1 Dogs On The Street, Pumas On Your Feet - How Cues in The Environment Influence Product Evaluation and Choice

Dogs on the Street, Pumas on Your Feet: How Cues in the Environment Influence Product
Evaluation and Choice

Author(s): Jonah Berger and Gráinne Fitzsimons
Source: Journal of Marketing Research, Vol. 45, No. 1 (Feb., 2008), pp. 1-14
Published by: Sage Publications, Inc.
Stable URL: https://www.jstor.org/stable/30162517
Accessed: 15-02-2019 12:52 UTC
REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/30162517?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
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 support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Sage Publications, Inc. is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Marketing Research
This content downloaded from 14.139.189.34 on Fri, 15 Feb 2019 12:52:28 UTC
All use subject to https://about.jstor.org/terms
JONAH BERGER and GRAINNE FITZSIMONS*
Little empirical research has examined the implicit effects of

environmental cues on consumer behavior. Across six studies using a
combination of field and laboratory methods, the authors find that
products are more accessible, evaluated more favorably, and chosen
more frequently when the surrounding environment contains more
perceptually or conceptually related cues. The findings highlight the
impact of frequent-in addition to recent-priming in shaping product
evaluation and choice: More frequent exposure to perceptually or
conceptually related cues increases product accessibility and makes the
product easier to process. In turn, this increased accessibility influences
product evaluation and choice, which are found to vary directly with the
frequency of exposure to conceptually related cues. These results
support the hypothesis that conceptual priming effects can have a strong
impact on real-world consumer judgments.
Keywords: environmental cues, priming, advertising, product choice,

product evaluation
Dogs on the Street, Pumas on Your Feet:

How Cues in the Environment Influence
Product Evaluation and Choice
On July 4, 1997, NASA landed the Pathfinder spacecraft In this article, we examine how repeated incidental expo-
on the surface of Mars. This "Mission to Mars" captured
sure to features of the everyday environment can influence
media attention worldwide over the course of the following
product evaluation and choice. Building on recent process-
ing fluency research (Lee and Labroo 2004; Whittlesea
months, and during this period, candy maker Mars Inc. also
noticed a rather unusual increase in sales (White 1997). 1993), as well as classic work on spreading activation
Although the Mars Bar takes its name from the company(Collins and Loftus 1975), we hypothesize that exposure to
founder and not from Earth's neighboring planet, con- environmental cues repeatedly "prime" perceptually or con-
ceptually related product representations in memory. In
sumers apparently responded to news about the planet Mars
by purchasing more Mars Bars. This was a lucky turn of turn, the resultant ease of processing the product represen-
events for the candy company, but what does it mean fortation can cause increases in product evaluation, purchase
understanding consumer choice? likelihood, and choice (Lee and Labroo 2004; Nedungadi
1990).
Using a combination of laboratory and field methods, we
investigate several hypotheses that further the understand-
*Jonah Berger is Assistant Professor of Marketing, The Wharton
ing of priming effects in the everyday consumer environ-
School, University of Pennsylvania (e-mail: jberger@wharton.upenn.edu).
ment and the processes underlying such effects. First, we
Grainne Fitzsimons is Assistant Professor of Psychology, University of
Waterloo (e-mail: grainne@uwaterloo.ca). Jennifer Aaker, Chip Heath, examine whether consumers whose environments contain
Itamar Simonson, and Christian Wheeler provided useful comments that
more perceptually or conceptually related cues evaluate
greatly strengthened the manuscript. The authors are also grateful for feed-
back from participants at the 2006 Society for Consumer Research Confer-
products more positively and choose them more often. Sec-
ence, the 2007 Judgment and Decision Making Preconference at the Soci- ond, we examine the role of frequency of cue exposure in
ety for Personality and Social Psychology conference, and Stanford's Cafepriming, hypothesizing that increased exposure frequency
Verona sessions. Tai Nicolopoulos provided helpful research assistance.leads to increases in evaluations of related products. By
The authors also thank the associate editor and the two anonymous JMR
reviewers for their constructive feedback. Chris Janiszewski served as doing so, we highlight the underlying role of fluency in
associate editor for this article. producing these effects, and we extend prior research by
demonstrating that an accumulation of previous exposures
(c) 2008, American Marketing Association Journal of Marketing Research

ISSN: 0022-2437 (print), 1547-7193 (electronic) 1 Vol. XLV (February 2008), l -14
2 JOURNAL OF MARKETING RESEARCH, FEBRUARY 2008
to related cues can increase product accessibility

(2004) suggest that these findings areand
driven by increases in
evaluation in the same way as recent conceptual fluency (ease of processing) that arise from
exposure.
We also pursue several smaller objectives that contribute
exposure to the predictive context.
to the understanding of priming effects Importantly,
inaccording
consumer to the discrepancy-attribution
envi-
ronments. We examine whether such effects can arise hypothesis, such processing ease positively influences judg-
through newly constructed links between previously unre- ment only when the fluency is unexpected (Whittlesea and
Williams 1998, 2001a, b). When people are aware of the
lated constructs (in addition to well-learned semantic links),
whether they occur for familiar brands (or only unfamiliarreason for processing ease, there is no discrepancy between
brands), and whether they can occur outside conscious how much fluency they feel and how much they expect to
awareness. Six studies examine these hypotheses and feel (because they can easily attribute the ease of process-
related objectives. ing to the source), and thus there is no reason to attribute
the fluency to positive qualities of the target. However, if
THEORETICAL BACKGROUND there is a discrepancy between expected and experienced
When concepts are activated through direct exposure, fluency, people are likely to attribute it to positive qualities
they are known to affect judgment and decision making. of the target.
For example, recently primed brands are more likely Antoimplication
be of this hypothesis is that obvious direct
included in and chosen from a consumer's consideration set exposure to an object may not always positively affect judg-
(e.g., Nedungadi 1990; Shapiro 1999). Similarly, repeated ment, because people may attribute their fluency experience
exposure to an object can produce more favorable ratings of to their recent observation of the object. In contrast, indirect
that object (Zajonc 1968), and such "mere exposure" effects exposure or conceptual priming-when people are exposed
have been shown to occur for everything from the attrac- to an object that is cognitively linked with the target-may
tiveness of people (Moreland and Beach 1992) to brands be especially likely to produce positive judgments, because
(Baker 1999; Janiszewski 1993). Thus, it has been well people are unlikely to attribute the fluency to exposure to a
established that people prefer objects they have previously seemingly irrelevant object. In this situation, people are not
encountered, but can similar effects emerge for objects expecting to experience fluency and thus are less likely to
correct for it.
related to those that were previously encountered?
Psychological research has demonstrated that situational DOWNSTREAM PRIMING EFFECTS IN THE EVERYDAY
cues or primes can automatically activate associated repre- ENVIRONMENT
sentations in memory, leading them to become more acces-
sible (e.g., Higgins, Rholes, and Jones 1977). This activa- The previously mentioned research presents a case for
tion can also spread to related constructs automatically the importance of cue exposure in constructing evaluation
through an associative network (Anderson 1983; Collins and making choices, but there have been few examination
and Loftus 1975; Neely 1977). According to this spreading- of such effects in real-world contexts. In everyday life-as
activation account, priming (or activating) a given construct opposed to the carefully controlled laboratory-consumers
in memory leads to the spontaneous activation of related are inundated with a seemingly infinite number of cue
constructs in memory. Thus, priming a given construct can (Peter and Olson 2005); how much influence can any on
have downstream effects on perceptually or conceptually cue really have? Indeed, Bargh (2006) recently noted that
related constructs. major remaining issue in the basic priming literature is how
Indeed, although most studies have focused on direct these effects play out in real-world environments, which ar
exposure, research also suggests that product choice and infinitely more complex than those in the lab. Simonson
evaluations can be influenced by exposure to perceptually (2005) suggests that until researchers can show that these
or conceptually related stimuli (Gordon and Holyoak 1983; automatic priming effects produce real-world change, the
Whittlesea 1993). For example, participants repeatedly impact and importance of these laboratory findings is unde
exposed to random polygons showed increased favorability termined. The current article contributes to growing effort
toward previously unseen polygons that were perceptually to understand how priming effects shape everyday judg-
similar (Monahan, Murphy, and Zajonc 2000). Exposure to ment and decision making (Berger and Heath 2005; Berger
a given stimulus can also increase the choice and evaluation Meredith, and Wheeler 2006; Kay et al. 2004).
of conceptually related targets (e.g., those with semantic or Furthermore, it is known that direct product exposure
conceptual links to the primed stimulus; Lee and Labroo (e.g., advertising, aisle displays) positively affects sale
2004; Whittlesea 1993). Nedungadi (1990) finds that brand (e.g., Baker 1999; Bemmaor and Mouchoux 1991), but
choice is affected by prior exposure not only to that specific what about exposure to perceptually or conceptually linke
stimuli? The more consumers see advertisements for Puma
brand but also to competing brands. He theorizes that acti-
vation of a brand spreads to related brands, causing them to brand sneakers, the more they should like and purchase
be more accessible, which leads to increased likelihood of Puma sneakers, but what about exposure to stimuli related
their inclusion in the consideration set. Predictive contexts to Puma? Might it be the case that the more consumers see
can also lead people to evaluate items more positively. Par- dogs, the more they will like and purchase Pumas?
ticipants rated a bottle of ketchup more favorably after
CURRENT RESEARCH
viewing a pictorial story about a fast-food restaurant than
about a supermarket, presumably because ketchup is more The current research investigates the impact of incidenta
closely linked to fast-food restaurants (Lee and Labroo exposure to everyday environmental cues on product
2004). Because ease of processing is often positively evaluation and choice. Because of the lack of empirical dat
valenced (Harmon-Jones and Allen 2001), Lee and Labroo on priming product representations in the everyday con-
Environmental Cues, Product Evaluation, and Choice 3
sumer environment, we first sought FIELD STUDY 1: THE HALLOWEEN

