4 images of success and the preference-for-luxury-brands 2006-journal-of-consumer-psychology

1. JOURNAL OF CONSUMERPSYCHOLOGY, 16(1),57-69
copyright O 2006, Lawrence Erlbaum Associates,Inc.
Images of Success and the Preference
for Luxury Brands
Naomi Mandel, Petia K. Petrova, and Robert B. Cialdini
Arizorza State University
This research examines the impact of media depictions of success (or failure)on consumers'
desirefor luxury brands.In apilot study and three additional studies,we demonstratethat read-
ing a story about a similar/successfulother,such as a business major from the sameuniversity,
increasesconsumers' expectationsabout their own future wealth, which in turn increases their
desire for luxury brands. However,reading about a dissimilarsuccessfulother, such as a biol-
ogy major,lowersconsumers' preferences for luxury brands.Furthermore,we examinethe role
of ease of imaginingoneself in the narrative as a mediator of the relation between direction of
comparison,similarity, and brand preference.
In the past 30 years, media portrayals of life in America have
frequently depicted images of wealth and success. Popular
magazines such as People and In Style offer readers the op-
portunity to emulate their favorite celebrities via clothing or
other consumer purchases. Business magazines such as For-
tune or Forbes often provide stories of success, for example,
profiling "America's Richest Under 4 0 or "The World's
Richest People." These images of success and wealth are
dominant on television and in popular music as well. The me-
dia help define consumers' worlds by sketching images in
their minds (Lippmann, 1922), biasing their views of reality
toward the norms, values, and social perceptions they present
(Gerbner, Morgan, & Signorelli, 1986). For example, people
who watch more television provide higher estimates of the
average level of affluence in the United States (O'Guinn &
Shrum, 1997): "The more people watch television, the more
they think American households have tennis courts, private
planes, convertibles,maids and swimming pools ... "(Schor,
1999,p. 80).
In this research, we investigate how stories about success-
ful others might impact one's expectations and consumption
preferences. Along with their utilitarian functions, products
may communicate information about the identities of their
owners (Belk, Bahn, & Mayer, 1982; Shavitt, 1990; Shavitt
& Nelson, 1999). Luxury products are often purchased sim-
ply because they cost more, without providing any additional
direct utility over their cheaper counterparts (Dubois &
Duquesne, 1993). Instead, luxury brands may be used to en-
hance one's social status. Americans are increasingly trading
up to luxury brands (Silverstein & Fiske, 2003), regardless of
their economic status, because these products provide an op-
portunity for the middle class to attain the perception of pros-
perity (Schwartz, 2002). Conspicuous consun~ptionof such
products displays individuals' wealth, differentiating them
from others (Liebenstein, 1950).
In a pilot study and three additional studies, we examine
the possibility that reading narratives about successful others
may increase consumers' own expectations about future
wealth, along with their preferences for luxury brands, by
causing them to fantasize about their own future successes.
We further examine variables that can reduce the ease with
which consumers can imagine themselves in the narrative
(e.g., the dissimilarity of the successful other) and conse-
quently diminish or even reverse the effects of success stories
on luxury brand preferences.
CONCEPTUAL DEVELOPMENT
Individuals have a drive to evaluate their own opinions and
abilities, which they satisfy by comparing themselves to oth-
ers (Festinger, 1954).These social comparisons may involve
not only directly assessing another person, but also reading
about others in the newspaper or hearing about them by word
of mouth (Wood, 1996):
Correspondenceshould be addressed to Naomi Mandel, Arizona State
University, W. P. Carey School of Business, P.O. Box 874106, Tempe,
AZ 85284. E-mail:naomi.mandel@asu.edu
It can be hard to hear an extremely intelligent person on the
radio,or see an extremelyhandsomeonein the grocery store,

2. or participate on a panel with an expert without engaging in
social comparison, no matter how much we would not like to.
(Goethals, 1986,p. 261)
Prior research has demonstrated both contrast and assimi-
lation effects that result from social comparison. In one ex-
ample of a contrast effect, Salovey and Rodin (1984) demon-
strated that individuals who received negative feedback on a
characteristic on which others performed well experienced
anxiety and depression. Similarly, college women who saw
an ad featuring attractive models were less satisfied with
their own level of attractiveness than were those who saw an
ad without a model (Richins, 1991). In the case of assimila-
tion, a successful person might serve as a role model, rather
than a source of envy. For example, women performed better
on a math test administered by a competent female rather
than an incompetent female (Marx & Roman, 2002). In such
a case, exposure to a successfulother can serve as a source of
inspiration or basking in reflected glory (Cialdini et al.,
1976),whereas exposure to an unsuccessful other can be dis-
heartening or threatening (Stapel & Koomen, 2001). Further-
more, Mills, Polivy, Herman, and Tiggemann (2002) demon-
strated that when dieters viewed pictures of thin models, they
enjoyed a self-enhancing "thin fantasy," in which they rated
their own bodies as thinner and thus consumed more cookies
than participants who viewed pictures of heavy models.
Building on the existing findings regarding contrast and
assimilation effects of social comparison, in the present re-
search we investigate these effects in a domain previously
unexplored by researchers. That is, if social comparison pro-
cesses can impact individuals' ratings of their own abilities
and traits, can such processes impact individuals' expecta-
tions for their future as well? We propose that when consum-
ers compare themselves to more successful others, they may
alter their expectations for their own future wealth, which can
consequently impact their brand preferences. We further pro-
pose that the direction of these effects will depend on
whether assimilation or contrast occurs. In the case of an as-
similation effect, when the comparison target is an academi-
cally or financially successful other, individuals will fanta-
size about their own successful futures. This expectation of
future wealth might allow them to afford more expensive
possessions, thereby increasing their desire for luxury
brands. On the other hand, comparison with unsuccessful
others might lower expectations for future income, thereby
lowering preference for luxury brands. In the case of contrast
effects, comparison with a successful other is expected to
lower one's expectations for success and luxury products
preferences, whereas comparison with an unsuccessful other
is expected to increase expectations for personal success and
preferences for luxury brands.
Previous social comparison research has suggested sev-
eral variables that may influence the conditions under which
contrast or assimilation occurs, including the self-relevance
of the comparison dimension (Tesser & Campbell, 1980);the
salience of a group identity (Brewer & Weber, 1994);
whether participants can draw an analogy between them-
selves and the target other, thereby impacting perceived vul-
nerability (Lockwood, 2002); the actual similarity or primed
similarity of the comparison target (Brown, Novick, Lord, &
Richards, 1992; Hafner, 2003); and the feasibility of the
comparison target's success (Lockwood & Kunda, 1997).
Some researchers have suggested that when the comparison
target is viewed as an extension of one's self construal, the
target's successes might be celebrated as one's own suc-
cesses (Gardner, Gabriel, & Hochschild,2002). For example,
McFarland, Buehler, and MacKay (2001) found that upward
social comparisons produced assimilation effects when the
participant shared an intimate relationship (such as spouse,
romantic partner, or parent-child) with the comparison tar-
get. Based on these findings, we propose that the comparison
target's similarity on an important dimension will moderate
the effect of depictions of successful or unsuccessful others
on individuals' expectations for success and brand prefer-
ences. For example, business students might increase their
expectations for future wealth when reading a story about a
fellow business major who has achieved success, thus lead-
ing them to express more interest in purchasing luxury items
than would someone reading a story about an unsuccessful
business major. However, reading about someone with a dif-
ferent college major who has achieved success should not
produce this effect and might even decrease expectations for
success, because the reader and comparison target do not
share the same destiny.
