarm. If, however, this conclu-
sion has been drawn because researchers have focused
their attention in the wrong areas of subliminal
perception, or have made demands that are greater
than normally expected of even supraliminal adver-
tising, then a problem may indeed exist.
Moore (21) has stated that the effects of subliminal
stimulation can be looked upon within two realms,
the strong claim and the weak claim. The difference
between the two is the type of behavioral response
attributed to the subliminal presentation. The strong
claim predicts direct behavioral consequences of
subliminal perception while the weak claim requires
only a positive affective reaction. Moore's (21)
position appears to be that any response that might
be measured, short of some physical action (as pur-
chasing), is not evidence for the efficacy of subliminal
stimulation. That is, only the strong claim is worthy
of interest. This is an untenable position given that
much of advertising is not directed toward this end.
If much of supraliminal advertising aims at image
creation, preference, or some other objective short
of actual purchase, it is unreasonable to require more
of subliminal presentations. In fact, in the Krugman
(15) sense of low involvement learning, the impact
JOURNAL OF ADVERTISING, Vol . 14, No. 1, 1985
36
may be neither a direct behavior nor an affective
change.
The position of this paper is not that past research
which indicates that supraliminal presentations have
more impact than subliminal ones is incorrect. Rather,
our interest is in whether it is possible to gain some
impact from subliminal stimulation. In this regard,
a small but important body of research tends to be
ignored or passed over.
There have been several studies which have shown
an impact from subliminal stimulation (13, 1, 16, 22).
In these studies, positive results have been achieved
through the use of visual stimuli of a graphic rather
than verbal nature. Klein, Spence, Holt, and Goure-
vitch (13), for instance, used sexuaUy neutral draw-
ings of people and had them reproduced and described
by the subjects. Subliminal presentations of genitals
were made and were found to affect the drawings
and descriptions given. This is a different type of task
than those found in the unsuccessful studies. The
successful studies tended not to rely on the additional
process of converting words into meaning and then
measuring impact. Rather, the visual stimuli had
meaning without the need of the additional step. The
use of words has also been shown to have an impact
under special circumstances.
McNulty, Dockrill and Levy (18) concluded after
using word lists that individuals must be aware of the
structure of a stimulus before any meaning can be
perceived. This emphasis on structure rather than
language may indicate why studies using words have
generally been unsuccessful while those relying on
a more gestalt view have tended to be positive in
results. It may be easier for the brain to derive mean-
ing from a partial picture than from a partial word.
Bricker and Champanis (2) and Weintraub and
McNulty (27) found that repeated exposure increased
the impact of the stimulus. This may again point to
the need to gain enough information to correctly
picture the whole stimulus. This type of structural
research is in contrast to the motivational reasoning
of Key (12). This indicates that subliminal percep-
tion need not be considered some type of black
magic but rather may be successfully investigated
in a scientific manner.
Complaints have been raised that some effects
attributed to subliminal stimulation have been arti-
factually produced (7). The position is often that the
contrived laboratory situation has caused the effects
noted. One may think of these situational influences
as being either a problem to be avoided in such
research or as indicative of circumstances that may
normally occur during, and impact upon the recep-
tion of, subliminal stimuli. That is, external cues may
be appropriate in an advertising context. An example
from the psychology literature using words in a con-
textual situation is the work of Mercel (19). Mercel
found that subliminal words could have an impact as
modifiers. These subliminal words helped to place in
context a supraliminal word that followed and had
an impact on the definition given to the supralimi-
nal word. While it may be more difficult to capture
the structure of a word by itself, the use of a contex-
tual stimulus may be a sufficient cue.
The above findings, indicating a general lack of
results from research requiring the decoding of a
subliminal presentation of words, and more positive
results from research based on the comprehension of
graphic presentation without need of words, suggests
a new theoretical view of subliminal stimulation and
its relevance to the marketplace. This view must also
take into account the need for comprehension of
structure before perception and the importance of
contextual effects with subliminal presentations.
MODEL
Brain research has indicated that there appear
to be two distinct modes of processing information.
Left brain processing appears to be language based
while right brain processing is graphically based (25).
Hansen (8) presents a hierarchial theory of percep-
tion which proposes that the first step in perception
of pictures is a holistic, right brain activity. This
type of activity would appear to be of greater impor-
tance in low involvement conditions. In cases where
pictorial information is an important means of
processing, the type of information presented may
have an impact upon its effectiveness. This indicates
that past research, studying only outputs of subliminal
presentations rather than inputs, has been theoret-
ically deficient.