evidence that whenSTUDY
consumers were in environments that provided many cueson priming
Because little research has been conducted
to a given product, the cognitive accessibility of that prod-
everyday environments, our first study simply sought t
uct would increase.
establish that real-world environmental cues can activate-
Hi: Products are more accessible when consumers are fre-or make more accessible-related product representations.
quently exposed to real-world stimuli with perceptual orTo do so, we took advantage of a natural temporary differ-
conceptual links to those products. ence in the prevalence of certain environmental cues (the
color orange) due to a holiday (Halloween). We examined
We examine this hypothesis in Field Study 1 and the pretest
the accessibility of various consumer products at two times:
to Experiment 4.
the day before Halloween and one week later. Right before
In turn, this increased accessibility can have positive
the holiday, pumpkins adorn the neighborhoods, and stores
downstream consequences. Environments that contain per-
and advertisements prominently feature orange displays. As
ceptual or conceptual cues to a given product should repeat-
of November 1, however, these orange cues disappear. If
edly prime the product representation in memory, leading
the prevalence of real-world environmental cues can influ-
the product to be processed more easily. This should result
ence product accessibility, products that are associated with
in increased product choice, greater purchase likelihood,
the color orange (e.g., Reese's, a product whose logo and
and more positive evaluations.
packaging are largely orange in color, or orange-colored
H2: Products are more likely to be chosen if consumers are fre-products, such as Sunkist or Orange Crush) should be rela-
quently exposed to real-world stimuli with perceptual ortively more accessible before Halloween than one week
conceptual links to those products. later.
We examine this hypothesis in Experiments 1 and 3. Method
H3: Products are evaluated more positively if consumers are Respondents (N = 144) were approached as they entered
frequently exposed to real-world stimuli with perceptual or
or exited a local supermarket and were asked to complete a
conceptual links to those products.
"Quick Thinking" questionnaire. All participants were
We examine this hypothesis in Field Study 2 and Experi- approached on a Saturday afternoon; half were approached
ments 2 and 4. the day before Halloween (target condition), and half were
Beyond examining these primary effects of conceptual approached a week later (control condition). All other
priming, we also extend prior research by emphasizing the aspects of the study were identical across the two
conditions.
hypothesis that conceptual fluency is the underlying mecha-
nism producing these effects. Prior research in the con- Respondents were told that the experimenters were
sumer domain has not directly tested the underlying mecha- "interested in what things come to mind when people think
nisms: Indeed, Lee and Labroo (2004) note the need for of different categories" and were asked to "list the first
work that tests the underlying role of ease of processing or things that come to your mind" in the categories of candy/
fluency. chocolate and soda. For the candy/chocolate category, they
The current research furthers this goal in two ways: First, were asked to "list eight types of candy/chocolate," and for
we investigate exposure frequency. Prior research has the soda category, they were asked to list six brands of soda.
focused on effects of recent incidental exposure on evalua- We measured product accessibility in two ways. First, we
tions, in which judgments immediately follow cue exposure examined the percentage of lists on which the product was
(i.e., within seconds or minutes; Lee and Labroo 2004; mentioned. Second, we examined the ease with which these
Whittlesea 1993). We examine whether similar effects products came to mind relative to other products in the
result from frequent exposure to conceptual cues. We category, using a "primacy of output" method (Higgins,
hypothesize that the more frequently consumers are King, and Mavin 1982). Items generated earlier in the list
exposed to conceptually linked stimuli in their everyday were considered more accessible in memory. For example,
lives, the higher are the resultant product evaluations, even if Reese's was listed fifth, it received a score of five for that
when judgments are not immediately preceded by cue respondent. If the product did not appear on the list, a score
exposure. was recorded equivalent to the number below the last one
on the list (i.e., nine for chocolates and seven for sodas).
114: As frequency of exposure to conceptually linked stimuli
increases, the positivity of product evaluations increase. Results
We both measure self-reported frequency of exposure As we predicted, orange-associated products were rela-
tively more accessible when the color orange was more
(Experiments 2 and 3) and manipulate it (Experiment 4).
Second, we explore the role of conscious awareness inprevalent in the environment. For the chocolate category,
this process. Because research has suggested a dissociation participants mentioned Reese's more on the day before Hal-
between explicit and implicit responses to prior exposure loween (54%) than one week later (30%; x2(1, N = 144) =
(e.g., Janiszewski 1993; Zajonc 2001), we expect that con- 3.97, p = .05). Similarly, among sodas, participants men-
ceptual priming effects on evaluation are unlikely to dependtioned orange sodas (e.g., Orange Crush, Sunkist) more on
on conscious processing or rehearsal. the day before Halloween (47%) than one week later (30%;
x2(1, N = 143) = 4.37, p < .04). The participants did not
H5: The effects of exposure to conceptually linked stimuli do
generate any other orange products.
not depend on conscious awareness or deliberate learning.
We also examined product ordering within consumers'
We examine this hypothesis in Experiment 4. lists. As we predicted, Reese's was more accessible-that
Participants
is, it was generated more easily than otherwere told that the experimenter was inter-
products-the
day before Halloween (M = 6.00) estedthan a week
in the types later
of things people like, (M
and to=
ensure that
7.18; t(142) = 2.34, p < .02). Participants also
they had been exposed generated
to the ink color, they were first
orange sodas more easily the day askedbefore
to write a few sentences about (M
Halloween a book=they read
5.84) than a week later (M = 6.21), butAfter
recently. this effect
writing, did
they were askednot
which option they
would
reach significance (t(141) = 1.49, p choose from 20 different choice pairs (e.g., bever-
= .14).
ages, detergents, candies). They were shown pictures of
Discussion
each option; some of the options were related to the color
Consistent with H1, in times when there should orange
have (e.g., Sunkist orange soda), some were related to
been more real-world orange cues in the environment, green (e.g., Lemon-Lime Gatorade), and some were related
orange-related products (e.g., Reese's, orange sodas) to
were
neither (e.g., All detergent).
more accessible. Our results demonstrate that the preva-
Results
lence of environmental cues as simple as color can affect
product accessibility, even in noisy real-world consumer We summed the number of orange-related and green-
environments. related products each participant chose and examined this
It is difficult to explain these results through patterns of in a 2 (environmental cues: orange versus green) x 2 (prod-
holiday consumption. Outside of color, Reese's is not uct color: orange versus green) repeated measures analysis
uniquely linked with Halloween. A separate set of respon- of variance (ANOVA). As we predicted, exposure to differ-
dents (N = 30) did not list Reese's more frequently than ent environmental cues influenced product choice. In addi-
other types of candy (e.g., Snickers, Kit Kat, Hershey's tion to a main effect of(Mproduct color
orange = 6.39 ver-
Kisses) when asked what candy they associated with Hal- SUS MGreen -
= 5.11; F(1, 27) = 6.57, p = .02), th
loween (ps > .20). Furthermore, other chocolate products significant environmental cue x product col
that were highly associated with Halloween (Snickers and (F(1, 27) = 7.23, p = .01). Participants who
Kit Kat) showed no significant boost in accessibility due to orange pen were more likely to choose or
the holiday (x2s(1, N = 144) < 1.00, ps > .20). (M = 7.15) than participants who used a g
As with most field studies, however, threats to internal 5.63; F(1, 27) = 4.46, p = .04). Similarly, pa
validity exist. Rather than increased exposure to the color wrote with a green pen were more likely to
orange, it could be argued that increased direct exposure to products (M = 5.69) than participants who u
the products (on the shelves of the store) or different active pen (M = 4.54; F(1, 27) = 5.05, p = .03). Pen
goals consumers had before Halloween (e.g., to buy some- affect choice of the control products(MOran
thing festive) may have driven the results. The fact that half MGreen 8.31; F < .5, not significant [n.s.]).
the participants completed the survey before entering the
Discussion
store casts doubt on the first possibility, but this study does
not allow us to rule out the second concern. Our next study In support of H2, the results illustrate that exposure to
uses an experimental design to hold constant other aspects perceptually related environmental cues can influence prod-
of the situation while directly manipulating environmental uct choice. Exposure to a colored pen led participants to
cues. choose more products of that same color. Participants chose
more orange (green) products when they were exposed to
EXPERIMENT 1: THE PEN COLOR STUDY
the color orange (green).
Experiment 1 used a more controlled design to examineField Study 1 and Experiment 1 both examine the effects
of existing perceptual links between products and environ-
whether exposure to perceptually related environmental
cues can influence product choice. Participants weremental
asked cues. Field Study 2 and Experiment 2 examine
to complete a survey and were randomly given either whether
an similar effects occur using newly constructed links
orange pen or a green pen to do so. The survey asked that are conceptual in nature. During an initial lab session,
them
to make choices between consumer goods, some of whichparticipants learned a slogan that linked a product to a pre-
viously
were related to the color orange (e.g., Fanta) or green (e.g.,unrelated everyday object. Because linking a prod-
Sprite) and others of which were not linked to eitheruct to a positive object could elicit increased choice or
color
evaluation through conditioning, we chose objects that were
(e.g., Pepsi). We predicted that exposure to perceptually
related environmental cues (in this case, pen color) neutral
would in valence (i.e., dining hall trays and luggage).
influence product choice. That is, participants who used an
FIELD STUDY 2: THE LUGGAGE STUDY
orange pen would be more likely to choose orange-related
products than participants who used a green pen, andField
vice Study 2 examined whether real-world primin
versa. effects extend to purchase likelihood and willingness
pay, in addition to considering conceptual (rather than pe
Method
ceptual) priming effects and newly constructed (rather t
Participants (N = 29) were approached on a university preexisting) links. We investigated whether consumers w
campus, asked to complete a short "Consumer Choice Sur- are frequently exposed to a given cue (a common, everyd
vey," and randomly assigned to a condition. The experi- object) would react to conceptually related products mor
menter carried ten pens in his pocket; half were orange and favorably. Exposure to conceptually related cues sho
wrote in orange ink, and half were green and wrote in green increase a product's fluency (Lee and Labroo 2004) a
ink. After giving them a survey, the experimenter reached lead to greater purchase likelihood and willingness to pay
into his pocket, randomly selected a pen, and gave it to the As in Field Study 1 and Experiment 1, this study take
participant to use to complete the survey. advantage of preexisting differences in exposure to certa
assessed both
environmental cues. We conducted our questions
study on seven-point
before scales).
the As we
beginning of the school year when
expected, these some students
two items were had
highly correlated (r = .81),
already arrived on campus and and others
we averagedhadthem not
to formyet traveled
a purchase likelihood index.
to campus (they would travel between Participants were
Timealso asked
1 to list how
and much they
Time 2 would
of our study). Thus, some participants be willing to pay for(who
the musichad not
player (in yet
dollars).
arrived) would be frequently exposed We also included a few final measures
to luggage (all on seven-point
through
the act of traveling to campus, scales). and Participants
others rated (who howhadmuch they
alreadywould like to
arrived) would not be frequently haveexposed
a digital musicto player, how much having
luggage during, one would
the week of our study. At the onset have made oftheir
the past study,
week more enjoyable, and how much
all partici-
pants were contacted by e-mail and
they liked their were
luggage and repeatedly
dining hall trays. Participants
exposed to a slogan that linked were the alsotarget
asked, "Howproduct
much attention(a did new
you pay to the
digital music player) to an environmental first portion of the study?"
cue. This enabled
Half the us par-to examine
ticipants learned a slogan that linked whether our the
results could be explained
product toby participants
lug- who
gage-a cue whose prevalence varies between these had not yet traveled somehow finding the luggage slogan to
groups-and the other half were exposed to a control slo- be more relevant and thus showing increased elaboration.
gan that linked the product to a cue that should not vary Finally, they were asked whether they had traveled to cam-
across groups (dining hall trays). One week later, all par- pus during the week of the study.
ticipants responded to a follow-up e-mail survey and indi- Results
cated their purchase likelihood and willingness to pay.
Thus, we used a 2 (slogan: luggage versus tray) x 2 (group: Preliminary analyses. Traveling had no influence on how
travelers versus nontravelers) between-subjects design. much participants reported wanting to have a digital music
We hypothesized that these variables would be affected player 4.90 versus MNo Travel = 5.10; F < .5, n.s.)
by our "environment" grouping, which was simply based or whether they thought it would have made their past week
on whether participants traveled between Time 1 (when more enjoyable (M Travel 4.63 versus M No Travel = 4.87; F <
they learned the slogan) and Time 2 (when they completed .5, n.s.). Similarly, traveling did not influence how much
the dependent variables). Participants who had not yet trav- participants liked their luggage (M = 4.45 versus
eled to campus should be exposed to luggage more fre- MNo Travel = 4.53; F < .5, n.s.) or dining hall trays (M Travel =
quently during the week of the study. Consequently, among 3.63 versus MNo Travel 3.78; F < .5, n.s.). Finally, although
participants who learned the slogan linking the product to participants who received the luggage slogan reported pay-
luggage, those who traveled during the study should show ing more attention to Part 1 of the study than participants
greater purchase likelihood and willingness to pay. Travel- who received the tray slogan (M = 5.17 versus 4.58;
ing should have no effect on participants who learned the F(1, 112) = 5.56, p < .02), the slogan x group interaction
control (dining hall tray) slogan. was not significant (F < .75, n.s.), indicating that travelers
did not report paying significantly more attention to the
Method luggage slogan relative to the tray slogan (M versus
4.31) than did nontravelers (M = 5.24 versus 4.84).
A week before the start of the school year, students (N =
116, 64% female) completed the first part of a two-part Main dependent measures. A 2 x 2 ANOVA examined
"Advertising Campaigns" study as part of a larger group of the effect of slogan and group on purchase likelihood and
studies. One participant did not complete all the measures willingness to pay (see Table 1). In addition to the main
and was removed from the analysis. effects of slogan (purchase likelihood: F(1, 112) = 5.68, p =
Slogan exposure. Participants were told that a company .02; willingness to pay: F(1, 112) = 4.06, p = .05), the
was about to release a new digital music player (named analyses revealed the predicted slogan x group interaction
ePlay) and were asked to provide feedback on the com- (purchase likelihood: F(1, 112) = 4.91, p < .03; willingness
pany's upcoming ad campaign. They read an in-depth prod- to pay: F(1, 112) = 3.71, p < .06). Specifically, among par-
uct description and were shown the campaign's slogan. ticipants who received the luggage slogan, those who had
Half received a slogan that linked the product to luggage traveled were more likely to purchase the product (M =
("Luggage carries your gear, ePlay carries what you want to 3.55 versus 2.55; F(1, 112) = 6.68, p = .01) and were will-
hear"), and the other half received a slogan that linked the ing to pay over $50 more to purchase it (M = $103.10 ver-
product to dining hall trays ("Dinner is carried by a tray,
music is carried by ePlay"). Participants then completed
Table 1
several tasks that were designed to link the digital music
EFFECT OF TRAVEL AND SLOGAN ON PURCHASE
player to the designated environmental cue (e.g., reciting
the slogan in their heads or evaluating the slogan in differ- LIKELIHOOD AND WILLINGNESS TO PAY (FIELD STUDY 2)
ent fonts). They also wrote the slogan from memory several
times. In all, they were exposed to the slogan (either Purchase Likelihood Willingness to Pay
directly or by rehearsal) 20 times. No Travel Travel No Travel Travel
Dependent measures. Four days later, participants were
Luggage slogan 2.55a 3.55b $68.16a $103.10b
contacted by e-mail and were asked to complete the second
Tray slogan 2.50a 2.22a $67.05a $50.31a
part of the study (dependent and ancillary measures). Par-
ticipants responded to two questions regarding purchase Notes: Within each dependent variable, means with
scripts differ significantly at p < .05. Participants repor
likelihood ("How interested are you in purchasing the ePlay
likelihood and greater willingness to pay for a digital m
digital music player?" and "How likely would you be to was linked conceptually to cues that were encountered
purchase an ePlay digital music playerT' participants everyday environment.
sus $68.16; F(1, 112) = 3.81, p = .05). There

