This article was downloaded by: [University of La Rioja]

On: 13 February 2014, At: 04:34

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Journal of Advertising
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/ujoa20

Missing Ingredients in Metaphor Advertising: The Right

Formula of Metaphor Type, Product Type, and Need for
Cognition
a b
Chun-Tuan Chang & Ching-Ting Yen
a
National Sun Yat-sen University , Kaohsiung City , Taiwan
b
Relmek Co. Ltd. , Kaohsiung City , Taiwan
Published online: 07 Feb 2013.

To cite this article: Chun-Tuan Chang & Ching-Ting Yen (2013) Missing Ingredients in Metaphor Advertising: The
Right Formula of Metaphor Type, Product Type, and Need for Cognition, Journal of Advertising, 42:1, 80-94, DOI:
10.1080/00913367.2012.749090

To link to this article: http://dx.doi.org/10.1080/00913367.2012.749090

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Journal of Advertising, 42(1), 80–94
Copyright C 2013, American Academy of Advertising

ISSN: 0091-3367 print / 1557-7805 online

DOI: 10.1080/00913367.2012.749090

Missing Ingredients in Metaphor Advertising:

The Right Formula of Metaphor Type, Product Type,
and Need for Cognition
Chun-Tuan Chang
National Sun Yat-sen University, Kaohsiung City, Taiwan

Ching-Ting Yen
Relmek Co. Ltd., Kaohsiung City, Taiwan
Downloaded by [University of La Rioja] at 04:34 13 February 2014

1999). This article focuses on visual metaphors and proposes

Metaphors appear frequently in contemporary advertising. two types of metaphor: explicit and implicit.
The purpose of this study is to distinguish between two types of The difference between explicit and implicit metaphors
metaphor (explicit versus implicit) and to further examine po- depends on whether or not the product’s likeness is incorporated
tential influences associated with product type and consumer dif-
ferences in need for cognition (NFC). Experimental results from
into the metaphoric picture. An explicit metaphor is clear about
two studies indicate that, regardless of metaphor type, ads with the subject, and the viewer can clearly see the product for
metaphors are more effective than nonmetaphor ads. An implicit what it is in the pictorial metaphoric display (e.g., a laptop
metaphor is more effective than an explicit metaphor in hedonic among flying papers; see http://www.adsora.com/electronics/
product advertising. In contrast, an explicit metaphor is more ef- sony-vaio-ads). In an implicit metaphor, the product is not in the
fective than an implicit metaphor in utilitarian product promotion.
metaphoric illustration and may be depicted in a less prominent
These results hold only for individuals with high NFC, not for those
with low NFC. Hence, in practice, marketers stand to gain not only place in the ad, such as the bottom corner. For example, the ad for
by matching the metaphor type with their advertised products or Ariel Remover Spray uses an image of a shark to show that the
brands but also by appropriately framing the products as hedonic spray will attack tough laundry stains (see http://www.coloribus.
or utilitarian. com/adsarchive/prints/ariel-liquid-washing-detergent-shark-
3888905/). Different from the explicit metaphor, an implicit
metaphor could have a higher propensity to stimulate the imagi-
Metaphor is the rhetorical icing on the cake of language. nation and may require more elaboration and inference drawing
on the part of the viewer. Given the prevalence of metaphoric
—MiriamVolkmann
advertising, understanding different types of metaphor can help
Among rhetorical devices, metaphors are the very heart of the marketers strategically manage their ad campaigns.
modern communication form in advertising. A metaphor claim This research demonstrates that the choice between an ex-
can be presented visually (in a picture) and/or verbally (via text). plicit or implicit metaphor should be influenced by two vari-
While similar to verbal metaphors, pictorial metaphors possess ables: product type and need for cognition (NFC). Although
extra advantages because the inferences are more likely to be these two variables have been identified as influential in previous
generated spontaneously (McQuarrie and Phillips 2005; Morgan metaphor research (e.g., Ang 2002; Ang and Lim 2006; Phillips
and Reichert 1999). Pictorial metaphor has become a kind of and McQuarrie 2004), they were examined separately and were
visual Esperanto in global marketing (Callow and Schiffman compared with a nonmetaphor condition. In the following sec-
tion, detailed interrelationships among researched variables are
proposed. The research method is then explained, and two ex-
Address correspondence to Chun-Tuan Chang, Professor, Depart-
ment of Business Management, National Sun Yat-sen University, No. periments that test the predictions being presented are described.
70, Lianhai Rd., Gushan District, Kaohsiung City 804, Taiwan. E-mail: The article concludes with a discussion of how these findings
ctchang@faculty.nsysu.edu.tw can be appropriately applied to advertising design.
Chun-Tuan Chang (PhD, University of Edinburgh) is a professor
in the Department of Business Management, National Sun Yat-sen
University.
THEORETICAL BACKGROUND AND HYPOTHESES
Ching-Ting Yen (MBA, National Sun Yat-sen University) is a re- Gkiouzepas and Hogg (2011) articulated a framework for vi-
gional manager at Relmek Co. Ltd., an agency for Sebamed. sual metaphors based on an object’s mode of representation. The