The necessity for pictorial subliminal presentations
has been reinforced by findings indicating the need
for comprehension of the structure of the stimulus.
Repetition in right brain subliminal presentations
can continue to add information until the brain has
successfully aggregated enough information to pro-
cess the stimulus. This may not be the case for a
stimulus requiring left brain processing where aggre-
gation may be much more difficult. Research has
indicated that the right brain tends to categorize
based on form while the left brain uses function (25).
This may mean that the two sides of the brain have
developed different strategies for the collection of
partial information and "quick looks" may be much
less efficient in left brain processing than in right
brain processing.
The need to make sense of the structure of a sub-
liminal presentation has implications for the type of
stimulus to use in the marketplace and for deciding
what products would most likely benefit from such
37
a presentation. The use of only the name of a product
may not provide sufficient cues for the brain to
subliminally process the stimulus while the presenta-
tion of a picture of the product may be much more
efficient.
At the same time this theoretical approach would
dictate that subliminal presentations would be
likely to have less impact for a new product than
for a product which was familiar in the marketplace.
A viewer would find it easier to form a complete
picture of a known product from a subliminal presen-
tation than to consiiuctthe entire subliminal stimulus
with no prior knowledge.
Another point to be considered in this model is
the context in which a subliminal stimulus is pre-
sented. The research indicating that when people are
in the state of attending to the stimulus there is more
likely to be a subliminal stimulus effect (16) suggests
that context may be very important. This conclusion
is strengthened by research showing results when
a subliminal stimulus is used to help the viewer inter-
pret a supraliminal presentation (19). This leads to
the possibility that subliminal stimulation may have
a greater potential impact (in the sense that con-
sumers would not be aware of any advertising being
directed toward them) within normal programming
on television rather than during commercials. During
regular programming it is likely that viewers will
have a large number of supraliminal contextual
cues that may help in the comprehension of any
subliminal stimulus presentation.
Classical conditioning would also indicate that the
presentation of a subliminal product graphic would
be most successful if paired with a positive product
use situation. This type of situation can easily be
found in normal non-commercial programming,
for example, a couple enjoying a bottle of wine
together coupled with the subliminal presenta-
tion for a brand of wine. If this is indeed possible,
a serious ethical issue exists that must be dealt with
by programming executives, public policy makers,
or both. Additionally, the work of Zajonc (28)
has indicated that mere exposure to a stimulus can
increase affect toward it. Multiple exposures to a
stimulus, once it is comprehended by the viewer,
may indeed provide for a positive gain for the sender.
Not all researchers agree with the mere exposure
hypothesis. Sawyer (23) has found fault with the
original methodology, yet there appears to be a gain
due to frequent exposure that is independent of the
design and measurement problems encountered in
the research (29, 9, 11). Kroeber-Riel (14) presents
the position that multiple exposures to pictorial
presentations may have an impact upon individuals'
attitudes toward products even though the picture
does not provide additional information about the
product. He also holds that this would be most likely
in low-involvement processing situations, such as
television advertising. Mitchell and Olsen (20) have
found empirical evidence to support the view that
pictorial stimulus can have an impact upon consumers,
even if the picture does not provide information
about attributes of the advertised product.
To summarize the theoretical perspective taken
in this paper, past research has lead to the following
model. Subliminal stimulation likely affects indivi-
duals mostly through right brain information proces-
sing. Due to the weakness of such stimuli and the
right brain's holistic and form-based processing,
multiple exposures to an appropriate type of stimulus
should help the viewer to aggregate that stimulus
mto a meaningful gestalt. Since this perspective has
taken the view that this will most effectively be
dealt with by right brain functioning, the form of
the stimulus should be a graphic respresentation of
the product and should be presented to the left
field of vision. Impact will be increased by the use
of brands that are not new to the marketplace.
Finally, impact will be increased by placing the
viewer in the right situation for processing the stimu-
lus. One such situation could be the positive use of
the product category within the normal programs
on television. This need not be the only such situa-
tion where subliminal stimulation may have an
impact however. Kroeher-Riel (14) has suggested that
any positive emotional scenes may impact upon
attitudes through the classical conditioning technique
of repeated exposure. This perspective was the basis
for the present pilot study.
METHOD
A laboratory study was conducted with three
conditions. In all conditions subjects viewed a two
minute "training film." The film was made by a local
retailer of wool and other yarns and concerned the
proper way to wash fine woolen products. Subjects
were volunteer university students. The reason given
for viewing the film was to determine if the film was
perceived as actually teaching how to wash woolens.