tray exposure was
should cause no to
the product corre-
be more accessi-
sponding difference among participants
ble, which (because
who it should
received
be unexpected) should
the lead to
positive
dining hall tray slogan (Fs < .75, n.s.).product
Whenevaluations
we(Whittlesea
considerand Williams
1998). We also measured
the data another way, among participants who reported exposure to trays to
had traveled,
those who saw the luggage slogan investigate
were more the role oflikely
frequency to
of cue exposure on
pur-
evaluation.
chase the product (M = 3.55 versus 2.22; F(1, 112) = 7.73,
p < .01) and were willing to pay more
Method than $50 more (M =
$103.10; F(1, 112) = 5.69, p = .02) than those who saw the
tray slogan (M = $50.31). There wasUndergraduate no correspondingstudents (N = 65, 55% female) completed
dif-
a
ference among participants who had not yet traveled over two-part "Advertising Study" as part of a group of studies
the period (Fs < .5, n.s.). for which they were paid $10. They were randomly
assigned to a condition, and approximately half ate in din-
Discussion
ing halls that used trays.
The results of Field Study 2 suggest that the prevalenceSlogan exposure. The slogan exposure procedure and the
of conceptually related environmental cues can influenceslogans themselves were identical to those used in Field
consumers' purchase likelihood and willingness to pay. As
Study 2.
we predicted, consumers whose environments provided Dependent measures. Ten days later, participants were
more conceptually related cues to a digital music playercontacted by e-mail and completed the dependent measures
(on seven-point scales). Two items measured participants'
reported being more likely to purchase the player and were
overall product evaluation: "How much do you like the
willing to pay more to get it. We believe that this is the first
ePlay digital music player?" and "How favorable are your
evidence to indicate that marketers can benefit by creating
novel conceptual links between their product and features attitudes toward the ePlay digital music player?" (r = .79,
of the consumer environment, a finding that could lead averaged
to to form product evaluation index). Participants
innovative marketing strategies. rated how much they thought they needed a digital music
The results help rule out several alternative explanations.
player, how much they liked dining hall trays, and how fre-
Because groups did not differ in how much they reported quently they had seen trays in the previous week. Finally,
needing a digital music player or liking their luggage, it is provided their dorm name and whether the dining hall
they
unlikely that our effects were due to travelers recognizing
useda trays.
greater need to transport music or to any kind of "affective
Results
contagion" spreading from luggage to the music player.
Travelers also did not report paying more attention to the Preliminary analyses. As we expected, participants who
luggage slogan during Time 1 of the study, which helps rulein dining halls with trays reported seeing trays more fre-
ate
out the alternative that differences in elaboration or atten- quently (M = 6.03) than those who ate in dining halls with-
tion to the slogan drove the effects. out trays (M = 1.79; t(63) = 12.61, p < .001). Furthermore,
As with most field studies however, threats to internal conditions did not differ in reported need for a digital music
validity remain. Participants in our nontraveling groupplayer (ps > .20). There was an unexpected marginal effect
arrived at school earlier than those in the traveling group,of slogan on tray liking (F(1, 65) = 3.07, p = .09), indicat-
and the type of students who arrive on campus early may be ing that participants who received the luggage slogan
different from those who arrive later (e.g., involved inreported liking trays somewhat more (M = 4.41) than those
sports or special academic programs). Although it is lesswho received the tray slogan (M = 3.77), but if anything,
clear how such differences between these groups would the direction of this effect should work against our
lead to the interactive pattern of results found, the next hypothesis.
study uses an experimental grouping variable (whether the Product evaluations. A 2 (slogan) x 2 (group) ANOVA
dining hall uses trays at meals) for which such differencesexamined the effects of slogan and group on product
should be less likely. Furthermore, even stronger support evaluations. There were no main effects of either slogan or
for our hypothesis would come from providing direct evi- group (Fs < 1, n.s.), but as we hypothesized, there was a
dence that exposure to related cues drove the effects. Thus, significant slogan x group interaction (F(1, 61) = 5.26, p <
Experiment 2 uses a more controlled design. .03; see Figure 1). Participants who received the tray slogan
evaluated the product more favorably if they ate in dorms
EXPERIMENT 2: THE TRAY STUDY
that used trays (M = 3.75) than if they ate in dorms that did
not use
Experiment 2 uses a 2 (slogan) x 2 (group) design totrays (M = 2.89; F(1, 61) = 4.49, p < .04). There
investigate whether the frequency of exposure to conceptu- was no difference between dorms for those who received
ally related cues increases product evaluations. We used the luggage
the slogan (F(1, 61) = 1.23, p > .25). If we consider
slogans from Field Study 2, but rather than examining thispar-
another way, participants who ate in dining halls with
ticipants who varied in their exposure to luggage, we trays
com-evaluated the product more favorably if they had
pared participants who varied in their exposure to dining received the tray slogan (M = 3.75) than if they had
hall trays. Half the participants ate their meals in dining received the luggage slogan (M = 2.87; F(1, 61) = 5.65, p <
halls that used trays, and half ate in dining halls that.02). did No
noteffects due to slogan emerged for participants who
use trays. ate in dining halls without trays (F < 1, n.s.).
We predicted a slogan x dining hall interaction. Partici- Moderated mediation analysis. We also conducted a
pants whose dining halls use trays and who are exposed to moderated mediation analysis (Preacher, Rucker, and Hayes
a slogan linking the product to trays should report greater 2006) to examine whether, as predicted, reported exposure
product evaluations. Repeated priming of the product by to dining hall trays would mediate the relationship between
Figure 1 vey was completed online (when conceptually related cues