Furthermore, we believe that the diagnosticity of the di-
mension of similarity might also play an important role
(Mussweiler & Strack, 2000b). For example, college stu-
dents might find a comparison target's college major to be
more relevant to predicting their own future career success
than the target's gender, making it easier to imagine them-
selves in the same situation. Therefore, we expect the com-
parison target's college major, but not the target's gender, to
moderate the effect of depictions of success or failure on fu-
ture expectations and brand preferences.
Our research also attempts to provide insight into the
cognitive processes underlying assimilation and contrast ef-
fects. We propose that when exposed to a comparison tar-
get, individuals spontaneously attempt to imagine them-
selves in the same position. To the extent that they can
easily imagine themselves in the place of the comparison
target, an assimilation effect is likely to occur. However,
when individuals experience difficulty imagining them-
selves in the place of the comparison target (for example,
because of dissimilarity to the target), a contrast effect is
likely to occur. This proposition stems from research on the
effects of imagery on likelihood judgments demonstrating
that individuals use the ease with which they can imagine a
possible outcome as an indicator of the likelihood of this
outcome (Carroll, 1978; Gregory, Cialdini, & Carpenter,
1982; Petrova & Cialdini, 2005; Sherman, Cialdini,

3. IMAGES OF SUCCESS 59
Schwartzman, & Reynolds, 1985). For example, partici-
pants who imagined winning a contest estimated the proba-
bility of this event happening to them as higher than partici-
pants who did not imagine the scenario (Gregory et al.,
1982). Moreover, the difficulty of imagining an event can
decrease the perceived likelihood of the event. For example,
the difficulty of imagining having a disease decreased par-
ticipants' estimates of the likelihood that they would con-
tract the disease (Sherman et al., 1985). Similarly, difficulty
of imagining a consumption experience decreases the per-
ceived likelihood of purchasing the product and product
evaluations (Petrova & Cialdini, 2005). Our proposition is
also consistent with recent research on metacognitive expe-
riences such as ease of retrieval and processing fluency
(Schwarz, 2004), suggesting that individuals are likely to
base their judgments not only on the content of relevant in-
formation but also on the ease with which this information
comes to mind. Although this research has provided fruitful
insights regarding the impact of metacognitive experiences
of ease in several domains, the role of ease of imagination
in the effects of narrative stories or social comparison pro-
cesses has not been investigated. As a step in this direction,
we propose that when consumers find it easy to imagine
themselves in a success story, they should be more likely to
expect themselves to be successful in the future and thus
more likely to desire luxury brands. However, when con-
sumers find it difficult to imagine themselves in a success
story, they should decrease their expectations for future
success and their preferences for luxury brands. Therefore,
we hypothesize that the effects of depictions of success on
participants' expectations for success and product prefer-
ences will be mediated by the ease with which participants
can imagine themselves in the story.
The aforementioned hypotheses are supported by existing
literature (Mussweiler, 2003), which suggests that when en-
gaging in social comparison, an individual searches his or her
memory for evidence of similarities or dissimilarities be-
tween the self and the target. Subsequent self-evaluation
should assimilate to the standard when it is based on highly
accessible standard-consistent information of the self and
contrast away from the standard when standard-inconsistent
information of the self is accessible. Consistent with
Mussweiler's framework, we propose that when a common
attribute is perceived as diagnostic to future success (such as
college major), individuals will find it easier to imagine
themselves in the same situation as the comparison target at
some point in the future than when they do not share that at-
tribute. However, we diverge from Mussweiler's model in
that our participants are searching for evidence of similarity
by testing the ease of imagining their future selves (Markus
& Nurius 1986), rather than searching for evidence of past
behavior. When it is easy to imagine a similar future self
(whether an ideal self, in the case of success, or a feared self,
in the case of failure), we expect participants to assimilate to
the comparison target. When it is difficult to generate a plau-
sible future self, we expect them to contrast away from the
target.
In addition to target similarity, we examine other factors
that may influence participants' ease of imagining them-
selves in a success or failure story. One variable that seems
particularly relevant to our model is the feasibility of the
comparison target's success or failure. For example, Lock-
wood and Kunda (1997) demonstrated that when the achieve-
ments of a "superstar" in the participant's domain of interest
seem attainable, an assimilation effect occurs, and the partic-
ipant is inspired by the star. On the other hand, when these
achievements do not seem feasible for the observer, a con-
trast effect occurs, and the person feels discouraged and
self-deflated. Research on the effects of imagery has also
demonstrated that imagining a scenario is more likely to im-
pact future intentions when the scenario is perceived as plau-
sible (Bone & Ellen, 1992).These findings suggest that when
individuals find the depicted story of success or failure to be
feasible, they will be likely to imagine themselves in a simi-
lar situation and assimilate their expectations for success and
desire for luxury brands upward or downward accordingly.
However, when the depicted success seems infeasible, two
outcomes are possible. First, individuals may find it difficult
to imagine themselves in the same position, and conse-
quently, contrast their expectations for success and desire for
luxury brands from the depiction. On the other hand, it may
also be possible that when the depicted success is infeasible,
individuals may not attempt to imagine themselves in the
story. In such a case, depictions of success would not influ-
ence individuals' expectations for success or brand prefer-
ences.
In sum, integrating previous research on social compari-
son and imagery, we propose a theoretical model to explain
the effects of depictions of success on preferences for luxury
brands. We propose that comparisons with successful others
can cause individuals to imagine themselves achievinga sim-
ilar level of success, which can alter their future expectations,
and consequently, their preferences for luxury brands. How-
ever, we also propose that these effects will be moderated by
the ease with which individuals can imagine themselves in
similar scenarios. We manipulate participants' ease of imagi-
nation via the similarity of the comparison target in terms of
college major (Study I), the feasibility of the outcome (Study
2), and by directly undermining the imagery effects by add-
ing unrelated statistics to the story (Study 3).
PILOT STUDY
We conducted a pilot study to measure the effects of our pro-
posed manipulations on participants' perceived similarity to
the comparison target, as well as the ease with which they
could imagine themselves in the same situation. Participants
were 394 business students who completed the study for par-
tial course credit. They first read a fictional news article de-

4. the materialism model, it yielded a highly significant effect
on materialism, F(l,241) =9.63,p <.005, whereas the direc-
tion of comparison x ease of imagination interaction became
nonsignificant, F(2, 241) = 1 . 3 0 , ~= .27. The highly consis-
tent pattern of results among the dependent variables of
brand preferences, materialism, and income expectations
provides further evidence that exposure to success stories can
have a profound impact on our consumption preferences.
GENERAL DISCUSSION
This research provides a necessary examination of how con-
sumers' purchasing habits are influenced by comparisons
with individuals who are wealthier and more successful than
themselves. Reading a story about a successful fellow busi-
ness major caused our participants to forecast higher annual
salaries for themselves and made them more likely to desire
luxury brands such as Rolex and Lexus, presumably because
they felt that they would be able to afford such luxury items
in the future. In contrast, reading a story about an unsuccess-
ful fellow business student caused participants to forecast
lower annual salaries, resulting in a decreased preference for
luxury brands. However, our results were reversed when par-
ticipants read stories about biology majors, with whom they
did not perceive the same shared destiny. After reading a
story about a successful biology student, participants low-
ered their expectations about their future prospects. Conse-
quently, when compared to those who read about an unsuc-
cessful biology student, they forecasted lower future salaries
for themselves, and thus displayed decreased interest in lux-
ury brands. This research demonstrates that comparisons to
successful others are likely to increase luxury brand prefer-
ences only when the depicted success is easy to imagine, a
factor that was manipulated by varying the college major of
the comparison target (Study l), varying the feasibility of the
depicted success (Study 2), measuring the perceived inter-
connectedness with the comparison target (Study 2), and in-
cluding distracting numerical information in the story (Study
3). When the story portrayed a success that was difficult for
the participant to imagine, it did not increase desire for lux-
ury brands, and in many cases actually decreased desire for
such brands.