80
 MISSING INGREDIENTS IN METAPHOR ADVERTISING 81

metaphorical objects could be depicted either in their entirety Influences of Need for Cognition
(juxtaposition) or in part (synthesis). Our taxonomy of visual Although NFC has been identified as influential in metaphor
metaphor depends upon whether or not the product is included advertising (Phillips and McQuarrie 2004), there is a lack of sup-
in the ad’s visual metaphoric illustration. An explicit metaphor porting empirical data from which to draw conclusions. NFC is
plays an active role in visuals because the product (brand) image defined as an individual tendency to engage in and enjoy effort-
is formally expressed and integrated into the visual metaphoric ful thinking (Cacioppo and Petty 1982). NFC also influences
picture. In contrast, an implicit metaphor maintains a passive one’s motivation to think. The elaboration likelihood model
role in visuals and uses only a small product (brand) image, (ELM; Petty, Cacioppo, and Schumann 1983; Petty and Ca-
perhaps at the bottom of the ad. cioppo 1986) provides a framework to understand the role of
Compared with direct claims, metaphors have the advan- NFC in metaphor’s persuasive effects.
tage because they evoke more cognitive elaboration than do lit- Consumers who are high in NFC tend to follow the ELM’s
eral messages (Toncar and Munch 2001) and make consumers central route to persuasion, forming their attitudes based on
more receptive to multiple, distinct, positive inferences about a rational evaluation of the message (Haugtvedt, Petty, and
the advertised product or brand (McQuarrie and Mick 1999; Cacioppo 1992). Although pictures are often considered to
McQuarrie and Phillips 2005). Metaphors also inject novelty be heuristic cues that lead to peripheral processing (Petty
and increase motivation to process the ad (Morgan and Re- and Cacioppo 1986), pictorial metaphors deliver arguments
ichert 1999). Compared with implicit or explicit metaphors, the
Downloaded by [University of La Rioja] at 04:34 13 February 2014