were unlikely to be present), we precluded the possibility
EXPOSURE TO ENVIRONMENTS THAT CONTAIN MORE
CONCEPTUALLY LINKED PRODUCT CUES LEADS TO MORE that intervening exposure to the conceptually related cue
FAVORABLE PRODUCT EVALUATIONS (EXPERIMENT 2) (dining halls trays) could influence product evaluations. As
we expected, participants who ate in dining halls with trays
reported seeing trays more frequently (M = 6.57) than par-
4- 3.75 ticipants in dining halls without trays (M = 2.06; t(37)
3.30
2.87 2.89 10.07, p < .001), but there were no corresponding differ-
3-
ences on product evaluation (M -Trays = 3.81 versus
2- MNo Trays = 3.43; t(37) < 1, n.s.). This suggests that any
Product
1
effects of experimental group in the main study are not
E
caused by preexisting group differences, and it provides
0
greater confidence that they are due to frequency of tray
Dorm With Trays Dorm Without Trays exposure.
DTray slogan o Luggage slogan Discussion
In support of H3, the results of Experiment 2 illustrate

that frequent exposure to conceptually related cues can
influence product evaluations. Participants evaluated a
group and product evaluation for people exposed to the tray
novel product more favorably when the product had a con-
slogan but not for people exposed to the luggage slogan. An
ceptual link to cues that were encountered frequently in
ordinary least squares regression found that the independ-
their everyday environment. The use of a control slogan and
ent variable (experimental group) predicted the mediator
control groups that evaluated the product in the absence of
(reported exposure to trays; t(1, 64) = 12.61, p < .001). We
exposure to environmental cues helps rule out alternative
used an ordinary least squares multiple regression model to
predict the product evaluations based on the mediator explanations.
(fre- A moderated mediation provided further sup-
port for H4 that differences in frequency of exposure to con-
quency of seeing trays), the moderator (slogan), the inde-
ceptually related cues drove the difference in evaluations.
pendent variable (group), and the interaction between the
mediator and the moderator. Reported exposure to trays mediated the relationship
between group and product evaluation but only for partici-
As we predicted, the results indicate the presence of
moderated mediation (for a full illustration, see the Web pants who had previously learned a conceptual link
Appendix at http://www.marketingpower.com/jmrfeb08);
between the product and that cue. For participants who
learned to link the product to a different cue (luggage), no
the effect of the mediator on the dependent variable
such pattern of mediation emerged.
depended on the moderator as evidenced by a significant
mediator x moderator interaction (t(1, 64) = 2.35, p < .02). EXPERIMENT 3: THE FRUIT AND VEGETABLE STUDY
Specifically, reported exposure to trays mediated the rela-
tionship between group and product evaluation for partici- Experiment 3 examines how everyday exposure to fea-
pants exposed to the tray slogan (indirect effect; z = 2.81, tures of the environment can influence actual consumption
p < .005) but not for participants exposed to the luggage We investigated whether students would eat more fruits an
slogan (indirect effect; z = .57, n.s.). vegetables if a slogan reminding them to do so was linked
Individual mediational analyses for each condition clar- to a common feature of their everyday environment (dinin
ify the results. For participants in the tray slogan condition, hall trays). Participants recorded what they ate over a two-
all four conditions for mediation were met. Dining hall type week period; half-way through, they learned one of two slo
was correlated with device liking .37, p < .04) and the gans. Because only some of our participants ate in dining
frequency of seeing trays (13 = .91, p < .001). When we halls that use trays, we again chose a slogan that involved
included both dining hall and frequency of trays in a regres- the word "tray." We theorized that this differential cueing
sion predicting device liking, the frequency of seeing trays by the environment would cause participants who wer
was significant (( p < .04), but dining hall was not exposed to trays in their daily environment to consum
((3 -.38, p > .20). This pattern of results did not emerge in more fruits and vegetables.
the luggage slogan condition. Dining hall type was corre- We created two control conditions to test this hypothesis
lated with the frequency of seeing trays ([3 .79, p < .001) First, we compared the effect of our tray slogan for partici
but not with device liking (p -.19, p = .28), and when we pants who eat in dining halls with trays (frequently cued
included both terms in a regression predicting product group) with those who eat in dining halls without tray
evaluation, neither dining hall ((3 = -.30, p = .30) nor trays (noncued group). Second, we compared the effects of the
((3 = .13, p = .64) were significant predictors. tray slogan on our frequently cued group with a group tha
Effect of experimental grouping in the absence of cue received a slogan that should not be cued as frequently by
exposure. We also conducted a pretest to examine whether the environment (competing-slogan group).
existing differences between the experimental groupings
Method
could have similar effects on product evaluation. Partici-
pants (N = 38, half from each type of dining hall) learned Slogan pretest. Before conducting the main study, we
just the tray slogan and filled out product evaluation meas- pretested possible slogans. Thirty-five respondents (all from
ures immediately, instead of ten days later. Because the dining halls with trays) rated several slogans. They were
dependent measures were collected right away and the sur- told that the National Board for Better Heath was designing
a campaign, and they were asked andtoexamined

