You are on page 1of 9

Effects of Brand Awareness on Choice for a Common, Repeat-Purchase Product

Author(s): Wayne D. Hoyer and Steven P. Brown

Source: The Journal of Consumer Research, Vol. 17, No. 2 (Sep., 1990), pp. 141-148
Published by: The University of Chicago Press
Stable URL:
Accessed: 29/09/2010 17:52

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at 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

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

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to The
Journal of Consumer Research.
Effects of Brand Awareness on Choice for a
Common, Product


Results of a controlled experiment on the role of brand awareness in the consumer

choice process showed that brand awareness was a dominant choice heuristic
among awareness-group subjects. Subjects with no brand awareness tended to
sample more brands and selected the high-quality brand on the final choice signifi-
cantly more often than those with brand awareness. Thus, when quality differences
exist among competing brands, consumers may "pay a price" for employing simple
choice heuristics such as brand awareness in the interest of economizing time and
effort. However, building brand awareness is a viable strategy for advertising aimed
at increasing brand-choice probabilities.

One of the majorgoals of advertisingin situations least, recognition of the brand name. Awarenessrep-
of low interestor involvement is to generateand resents the lowest end of a continuum of brand
maintain brand awareness (McMahon 1980). Ac- knowledge that rangesfrom simple recognition of the
cording to Bogart(1986, p. 208), advertisersuse repe- brand name to a highly developed cognitive structure
tition to impress "the advertised name upon the con- based on detailed information. Recognition is taken
sumers' consciousness and make them feel comfort- here to be the process of perceiving a brand as pre-
able with the brand." Presumably, advertisersexpect viously encountered (Mandler 1980). Thus, the dis-
that awarenesswill keep the brand in the consumer's tinction between awareness and recognition is a sub-
evoked set, therebyincreasingthe probabilitythat the tle one, the former denoting a state of knowledge pos-
brand will subsequentlybe purchased. sessed by the consumer and the latter a cognitive
However, researchfindings on this issue have been process resulting from awareness.
mixed. Some studies have demonstrated a positive
correlation between advertising and sales (e.g., Bass Effects of Brand Identification
and Clarke 1972; Bass and Leone 1983). Others have
found no relationship(Bogart 1986). Partof the prob- Previous researchhas consistently shown that con-
lem is that these studies are purely correlational ex- sumers in blind taste tests are unable to detect their
aminations at an aggregate level. Studies that pin- own preferredbrand. Allison and Uhl (1964) had beer
point the impact of brand awareness on the individu- drinkers with well-established preferences rate sev-
al-level choice process are badly needed. This article eral brands in a blind taste test and again when the
takes a first step towardfilling this void by examining brands were identified. They found that the beer
the nature of brand-awarenesseffects on the purchase drinkers tended to rate the taste of their preferred
of a common household product. brand significantly higher when it was identified than
they did in the blind taste test. Moreover, the beer
drinkers were unable to distinguish their preferred
BACKGROUND LITERATURE brand from the others tasted in the blind taste test.
In this research,brandawarenessis defined as a ru- However, this study did not investigate questions re-
dimentary level of brandknowledge involving, at the lated to choice and sampling of brands.
The present study explores the effects of brand
awareness on choice, brand sampling, and the fre-
"Wayne D. Hoyer is Zale Centennial Fellow in retail merchan-
dising and associate professor of marketing at the University of quency with which the highest-quality brand is cho-
Texas at Austin, Austin, TX 78712, and Steven P. Brown is assis- sen following a series of trials. Furthermore,the con-
tant professor of marketing and distribution at the University of sumers in the Allison and Uhl study were highly expe-
Georgia, Athens, GA 30602. The authors would like to thank rienced (drankbeer at least three times a week). In an
Douglas M. Stayman and the reviewers for their insightful com-
ments on earlier drafts of this article.
attempt to isolate the effect of simple brand aware-
ness, the present study employs novice consum-
? 1990 by JOURNAL OF CONSUMER RESEARCH, Inc. * Vol. 17 0 September 1990
All rightsreserved.0093-5301/91/1702-0003$02.00