visually related to the central message and can result in

sparseness of the nonmetaphoric execution will not be inter- systematic processing (Jeong 2008). Higher NFC implies more
esting enough to generate favorable responses. The following developed and differentiated knowledge schemata relative to
hypothesis is proposed. ad processing. In turn, it may increase the possibility that
H1: Measures of positive attitude toward the advertised brand and
the ad viewer will appreciate the meanings illustrated by the
purchase intentions will be higher for both the conditions of implicit visual figures and experience the pleasure of information
and explicit metaphors than for a nonmetaphor ad. elaboration (Putrevu, Tan, and Lord 2004). Compared with
those low in NFC, people with high NFC are less likely to
Influences of Product Type suffer from information overload (Kassin, Reddy, and Tulloch
Product type is an influential factor in determining the ef- 1990). Complexity attracts individuals with high NFC because
fectiveness of metaphor advertising (Ang 2002; Ang and Lim complex visuals are more challenging to comprehend (Petty
2006). For a utilitarian product, metaphoric headlines enhance and Cacioppo 1986; Phillips and McQuarrie 2004). NFC also
attitude and behavioral intentions more than nonmetaphoric moderates appreciation for rhetorical figures (Lagerwerf and
headlines do, but the use of metaphors in headlines for a sym- Meijers 2008). We expect that consumers in high-NFC groups
bolic product leads to less favorable attitude and behavioral can differentiate between an explicit and implicit metaphor.
intentions than do nonmetaphoric headlines (Ang 2002). When Consumers with low NFC often fail to follow the
comparing cologne and designer watches with mineral water logic/meaning of complex messages (Frey and Eagly 1993).
and toothpastes, Ang and Lim (2006) further suggested that They tend to refrain from attending to and processing the ad
metaphors affect brand personality perceptions of symbolic systematically. Those individuals usually have limited abilities
and utilitarian products. However, the products chosen for their and a lower threshold for boredom, and they seek to minimize
study might have confounded the experiment’s results because cognitive efforts while processing ads (Putrevu, Tan, and
of the wide price range. In addition, although symbolic and he- Lord 2004). Therefore, low-NFC consumers are less likely to
donic products share some similarities, they are still different differentiate between the two types of metaphor, which will
when it comes to consumer purchasing behaviors (Wood 1960). lead to minimal differences in the effects of ads using explicit
In this research, we considered similarly priced hedonic and versus implicit metaphors.
utilitarian products.
Current marketing literature categorizes products into two Interrelationships Among Metaphor Type, Product Type,
major types based on the affective states evoked (Dhar and and NFC
Wertenbroch 2000; Okada 2005). Hedonic products are plea- As noted in the preceding discussion, product type and
sure oriented and consumption is primarily induced by the de- the consumer’s NFC are likely to influence the effects of
sire for sensory experience, fantasy, and fun. Products with such metaphor advertising. We expect to find a three-way interaction
characteristics can cause consumers to experience a feeling of among metaphor type, product type, and NFC. The following
self-indulgence and pleasure. In contrast, utilitarian products arguments are developed based on different levels of NFC
appeal to the rational side as functional and down-to-earth, of- (high versus low). Because of their effortful processing
fering cognitively oriented benefits. Utilitarian consumption re- style and tendency toward critical comprehension, high-NFC
sults in less sensual pleasure. Generally speaking, a hedonic individuals will be able to differentiate between explicit and
product speaks to the heart while a utilitarian product speaks to implicit metaphors. Resource-matching theory (Anand and
the head. Sternthal 1989; Martin, Sherrard, and Wentzel 2005; Peracchio
 82 CHANG AND YEN

and Meyers-Levy 1997) also can be applied to explain how about the product in the metaphoric image. The use of an ex-
individuals evaluate explicit and implicit metaphors differently plicit metaphor should thus be more effective when promoting
based on product type. Resource-matching theory states that utilitarian products.
the persuasion of the message is maximized when the resources
H3: For individuals with high NFC, purchase intentions and positive
allocated to processing the information match those required
attitude toward the advertised brand will both be higher when an
for the task. The salient features in this study are hedonic versus explicit metaphor, rather than an implicit metaphor, is used to present
utilitarian in products and explicit versus implicit with regard an ad for a utilitarian product.
to metaphor. The right match will determine which is more
effective under which circumstances. As alluded to previously, consumers with low NFC are less
An important characteristic of implicit metaphor is the likely to differentiate the two types of metaphor because both are
greater degree of imagination involved in information process- a challenge to understand. Those low-NFC individuals have less
ing. Compared with an explicit metaphor, an implicit metaphor motivation to be creatively analytical. Therefore, it is expected
is more likely to generate artful deviations and provides intrinsic that the interaction between product type and metaphor type
rewards that come from processing the interpretations of the text will disappear in the case of consumers with low NFC.
and picture (McQuarrie and Mick 1999). In addition, an implicit
H4: In the case of both product types (hedonic and utilitarian), the
metaphor may have a stronger ability to decorate and disguise type of metaphor will have no influence on low-NFC individuals’
and thus increases a viewer’s motivation for ad processing. The
Downloaded by [University of La Rioja] at 04:34 13 February 2014

attitudes toward the advertised brand and purchase intentions.