Our research contributes to the existing literature in sev-
eral ways. Although some prior research has established as-
similation effects of social comparisons (e.g., Mills et al.,
2002), these researchers did not establish the cognitive pro-
cesses underlying their effects. Other researchers (e.g.,
Mussweiler, 2003; Mussweiler & Strack, 2000a) proposed
and tested a selective accessibility explanation, in which in-
dividuals select evidence in memory that is consistent with
their initial judgment of whether the target is similar or dis-
similar. We propose a related mechanism, in which individu-
als test the similarity or dissimilarity hypothesis by trying to
imagine themselves in a similar, future situation, and making
subsequent judgments based on the ease of imagination.
Whereas Lockwood (2002) suggested ease of imagination as
an explanation for why downward comparisons might
threaten one's own self-esteem, she did not explicitly mea-
sure or test the mediating role of this variable. In addition,
our studies differfrom most prior research on social compari-
son (e.g., Mussweiler & Strack, 2000a) in that we do not ex-
plicitly ask participants to make a comparative judgment be-
tween themselves and the comparison target prior to
subsequent evaluations. When reading a newspaper or maga-
zine article, individuals usually compare themselves with
others in an implicit and spontaneous manner, without being
instructed to do so (Stapel & Suls, 2004). Finally, this re-
search also expands on Richins's (1991) and others' findings
by investigating whether social comparisons impact not only
one's feelings of self-satisfaction, but also one's brand pref-
erences, particularly for luxury brands. Moreover, we find
that exposure to success or failure stories can temporarily in-
fluence one's level of materialism (Richins, 2004), a measure
typically assumed to remain stable over time and across situ-
ations.
An alternative explanation for our results, not discussed
previously, is that business students in the study liked to see
themselves as similar to successful business people and con-
sequently modeled their behavior by increasing their prefer-
ences for luxury brands. That is, successful business profes-
sionals are expected to own such brands as Lexus and Rolex,
and although these brands were never mentioned in the sto-
ries, it is possible that activation spread automatically in par-
ticipants' memories to constructs frequently associated with
business professionals. However, this possibility would not
explain the reversed effects of a story of success versus fail-
ure when the character was a different major. In other words,
a simple activation (or modeling) explanation would not ac-
count for the increased preference for luxury brands after
reading about an unsuccessful biology major (in Studies 2
and 3).
A possible limitation of this research is that it only exam-
ines the effects of a single medium, the newspaper article,
whereas consumers might be more heavily influenced by so-
cial comparisons with television characters. Future studies
might examine the effect of social comparison in other com-
mon marketing contexts, such as television or the Internet.
Moreover, participants were all undergraduate students who
have not yet embarked on their careers and thus have had un-
limited opportunities for success. Our results might be lim-
ited to situations in which exposure to successful others acti-
vates an image of the participant's future self, rather than an
ideal (yet unattained) present self. In other words, if the com-
parison target were a current colleague who had achieved a
higher level of success, participants might feel more threat-
ened than encouraged, possibly reversing our results. Fur-
thermore, our participants were limited to business majors,
who may have a more market-driven definition of success
than biology majors, not to mention students of the arts, so-

5. 68 MANDEL, PETROVA, CIALDINI
cia1 work, nursing, or theology. Nonbusiness majors might
have an entirely different model of success, measured in per-
sonal or emotional terms rather than financial terms, which
might lead to the purchase of a vehicle such as the environ-
mentally friendly Prius, rather than the status-enhancing
Lexus. In addition, another line of research might examine
the effect of social comparison not only on expectations of
success,but also on the motivation to work hard or make sac-
rifices to achieve the high standardsreflected in comparisons
to successful others.
REFERENCES
Aron, A,, Aron, E. N., & Smollan, D. (1992). Inclusion of other in the self
scale and the structure of interpersonal closeness. Journal of Personaliry
and Social Psychology, 63, 596-612.
Baron,R. M., & Kenny,D. A. (1986). The moderator-mediator variable dis-
tinction in social psychological research: Conceptual, strategic, and statis-
tical considerations. Journal of Personality and Social Psychology, 52,
1173-1 182.
Belk, R. W., Bahn, K. D., & Mayer, R. N. (1982). Developmental recogni-
tion of consumption symbolism.Journal of ConsumerResearch, 9,4-17.
Bone, P. F., & Ellen, P. S. (1992). The generation and consequences of com-
munication-evoked imagery. Journal of ConsumerResearch, 19,93-103.
Brewer, M. B., & Weber, J. G. (1994). Self evaluation effects of interper-
sonal versus intergroup social comparison. Journal of Personality and So-
cial Psychology, 66, 268-275.
Brown, J. D., Novick, N. J., Lord, K. A,, & Richards, J. M. (1992). When
Gulliver travels: Social context, psychological closeness, and
self-appraisals. Journal of Personality and Social Psychology, 62,
717-727.
Carroll,J. S.(1978). The effect of imagining an event on expectations for the
event: An interpretation in terms of the availability heuristic. Journal of
Experimental Social Psychology, 14, 88-96.
Cialdini, R. B., Borden, R. J., Thorne, A,, Walker, M. R., Freeman, S., &
Sloan, L. R. (1976). Basking in reflected glory: Three (football) field
studies. Journal of Personality and Social Psychology, 34, 366-375.
Dubois, B., & Duquesne, P. (1993). The market for luxury goods: Income
versus culture. European Journal of Marketing, 27, 3544.
Festinger, L. (1954).A theory of social comparison processes. Human Relu-
tions. 7, 117-140.
Gardner, W. L., Gabriel, S., & Hochschild, L. (2002). When you and I are
"we," you are not threatening: The role of self-expansion in social com-
parison. Journal of Personality and Social Psychology, 82, 239-25 1.
Gardner, W. L., Gabriel, S., & Lee, A. Y. (1999). "I"value freedom, but "we"
value relationships: Self-construal priming mirrors cultural differences in
judgment. Psychological Science, 10, 321-326.
Gerbner. G. L., Morgan, M. G., & Signorelli, N. (1986). Living with televi-
sion: The dynamics of the cultivation process. In B. Jennings & D.
Zillmann (Eds.), Perspectives on media effects (pp. 1748). Hillsdale, NJ:
Lawrence Erlbaum Associates, Inc.
Goethals, G. R. (1986). Social comparison theory: Psychology from the lost
and found. Personality and Social Psychology Bulletin, 12, 261-278.
Green, M. C., & Brock, T.C. (2000).The role of transportation in the persua-
siveness of public narratives. Journal of Personality and Sociul Psychol-
ogy, 79, 701-721.
Gregory, L. W., Cialdini, R. B., & Carpenter, K. M. (1982). Self-relevant
scenarios as mediators of likelihood estimates and compliance: Does
imagining make it so? Journal of Personality and Social Psychology, 43,
89-99.
Hafner,M. (2003). How dissimilar others may stillresemble the self:Assim-
ilation and contrast after social comparison. Journul of Consumer Psy-
chology, 14, 187-196.