ers with no prior purchase experience with the test stimuli tend to be better liked than unfamiliar ones,
product. even in the absence of recognition. The results of Za-
jonc and his colleagues, which have been extended in
Effectsof Prior Exposure of Stimuli consumer research by Janiszewski (1989), suggest
that familiarity leads to greater liking, even without
Very little empirical work has been done on the the mediation of conscious awareness. It might thus
effects of brand awareness on consumer choice. A be arguedthat the effects of awarenesson choice can-
considerably greater amount of work has examined not be separated from those of affect. Although it is
the effects of more elaborate knowledge structures clear that attempting to separate affect from aware-
(see Alba and Hutchinson 1987). A growing body of ness requiresextremely tight internal laboratorycon-
evidence, however, suggests that the consumer in trols, such experimental controls were not feasible in
many purchase situations is, at best, a passive recipi- this study of consumer choice and brand sampling.
ent of product information and one who tends to The questions of interest in this study are how brand
spend minimal time and cognitive effort in choosing awareness affects choice probability and sampling of
among brands (Hoyer 1984). In situations involving a common, repeat-purchase product under varying
common, repeat-purchase products, the consumer conditions of brand quality and awareness. Recent
may choose a brand on the basis of a simple heuristic work by Smith and Swinyard (1983) suggests that
(e.g., brand awareness, pricing, packaging) and then consumers tend not to form strong brand affect until
evaluate the brand subsequent to purchase(Ray et al. after a product trial experience. Hence, steps were un-
1973). In such instances, awareness results mainly dertaken to ensure that subjects in the study to be de-
from exposure to advertisingand other sources of in- scribed had not actually purchased or consumed the
formation. brands used as experimental stimuli.
Although prior experimental research has not spe- One objective of this study was to investigate the
cifically examined the effects of brand awareness on extent to which a simple heuristic based on aware-
consumer choice, several lines of investigation pro- ness, such as "buy the best known brand," was uti-
vide reason to expect a positive relationship. Several lized in a simple choice task. Specifically, the follow-
streams of research have reported perception of fa- ing hypothesis is offered.
miliar stimuli to be associated with positive affect. HI: Brand awareness serves as a dominant
For example, Titchener (1912, p. 408) comments, choice tactic among inexperienced consum-
"What . . . is the feeling [i.e., that experienced upon ers presented with a brand-selectiontask.
recognition]? In experiments upon recognition it is
variously reported as a glow of warmth, a sense of The awareness heuristic should be particularlyim-
ownership, a feeling of intimacy, a sense of being at portant in the first purchase of a product. In accor-
home, a feeling of ease, a comfortable feeling. It is a dance with the model of Ray et al. (1973), who pos-
feeling pleasurable in its affective quality, diffusively ited a hierarchy of effects when involvement is low,
organic in its sensory character." the first purchase of the product should be followed
According to Hasher and Zacks (1984), an auto- by trial and evaluation. Subsequent purchases from
matic frequency-counting mechanism records rela- the product category should be based on the outcome
tive frequency information regarding the instantia- of the trial-and-evaluation process (Smith and Swin-
tion of various phenomena. This relative frequency yard 1983). Thus, the influence of awareness as a
information can be used as the basis for making infer- choice tactic should decline on choices subsequent to
ences regarding product quality (Baker et al. 1986). the first. The known brand may be chosen on the first
For example, if the automatic frequency-counting occasion because of a belief that it is probably the
mechanism counts substantially more instances of best. Then, if the experience is judged satisfactory,the
communications about brandA than about brand X, decision heuristic may shift to "buy the brand I
then an inference may be made to the effect that bought last time because it was satisfactory." Under
brand A is better known, so it must be popular and these circumstances, consumers will have little moti-
probablybetter (Baker et al. 1986; Hasher and Zacks vation to sample unknown brands and will be un-
1984). likely to discover any quality differencesthat may ex-
Berlyne(1970) argues that novel stimuli tend to be ist between competing brands.
highly arousing and trigger aversive reactions. As a When the consumer has no prior knowledge or
person gains familiarity with a stimulus through re- awareness of any brand in a choice set, simple and
peated exposure, however, positive affect and an ap- reliable choice tactics may be less immediately avail-
proachtendency form. Perceivedrisk tends to decline able. This would most likely result in a more effortful
and positive affecttends to increase with repeated ex- decision-making process, but one that would perhaps
posure (Baker et al. 1986; Obermiller 1985). be relatively free of any biases and distortions caused
Research by Zajonc (1980) and his colleagues on by use of a simple heuristic in the interest of cognitive
the mere exposure effect also indicates that familiar economy (Hogarth 1980; Tversky and Kahneman