During the film the proprietress of the local retail
outlet actually demonstrated the washing of a woolen
sweater. She used a soap in a plain white package.
Subjects were run in small groups ranging from one
to eight. Sample sizes were, for the Zero slide condi-
tion (Z) 16, for the Woolite brand condition (W)
18, and for the control condition (C) 14.
The only difference in the conditions was the
presence or absence of a subliminal presentation.
In two treatments a subliminal message was used,
while the third was a control condition. Both pre-
sentation conditions had five tachistoscopic presen-
tations in the left field of vision of a picture of a
38
product package during the film. The film itself
had only one person on screen, situated in the center
acting as a fixation point for the viewers. Five repe-
titions were selected based upon the work of Wein-
traub and McNulty (27). One subliminal group saw
a slide for Zero, a woolen cleaning product well
known in Maritime Canada and widely available.
The second group was presented with a slide of
Woolite, not widely available at the time and adver-
tised only on cable and not the local network televi-
sion stations.
Each presentation of the subliminal slide lasted
for l/60th of a second. Also, filtering was used to
reduce the intensity of the presentations. When asked,
none of the subjects reported seeing any of the exper-
imental presentations, nor did any of the subjects indi-
cate awareness of the actual intent of the experiment.
After viewing the film, subjects were asked several
questions. Included in tbe questionnaire were two
questions of interest for this researcb. Tbe subjects
were shown pictures of five different soaps and asked
to rank order tbe soaps in terms of wbicb would
be tbe best soap for washing fine washables. Tbe
second question indicated tbat different products
are appropriate in different situations. Subjects were
asked to once again rank order tbe products in terms
of wbicb would be most appropriate in the specific
situation sbown in the film.
Subjects were also asked their opinion about tbe
training quality of the film. Demographic data were
also collected.
RESULTS
In addition to the random assignment of subjects
to experimental treatments, several check variables
were collected to be sure tbat the tbree groups were
comparable in possibly important external conditions.
TABLE 1
DIFFERENCES BETWEEN EXPERIMENTAL
GROUPS ON SELECTED VARIABLES
Variable
F Ratio Significance
Age
Gender
Do you do your own laundry?
How long have you been doing
your own laundry?
Do you consider doing laundry,
(a 5 point scale from hated
task to fun chore)
1.89
1.13
.56
1.60
.056
N.S.
N.S.
N.S.
N.S.
N.S.
Table One presents the results of a one-way analysis
of variance on tbese cbeck variables. As can be seen,
there were no significant differences between the
groups.
Additional measures were taken on viewers' opin-
ions of tbe film in terms of liking, wbetber tbey felt
tbe film bad taugbt tbem bow to wasb fine wasbables,
wbat kind of soap and wbat kind of water was used
in tbe film. Results indicated no differences between
groups on any of these variables.
TABLE 2
MEANS BY TREATMENT CONDITION
Ranking of Zero as
best for fine woolens
Zero
1.56
Ranking of Zero as most
appropriate in present
situation 1.56
Ranking of Woolite as
best for fine woolens 1.63
Ranking of Woolite as
most appropriate in
present situation 1.75
Woolite Control
2.11 1.79
2.06 2.07
1.83 1.64
1.94 1.71
Based upon tbe perspective taken, tbis experiment
was looked upon as three sets of comparisons. Eacb
condition was compared against tbe otber two
conditions. T-tests were conducted on tbe two ques-
tions in each of these comparisons. While the use
of rank order data in parametric tests is somewhat
controversial, tbere is evidence supporting sucb pro-
cedures (5, 17). As tbe model predicts positive
results, one-tailed tests were used. Table Two presents
tbe mean rankings (one indicates most preferred).
Tbe results of tbe T-tests are presented in Table
Tbree.
DISCUSSION
Tbe above results are in general agreement with
tbe model presented above. Subliminal stimulation
can have an impact on stated preference. In all of
tbe measures concerning Zero the ranking of Zero
was bigber in the condition when the subliminal
picture of the Zero package was presented to tbe
subjects tban the otber two conditions (Questions
1, 2, 9, and 10). Tbree of tbese measures were signi-
ficant at tbe .05 level and tbe fourtb at tbe .1 level.