Liebenstein, H. (1950). Bandwagon, Snob and Veblen effects in the theory
of consumers' demand. Quarterly Journal of Economics, 64, 183-207.
Lippman, W. (1922). Public opinion. New York: Macmillan.
Lockwood, P. (2002). Could it happen to you? Predicting the impact of
downward comparisons on the self. Journal of Per.unality and Social
P.~ycholog)182, 343-358.
Lockwood, P., Dolderman, D., Sadler, P., & Gerchak, E. (2004). Feeling
better about doing worse: Social comparisons within romantic relation-
ships. Journal of Personality and Social Psychology, 87, 80-95.
Lockwood, P., & Kunda, Z. (1997). Superstars and me: Predicting the im-
pact of role models on the self. Journal ofPersonalil)and Social Psychol-
og): 7.3, 91-103.
Markus, H., & Nurius, P. (1986). Possible selves. American Psychologist,
41, 954-969.
Marx, D. M., & Roman, J. S. (2002). Female role models: Protecting
women's math test performance. Personality and Social Psychology Bul-
letin, 28, 1183-1 193.
McFarland, C., Buehler, R., & MacKay, L. (2001). Affective responses to
social comparisons with extremely close others. Social Cognition, 19,
547-586.
Mills, J. S., Polivy, J. C., Herman, P., & Tiggemann, M. (2002). Effects of
exposure to thin media images: Evidence of self-enhancement among re-
strained eaters. Personality and Social Psychology Bulletin, 28,
1687-1699.
Mussweiler, T. (2003). Comparison processes in social judgment: Mecha-
nisms and consequences. P.rychologica1Review, 110, 472-489.
Mussweiler, T., Gabriel, S., & Bodenhausen, G. V. (2000). Shifting social
identities as a strategy for deflecting threatening social comparisons.
Journal of Personality and Social Psychology, 79, 398409.
Mussweiler, T., & Strack, F. (2000a). Consequences of social comparison:
Selective accessibility, assimilation, and contrast. In J. Suls& L. Wheeler
(Eds.), Handbook of social comparison: Theory and research (pp.
253-270). New York: Kluwer.
Mussweiler, T., & Strack, F, (2000b). The "relative self': Informational and
judgmental consequences of comparative self-evaluation. Journal of Per-
sonality and Sociul Psychology, 79, 23-38.
O'Guinn, T. C., & Faber, R. J. (1987). Mass mediated consumer socializa-
tion: Non-utilitarian and dysfunctional outcomes. Advances in Consumer
Research, 14, 473477.
O'Guinn, T.C., & Shmm,L. J. (1997).The role of television in the construc-
tion of consumer reality. Journal of Consumer Research, 23, 278-294.
Petrova, P. K., & Cialdini, R. B. (2005). Fluency of consumption imagery
and the backfire effects of imagery appeals. Journal of Consumer Re-
search, 32, 442-452.
Richins, M. (1991). Social comparison and the idealized images of advertis-
ing. Journal of Consumer Research, 19, 71-83.
Richins, M. (2004). The material values scale: Measurement properties and
development of a short form. Journal of Consumer Research, 31,
209-2 19.
Salovey, P., & Rodin, J. (1984). Some antecedents and consequences of so-
cial-comparisonjealousy. Journal of Personality and Social Psychology,
47,780-792.
Schor, J. (1999). TheoverspentAmerican: Whywe want what we don't need.
New York: Basic Books.
Schwartz, J. (2002, December 16).Supersize American dream: Expensive?
I'll take it. The New York Times, p. 8.
Schwarz, N. (2004). Metacognitive experiences in consumer judgment and
decision making, Journal of Consumer Psychology, 14, 332-348.
Shavitt, S. (1990). The role of attitude objects in attitude functions. Journal
of Experimental Social Psychology, 26, 124-148.
Shavitt, S., & Nelson, M. R. (1999). The social identity function in person
perception: Communicated meanings of product preferences. In G. R.

6. Maio & J. M. Olson (Eds.), Why we evaluate: Function of attitudes (pp.
37-57). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Sherman, S. J., Cialdini, R. B., Schwartzman, D. F., & Reynolds, K. D.
(1985). Imagining can heighten or lower the perceived likelihood of con-
tracting a disease: The mediating effect of ease of imagery. Per,sonality
and Social Psychology Bulletin, 11, 1 18-1 27.
silverstein, M. J., & Fiske, N. (2003). Trading up: Thenew American luxury
New York: Penguin.
Stapel, D. A., & Koomen, W. (2001). I, we and the effects of others on me:
How self-construal levelmoderates the social comparisoneffects. Journal
of Personality and Social Psycholog) 80, 766-78 1.
Stapel, D. A., & Suls, J. (2004). Method matters: Effects of explicit versus
implicit social comparisons on activation, behavior, and self-views. Jour-
nal of Personality and Social Psycholog): 87, 860-875.
Tesser, A. (2001). On the plasticity of self-defense. Current Directions in
Psychological Science, 10, 6649.
Tesser, A,, &Campbell, J.D. (1980). Self-definition: The impact of the rela-
tive performance and similarity of others. Social P,sychology Quarter!:
43, 341-347.
Tropp, L. R., & Wright, S. C. (2001). Ingroup identitication as the inclusion
of the ingroup in the self. personal it^ and Socitrl Psychology Bulletin. 5,
585-600.
Watson, D., Clark, L. A,, & Tellegen, A. (1988). Development and valida-
tion of brief measures of positive and negative affect: The PANAS scales.
Journal of Personality and Social Psychology, 53, 1063-1070.
Wood,J. V. (1996). What is social comparison and how should we study it?
Personality and Social Psychology Bulletin, 22, 520-537.
Received: September 13,2003
Revised: May 12, 2005
Accepted: June 16, 2005

7. JOURNAL OF CONSUMERPSYCHOLOGY, 16(1),70-78
Copyright O 2006,Lawrence Erlbaum Associates,Inc.
Evaluations of Moderately Typical Products: The Role
of Within- Versus Cross-ManufacturerComparisons
Hyeong Min Kim
Baruch College
Two studies examined how the valence of a core brand category and the typicality of a new
product combine to influence product evaluations.The results show that when the product is
moderately typical of the core brand category, the accessibility,relevance, and distinctness of
the manufacturer's core brand prompt participantsto use the manufacturer's product line as a
standardof comparison.Thus, they evaluatean extensionproduct less favorably if it is moder-
ately typical of a favorable core brand than if it is moderately typical of an unfavorable one.
When the product description contains additional information about the extension, however,
this difference is reversed among consumerswho are knowledgeable about the generalproduct
category.Theseconsumersuse the generalproduct category as a standard of comparisonrather
the corebrand per se.
The use of brand or product line extensions to expand the
market has stimulated a considerable body of research on
how consumers evaluate such products (see Bottomley &
Holden, 2001, for a recent review). Consistent with research
on categorization processes (e.g., Fiske & Pavelchak, 1986;
Herr, 1986, 1989; Srull & Wyer, 1989), extension research
has shown that the effect of associating a new extension prod-
uct with a core brand category is determined by its typicality,
or, in other words, its physical and functional similarity to the
products that are normally produced by the particular manu-
facturer in question (e.g., Aaker & Keller, 1990; Boush &
Loken 1991; Park, Milberg, & Lawson 1991).' That is, the
evaluation of the manufacturer's core brand is likely to be
transferred to an extension if it is perceived to be typical of
the core brand category. If an extension is viewed as atypical,
however, the evaluation of it will not reflect the core (e.g.,
Aaker & Keller, 1990; Boush & Loken, 1991) and can even
be displaced away from the core evaluation (e.g., Wanke,
Bless, & Schwarz, 1998).