1974). With a simple rule such as "buy the best- in the pool, subjects had to have never purchasedany
known brand" unavailable, the decision maker must of the brands used in the experiment.
resort to other selection criteria. By providing an op-
portunity to sample among brands across several se- Test Product
lection experiences, this study tested the prediction
that subjects who cannot rely on the awareness heu- Several factors led to the selection of peanut butter
ristic will take greater advantage of the opportunity as the test product. A pretest study revealed that
to sample. many freshmen college students had not purchased
the product before. Also, the product category con-
H2: Consumers choosing among a set of un- tained a number of well-known, easily identifiable
known brands are likely to sample more brands and several lesser-known, unadvertised
brandsacross product trials than consumers brands. This enabled a convenient manipulation of
who choose among a set of brands that in- the awareness construct. Additionally, the pretest
cludes one well-known one. study showed wide variations in perceived quality
When no known brand is available in a choice set across differentbrands of peanut butter, which facili-
and consumers are given an opportunity to sample tated manipulation of product quality. Finally, pea-
brands, their perceptions of quality are likely to be nut butter can be tried or tasted quite easily in an ex-
unaffected by the biases and distortions that previous perimental situation, permitting an assessment of the
exposure to the brand through advertising or word of influence of awarenesson brand evaluations.
mouth may create (Deighton 1984; Hoch and
Deighton 1989; Hoch and Ha 1986). Thus, they may Independent Variables
be more likely to detect true quality differences
Awareness. Awareness was operationalized as a
among brandsand select the highest-qualitybrand on two-level blocking factor consisting of awarenessand
a final choice following a series of trials.
no-awarenessconditions. In the awarenesscondition,
H3: Aftera series of product choices, consumers subjects were presented with three brands of peanut
choosing among a set of three totally un- butter. The first of these was a well-known national
known brands are more likely to choose the brand that had been advertised heavily. A pretest
high-qualitybrand than are consumers who demonstrated high recognition for this brand. Two
choose among a set of brands that includes unknown brands from other regions of the country
one well-known and two unknown brands, completed the three-brandset. To be included in the
especiallywhen the well-known brand is not awareness condition, subjects had to be aware of the
the high-qualitybrand. well-known brand (i.e., through advertising)without
ever having purchased or tried it (as determined by
a pretest questionnaire). They also had to exhibit no
METHOD awarenessof the other two brands in the set. 1
In the no-awareness condition, subjects were pre-
The hypotheses were tested with an experimental sented three totally unknown brands. Two of these
procedurethat asked subjects to make a series of deci- brands were the same as those utilized in the aware-
sions regardinga brand choice for peanut butter. Af- ness condition. The third was another unknown
ter each selection, subjectswere asked why they chose brand from another region of the country. To be in-
a particularbrandand then were permitted to taste it. cluded in the no-awarenesscondition, subjects had to
Following each trial, subjects answered several short be totally unfamiliar with all three brands in the set.
diagnostic questions. After five trials, a series of prod- Thus, these subjects faced the task of choosing among
uct-usage and experiencequestions were presentedto and evaluating a set of brands that were totally unfa-
complete the session. miliar to them.
Subjects Quality. To identify high- and low-quality
brands, eight brands of peanut butter were evaluated
Assessing the effects of brand awareness in a novel
choice context requiredfindingexperimental subjects
'Price of the brands was also manipulated to enhance the realism
with no previous purchase or usage experience with of the choice task. For example, one of the brands was marked with
the test brands. Therefore,freshman college students a higher price than the other two ($1.79 vs. $1.29). A pretest indi-
at a large southwestern university (N = 173) who cated that this difference was perceived to be significant. A counter-
were inexperienced in the purchase of peanut butter balancing procedure determined which brand was marked as high
participated in the experiment. A majority of these priced, so that each brand (within both awareness and no-aware-
ness conditions) was marked as high priced approximately the same
subjectshad never purchasedpeanut butter for them- number of times. This manipulation had little effect on subjects'
selves, and the rest indicated they had purchasedpea- choices and was orthogonal to the other manipulations. Hence, it
nut butter only a few times at most. To be included is not discussed further.