heightened elaboration in play may foster a pleasurable aesthetic
experience (McQuarrie and Mick 2003). Aristotle (384–322
BC) said that “to learn gives the liveliest pleasure.” This sug- OVERVIEW OF STUDIES
gests that learning from ads can bring pleasure and enjoyment. Two experiments were designed to test the hypotheses. Both
Hedonic consumption is tied to imaginative constructions of are methodologically similar, and the general similarity of re-
reality and provides opportunities for self-expression, entertain- sults affords some modest degree of generalizability. We provide
ment, and exploration (Singer 1966). Because the product is not a detailed description of Study 1 and then brief commentary on
shown in the metaphor, consumers can use imagination freely to Study 2.
connect the product with the ad. The elaboration will generate
more pleasure and enjoyment. It is thus expected that an implicit Study 1
metaphor has stronger “heart” aspects because of its novel and A 3 (ad type: implicit metaphor versus explicit metaphor
imaginative qualities. Such pleasures match well with the nature versus nonmetaphor) × 2 (product type: hedonic versus
of hedonic consumption. Therefore, hedonic products are more utilitarian) × 2 (NFC: low versus high) between-subjects exp-
likely to benefit from the use of an implicit metaphor than from eriment was designed wherein individual differences in NFC
an explicit metaphor. were measured. The nonmetaphor condition served as a
baseline to compare the effects of two metaphoric types.
H2: For individuals with high NFC, purchase intentions and positive
attitude toward the advertised brand will both be higher when an
Test Product Selection
implicit metaphor rather than an explicit metaphor is used to present
an ad for a hedonic product. A pretest of 56 undergraduate students at a large university
in southern Taiwan was conducted to select the products to be
Utilitarian products are expected to be better paired with an featured in our advertising copy. Participants received a list of
explicit metaphor. An explicit metaphor incorporates a likeness 10 items being considered for the experiment along with the
of the product in the metaphoric illustration and may be eas- definitions of hedonic and utilitarian products derived from pre-
ier to understand than an implicit metaphor. As it is explicit, vious literature (e.g., Okada 2005). To rate the hedonic value of
it may reduce part of the “work” of comprehending the com- each product, the participants used a 7-point Likert scale from
parison being implied. In addition, explicit execution can be Strongly disagree to Strongly agree. Similarly, they rated the
deemed to be more visually prominent than implicit presenta- utilitarian value of each product. Shampoo (Mutilitarian = 6.47;
tion. Ang (2002) suggested that a rational presentation is needed Mhedonic = 2.60) and chocolates (Mutilitarian = 3.91; Mhedonic =
to persuade consumers of the veracity of the product claims, 5.80) were chosen to be the utilitarian and hedonic products re-
while having a metaphoric spin to the picture is sufficient for spectively. No gender differences were found in shampoo (T =
providing the necessary emotional appeal. Because utilitarian 0.39) or chocolates (T = 0.05). When prompted to rate the util-
products are generally perceived to be associated with sincer- itarian value of the products, participants gave a higher rating to
ity and competence (Ang and Lim 2006), an explicit metaphor shampoo than to chocolates (T = 10.51; p < .001). Conversely,
that incorporates the product can emphasize its function and they perceived chocolates as more hedonic than shampoo (T =
physical appearance, and consumers can securely evaluate the 14.44; p < .001). They were found to be similarly familiar to
product. Contrarily, consumers may not be easily persuaded by participants. Respondents in the pretest did not participate in the
an implicit metaphor because they cannot find any information main study.
 MISSING INGREDIENTS IN METAPHOR ADVERTISING 83

Stimulus Materials examine the influences of NFC (Brennan and Bahn 2006;
We used verbal-visual pun in our experimental ads. Partici- McQuarrie and Mick 2009; Putrevu 2008; Zhang 1996).
pants were presented with one ad selected from a pool of six. 2. Attitude toward the advertised brand. The participants were
Two fictitious English brand names Kerason and Brescia were asked what they thought of the advertised brand on four-
used respectively for shampoo and chocolates to prevent trig- item 7-point semantic differential scales. The items were
gering brand associations and emotions that might bias partici- bad/good, unfavorable/favorable, unpleasant/pleasant, and
pants’ responses; a pretest showed no differences on the brand unappealing/appealing (Jeong 2008). A composite rating was
familiarity and likeability of the two selected brands. To make created by calculating the mean of the items (α = 0.92).
participants feel the premium image of the advertised product, 3. Purchase intentions. Intentions were assessed in terms of
American package design or/and Western models were used. the worthiness of the promoted item, the likelihood of pur-
The visual images were manipulated to create explicit chasing it, and the probability of recommending it to fam-
metaphor, implicit metaphor, and nonmetaphor conditions (see ily or friends, and were based on 7-point Likert scales
Appendix 1). “Kerason erases your dandruff” was a metaphoric from Strongly disagree to Strongly agree (Maheswaran and
headline for shampoo. In the explicit metaphor, an illustration Meyers-Levy 1990). The three items were averaged to derive
showed the bottle of shampoo as an eraser on the blackboard. a composite rating (α = 0.88).
In the implicit metaphor, an illustration highlighted the eraser 4. Manipulation check on type of metaphor. Participants were
on the blackboard. “Say bye-bye to your dandruff” was non- first given three descriptions and asked to identify the vi-
Downloaded by [University of La Rioja] at 04:34 13 February 2014