rate consumption
how much using a 2 (week)
they x 3 (condi-
liked each of ten slogans (1 = "not at
tion) all,"
repeated and
measures 7 = "a great
ANCOVA.
deal"). They also rated how much theyFruit and vegetable consumption.
thought each There slogan were no main
would influence their fruit and vegetable
effects of week orconsumption
condition (ps > .20), but as (1 =
we predicted,
"not at all," and 7 = "a great deal"). Weweek
a significant selected two slo-
x condition interaction emerged
gans for our main experiment: one (F(2, 55)
that= 3.14, would
p = .05; see Figure
not2).be Participants
cued in the
by the environment (competing slogan:
frequently cued "Live
group were the healthy
more affected by the slogan
way, eat five fruits and veggies a than
day")those in the noncued
and one (F(1,that55) = 4.02,
would p < .05) and
competing-slogan
be frequently cued for half the sample (target (F(1, 55) = 4.89,"Each
slogan: p < .03) groups.
and every dining hall tray needs Whereas
five fruitsparticipants and veggies
in the control groupsadid not change
day"). Participants actually liked the
their competing
consumption slogan
(noncued group: -Preslogan = 2.31 versus
(M = 4.11) more than the targetMPostslogan
slogan (M = 1.91;
competing-slogan t(34)
group:=-Preslogan
8.85, p < .001) and believed that it 2.60
was more
versus likely
MPostslogan to
2.34; Fs influ-
< 1, n.s.), those in the
ence their consumption (M = 3.54 frequently
versus cued group
2.27; increased
t(34) consumption by 25%
= 5.15,
p < .001). (MPreslogan 2.16 versus Mpostslogan 2.69; F(1, 55) = 5.23,
p < .03).
Main study procedure. Undergraduate students (N = 59,
56% female) completed a group of Relation
studies and were
to perceptions com- We also exam-
of cue frequency.
pensated $20 for their time. Participants
ined whetherwere told
participants' that the
self-reports of how frequently
experimenter was interested in howthey eating
were exposed habits varied
to trays were bywith product
correlated
day of the week, and they were asked to record
consumption. what
Consistent with they the more tar-
our hypotheses,
ate over a two-week period. Each evening,
get slogan participants participants
reported seeing trays, the more they
received an e-mail directing them increased
to a Web site where
their consumption they
of fruits and vegetables (r =
.33, p = .03).
recorded their meals. They were asked When we entered
to complete thebothsur-
frequency of tray
vey as soon as they received the e-mail
exposure andanddorm totype be specific
in a mediational analysis, however,
regarding what they ate. neither remained significant predictors of consumption
After one week, participants came to the
behavior ((3s < lab
.19, psfor
> .50). an osten-
sibly unrelated study. They were told that the National
Discussion
Board for Better Heath wanted feedback about a new cam-
paign aiming to increase student health. The slogan expo- In support of H1, the results of Experiment 3 illustrate
sure materials were similar to those used in the previous that the prevalence of everyday stimuli affected product
choice outside of the laboratory setting. Creating a concep-
two studies but were modified to involve fruits and vegeta-
bles. Approximately half the participants whose diningtual link between a reminder to eat more fruits and vegeta-
halls used trays received the target slogan (frequently cuedbles and a common real-world object increased fruit and
group), and the other half received the competing slogan vegetable consumption. Comparison groups and pretests
enable us to eliminate several alternative explanations.
(competing-slogan group). All participants whose dining
halls did not use trays received the target slogan (noncuedThese effects cannot be attributed to qualities of the slogan
group). itself because the same slogan did not change consumption
After two weeks, participants completed the final meas- for participants in the noncued group. The findings also
ures. They were asked how positively they felt about dining cannot be attributed to the particular dining halls. Con-
hall trays and how frequently they had seen trays in the past sumption did not increase for participants in the competing-
week (on seven-point scales). Participants recorded their slogan group. Thus, neither the slogan nor the environment
gender and whether their dining hall used trays. themselves caused the consumption change but rather the
Results
Figure 2
Preliminary analyses. Participants who ate in dining CONCEPTAULLY LINKING A PRODUCT TO A CUE THAT
halls that used trays reported seeing trays more frequently
FREQUENTLY APPEARS IN THE ENVIRONMENT CAN
in the past week than participants whose dining halls did
INCREASE PRODUCT CHOICE (EXPERIMENT 3)
not use trays (M = 5.89 versus 1.42; t(57) = 13.02, p <
.001). There were no differences in liking for trays across
the conditions (M Frequently Cued = 4.43 versus M -Noncued =- 3 2.69
2.60
3.96 versus M Competing Slogan = 4.13; F < 1, n. s . ). 2.31 2.34
2.16
Using federal serving-size guidelines, a coder (blind to ,2.18
the conditions) recorded the number of fruit and vegetable 2
servings each participant consumed each day before and

after the slogan manipulation. Because gender has been pd 1
shown to have a significant impact on eating behaviors
(Roos et al. 1998), our main analysis included gender as a
0
covariate. A 2 (week) x 4 (day) x 3 (condition: noncued
CV
aFos
versus frequently cued versus competing slogan) repeated High Exposure Low Exposure Competing Slogan
measures analysis of covariance (ANCOVA) found no sig-
Preslogan Postslogan
nificant week x day x condition interaction (F(6, 165) =
1.05, p > .5); thus, we collapsed across days of the week
interaction between the slogan alternative explanation cues

and prevalent for ourin results.
theIt could be arg
environment. that the results of Field Study 2 and Experiments 2 an
were tray
The correlation between reported due to cues priming the
exposure slogan
and over time, leading
con-
sumption is consistent with our slogan to become more
hypothesis that fluent and, thus, more effe
product
evaluations (measured here by consumption)
(rather than the cues increase
influencing thewith
fluency of the pr
itself).in
exposure frequency (H4). However, By contrast,
relying on existing conceptual links, we
the media-
tional test did not reach significance,
avoid this possibility.
as it did A funneled
in Experi-debriefing task (Bar
ment 2. Although exposure to andconceptually
Chartrand 2000) also enables
relatedus to examine
cues the
should increase fluency and, involvement
thus, product
of consciousness evaluation,
(H5).
choice and consumption are basedWe hypothesize
on many that participantsin
factors willaddi-
evaluate Puma
tion to evaluation. In this case, products
whether more favorably
participants
when they have been choseexposed to
fruits and vegetables was likely dog imagesdue not only
more frequently. It is wellto their
established that cats
evaluations of those foods but alsoand dogs
tohave
whata strong cognitive
they wereassociation
doingin memory
before the meal, who they were because
talking
of their many
to feature
while similarities
ordering,as domestic pets
and so on. In addition, both the (Smith,
priming Shoben, and
and Ripsthe
1974) and
meretheir frequency
expo-of co-
sure literature streams have repeatedly emphasized
occurrence in the lexicon (e.g., rainingthe dis-
cats and dogs, fight-
ing like cats awareness
connect between explicit or conscious and dogs; Lucas 1999; of O'Seaghdha
expo- 1989).
sure to the cue/prime and various Because downstream
of the strong links between effects
cat and dog,onthe activa-
everything from judgment to complextion of "dog" constructs
behaviorsin memory(Barghshould spreadet to the
al. 2001; Fitzsimons and Bargh related construct
2003; "cat." Indeed, research has shown
Kunst-Wilson andthat
Zajonc 1980). Consequently, it should
when people not
are askedbefor entirely
the first word that sur-
comes to mind
prising that participants' self-reported
when they hear theperceptions
word "dog," 75% respond of withcue the word
"cat" (Moss and Older
exposure are not always strong predictors of1996).
the Consequently,
effects we of suggest
this exposure on evaluations or thatchoice.
when peopleIndeed,
are exposed to we dog failed
images, the tocat cate-
find mediation by explicit self-report
gory will becomein active.
Field Study 2 as
Research has
well. It is possible that our self-report also shown that
measure activation
was not naturally
sen- spreads
sitive enough to pick up small
fromvariations in
the category label to exposure,
members of that category
though it significantly mediated(Collins
the and Loftus 1975;
effects ofCollins
trayand Quillian 1969, 1972),
exposure
on liking in Experiment 2. and thus we assume that when the cat category is primed,
members of that
Thus far, our studies have illustrated category
that (e.g., lions, pumas)
exposure to will also
become
everyday objects in real-world more accessible. Becausecan
environments the Puma brand is
have a
strong impact on accessibility, choice,
strongly linked to and evaluation.
cats both directly (there is a picture
Experiment 4 builds on these results
cat on theby directly
logo) testing
and indirectly the
(the brand name is a me
role of our proposed underlying of the mechanism, conceptual
cat category), we assume that the Puma brand w
fluency. Prior research has foundmore that fluency
accessible in memorygenerated
after dog priming. The re
from exposure to conceptually of related primes
a pretest also support (i.e., predic-
this logic.
tive contexts) can have a positive The increased
impact onaccessibility
product of the Puma brand should
evalua-
tions (Lee and Labroo 2004). Experiment 2 found
unexpected and unpredictable initial
to participants, because
evidence for the role of frequency of
have no exposure,
recall of exposure as measured
to the brand. Thus, followin
by explicit self-report. However, Field Study 2 hypothesis
discrepancy-attribution and Experi- (Whittlesea and Wil
ment 3 did not find evidence for 1998), we hypothesize
mediation bythat participants will
frequency of attribute
exposure, perhaps because offluency
the disconnect between
to positive feelings toward the Puma brand
explicit memory and the effects of prior
respond exposure
with higher (Kunst-
evaluations of Puma products.
Wilson and Zajonc 1980). Experiment 4 an
We also provide adopts a more
even stronger test of our theoriz
objective method of measuring the effects
by examining of frequency
individual by
differences in recognition of P
manipulating the number of exposures directly.
products. Exposure to dog images may increase the acc
bility of the Puma brand, but this accessibility should t
EXPERIMENT 4: THE PUMAS STUDY late into a boost in evaluation of Puma products only if
Experiment 4 examines the role of frequency in ticipants
produc- have knowledge that those products are mad
ing conceptual priming effects by investigating how Puma. Consequently, we expect that exposure to d
evalua-
images should increase evaluations of Puma sneakers o
tions are affected by varying the exposures to conceptually
for participants
related cues. By conducting this study in a lab setting, we who recognize the sneakers as the
brand.
can manipulate frequency of exposure (versus relying on
self-report), which enables a stronger test of the role of flu-
ency in conceptual priming effects.
Pretest: Effect of Exposure to Dogs on Puma Brand
Accessibility
Experiment 4 also tests the boundaries of conceptual
priming effects on product evaluations. We return to Weusing
first examined the assertion that varying the ex
preexisting conceptual links (in this case, betweensures to dog images would affect the ease of proces
cats and
dogs) to examine further whether our effects are (i.e.,
limited to
accessibility) of the Puma brand. If we are correc
novel or unfamiliar products and whether they are suggesting
depend- that frequent cue exposure should increase
accessibility
ent on deliberate learning and practice. Using existing links of conceptually related products, a gr
rather than slogans also enables us to rule out number
a possible
of exposures to dog images should increas
accessibility of the Puma brand andcompleted some ancillary