by a group of pretest subjects (N = 62) in a blind taste the present research, subjects had little reason to be-
test. These brands ranged from presumably high- lieve that accurate self-reportswould cause them any
quality brands (e.g., well-known national brands) to embarrassmentor have any other detrimental effect
those of presumablylower quality (e.g., store and ge- and should have had no reason to intentionally mis-
neric brands). Pretest subjects rated the brands on a lead the experimenter. Thus, while self-reports may
five-point scale anchored by "very low quality" and not be an ideal measure for capturing precise choice
"very high quality." tactics, they have proved useful in describing con-
Two brandswere selected for the quality manipula- sumer decision-making processes (Bettman 1971;
tion based on the pretest results. The brand with the Rip 1980; Wrightand Rip 1980) and are particularly
highest pretest rating (a well-known national brand, useful when supported by converging evidence from
X = 4.09) and the brand with the second lowest rating other measures.
(a generic brand,X = 2.32) were selected to represent
high- and low-quality brands, respectively. Testing Number of Brands Sampled. The number of
indicated a significantdifferencein perceived quality brands sampled was measured by counting the num-
between these two brands (t = 10.73, p < .00 1). The
ber of brands from a set of three alternativesthat sub-
lowest-ratedbrand in the pretest was not selected be- jects sampled across a series of five trials.
cause its extremely low rating raised apprehension Choice of Quality Brand. Choice of quality brand
that its use would make the manipulation trans- was operationalized as whether a subject's final
parent. choice following a series of five trials was actually the
The two brands selected were then used to create high-quality brand. A "hit" was recorded if the sub-
two levels of the quality manipulation. In each condi- ject finally chose the brand that contained the quality
tion, subjects were presented with a set of three peanut butter, and a "miss" was recorded if the sub-
different brands of peanut butter. In the quality- ject finally chose a brand other than the one that con-
differencecondition, one of the brands actually con- tained the high-quality peanut butter. This analysis
tained the high-quality brand, while the other two involved only subjects in the quality-differencecon-
brands contained the low-quality brand. A counter- dition (N = 88).
balancing procedure ensured that each of the three
brandscontained the quality brand an approximately Procedure
equal number of times. This was accomplished by al-
tering the contents of the jars. Thus, the well-known Experimentalsessions were conducted individually
brand'sjar did not always contain the high-quality for each subject. Upon arriving, they read a short in-
peanut butter;rather,two-thirdsof the time, the high- troduction to the experiment and then were told that
quality peanut butter was contained in the jars of the they would be asked to make a choice between differ-
unknown brands. ent brands of peanut butter. They were asked to pre-
In the no-quality-differencecondition, all threejars tend that they were in a grocerystore and to make the
contained the low-quality peanut butter. This group choice exactly as they would if they were in the store.
served as a control against which to compare the re- The experimenter emphasized that there were no
sults of the quality-differencecondition to assess the right or wrong answers and that there was no one
influence of brand awarenesson quality judgments in brand that the experimenter wanted them to choose.
a situation in which no actual quality differencesex- Rather, all that was asked was that they make a choice
isted. of peanut butter.
Subjectswere then led over to the peanut butter dis-
Dependent Variables play. The display consisted of a shelf on which the
three brands of peanut butter were positioned (four
Choice Tactics. To assess the extent to which the jars of each brand). In the awareness condition, the
awarenessheuristic was utilized in making an initial well-known brandand two unknown brandswere dis-
choice, subjects were asked, "Can you tell me why played; in the no-awareness condition, the three un-
you selected the brand you have chosen?" The ques- known brands were displayed. Subjects were then
tion was intended to elicit immediate postdecision re- told to take as little or as much time as they wanted
sponses regarding why a particular brand was se- and to select whichever brand they desired. After se-
lected. In accordance with Hoyer (1984), it was lecting a brand, subjects were immediately asked to
strictly "freeresponse," with no probes, to allow sub- indicate which brand they had chosen and why. The
jects to describetheir choices in their own terms. experimentertook a sample of peanut butter from the
Although Nisbett and Wilson (1977) have argued designated jar, placed it on a cracker, and gave it to
stronglyagainstthe ability of subjectsto give accurate the subject to taste. Subjects were asked to repeat the
reasons for judgments and choices in experimental selection and sampling process four more times. (A
tasks, Wrightand Rip ( 1980) and Rip ( 1980) have ar- pretest of the experimental procedure had demon-
gued just as strongly for the value of self-reports. In stratedthat five trials were sufficientto allow subjects