At tbe same time it must be noted tbat none of tbe
measures for Woolite were significant. Tbere are two
39
TABLE 3
T-TESTS BETWEEN EXPERIMENTAL GROUPS
Zero vs. Control
Questions
1) Ranking of Zero as
best for fine woolens 1.56
2) Rankingof Zero as most
appropriate in present
situation 1.81
3) Ranking of Woolite as
best for fine woolens 1.05
4) Ranking of Wooiite as most
most appropriate in present
situation 1.14
Questions
Woolite vs. Control
t
5) Ranking of Zero as
best for fine woolens 1.59
6) Rankingof Zero as most
appropriate in present
situation 1.14
7) Rankingof Woolite as
best for fine woolens 1.06
8) Rankingof Woolite as most
appropriate in present
situation 1,18
Zero vs. Woolite
Questions
9) Ranking of Zero as
best for fine woolens 2.49
10) Rankingof Zero as most
appropriate in present
situation 2.07
11) Rankingof Woolite as
best for fine woolens 1.12
12) Ranking of Woolite as most
appropriate in present
situation 1,34
Significance
<.O5
NJS.
Significance
N.S.
N.S.
N.S.
Significance
<.O5
<.O5
N.S.
N.S.
possible reasons for this. The first, in keeping with
the theory presented here is that Woolite was not well
enough known in the marketplace for viewers to
build a satisfactory impression from the subliminal
stimulus. The second possible explanation is one of
absolute perceptual ability. That is, as intensity of
a stimulus decreases, it will go from supraliminal
to subliminal to no perception of any sort. Since
the intensity of the slides was reduced in the experi-
ment through filtering, it is possible that it was
reduced satisfactorily for the Zero package (a red
color) but reduced below even subliminal perception
for the Woolite package (a lighter colored package).
The above results indicate that it is possible to have
an impact on viewers' stated preferences without
their even being aware that their preferences are being
addressed. This has significant implications for the
necessity of control of subliminal stimulation. In line
with the model, the stimuli were presented within
the context of an action that is in keeping with the
use of the product stimulated. This may mean that
there is greater potential for the improper use of
subliminal stimulation during regular programming
rather than simply during commercial breaks.
The research presented here indicates that aca-
demic marketers may have been too quick to discount
the ability of subliminal presentations to impact
upon consumer decision making. Further research
is needed in several areas. Besides replication of the
present finding of differential impact depending upon
whether the product is known in the marketplace,
additional work should be done with right and left
fields of vision to help determine if subliminal percep-
tion is really a right brain phenomenon. Context
effects should also be studied. Studies along this
line would include looking at the differential impact
of appropriate versus inappropriate supraliminal
presentations while presenting the subliminal stimu-
lus. Finally, further research ought to be conducted
looking toward determining what type of input
stimulus is in fact most effective in subliminal per-
ception. If these studies are not carried out in an
open forum, surely there will be cries that they are in
fact being conducted in secret.
REFERENCES
1. Bressler, J. "Illusion Jn the Case of Subliminal Visual
Stimulation." Journal of General Psychology, 5 (1931) DD
244-50.
2. Bricker, P.D. & A. Champanis. "Do Correctly Received
Tachistiscopic Stimuli Convey Some Information?," Psychol-
ogy Review, 60 (3, 1953), pp. 181-88.
3. Ericksen. C.W., N. Azuma & R.B. Hicks. "Verbal Discrimi-
nation of Pleasant and Unpleasant Stimuli Prior to Specific
Identification," Journal of Abnormal-Social Psychology
59(1959), pp. 114-19.
4. Financial Post Magazine, September 27. 1980, p. 6.
5. Gaito, J. "Scale Classification and Statistics." in R.E, Kirk
{ed.). Statistical Issues, Monterey: Brooks/Cole, 1972.
6. George, S.G. & L.B. Jenning. "Effect of Subliminal Stimuli
on Consumer Behavior: Negative Effect." Perceptual and Motor
Skills, 41 (3, December 1975). pp. 847-54.
7. Goldberg, F,H., & H. Fiss, "Partial Cues and the Phenom-
enon of 'Awareness Without Discrimination'," Perceptual
and Motor Skills, 9 (1959). pp, 243-51.
8. Hansen, Flemming. "Hemispheral Lateralization: Impli-
cations for Understanding Consumer Behavior," Journal of
Consumer Research, 8 (1. 1981) pp, 23-36.
40
9. Heingartner, A. & J.V. Hall. "Affective Consequences in
Adults and Children of Repeated Exposure to Auditory Stim-
uli." Journal of Personality and Social Psychology, 29 (6,
1974), pp. 719-23.
10 Hovsepian. W. & G. Qualman. "Effects of Subliminal
Stimulation on Masculinity-Femininity Ratings of a Male
Model," Perceptual and Motor Skills, 46 (1, February 1978),
pp. 155-61.