lead participants to rec- measures an
ognize Puma as a brand of sneakers ing
moretask. To examine whether particip
quickly.
Participants (N = 46) completed two
Puma ostensibly
sneakersunrelated
as belonging to the Pum
shown
studies. The first "study" manipulated severalto
exposure sneaker
images images (inclu
Puma to
under the guise of examining the ability sneakers)
perceiveand
huewere
dif- asked to writ
ferences. Participants were shown line with
several Bargh told
photos, and that
Chartrand (2000),
some of the photos had been altered to have
asked themgreen or red on the purpose
to speculate
hues, and asked to rate the photos whether the two
on the basis studies
of the spe- were connect
cific color properties displayed. response to the second was affected by
The purpose of this task was tohow. expose participants
Finally, they reported to their mood
scales (happy,
varying numbers of exposures to conceptually excited,
related cues and upbeat; a =
(i.e., photos of dogs). Participants were shown 20 images.
Results
Those in the zero-exposure condition saw 20 images unre-
Preliminary
lated to the Puma brand (e.g., a stapler), analyses.in
those Exposure
thetolow- product-related cues
exposure condition saw 5 images of did dogs
not affect participants'
(and reported mood (Mzero = 3.96
15 unrelated
versus MFive
images), and those in the high-exposure = 4.07 versus Mien
condition saw = 3.89;
10 F < .07, n.s.). No
dog images (and 10 unrelated images).participants guessed the connection between the two stud-
ies. Even when
Participants then began a response-time asked (for
task to imagine
use howofsuchainfluence might
similar task to examine conceptualhave fluency,
occurred, mostsee Whittlesea
participants denied the possibility. We
averaged both
1993). They were asked to decide quickly evaluations ofthe
whether the Puma
pre- sneakers and the
sented stimulus (words or images) sneakers from other brands
represented a to create an evaluation
brand of index for
sneakers and to press one of two keysthe cue-related and cue-unrelated
to indicate theirproducts,
deci- respectively.
Participants
sion. In each of 40 trials, a word or picture wereappeared
split by whetherinthey
thecorrectly identified
center of the screen and remained until
all Puma sneakersthe participant
as being from the Puma brand. We then
pressed one of the designated keys.examined
Stimuli evaluations
appearedusing a 3in(cue-exposure
ran- frequency:
dom order. Six were related to zero Puma (e.g.,
versus low versus the
high) x word
2 (product recognition: high
"Puma" or the Puma logo); fillers were
versus low) xfrom
2 (sneaker other cate- related versus
brand: conceptually
gories (e.g., the word "Plank" unrelated)
or a ToyotaANOVA. logo). The
dependent variable was response Main analyses. Theto
latencies analysis
therevealed
Puma a main effect of
brand. We averaged the response timesneaker brand
for (F(1, 106)
Puma = 4.30, p = and
stimuli .04), a two-way
sneaker brandof
examined this index in a one-way (number x product
product recognition
cues: interaction (F(1,
106) = 4.63, p = .03), and the predicted three-way cue-
none versus small versus large) ANOVA.
exposure frequency
As we predicted, exposure to conceptually x product recognition
related cues x sneaker brand
influenced product accessibility interaction
(F(2, 44) (F(2, =106)3.60,
= 5.72, p p
< .005;
= see Table 2). To bet-
.04).
Response latencies indicated thatter understand this interaction,
compared with partici-we split the data in two
pants exposed to no images of dogs ways.
(M First,
= 1118we ran separate cue-exposure frequency x
ms), partici-
pants were able to identify Puma sneaker
stimuli brand more
ANOVAs for participants if
quickly who did versus did
they had been exposed to five (M not =recognize
813 Puma ms; sneakers.
t(44) As =we predicted,
2.56, among par-
p = .01) or ten (M = 849 ms; t(44) = 1.99,
ticipants p = the
who recognized .05)
Puma dog
sneakers, the analysis
images. Participants exposed to fiverevealed
versusa significant
ten cue-exposure
images frequency
did x sneaker
brand to
not differ (t < .5, n.s.). We now turn interaction (F(2, 66) = 5.42, p < of
an examination .01). Specifically,
the effect of exposure to dogs on evaluations of the Puma
brand.
Table 2
Main Study Method
EFFECT OF FREQUENCY OF EXPOSURE TO INDIRECT
Participants (N = 109, 71% female) completed two
PRODUCT CUES AND RECOGNITION OF PRODUCT MAKER
ostensibly unrelated studies. Participants were randomly
ON EVALUATION OF DIFFERENT TYPES OF PRODUCTS
assigned to a condition and were entered into a drawing for
(EXPERIMENT 4)
$25 gift certificates as compensation. The first study was
identical to the image exposure task used in the pretest; it
Low Puma High Puma
exposed participants to varying numbers of conceptually
Recognition Recognition
related cues (dog photos).
Puma Non-Puma Puma Non-Puma
The second study examined product evaluation. Partici-
Sneakers Sneakers Sneakers Sneakers
pants evaluated products from the target brand (four pic-
tures of Puma sneakers, a = .82) and products
Zero dog imagesfrom other
3.39a 3.55a 2.85a 3.67a
Five dog Participants
sneakers brands (e.g., Reebok sneakers).1 images 3.17a 4.00a 3.67b 3.84a
then
Ten dog images 2.93a 4.18a 4.53c 3.50a
Notes: In a given column, means with different supersc
nificantly at p < .05. Participants evaluated Puma sneaker
1A few pairs of the sneakers contained small pictures of cats next to the
the more frequently they were exposed to conceptu
Puma insignia. We digitally removed these images to ensure that we were
(images of dogs) but only if they recognized the sneak
measuring evaluations of the Puma brand itself, removing any possible
by the Puma brand. Similar effects were not found for s
interfering role of cat images alone. Participants could still recognize the
were conceptually unrelated to the cue.
sneakers as Pumas (based on the insignia and brand-typical features).
exposure to a greater number of dogs