to make a final selection.) Subjects were provided when a clear distinction between brandsexists on this
with drinkingwaterbetween selections. dimension. However, when no brands are known,
To conclude the session, subjects completed a se- other criteria such as packaging, ingredients, and
ries of posttask questions. These included items de- price (in combination with other factors) are likely to
signed as checks on the "purity" of the sample: (1) be employed.
whether subjects had ever purchased peanut butter, The results reported in Table 1 also reveal differ-
(2) whetherthey had heard of any of the brands prior ences in choice criteria between awareness-condition
to the experiment, and (3) whether their families had and no-awareness-condition subjects for the final
ever used any of the brands in the experiment. Also, choice. A significantly higher proportion of subjects
to check for possible demand characteristics,subjects in the awareness condition continued to report
were asked what they thought was the purpose of the awarenessas the basis for their decision ( 17.8 percent
study. None indicated a suspicion that the study con- vs. 5.4 percent, Z = -2.76, p < .05). The fact that 17.8
cerned the effects of brand awareness. Subjects were percent of the subjects reported basing their decision
then debriefedand thanked for their participation in on awarenessaftertrying other brandssuggestsits im-
the experiment. portance in brand-choice decisions among inexperi-
enced decision makers. It also suggests that the in-
RESULTS fluence of awareness may persist across a series. of
choices and trials. Another interesting result is that
Hypothesis 1 no-awareness-condition subjects reported basing
Hypothesis 1 held that when brand awareness is their final decision on taste significantly more often
present in an initial choice task subjects will be sig- than awareness-condition subjects (62.4 percent vs.
nificantly more likely to base their decision on brand 41.1 percent, Z = 3.04, p < .01).
awareness.This hypothesis was examined by observ- Some noteworthy changes occurredin the reported
ing actual choices and by eliciting subjects' free re- choice tactics between the first and the final trial. For
sponses about their choice strategies. Subjects'initial the awarenessgroup, use of awarenessas a choice cri-
choices (in a series of five) provide strong support for terion decreased substantially (82.2 percent vs. 33.3
the hypothesis. Specifically, 93.5 percent of the sub- percent, Z = 9.83, p < .00 1). Thus, although aware-
jects in the awarenesscondition selected the familiar ness remained an important choice criterion, other
brand, with only 1.1 percent choosing unknown factors such as taste became more important as sub-
brand A and 5.4 percent choosing unknown brand B. jects gained experience with the decision task. The
Thus, it appears that when inexperienced decision most notable change in tactics for the no-awareness-
makers are faced with a choice situation in which a condition subjectswas a largeincrease in the reported
known brand competes with unknown brands they use of taste as a decision criterion on the final choice.
are considerably more likely to choose the known On the final choice, 62.4 percent of the no-awareness
brand. subjects reportedtaste as the basis for their final selec-
To directly assess the extent to which subjects used tion. The reported use of package and of ingredients
awareness as a choice heuristic, they were asked to as decision criteria by the no-awareness subjects de-
give their reasons for selecting a particularbrand fol- clined to zero by the final trial. Thus, when no aware-
lowing their first trial and their final choice. Table 1 ness exists initially, subjectsreportgreaterreliance on
presents the results of the open-ended question re- taste perceptions than they do when brand awareness
gardingchoice criteriafor the two points in time. The is present.
table shows substantial reported reliance on aware-
ness as a choice tactic by subjects in the awareness Hypothesis 2
condition on the first trial, with 60 percent reporting
use of this tactic. Another 22 percent reportedbasing Hypothesis 2 predicted that subjects in the no-
their decision on a combination of awareness and awareness condition would tend to sample more
some other tactic. The remaining portion of aware- brandsacross a series of five trials than subjects in the
ness-condition subjects reported using a variety of awareness condition. This hypothesis would be sup-
other evaluative criteria. ported by a significant main effect of awarenessin the
By contrast, the largestproportion of no-awareness two-way awareness-by-qualityANOVA. A significant
subjects (45.2 percent)reporteda choice based on lik- main effect for awarenessdoes in fact emerge from the
ing of the package,while the second-largestgroup em- analysis (F(3,169) = 7.87, p < .01). Neither the main
ployed a combination of price and some other crite- effect of quality nor the two-way interaction ap-
rion (14 percent). Another 10.8 percent of the no- proached significance. This result supports Hypothe-
awareness subjects reportedbasing their decision on sis 2. Comparisonof the mean number of brandssam-
the ingredients. pled between awarenessand no-awarenessconditions
Overall, these results suggestthat first-time buyers (X = 2.67 vs. X = 2.29, t = 3.49, p < .001) indicates
may rely on awareness as a cue for choosing a brand that subjects in the no-awarenesscondition tended to


First selection Final selection

No awareness Awareness No awareness Awareness

Criterion (%) (%) z (%) (%) z

Known brand 0 60.0 11.11** 5.4 17.8 2.76*

Taste texture 4.3 0 NS 62.4 41.1 3.04*
Lowest price 2.2 0 NS 3.2 1.1 NS
Ingredients 10.8 3.3 2.16* 0 1.1 NS
Package 45.2 4.4 6.95* 0 3.3 NS
Try new brand 1.1 0 NS 4.3 7.8 NS
Known brand and taste 0 3.3 NS 6.5 13.3 NS
Known brand and other 0 18.9 4.41 0 2.2 NS
Price and taste 1.0 0 NS 5.4 3.3 NS
Price and other 14.0 4.4 2.63* 1.1 3.3 NS
Other 21.4 5.7 ... 11.7 5.7 ...