In some cases, however, differences in the products they
normally produce may both introduce a similar extension
product. For example, a manufacturer of highly regarded lux-
ury automobiles (e.g., BMW) and a manufacturer of medio-
cre economy cars might both introduce a midpriced product
Correspondenceshould be addressed to Hyeong Min Kim, Ziklin School
of Business, B12-240 Baruch College, One Bernard Baruch Way, New
York, NY 10010.E-mail: hyeong-min-kim@baruch.cuny.edu
lSomeauthors use the word congruity or$t insteadof t~vpiculify,but they
are similar concepts.
(e.g., a $16,000 car). Such an extension might be viewed as
moderately typical of those that are produced by each manu-
facturer. In this situation, two possibilities arise. On one
hand, consumers might use each manufacturer's general ex-
pertise or the favorableness of each manufacturer as a judg-
mental criterion. In this case, they might evaluate the product
from the luxury car manufacturer more favorably than the
same product from the economy car manufacturer. On the
other hand, they might use the quality of cars that are typi-
cally produced by the manufacturer as a standard of compari-
son. In the latter case, the extension of the luxury car manu-
facturer would be evaluated less favorably than that of the
economy car manufacturer.
The conditions in which these possibilities are likely to
occur are unclear. Past research has paid little attention to
evaluations of extensions that are moderately typical of dif-
ferent manufacturers. These experiments provided insights
into these evaluations and the processes that underlie their
occurrence. They show that evaluations of such extensions
are a function of both comparative evaluation processes and
the availability and distinctness of relevant product informa-
tion. Specifically, when consumers evaluate a moderately
typical product, they tend to use the manufacturer's core as a
standard of comparison,judging an extension less favorably
when it is from a favorable manufacturer than when it is from
an unfavorable manufacturer. However, by providing con-
sumers with information about the general product category
(i.e., a superordinate category to which an extension product
belongs, e.g., midpriced cars made by other manufacturers),
it is sometimes possible to reverse this effect.

8. CATEGORIZATIONOF NEW PRODUCTS ing its category membership. To this extent, the core brand
per se may have little influence on evaluations of the exten-
Evaluationsof HighlyTypical sion.
or Atypical Products Another conception (e.g., Herr, 1986, 1989) suggests that
The effect of identifying an object as a member of a category
(e.g., a brand) has been extensively investigated. Theories of
categorization often assume a two-stage process. Fiske and
Pavelchak (1986), for example, proposed that people initially
compare a new object to the category with which it is associ-
ated. If there is a clear match, they assign the object to this
category and infer that the object has other features that char-
acterize category members. Furthermore, they may evaluate
the object as similar in favorableness to the category as a
whole. If there is a clear mismatch, but the object can be eas-
ily assigned to an alternative category, it may be evaluated on
the basis of this alternative category instead. This evaluation
may be made just as quickly and easily as the judgment of a
highly typical object (Boush & Loken, 1991; Mervis &
Rosch, 1990). Finally, if the object's category membership is
unclear, people may engage in piecemeal processing and
evaluate the object on the basis of its individual attributes.
This takes more cognitive effort.
Research has consistently shown that the more typical a
product is of the core brand, the more likely the core image is
transferred to it (Aaker & Keller, 1990;Bottomley & Holden,
2001; Boush & Loken, 1991;Parket al., 1991).2Because this
research has mainly examined favorable core brands, how-
ever, there is little evidence of the relation between typicality
ratings and extension evaluations when the extension is asso-
ciated with an unfavorable core brand. This is understandable
because the rationale behind extensions is to take advantage
of a favorable core image (Bottomley & Holden, 2001).
However, extant findings imply that given the association of
an extension with an unfavorable manufacturer, evaluations
of extensions may be inversely correlated with typicality
ratings.
Evaluationsof ModeratelyTypical Products
Ambiguities arise in judgment of a moderately typical new
product. Because its category membership is unclear, people
may engage in piecemeal processing (Fiske & Pavlchak,
1986). That is, they may compute an evaluation of the prod-
uct on the basis of its individual features rather than consider-
2There is a possibility that an extremely atypical product may be con-
trasted with the core brand (e.g., Herr, 1986, 1989). However, contrast is
more cognitively demanding than assimilation (e.g., Meyers-Levy &
Tybout, 1997) and hence rapid evaluations of an extremely atypical prod-
uct is not likely to lead to contrast unless the product is highly simple to
evaluate. Supporting this argument, research on extension products gener-
ally shows that evaluations of extremely atypical extensions reflect the ab-
sence of the core brand image rather than contrast with the core (e.g.,
Aaker & Keller, 1990; Bottomley & Holden, 2001; Boush & Loken,
consumers may contrast an atypical object with the core.
Thus, this conception would predict contrast effects when ex-
tension products are sufficiently different from the manufac-
turer's core to be judged as nonmembers of the category.
However, one does not know a priori how atypical an exten-
sion has to be to induce such contrast (i.e., the use of the core
as a standard of comparison). In this context, therefore, it is
not clear how consumers would view moderately typical ex-
tension products, and therefore, whether contrast effects
would be evident.
The conceptualization proposed in this research takes
both of the aforementioned conceptualizations into account.
When consumers are unclear about an extension's member-
ship in the core brand category, they are likely to engage in
piecemeal processing of the extension's attributes, as Fiske
and Pavelchak (1986) suggested. In doing so, they compare
the extension's individual features to those of the products
that are typically made by the core brand's manufacturer to
assess the manufacturer's ability to make the product being
evaluated. Thus, they use the individual attributes of the core
brand as standards of comparison in evaluating the exten-
sion's attributes. To this extent, their evaluations of the exten-
sion, based on an integration of these attributes, may be dis-
placed away from the favorableness of the core brand,
leading to a contrast effect. This contrast effect, however, un-
like the contrast effect of the core brand category on evalua-
tions of extremely atypical extensions, is mediated by more
extensive, piecemeal processing of the extension's individual
features and not by a more global comparison of the exten-
sion with the core brand category per se. Study I evaluated
this hypothesis.
A contingency in this hypothesis should be noted, how-
ever. The hypothesis assumes that individuals compare the
extension product's attributes to the attributes of products
that are typically made by the core brand manufacturer. This
is presumably because the core brand is particularly salient to
them at the time of evaluations. In other circumstances, how-
ever, people might evaluate the extension's attributes with
reference to those of other products in the same general cate-
gory rather than products associated with the particular man-
ufacturer in question. Thus, for example, they might evaluate
attributes of a midpriced car that is made by a sports car man-
ufacturer in relation to the attributes of other midpriced cars
that are available on the market rather than attributes of the
type of sports car with which the manufacturer is identified.
These latter comparisons, of course, require prior knowledge
about characteristics of the extension product category. To
the extent this knowledge exists and is accessible in memory,
however, the effects implied by the aforementioned hypothe-
sis might not occur. This possibility was examined in
1991). Study 2.

9. STUDY 1
Method
Participants and design. A total of 189 undergradu-
ate business students at a large U.S. university received $5 or
course credit for participating in this experiment. They were
randomly assigned to small group sessions in a 3 (manufac-
turer: sports car vs. compact car vs. control) x 2 (extension
product type: sports car vs. midpriced car vs. compact car)
between-subjects design. It was assumed that the sports car
would be typical of the sports car manufacturer and atypical
of the compact car manufacturer, whereas the compact car
would be atypical of the first manufacturer and typical of the
second. The midpriced car, on the other hand, was expected
to be moderately typical of both manufacturers. This as-
sumption was validated on the basis of pretesting in a manner
to be described.