11. Kail Jr., R.V. "Familiarity and Attraction to Stimuli:
Developmental Change or Methodological Artifact?," Journal
of Experimental Child Psychology, 18 (1974). pp. 504-11.
12. Key, W.B. Subliminal Seduction, New York: Signet Books,
1973.
13. Klein, G.S.. D.P. Spence, R.R. Holt, & S. Gourevitch.
"Cognition Without Awareness; Subliminal Influences Upon
Conscious Thought," Journal of Abnormal and Social
Psychology, 57 (1958), pp. 255-66.
14. Kroeber-Riel, Werner. "Emotional Product Differentiation
by Classical Conditioning (With Consequences for the 'Low-
Involvement Hierarchy')", in T.C. Kinnear (ed.). Advances
in Consumer Research, 9 (1984), pp. 538-43.
15. Krugman. HE. "The Impact of Television Advertising:
Learning Without Involvement," Public Opinion Quarterly,
29 (Autumn 1965). pp. 349-56.
16 Kunst-Wilson, W. & R.B. Zajonc. "Affective Discrimination
of Stimuli That Cannot Be Recognized," Science. 207 (1,
1980). pp. 557-58.
17. Lord, F.M. "On the Statistical Treatment of Football
Numbers," in R.E. Kirk (ed.), Statistical Issues, Monterey:
Brooks/Cole, 1972.
18. McNulty, J.A,, F.J. Dockrill. & B.A. Levy. "The Sublimi-
nal Perception of Stimulus Meaning." American Journal
of Psychology, 80 (March 1967), pp. 28-40.
19. Mercel, T. 'Conscious and Preconscious Recognition of
Polysemous Words: Locating the Selective Effects of Prior
Verbal Context." in R.S. Nickerson (ed.). Attention and
Performance VII, Hillsdale, NJ: Eartbaum. 1980.
20. Mitchell. A.A. & J.C. Olsen. "Are Product Attribute
Beliefs the Only Mediator of Advertising Effect on Brand
Attitude?." Journal of Marketing Research. 18 (August
1981), pp. 318-32.
21 Moore. T,E. "Subliminal Advertising: What You See Is
What You'Get." Journal of Marketing, 46 (2. Spring 1982).
pp. 38-47.
22. O'Grady. M, "Effect of Subliminal Pictorial Stimulation
on Skin Resistance." Perceptual and Motor Skills, 44 (3, Pt. 2.
June 1977). pp. 1051-56.
23. Sawyer, Alan G. "Demand Artifacts in Laboratory Experi-
ments in Consumer Research," Journa/ of Consumer Research,
I (1975). pp. 20-30.
24. Severance, L.J. & F.N. Dyer. "Failure of Subliminal
Word Presentations to Generate Interference to Color Nam-
ing." Journal of Experimental Psychology, 101 (1. 1973).
pp. 186-89.
25. Springer. S.P,. & G. Deutch. Left Brain Right Brain,
San Francisco: W.H. Freeman and Co., 1981.
26. Weiner, A. "The World Beyond Words." Financial Post
Magazine, September 1973, pp. 12-20.
27. Weintraub. D.J. & J.A. McNulty. "Clarity versus Identifi-
ability of Repeatedly Flashed Patterns," Journal of Experi-
mental Psychology, 99 (3. 1973). pp. 293-305.
28. Zajonc, R.B. "Attitudinal Effects of Mere Exposure,"
Journal of Personality and Social Psychology Monograph
Supplement. 9 (1968. part 2), pp. 1-27.
29. Zajonc. R,B., W.C. Swap. A.A. Harrison, & P. Roberts.
"Limiting Conditions of the Exposure Effect: Satiation and
Relativity." Journal of Personality and Social Psychology,
18(3. 1971). pp. 384-91.
30, Zanot. Eric J., J. David Pincus. & E. Joseph Lamp. "Public
Perceptions of Subliminal Advertising," Journal of Adver-
tising, .12 (I, 1983). pp. 39-45.
The American Academy of Advertising
(AAA) is an organization of advertising pro-
fessionals and scholars with an interest in
advertising and education. If you are an adver-
tising professional or an academic in Advertising
or a related field, the Academy
INVITES YOUR MEMBERSHIP !
Membership dues are $25 per year (students
$12.50) and include a subscription to the
Journal of Advertising.
Send a check, payable to the American Academy
of Advertising, and your name, address or
business card t o:
Professor H. Keith Hunt
Executive Secretary
632 TNRB, Graduate School of Management
Brigham Young University
Provo, UT 84602
41