NOTE.-NO awareness, n = 90; awareness,n = 83.

p< .05.
**p <.01.

sample more brands than those in the awarenesscon- brand jar contained the high-quality peanut butter
dition. are excluded, the known brand was still selected 73.3
percent of the time (22/30).
Hypothesis 3 When awarenesswas present but no quality differ-
ences existed between brands, 74.5 percent (35/47) of
Hypothesis 3 predicted that when quality differ- the subjects selected the known brand, with 17 per-
ences existed between brands in a choice set subjects cent (8/47) selecting unknown brand A and 8.5 per-
in the no-awarenesscondition would choose the qual- cent (4/47) choosing unknown brand B. These results
ity brand on the final selection more often than sub- provide furtherevidence of the effect of brand aware-
jects in the awareness condition, especially when the ness on choice.
well-known brand was not the high-quality brand. Of Thus, the relative quality of the known brand vis-
the subjects in the no-awareness/quality-difference a-vis the unknown brand appeared to have little in-
condition, 59 percent (26/44) selected the high-qual- fluence on the final choices of awareness-condition
ity brand on the final trial. This proportion signifi- subjects. Whetherthe quality of the known brandwas
cantly exceeds the 33 percent hit rate that would be the same as, better than, or worse than the unknown
expected owing to chance alone. brands,it was still chosen by a substantial majority of
In the awareness/quality-differencecondition, only subjects,even afterthey had sampled other brands.In
41 percent (18/44) of the subjects chose the high- the control condition (i.e., when no quality differ-
quality brand. Of this proportion of "hits," however, ence existed), 74.5 percent (35/47) of the subjects
a large majority (67 percent) were attributable to se- chose the known brand. When quality differencesdid
lections of the known brand when it actually con- exist and the known brandwas actually the high-qual-
tainedthe high-quality peanut butter. When the high- ity brand, awareness-condition subjects chose the
quality peanut butter was in the jar of an unknown known brand 85.6 percent of the time (12/14). When
brand(which it was on two-thirds of the trials), only quality differencesexisted and the known brand was
20 percent(6/30) of the awareness-conditionsubjects not the high-qualitybrand, awareness-conditionsub-
selectedthe brand containing the high-quality peanut jects chose the known brand 73.3 percent of the time
butter. Seventy-three percent of them (22/30) chose (22/30).
the well-knownbrand, even though it did not contain It was also possible that prior usage of the known
the high-qualityproduct. Thus, awareness-condition brand by subjects' families might have affected the
subjects were considerably less likely to choose the observedresults. To assessthis possibility, an analysis
high-quality brand, especially when an unknown was conducted to compare the choice results of those
brand'sjar contained the high-qualityproduct. awareness-conditionsubjects who reportedthat their
To further emphasize the point, across all trials families had previously used the known brand with
within the awareness/quality-differencecondition, those of awareness-condition subjects who reported
the known brandwas chosen 77.3 percent of the time no such previous usage. Subjects whose families had
(34/44). When the trials during which the known previously used the known brand selected the high-