Selection of stimulus materials. Two pretests were
conducted to select manufacturers and extension products.
To avoid potential confounds that might stem from using real
manufacturer names, two hypothetical manufacturer names
(Trinity and Triton) were used. To ensure the similarity of
these names, 30 business students rated the favorableness of
each as the name of an automobile along a scale from 1 (not
at all) to 7 (very much). No difference in favorableness was
observed (Ms =4.03 vs. 4.07, respectively, t < 1).Further pre-
testing established that the favorableness of the names did
not vary over car types (sports, midpriced, or compact,
Fs < 1).
Descriptions of a sports car, a midpriced car, and a com-
pact car were also constructed on the basis of pretesting.
These descriptions were prepared in a fashion similar to
those published in automobile magazines. The sports car was
described as follows:
The ZX-2000 is perfect for drivers who want a sports
look, not the most powerful performance. The
ZX-2000, encased in an aero-dynamic frame, is a good
looking entry-level sports car priced at $25,000. The
performance of the ZX is good: it boasts top speed of
160 rnph and can reach 60 rnph from 0 in 8 seconds.
But that's enough for most people. You will love
drivability and safety of the ZX. It comes with all the
standard features like CD player, power windows,
power lock, power steering wheel, etc. Don't expect a
comfy ride, though: You will find the ZX noisy, stiff
riding. The rear seats are there just in case you are too
lazy to stow luggage in the trunk.
The midpriced car was described as follows:
The ZX-2000 is a lot of car for not a lot of money. It's
affordable, yet very well made. You will love
drivability and safety of the ZX. The retail price of the
ZX is $16,000, and the affordable car comes with all
the standard features like CD player, power windows,
power lock, power steering wheel, etc. It's safe. The
ZX is equipped with air bags and anti-lock break sys-
tem. The performance of the ZX is excellent, too: it
boasts top speed of 160 rnph and can reach 60 rnph
from 0 in 8.4 seconds. Don't expect a fancy look,
though: the exterior design is nothing to write home
about.
And the compact car was described as follows:
The ZX-2000 is very affordable and offers good value.
You will love reliability. It's reliable: it'll start first time
after being stood in the rain for a week. If anything
goes wrong on it (which isn't often) you can pop to the
local scrap yard and get the bits you need for a few
quid, instead of spending a fortune. The retail price of
the ZX is only $10,000, and the affordable compact car
comes with usual features like cassette player, etc. The
performance of the ZX is good: it boasts top speed of
140rnph and can reach 60 rnph from 0 in 12.0 seconds.
Don't expect a luxury ride and fancy gadgets. That's
not a forte of the ZX. For those who look for a reliable
driving machine, the ZX will keep you moving without
worries.
These descriptions, evaluated by pretest participants along a
scale from 1 (not at allfavorable) to 7 (veryfavorable), re-
ceived mean ratings of 4.81 (n = 16), 4.36 (n = 14),and 4.27
(n = 15), respectively.
To confirm differences in the extent to which these prod-
ucts were typical of the manufacturer's core brand, another
15 business students first read a short description of both a
manufacturer of highly regarded sports cars and a manufac-
turer of standard compact cars and then estimated how typi-
cal the three products would be of these manufacturers along
a scale from 1 (very atypical) to 7 (very typical). The sports
car description was rated as much more typical of the sports
car manufacturer (M=6.1) than of the compact car manufac-
turer (M = 1.8), and the compact car description was seen as
more typical of the compact car manufacturer (M= 5.9) than
of the sports car manufacturer (M = 1.7). However, the de-
scription of the midpriced car was rated as moderately typi-
cal of both manufacturers (3.5 vs. 3.8, respectively).
Procedure. Participants were told that they would be
asked their opinions of a new product. They received a book-
let that contained a description of a fictitious new automo-
bile, the ZX-2000. Participants in experimental conditions
first read a general description of either a sports car manufac-
turer (Trinity) or a compact car manufacturer (Triton), and
were asked to indicate their general evaluations of the manu-
facturer on a scale from 1 (unfavorable) to 7 (favorable).

10. WITHIN- VERSUS CROSS-MANUFACTURER COMPARISONS 73
Then, participants were exposed to the information about the
ZX-2000, describing it as a sports car, a midpriced car, or a
compact car, and attributing it to the manufacturer. In control
conditions, participants received information about one of
the cars without being given any indication of the manufac-
turer.
After reading the description, participants evaluated the
product along scales from 1 (negative) to 7 (positive) and
from 1 (not at all favorable) to 7 (very favorable). These
items were averaged to form an evaluation index (r = 31).
Then, participants in experimental conditions rated the typi-
cality of the new product along a scale from 1(very atypical)
to 7 (very typical), and estimated the difficulty to evaluate the
product along a scale from 1 (not much dificult) to 7 (very
dificult). Finally, they were given 3 min to list thoughts that
came to their mind while reading the material. The total num-
ber of thoughts generated by each participant was counted as
a measure of elaboration (e.g., Giirhan-Canli & Maheswaran,
1998). (In the control conditions, estimates of typicality,
judgment difficulty, and thoughts were not collected.)
Results
Manipulation checks. As intended, the sports car
manufacturer was evaluated more favorably (M= 5.84) than
the compact car manufacturer (M = 3.27), F(1, 125) =
261.80, p < .01. Furthermore, estimates of the typicality of
the extension products varied in the manner expected. These
estimates, summarized in Table 1, indicate that the sports
(compact) car was much more typical of the sports (compact)
car manufacturer than of the compact (sports) car manufac-
TABLE 1
Extension Evaluations,Typicality,and Responses
to Stimulus Information As a Function
of Extension Typicality and Manufacturer-Study 1
Manufacturer (Core Brand Category)
Dependent Variables Sports Car Compucr Car Control
Extension evaluation
Sports car 5.07, 3.73, 4.61h
Midpriced car 3.80, 4.74b 4.34
Compact car 4.07, 4.14, 4.12,
Typicality
Sports car 6.09, 1.69h
Midpriced car 3.43, 3.81,
Compact car 1.71b 6.00,
Number of thoughts
Sports car 1.43, 1.55,
Midpriced car 2.48b 2.61b
Compact car 1.33, 1.48,
Judgment difficulty
Sports car 1 37, 2.18,
Midpriced car 4.87b 4.71h
Compact car 2.05, 1.95,
Note. Cells of the same dependent variable with unlike subscripts dif-
fer at p < .05.
turer, whereas midpriced car was seen as moderately typical
of both manufacturers. The interaction of manufacturer and
extension type was quite significant, F(2, 125)= 310.00,p >
.01. More important, the difference between typicality rat-
ings of the midsized cars was not significant, confirming that
the moderately typical product's typicality did not depend on
the manufacturer's core brand.
Evaluations. Extension evaluations are shown in the
top section of Table 1 as a function of the type of extension
and the core brand manufacturer. The effects of these vari-
ables, which interacted significantly, F(4, 180) = 13.44,p <
.01, were evaluated in two sets of planned contrasts. First, to
obtain an overall indication of the effects of typicality, data
were pooled over conditions in which the extension was typi-
cal of the manufacturer, moderately typical of the manufac-
turer, and atypical of the manufacturer. As expected, exten-
sions were evaluated more favorably if they were typical (M
= 4.60) than if they were moderately typical (M = 4.25), and
more favorably in the latter case than if they were atypical (M
= 3.86), F(2, 128) = 8 . 4 6 , ~< .01.