Labroo increased
2004). Participants wereevaluations
quicker to recognize Puma
of the conceptually related products (i.e.,
as a sneaker brand whenPuma brand
they had been previously exposed
sneakers; F(2, 66) = 7.43, p = .001), but of
to pictures itdogs.
had no effect
Together, on underscore our
these findings
evaluations of unrelated products (i.e., that
suggestion other sneakers;
conceptual F <positive evalua-
priming elicits
.25, n.s.). Among participants who tions by did not
increasing recognize
fluency. the frequency of
By manipulating
Puma sneakers, there was no frequency
exposure, we werex ablesneaker brand
to avoid the problems of measuring
interaction (F = 1.38, p > .25); there
exposure was only
through a main
explicit effect self-report
and memory-based
of sneaker brand (F(1, 41) = 7.58, p =
items. .01).
Furthermore, the finding that directly manipulating
Second, we ran separate frequency exposurex frequency
product recognition
produced a linear pattern of results on
ANOVAs for each sneaker brand. For Puma
evaluations sneakers,
provides strong support forthere
the role of exposure
was a significant frequency x product
frequency in recognition
producing conceptual inter-
priming effects, which
action (F(2, 106) = 4.64, p = .01). helps
Frequency
answer some of of theexposure
questions raisedtoby the inconsis-
dogs influenced the evaluationtent ofresults
Puma using sneakers
reported exposure among
as a mediator in our
participants who recognized the prior
products as being
studies. Importantly, made
although this by
lab experiment pro-
Puma (F(2, 106) = 7.20, p = .001), videdbut ittest
a better had
of ourno effect
proposed on by manipu-
mechanism
participants who did not recognize the products as being
lating exposure in a controlled setting, it used repeated
made by Puma (F < .5, n.s.). Specifically
exposures occurring asover
we hypothe-
a brief period. The types of cue
sized (H4), among the people who should
exposure have
we are most had in
interested a occur
con- over days and
ceptual connection between the weeks,
cue and the If
not seconds. product, the
accessibility were measured after a
effects of cue frequency on product
week in evaluations
which participants were linear
were exposed to different num-
in nature. Compared with the evaluations of related
bers of conceptually participants
cues, accessibility increases
who saw zero pictures of dogs (M
should = visible.
be more 2.85), participants
evaluated Puma sneakers more favorably if they saw five
pictures (M 3.67; t(66) = 2.11, p = .04)GENERAL
or tenDISCUSSION
pictures
(M = 4.53; t(66) = 16.11, p < .001) of dogs.
Researchers Participants
have argued that "consumer behavior
who saw ten pictures of dogs also evaluated
strongly influencedPuma
by subtle sneakers
environmental cues" (Dijk
more favorably than participantssterhuis who et saw only five pictures
al. 2005, p. 193), but few studies have empiri
(t(66) = 2.00, p = .05). For unrelated cally investigated this argument.were
products, there The current research co
no significant effects (Fs < .75, n.s.).
tributes to this goal by examining how everyday exposu
to conceptually related cues can affect product accessibili
Discussion
evaluation, and choice. The following sections review the
key findings and implications.
In support of H3 and H4, Experiment 4 illustrates that flu-
ency caused by repeated exposure to a conceptually related
Summary of Findings and Contributions
stimulus can increase product evaluations. Participants eval-
uated Puma sneakers more favorably the more frequently Six studies illustrate that the prevalence of perceptual
they were exposed to pictures of dogs, but only if and theyconceptually related stimuli can shape real-world jud
should have had a conceptual link between the cue (dogs) ment and decision making. Experiment 1 found that mer
and the product (Puma sneakers). We did not find similar using a different color of pen to complete a survey l
effects on evaluations of products that were less related to to choose more perceptually related products; par
people
the cue (i.e., sneakers from other brands). These findingsticipants who used an orange (green) pen chose mor
extend the findings of the previous experiments to a context
orange (green) products. We found similar effects usi
in which cue exposure is manipulated rather than measured.conceptually related cues and actual product choice in th
In support of H5, these results also illustrate that concep-
real-world environment. People who learned a slogan lin
ing fruits and vegetables to an environmental cue an
tual priming effects can emerge without deliberate learning
and can occur outside of conscious awareness. Although whose daily environment contained more of those cues co
participants in prior studies learned to link the productsumed
and more fruits and vegetables (Experiment 3). Th
cue and thus could have elaborated on it consciously when results showed that the prevalence of conceptual prim
making product evaluations, the preexisting conceptual alsolink influences product evaluations and purchase like
between dogs and cats produced the same pattern of resultshood, whether those primes were measured in participant
in Experiment 4. Responses to the funnel debriefing task everyday environments (Field Study 2, Experiment 2)
further support the lack of conscious involvement in thesemanipulated in the lab (Experiment 4).
conceptual priming effects, indicating that participants pos-The data further suggest that these effects reflect an
sessed no awareness of how the dog photos affected their increase in fluency. Field Study 1 illustrated that produ
ratings of Pumas. Given this lack of awareness, it is safe to more accessible during times of the year when the
were
presume that participants were not consciously thinking were more perceptually related cues in the surroundi
about the conceptual connection between dogs and Pumas environment. Participants' reported exposure to concept
and that these effects resulted from the automatic spreading
ally related cues was also correlated with product evalua-
of activation from one construct to another. tions (Experiment 2) and actual consumption (Experimen
Finally, the accessibility pretest supports the suggestion
3). We found similar results when we directly manipulat
that frequent exposure to conceptually related cues affects a
exposure to conceptual primes (Experiment 4), and th
product's conceptual fluency or ease of processing (Leepretest
and to Experiment 4 illustrates that exposure to conce
ciations withThe
tual primes increased ease of processing. shopping),
setand ofthus the same cue may have
results
also suggests that conceptual fluency effects
different effects are
on different not
groups (Wheeler and Berger
dependent on deliberate learning of the
2007). conceptual
Consequently, link,
marketers can
should be aware of the spe-
operate outside of awareness, and cific
canlinks occur for
among the novel
group they are and
trying to reach.
familiar brands. These results also speak to marketers' efforts to make
These findings contribute to thetheir
understanding ofsales.
slogans catchy to increase the In Experiment 3, par-
influence of real-world environments on consumer behav- ticipants liked the comparison slogan more and found it
ior. People do not choose products in a vacuum, but little ismore persuasive. Nonetheless, the slogan that was more
known about the role of daily environments in shaping con- effective was the one that linked the product to a commonly
sumer choice. By showing that cues have an impact even inencountered feature of the participants' everyday environ-
noisy real-world situations, this work extends psychological ment. Thus, although catchiness may indeed be important,
research on priming (e.g., Dijksterhuis and Bargh 2001) this result suggests that marketers should also pay greater
and marketing research on the effects of cue exposure (e.g.,
attention to whether their slogans (and products) will be
Lee 2002; Shapiro 1999). The ecological validity of prim- cued by the environment.
ing has recently attracted interest in both marketing and More generally, these results speak to the importance of
psychology literature streams (Bargh 2006; Dikjsterhuis etexamining how the distribution of cues in different environ-
al. 2005; Simonson 2005); our findings support the rele-
ments affects consumer behavior. Although laboratory-
vance of this research for real marketing contexts. based studies have made great strides in understanding the
These findings also support Lee and Labroo's (2004) mechanisms behind the effects on primes on behavior,
argument that exposure to related cues can influenceresearch is just beginning to examine how these effects play
responses to a stimulus through conceptual and perceptualout in real-world consumer environments (Berger and
connections. Most research on cue exposure has focused onHeath 2005; Berger, Meredith, and Wheeler 2006; North,
the influence of direct exposure to a stimulus (e.g., ZajoncHargreaves, and McKendrick 1997). Environmental cues
1968). Our findings show that cue exposure can affect atti-influence the success of implementation intentions (Goll-
tudes not only toward the exposed object but also toward witzer 1999), trigger addictive behaviors (Bernheim and
any object that shares a conceptual relationship (see also
Rangel 2004), and influence the salience of cultural identi-
Lee and Labroo 2004; Nedungadi 1990; Whittlesea 1993).ties (Hong et al. 2000). Consequently, the distribution of
By moving beyond the use of preexisting conceptual linkscues in different environments should have important
by constructing novel associations, the results provide affects
a on the prevalence of different behaviors in those
powerful demonstration of the impact of conceptual prim-
environments (for methods of measuring cue distribution,
ing effects on consumer behavior.
see Berger and Heath 2005; Saiz and Simonsohn 2007).
Finally, these findings contribute to recent research on In conclusion, marketers should consider the nature of
conceptual fluency. Prior research has focused on the
consumer environments when designing product names,
effects of recent exposure on judgment and decision mak-
packages, and advertising campaigns. A car dealership in
ing, showing that conceptually related stimuli are evaluated
Minnesota might consider linking itself to cold weather or
more positively immediately after priming. Our findings
mittens, whereas a restaurant in Arizona might want to con-
extend this work by suggesting that similar effects can
sider links to the dry climate. Depending on what planet
result from frequent exposure to a related stimulus. This
NASA decides to go to next, the Mars candy company
finding suggests that conceptual fluency can "accumulate"
might even want to think about introducing a new candy
in some sense over time. In support of previous research on
bar.
the role of fluency in cue-exposure effects (Lee and Labroo
2004; Whittlesea 1993), our findings also show that an REFERENCES
underlying mechanism of conceptual priming effects is ease
Anderson, John R. (1983), The Architecture of Cognition.
of processing. bridge, MA: Harvard University Press.
Baker, William E. (1999), "When Can Affective Conditioning
Implications
Mere Exposure Directly Influence Brand Choice?" Journa
Our results suggest that marketers will be more effective Advertising, 28 (Fall), 31-46.
to the extent that they link their product to prevalent envi-
Bargh, John A. (2006), "What Have We Been Priming All T
ronmental cues. In addition to relying on existing concep- Years? On the Development, Mechanisms, and Ecology of N
tual relationships, the data indicate that marketers can also conscious Social Behavior," European Journal of Social P
create novel links between their product and a commonly chology, 36 (2), 147-68.
encountered feature in the consumer environment. This and Tanya L. Chartrand (2000), "The Mind in the Midd
A Practical Guide to Priming and Automaticity Research
finding could lead to innovative marketing strategies, in
Handbook of Research Methods in Social and Personality
which marketers customize slogans, brand names, and
chology, Harry T. Reis and Charles M. Judd, eds. New Yo
advertising messages on the basis of the specific features
Cambridge University Press, 253-85.
common to various environments, whether those be specific , Peter M. Gollwitzer, Annette Lee-Chai, Kim Barndollar
geographic regions (e.g., palm trees for West Coast con- and Roman Troetschel (2001), "The Automated Will: Nonc
sumers) or even specific demographic groups (e.g., lockers scious Activation and Pursuit of Behavioral Goals," Journ
for students, toys for new parents). Notably, though, a given Personality and Social Psychology, 81 (December), 1014-10
cue may have different conceptual links for different sub- Bemmaor, Albert C. and Dominique Mouchoux (1991), "Mea
populations (e.g., men and women may have different asso- ing the Short-Term Effect of In-Store Promotion and Re
and Aparna
Advertising on Brand Sales: A Factorial Labroo (2004), "Effects
Experiment," of Conceptual and
Journal
of Marketing Research, 28 (May), 202-214.
Perceptual Fluency on Affective Judgment," Journal of Market-
Berger, Jonah and Chip Heath (2005), "Idea
ing Research, 41 Habitats:
(May), 151-65. How the
Prevalence of Environmental Cues Influences the Success of Lucas, Margery (1999), "Context Effects in Lexical Access: A
Ideas," Cognitive Science, 29 (2), 195-221. Meta-Analysis," Memory and Cognition, 27 (3), 375-98.
, Marc Meredith, and S. Christian Wheeler (2006), "Can Monahan, Jennifer L., Sheila T. Murphy, and Robert B. Zajonc
Where People Vote Influence How They Vote? The Influence of(2000), "Subliminal Mere Exposure: Specific, General, and Dif-
Polling Location Type on Voting Behavior," Stanford University fuse Effects," Psychological Science, 11 (6), 462-66.
Graduate School of Business Research Paper 1926. Moreland, Richard L. and Scott R. Beach (1992), "Exposure
Bernheim, B. Douglas and Antonio Rangel (2004), "Addiction and Effects in the Classroom: The Development of Affinity Among
Cue-Triggered Decision Processes," American Economic Students," Journal of Experimental Social Psychology, 28 (3),
Review, 94 (December), 1558-90. 527-55.
Collins, Allan M. and Elizabeth F. Loftus (1975), "A Spreading- Moss, Helen and Liann Older (1996), Birkbeck Word Association
Activation Theory of Semantic Processing," Psychological Norms. Hove, UK: Lawrence Erlbaum Associates.
Review, 82 (6), 407-428. Nedungadi, Prakash (1990), "Recall and Consumer Consideration
and M. Ross Quillian (1969), "Retrieval Time from Sets: Influencing Choice Without Altering Brand Evaluations,"
Semantic Memory," Journal of Verbal Learning and Verbal Journal of Consumer Research, 17 (December), 263-76.
Behavior, 8 (2), 240-48. Neely, James H. (1977), "Semantic Priming and Retrieval from
and (1972), "How to Make a Language User," in Lexical Memory: Roles of Inhibitionless Spreading Activation
Organization and Memory, Endel Tulving and Wayne Donald- and Limited Capacity Attention," Journal of Experimental Psy-
son, eds. New York: Academic Press, 310-51.
chology: General, 106 (3), 226-54.
Dijksterhuis, Ap and John A. Bargh (2001), "The Perception- North, Adrian C., David J. Hargreaves, and Jennifer McKendrick
Behavior Expressway: Automatic Effects of Social Perception (1997), "In-Store Music Affects Product Choice," Nature, 390
on Social Behavior," in Advances in Experimental Social Psy- (November), 132.
chology, Vol. 33, Mark P. Zanna, ed. San Diego: Academic O'Seaghdha, Padraig G. (1989), "The Dependence of Lexical
Press, 1-40.
Relatedness Effects on Syntactic Connectedness," Journal of
, Pamela K. Smith, Rick B. van Baaren, and Daniel. H.J.
Experimental: Psychology: Learning, Memory, and Cognition,
Wigboldus (2005), "The Unconscious Consumer: Effects of 15 (1), 73-87.
Environment on Consumer Behavior," Journal of Consumer
Peter, J. Paul and Jerry C. Olson (2005), Consumer Behavior and
Psychology, 15 (3), 193-202.
Marketing Strategy. Chicago: McGraw-Hill/Richard D. Irwin.
Fitzsimons, Grainne M. and John A. Bargh (2003), "Thinking of
Preacher, Kristopher J., Derek D. Rucker, and Andrew F. Hayes
You: Nonconscious Pursuit of Interpersonal Goals Associated
(2006), "Suggested Procedures for Addressing Moderated
with Relationship Partners," Journal of Personality and Social
Mediation Hypotheses," working paper, Psychology Depart-
Psychology, 84 (1), 148-64.
ment, University of Kansas.
Gollwitzer, Peter M. (1999). "Implementation. Intentions: Strong
Roos, Eva, Eero Lahelma, Mikko Virtanen, Ritva Prattala, and
Effects of Simple Plans," American Psychologist, 54 (July),
493-503. Pirjo Pietinen (1998), "Gender, Socioeconomic Status and Fam-
ily Status as Determinants of Food Behaviour," Social Science
Gordon, Peter C. and Keith J. Holyoak (1983), "Implicit Learning
and Medicine, 46 (12), 1519-29.
and Generalization of the 'Mere Exposure' Effect," Journal of
Saiz, Albert and Uri Simonsohn (2007), "Downloading Wis-
Personality and Social Psychology, 45 (September), 492-500.
Harmon-Jones, Eddie and John J.B. Allen (2001), "The Role ofdom from Online Crowds," working paper, Department of
Affect in the Mere Exposure Effect: Evidence from Psycho-Operations and Information Management, University of
Pennsylvania.
physiological and Individual Differences Approaches," Person-
ality and Social Psychology Bulletin, 27 (7), 889-98. Shapiro, Stewart (1999), "When an Ad's Influence Is Beyond Our
Higgins, Tory E., Gillian A. King, and Gregory H. Mavin (1982), Conscious Control: Perceptual and. Conceptual Fluency Effects
Caused by Incidental Ad Exposure," Journal of Consumer
"Individual Construct Accessibility and Subjective Impression
on Recall," Journal of Personality and Social Psychology. 43Research, 26 (1), 16-36.
(3), 35-47. Simonson, Itamar (2005), "In Defense of Consciousness: The
, William S. Rholes, and Carl R. Jones (1977), "Category Role of Conscious and Unconscious Inputs in Consumer
Choice," Journal of Consumer Psychology, 15 (3), 211-17.
Accessibility and Impression Formation," Journal of Social
Psychology, 13 (March), 141-54. Smith, Edward E., Edward J. Shoben, and Lance J. Rips (1974),
Hong, Ying-Yi, Michael Morris, Chi-Yue Chiu, and Veronica"Structure and Process in Semantic Memory: A Featural Model
for Semantic Decisions," Psychological Review, 81 (May),
Benet-Martinez (2000), "Multicultural Minds: A Dynamic Con-
214-41.
structivist Approach to Culture and Cognition," American Psy-
chologist, 55 (7), 709-720. Wheeler, Christian S. and Jonah Berger (2007), "When the Same
Prime Leads to Different Effects," Journal of Consumer
Janiszewski, Chris (1993), "Preattentive Mere Exposure Effects,"
Journal of Consumer Research, 20 (December), 376-92. Research, 34 (3), 357-68.
White, Michael (1997), "Toy Rover Sales Soar into Orbit: Mars
Kay, Aaron C., S. Christian Wheeler, John A. Bargh, and Lee Ross
(2004), "Material Priming: The Influence of Mundane Physical Landing Puts Gold Shine Back Into Space Items," Arizona
Objects on Situational Construal and Competitive Behavioral Republic., (July 12A), El.
Choice," Organizational Behavior and Human Decision Pro- Whittlesea, Bruce W.A. (1993), "Illusions of Familiarity," Journal
cesses, 95 (1), 83-96. of Experimental Psychology: Learning, Memory, and
Kunst-Wilson, William R. and Robert B. Zajonc (1980), "Affec- Cognition, 19 (6), 1235-53.
tive Discrimination of Stimuli That Cannot Be Recognized," and Lisa D. Williams (1998), "Why Do Strangers Feel
Science, 207 (February), 557-58. Familiar, but Friends Don't? The Unexpected Basis of Feelings
Lee, Angela Y. (2002), "Effects of Implicit Memory on Memory- of Familiarity," Acta Psychologica, 98 (2), 141-66.
Based Versus Stimulus-Based Brand Choice," Journal of Mar- and (2001a), "The Discrepancy-Attribution
keting Research, 39 (November), 440-54. Hypothesis I: The Heuristic Basis of Feelings of Familiar
Journal of Experimental Psychology: Learning,