quality brand only 29 percent of the time compared awarenessand in part may have accounted for the ob-
to 50 percent among those whose families had never served pattern of results. For instance, although an
used the known brand. Apparently, then, previous attempt was made to control for initial brand affect,
brand usage affectedchoice of the high-quality brand it was allowed to vary freely over the course of the five
to an even greaterextent (i.e., was even more biasing) trials. Also, it is possible that the choice tactics used
than did simple brand awareness. by subjects may have been subject to some desirable
responding or self-presentation bias. However, the
DISCUSSION measuresused in this study should not have been any
more prone to these types of distortions than those
Overall, the results indicate that brand awareness most commonly employed in studies of consumer
may have considerable effect on consumer choice. choice. Finally, only a single product category, pea-
Subjectswho were awareof one brand in a set of three nut butter, was investigated. It remains to be seen
sampled fewer brands over a series of four trials and whether the findings can be replicated in other prod-
were considerablyless likely to select the high-quality uct categories.
brand on a final choice than subjects who were not
awareof any of the brands in the set. When awareness CONCLUSIONS
was presentin an initial choice situation, subjects re-
portedusing it as a decision criterion in a high propor- Despite these limitations, the study demonstrates
tion of cases. After subjects had gained some experi- that brand awareness exerts an influence on choice
ence with the choice task and had an opportunity to and brand sampling. In particular, it has shown that
sample several brands, the use of awareness as a (1) brand awareness is a prevalent choice tactic
choice tactic declined in importance. A significantly among inexperienced consumers facing a new deci-
greater proportion of subjects in the no-awareness sion task, (2) subjects who are aware of one brand in
condition reportedusing taste as a choice tactic on the a choice set tend to sample fewer brands across a se-
final trial than did those in the awareness condition. ries of product trials, and (3) subjects who are aware
Concomitantly, perceived product quality increased of one brand in a choice set tend to choose the known
in importance as a choice tactic as novice consumers brand even when it is lower in quality than other
gained more familiaritywith the task. The behavioral brandsthey have had the opportunity to sample.
measureof observedbrand choice confirmedthe self- Other possible effects of brand awarenessmerit fu-
reporteduse of awarenessas a choice tactic, as a large ture investigation. The results of this study suggest
majority of subjects in the awareness condition se- that brand awarenessmay triggerdifferencesin infor-
lected the known brand on the final trial. mation processing. In particular, it appears possible
It would appear from these results that brand that the use of an awarenessheuristic may be associ-
awarenessprovides a convenient cue for choice, but ated with a top-down, concept-driven mode of infor-
that its use as a simplifying heuristic is not without mation processing, whereas the absence a known
cost to the consumer. Use of brand awareness as a brand in the choice set may evoke an inductive pro-
choice tactic appearedto introduce a bias into the se- cess of information search and trial. The top-down,
lections of the awareness subjects that was not ob- concept-driven mode may reflect consumer hypothe-
served in those of the no-awarenesssubjects. The lat- sis testing of the type described by Deighton (1984),
ter appeared to engage in a somewhat more effortful Hoch and Ha (1986), and Hoch and Deighton (1989).
selection process characterized by more intensive Research utilizing process measures to probe for
brand sampling. Without reliance on an awareness these types of differencescould make interesting and
heuristic, these subjects more often chose the high- valuable contributions.
quality brand on their final selection. These results If, as appearslikely, prior exposure to brand infor-
suggestthat the presenceof a known brandin a choice mation throughadvertisingor other sources resultsin
set may have some negative effects on the consumer's the formation of a brand-relatedschema, it would be
ability to detect differencesin product quality across interesting to explore the effects of this schema on
brands. Moreover, these effects were observed after consumer expectations regardingthe brand.A closely
subjects had an opportunity to sample and compare related question would be how these schema-based
severalbrands, suggestingthat awarenessnot only in- expectations affectbrandevaluations and satisfaction
fluences the first choice, but that it may also affect judgments. Such research might take a step toward
choice on subsequent selections. integratedresearchon brand awarenesswith the long
tradition of researchon consumer satisfaction.
Limitations The results of the present study suggest that con-
sumers' choice tactics in purchase of a common, re-
The fact that awarenesswas a measuredratherthan peat-purchaseproducttend to change with increasing
manipulated variable leaves open the possibility that experience in making selections. It may be of interest
other psychological variables were correlated with in futureresearchto furtherexplore changes in choice