Of greater relevance to the hypothesis, the effect of the
core brand manufacturer on the evaluations of moderately
typical extensions was investigated alone. This effect was
significant, F(2, 180)= 8.97,p < .01,and is attributable to the
fact that the midpriced extension was evaluated less favor-
ably when it was made by the sports car manufacturer (M =
3.80), and more favorably when it was made by the compact
car manufacturer (M=4.74), than when its manufacturer was
unspecified (M=4.34). Thus, as expected, the core brand cat-
egory appeared to have a contrast effect on evaluations of the
moderately similar extension.
Elaboration. Participants' estimates of judgment diffi-
culty and also their thought listings are summarized in the
last two sections of Table 1.Participants reported greater dif-
ficulty and expended more cognitive resources when evaluat-
ing the midpriced extension than the typical or atypical ex-
tensions [in each case, F(2, 125) > 27.12, p < .01], and this
was true regardless of whether the manufacturer specialized
in sports cars or compact cars. This result is consistent with
the assumption that consumers engage in more effortful pro-
cessing in this condition than in typical or atypical condi-
tions.
Discussion
Two considerations arise in evaluating the generalizeability
of these effects. First, the two core brand manufacturers dif-
fered in favorableness. Therefore, it is unclear on the basis of
these data alone whether the more favorable (unfavorable)
evaluation of the sports (compact) car from a sports (com-
pact) car mallufacturer was either due to its greater typicality
or simply due to its association with a more favorable (unfa-
vorable) manufacturer, independent of typicality. The results

11. ofthe rnidpriced car condition argue against the latter inter-
pretation. In other words, if judgments were simply a result
of core brand favorableness, judgments of this extension
would be evaluated similarly and they were not.
A second consideration is also worth mentioning. Al-
though the difference in typicality ratings of the midpriced car
did not significantly differ in experimental conditions, it was
in the same direction as the difference in evaluations under
these conditions (see Table 1).One could therefore argue that
typicality played a similarroleinthisconditionthat itplayed in
the caseofthe sportscarcondition. Toexamine this possibility,
evaluations of each product were separately analyzed as a
function of the manufacturer (i.e., corebrand), c'ovaryingtypi-
cality. That is, covarying typicality should decrease or elimi-
nate the effect of corebrand onevaluations of the sportscarbut
should not influence the effect of core brand on evaluations of
the midpriced car. As expected, when typicality was used as a
covariate inthe sportscarcondition,theeffectof corebrandon
product evaluations became nonsignificant, F(l,42)=.68,p<
.42. In the midprice car condition, however, the effect of core
brand on product evaluations did not depend on whether typi-
cality was used as a covariate,F(l,42) =2 0 . 0 8 , ~<.0l,ornot,
F(l,41)= 19.44,p<.Ol. Thus,participants didnotevaluatethe
product on the basis of its typicality per se.
Note that considered in isolation, the alternative concep-
tualizations discussed earlier in this research (e.g., Fiske &
Pavelchak, 1986; Herr; 1986, 1989) do not clearly account
for the results of Study 1. That is, Fiske and Pavelchak's
(1986) conceptualization would predict no differences in
evaluations of the midpriced car (i.e., a moderately typical
product). Furthermore, Herr's (1986, 1989) conceptualiza-
tion predicted a contrast effect only if an extension product is
sufficiently different from the manufacturer to be judged as a
nonmember of the category. However, the typicality of the
midpriced car was moderate, not extreme.
In contrast, the results of Study 1 provide strong support
for the hypothesis proposed by this research and the concep-
tualization underlying it. That is, participants who consid-
ered a very typical or very atypical extension evaluated the
extension on the basis of its similarity to the core brand cate-
gory, judging it more favorably when it was similar than
when it was dissimilar regardless of its specific attributes.
When participants evaluated moderately typical products,
however, they engaged in piecemeal processing of the exten-
sion's attributes, and, in so doing, compared these attributes
to those that characterized the core brand category. This led
the extension's attributes, and consequently the extension it-
self, to be evaluated as less similar in favorableness to the
core brand than would otherwise be the case.
STUDY 2
The product descriptions conveyed in Study 1did not provide
any explicit basis for choosinga comparison standard. Partic-
ipants in the moderately typical condition spontaneously
chose the manufacturer's core as a standard (i.e.,
within-manufacturer comparisons). This is consistent with
research evidence that consumers' choice of a comparison
standard is spontaneous rather than planned (Adaval & Mon-
roe, 2002). Therefore, if a product description contains addi-
tional information, consumers may pay attention to this in-
formation and use it as a basis for judgment (Feldman &
Lynch, 1988; Stapel, Koomen, & Velthuijsen, 1998; Yi,
1990).
For this additional information to be used effectively,
however, consumers must be able to understand the informa-
tion and its implications. Thus, consumer knowledge should
influence the effect of additional information. For example,
Mandel and Johnson (2002) demonstrated that providing
safety information caused novice consumers to search for
safety information when choosing a car, but did not influence
experts' search behavior. That is, experts, unlike novices, al-
ready knew the importance of safety, and thus they searched
for safety information whether such information was pro-
vided.
On the other hand, Herr (1989) found that experts' judg-
ments about a price were more influenced by the contextual
information than were novices' judgments. For example, af-
ter exposed to an inexpensive real car name (e.g., Chevette),
expertsjudged a hypothetical car as more expensive than did
novices. Expert consumers presumably knew that the price of
the real car was low, and used this price as a standard of com-
parison. In contrast, consumers with little knowledge did not
know the price of the real car and hence their subsequent
judgment was not influenced by it (for more discussion, see
Herr, 1989, and Lynch, Chakravarti, & Mitra, 1991). Thus,
additional information may influence judgments of only
those persons who can use it effectively.
The description of the products used in experimental con-
ditions of Study 1 began with the manufacturer's name, and
therefore, was likely to activate thoughts about the manufac-
turer and the products it typically produced. Thus, it was par-
ticularly likely to stimulate the use of these products as a
comparison standard. If the description begins with informa-
tion that prompts consumers to attend to the general product
category instead, it might often offset this tendency. On the
other hand, this may only be true of consumers who have a
larger amount of knowledge about the product category at
their disposal (Alba & Hutchinson, 1987; Bettman & Sujan,
1987). Furthermore, the knowledge must be accessible in
memory at the time of judgment (Quinn & MacCrae, 2005).
If experts are exposed to the general product category infor-
mation before evaluating a product, however, they may draw
on this information for use in making judgments (Feldman &
Lynch, 1988; Stapel et al., 1998).
In this context, therefore, experts who have been induced
to think about midpriced cars may compare attributes of the
midpriced extension product they are asked to consider with
those of the usual midpriced cars on the market, and may

12. WITHIN- VERSUS CROSS-MANUFACTURER COMPARISONS 75
evaluate attributes of a midpriced car from a manufacturer
with a reputation for high quality products quite favorably in
relation to this standard. This may be particularly true if they
have a concrete mental representation of the typical product
in mind (Alba & Hutchinson, 1987;Bettman & Sujan, 1987).
However, this might not be true of experts who are not stimu-
lated to access and apply this representation or of novices
who do not have sufficient knowledge about the product cate-
gory to access such a representation. Thus, the latter individ-
uals should continue to use the manufacturer's core brand as
a comparison standard, as in Study 1.