Zajonc, Memory,
Robert B. (1968), and
"Attitudinal Effects of Mere Exposu
Cognition, 27 (1), 3-13. Journal of Personality and Social Psychology: Monograph Su
- and - (2001b), "The Discrepancy-Attribution plement, 9 (2), 1-27.
Hypothesis II: Expectation, Uncertainty, Surprise, and (2001), "Mere Exposure: A Gateway to the Subliminal,"
Feelings
Current Directions in Psychological Science, 10 (6), 224-28.
of Familiarity," Journal of Experimental Psychology: Learning,
Memory, and Cognition, 27 (1), 14-33.

1 Dogs On The Street, Pumas On Your Feet - How Cues in The Environment Influence Product Evaluation and Choice

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

1 Dogs On The Street, Pumas On Your Feet - How Cues in The Environment Influence Product Evaluation and Choice

Загружено:

Авторское право:

Доступные форматы

Dogs on the Street, Pumas on Your Feet: How Cues in the Environment Influence Product

Evaluation and Choice

Little empirical research has examined the implicit effects of

Keywords: environmental cues, priming, advertising, product choice,

Dogs on the Street, Pumas on Your Feet:

(c) 2008, American Marketing Association Journal of Marketing Research

to related cues can increase product accessibility

sumer environment, we first sought FIELD STUDY 1: THE HALLOWEEN

We examine this hypothesis in Experiments 1 and 3. Method

sus $68.16; F(1, 112) = 3.81, p = .05). There

Figure 1 vey was completed online (when conceptually related cues

DTray slogan o Luggage slogan Discussion

In support of H3, the results of Experiment 2 illustrate

a campaign, and they were asked andtoexamined

servings each participant consumed each day before and

interaction between the slogan alternative explanation cues

accessibility of the Puma brand andcompleted some ancillary

exposure to a greater number of dogs

Journal of Experimental Psychology: Learning,

Вам также может понравиться