cues over time to determine whether intrinsic attri- Hasher, Lynn and Rose T. Zacks (1984), "Automatic Pro-
butes take on increasingimportance as experience in- cessing of FundamentalInformation:The Case of Fre-
creases. Also, studies providing stronger manipula- quency of Occurrence," American Psychologist, 39
tions of consumer motivation to detect quality (12), 1372-1388.
Hoch, Stephen J. and John Deighton (1989), "Managing
differences across brands may reveal additional in- What Consumers Learn from Experience,"Journal of
sights into the strength of effects of brand awareness. Marketing,53 (April), 1-21.
It would be interestingto know how strong such moti- and Young-Won Ha (1986), "Consumer Learning:
vation has to be to override the type of brand-aware- Advertising and the Ambiguity of Product Experi-
ness effectthat has been demonstrated in this study. ence," Journal of Consumer Research, 13 (October),
In conclusion, this study has presented empirical 221-233.
resultssuggestingthat brandawarenesshas important Hogarth, Robin (1980), Judgment and Choice. The Psy-
effects on consumer choice. The data suggest that ad- chology of Decision, New York: Wiley.
vertising targetedat increasing brand awareness may Hoyer, Wayne D. (1984), "An Examination of Consumer
be effective in increasing choice probabilities. Fur- Decision Making for a Common Repeat Product,"
Journal of ConsumerResearch, 11 (December), 822-
ther, the results suggest that the effects of awareness 829.
on choice may persist beyond the consumer's first Janiszewski, Chris (1988), "Preconscious Processing Ef-
choice from a product category. Although the present fects: The Independence of Attitude Formation and
study has shown the presence of brand-awareness Conscious Thought," Journal of ConsumerResearch,
effects on choice, the task of describing the processes 15 (September), 199-209.
by which these effects operate remains for future re- Mandler, George (1980), "Recognizing: The Judgment of
search. Previous Occurrence," Psychological Review, 87 (3),
McMahan, HarryW. (1980), "TV Loses the 'Name Game'
[ReceivedOctober1989. Revised April 1990.] but Wins Big in Personality,"AdvertisingAge (Decem-
ber 1), p. 54.
Nisbett, R.E. and T.D. Wilson (1977), "Telling More than
REFERENCES We Can Know: Verbal Reports on Mental Processes,"
PsychologicalReview, 84 (May), 231-259.
Alba, Joseph W. and J. Wesley Hutchinson (1987), "Di- Obermiller,Carl (1985), "Varieties of Mere Exposure:The
mensions of Consumer Expertise," Journal of Con- Effects of Processing Style on Affective Response,"
sumerResearch, 13 (March),411-454. Journal of ConsumerResearch, 12 (June), 17-30.
Allison, Ralph I. and Kenneth P. Uhl (1964), "Influenceof Ray, Michael L., Alan G. Sawyer, Michael L. Rothschild,
Beer Brand Identificationon Taste Perception," Jour- Roger M. Heeler, Edward C. Strong, and Jerome B.
nal of MarketingResearch, 1 (August), 36-39. Reed (1973), "MarketingCommunication and the Hi-
Baker,William, J. Wesley Hutchinson, Danny Moore, and erarchyof Effects,"in New Modelsfor Mass Communi-
Prakash Nedungadi (1986), "Brand Familiarity and cation Research, ed. Peter Clarke, Beverly Hills, CA:
Advertising:Effectson the Evoked Set and BrandPref- Sage, 147-176.
erences,"Advancesin ConsumerResearch,Vol. 13, ed. Rip, Peter (1980), "The Informational Basis of Self-Re-
RichardJ. Lutz, Provo, UT: Association for Consumer ports: A Preliminary Report," in Advances in Con-
Research, 146-147. sumerResearch,Vol. 7, ed. JerryC. Olson, Provo, UT:
Bass, Frank M. and DarralG. Clarke(1972), "Testing Dis- Association for Consumer Research, 140-145.
tributed Lag Models of AdvertisingEffect,"Journal of Smith, Robert E. and William R. Swinyard (1983), "Atti-
MarketingResearch,9 (August), 298-308. tude-Behavior Consistency: The Impact of Product
and Robert P. Leone (1983), "Temporal Aggrega- Trial versus Advertising," Journal of Marketing Re-
tion, the Data IntervalBias, and EmpiricalEstimation search, 20 (August), 257-267.
of Bimonthly Relations from Annual Data," Manage- Titchener, EdwardB. (1912), A Text-Book of Psychology,
mentScience, 29 (January), 1- 11. New York: Macmillan.
Berlyne, Daniel E. (1970), "Novelty, Complexity and He- Tversky, Amos and Daniel Kahneman (1974), "Judgment
donic Value," Perception and Psychophysics, 8 (No- under Uncertainty: Heuristics and Biases," Science,
vember),279-285. 185,1124-1131.
Bettman, James R. (1971), "The Structure of Consumer Wright,Peterand PeterRip (1980), "RetrospectiveReports
Choice Processes," Journal of Marketing Research, 8 on ConsumerDecision Processes:'I Can Remember If
(November),465-471. I Want to, but Why Should I Bother Trying?'" in Ad-
Bogart,Leo (1986), Strategy in Advertising:Matching Me- vances in Consumer Research, Vol. 7, ed. Jerry C.
dia and Messages of Markets and Motivations, Lin- Olson, Provo, UT: Association for Consumer Re-
coln, IL:NTC Business. search, 146-147.
Deighton, John (1984), "The Interaction of Advertising Zajonc, Robert B. (1980), "Feeling and Thinking: Prefer-
and Evidence,"Journal of ConsumerResearch, 11 (De- ences Need No Inferences," American Psychologist,
cember),763-770. 35,151-175.