Method
Participants. A total of 162 undergraduate business
students at a large U.S. university participated in the study to
earn course credit. They were randomly assigned to small
group sessions in a 2 (core brand: sports car vs. compact car)
x 2 (category information: provided vs. not provided) x 2
(knowledge level: high vs. low) factorial design. The first two
variables were between-subjects factors and the last factor
was determined by a median split on knowledge scores.
Procedure and material. The procedure was the same
as that of Study I. Participants were told that they would pro-
vide opinions about a new product. They first read a general
description of Trinity (a sports car manufacturer) or Triton (a
compact car manufacturer) and were asked to indicate their
general evaluations of the manufacturer on a scale from 1
(unfavorable) to 7 (favorable). Then, participants were ex-
posed to a description of the ZX-2000, which was identical to
that of the moderately typical car in Study 1. When category
information was provided, however, the following was added
to the beginning of the description:
There's one reason why midpriced cars are the biggest
sellers in the auto market. They're economical enough
and maneuverable enough for daily commuting and
running errands around town, yet roomy enough for
cross-country family jaunts. Nearly every auto maker
has at least one or two entries in this class. In this
crowded market for midpriced cars, the Trinity (Triton)
ZX-2000 is a new player.
After reading the product description, participants evalu-
ated the product along two scales from 1(unfavorable or neg-
ative) to 7 (favorable or positive), rated the typicality of the
product with respect to the core brand along a scale from 1
(a~pical)to 7 (typical), and estimated the difficulty of mak-
ing judgments along a scale from 1 (not much difJicult) to 7
(very dificult). Then, they reported their knowledge of cars
on two scales adopted from Bloch, Ridgway, and Sherrell
(1989), concerning their knowledge of midpriced cars rela-
tive to other people and how much they knew about
midpriced cars in general. They also indicated how often they
sought information about midpriced cars during the last 6
months along a scale from 1 (not at all) to 7 (very much).
These items were highly correlated (a= 37) and were aver-
aged to form a single index of product knowledge. Based on
this index, participants were categorized as high versus low
knowledge consumers by a median split (Mdn = 4.00).
Finally, participants were given 3 min to list thoughts that
came to their mind while reading the material. If any of a par-
ticipant's thoughts reflected an explicit comparison with
other manufacturers' midpriced cars and if none of his or her
other thoughts reflected other types of comparisons (e.g.,
"This car looks bad compared with my dream car, the Ford
Mustang."), the participant was categorized as making a
cross-manufacturer comparison. Examples of cross-manu-
facturer comparison thoughts include the following: "This
car seems better than a Corola," "I'd be more excited to drive
this car than midpriced cars out there," and "I'd rather buy
this car from a sports car manufacturer than usual midpriced
cars." Conversely, if any of a participant's thoughts reflected
an explicit comparison with the core brand and if none of his
or her other thoughts reflected other types of comparisons,
the participant was identified as making a within-manu-
facturer comparison. Examples of these thoughts include the
following: "The new ZX looks like a huge improvement from
regular Triton (i.e., compact car manufacturer) cars," and
"Trinity (i.e., sports car manufacturer) cars should be much
better than this car." Two raters who were blind to the re-
search objective independently categorized participants, the
interrater agreement was 84%, and the discrepancies were re-
solved through discussion.
Results and Discussion
Manipulation checks. Three participants' responses
were not complete and they were dropped from consider-
ation. Analysis of ratings of the core brand category indi-
cated that participants rated the sports car manufacturer more
favorably than the compact car manufacturer (5.83 versus
3.11), F(1, 151) = 324.87, p < .01. However, no effects in-
volving either product knowledge or the availability of cate-
gory information were significant,^ > .30. Analyses of typi-
cality ratings as a function of the core brand category,
category information, and knowledge yielded no significant
results whatsoever, p > .10 in all cases. A5 shown in Table 2,
typicality ratings were moderate in all conditions.
Evaluations. Highly knowledgeable participants were
expected to use the product category as a standard of compar-
ison when they were provided with product category infor-
mation. Data bearing on this hypothesis are summarized in
Table 2. When no category information was provided, both
experts and novices evaluated the extension significantly less
favorably when it was made by the sports car manufacturer
(M =3.92) than when it was made by the compact car manu-
facturer (M = 4.82), F(1, 151) = 7.5 1, p < .Ol. This was also

13. 76 KIM
TABLE 2
Extension Evaluations, Thoughts, and Judgment Difficulty As a Functionof Core Brand Category,
the Availability of Product Category Information,and Participants' Expertise-Study 2
Product Category Product Category
Information Provided InformarionNot Provided
Sports Car Compact Car Sports Car Compact Car
Core Brand Core Brand Core Brand Core BrandDependent Variables
Evaluation
Low knowledge
High knowledge
Typicality
Low knowledge
High knowledge
Mean number of thoughts
Low knowledge
High knowledge
Number of participants who did cross-manufacturer comparisons
Low knowledge
High knowledge
Number of participants who did within-manufacturer comparisons
Low knowledge
High knowledge
Judgment difficulty
Low knowledge
High knowledge
Note. Cells of the same dependent variable with unlike subscripts differ at p < .05.
true of novices even when category information was pro-
vided (3.92 vs. 4.66), F(1, 151) = 5.01, p < .03. However,
knowledgeable participants evaluated the extension more fa-
vorably in this condition when it came from the sports car
manufacturer (M=4.93) than when it came from the compact
car manufacturer (M = 4.29). These effects were confirmed
by significant interactions of the core brand with partici-
pants' knowledge level, F(1, 151) = 4.29, p < .04; the avail-
ability of category information, F(1,151)=6.87, p <.Ol; and
a three-way interaction involving all of these variables, F(1,
151) = 4.90,p < .03.
as making cross-manufacturer comparisons, with the re-
maining 35% of participants being unclassified. These
percentages did not depend on whether participants' knowl-
edge level was high (55% vs. 12%) or low (55% vs. 8%).
When product category information was presented, partici-
pants with little knowledge were again more likely to make
within-manufacturer comparisons than cross-manufacturer
comparisons (47% vs. 13% of the 81 participants). In con-
trast, highly knowledgeable participants were more likely
to make cross-manufacturer comparisons than within-
manufacturer comparisons (63% vs. 7%). The effect of
knowledge level on the proportion of participants who made
cross-manufacturer comparisons was highly significant,
~ ( 1 ) ~= 2 5 . 0 2 , ~< .01.Thoughts. In principle, the effects of the core brand
category on experts' extension evaluations could occur even
if these participants simply used the favorableness of the
manufacturer's reputation as a basis for judgment, and did
not make cross-category comparisons at all. However, analy-
ses of participants' thought protocols suggest that these com-
parisons were in fact made. The numbers of novices and
experts who made within-category and cross-category com-
parisons in each condition are summarized in Table 2 along
with their estimates ofjudgment difficulty. Analyses of judg-
ment difficulty and the total number of thoughts generated
yielded no significant effects, ps > .18. However, the type
of thoughts that participants generated varied over con-
ditions. When product category information was not pro-
vided, 55% of the 78 participants were classified as making
within-manufacturer comparisons and 10% were classified
GENERAL DISCUSSION
This research demonstrates that a moderately typical exten-
sion product from a favorable manufacturer often receives
lower evaluations than the same product from an unfavorable
manufacturer. As Study 2 indicates, however, this may only
be true if participants have little prior knowledge of the gen-
eral type of product to which the extension belongs and if this
knowledge is not easily accessible in memory at the time of
judgment. When the product description contained product
category information that highly knowledgeable participants
could utilize, the evaluation pattern was reversed only for