metaphoric, as the statement described the product benefit in sual image they just viewed.2 Participants then rated the
a direct and straightforward way.1 The illustration presented a metaphoric level of headline on four 7-point scales (1 =
couple with nice hair smiling at each other; thus the ad was Strongly disagree, 7 = Strongly agree). These items included
similar to real ads for antidandruff shampoo, such as Head and questions about the extent to which participants thought the
Shoulders. ad (a) was metaphoric, (b) encouraged imagination, (c) tried
“Brescia is your sweet commitment” was a metaphoric head- to suggest the qualities of the product by associating it with
line for chocolates. The comparison to a diamond on a finger another object, and (d) was direct in saying what the prod-
provides an image that the chocolate is something to be cher- uct could do (reverse coded) (Ang 2002) (α = 0.85). The
ished. The explicit metaphor showed the chocolate in place of higher the number, the more implicitly metaphoric the ad
a diamond on the model’s finger. The implicit metaphor high- was perceived.
lighted a diamond on the finger. Finally, “Enjoy the rich and 5. NFC measure. NFC was assessed using Cacioppo, Petty,
creamy chocolates” was the nonmetaphoric statement describ- and Kao’s (1984) abbreviated 18-item NFC scale (α =
ing the taste of chocolates in a direct and straightforward way. 0.87). The sample was divided into high and low groups
Adjectives rich and creamy are common terms to represent the based on a median split on the NFC scores (M high NFC =
hedonic nature of the high-quality luxurious chocolates (Kahn, 4.87, M low NFC = 3.62; F = 311.13, p < .001).
Dhar, and Wertenbroch 2005). The illustration presented an at- 6. Background demographics. Demographics including age,
tractive young woman gently holding her chin and smiling, gender, academic major, and disposable income were as-
mimicking ads for luxurious chocolates. Each ad was A4 size. sessed. Related demographics were considered as potential
We asked three marketing experts to review the ads to see if variables that might confound the experiment’s results.
they looked realistic and professional. It was agreed that the ads
were plausible and representative of the kind of ads that may be
expected to be seen in magazines, which gave us confidence in Participants and Procedure
the generalizability of the results. The translation and back-translation procedure (Brislin 1987)
was adopted to create the Chinese version of each measure. We
Premanipulation Measure also piloted the Chinese version of the survey on a sample of
Manipulation check on product type. Similar to the pretest, 20 master’s in business administration (MBA) students from a
respondents were asked to rate the hedonic and utilitarian value marketing research course to assure the reliability of the mean-
of shampoo and chocolates based on 7-point Likert scales with ing of the measures. Participants consisted of 414 undergradu-
ends anchored at Strongly disagree and Strongly agree. ate students from courses across a variety of disciplines at the
aforementioned university. Six answer booklets were discarded
Postmanipulation Measures because of excessive missing data. In the final sample of 408
1. Thought-listing task. After exposure to the ad, participants respondents (210 males, 194 females, and 4 who failed to indi-
were asked to write down their thoughts. The thought-listing cate sex), age range was between 19 and 31 (M = 23.83, SD =
measures lower the likelihood of respondents’ boredom, 2.77).
capture a broad range of responses, and reduce the com- Prior to the experiment, the treatment booklets were ran-
mon method covariation with other objective scale measures domized. The experiment was conducted by the researcher at
(Homer and Yoon 1992). This technique has been used to the beginning of classes with the permission of the lecturers.
 84 CHANG AND YEN

Participants were told that the purpose of this study was to eval- utilitarian product (M = 6.23) was perceived to be more utilitar-
uate an ad intended for use in a forthcoming campaign. They ian than the hedonic product (M = 3.66) (F (1,407) = 186.58,
received a booklet containing a test ad and questionnaires. Par- p < .001). The manipulation of product type was confirmed.
ticipants did not interact throughout the procedure and were
debriefed after booklet collection. The experiment lasted about Results of Multivariate Analysis of Variance
20 minutes. and Univariate Results
To assess these predictions regarding the moderating effects
Results of Study 1 of product type and NFC on the metaphor manipulation used
No significant difference with regard to the distribution of in the ad, we conducted a multivariate analysis of variance
the background demographics or NFC among the experimental (MANOVA). The dependent variables (i.e., attitude toward the
groups was detected. No moderating effect of the demographics advertised brand and purchase intentions) were first confirmed
on purchase intentions or attitude toward the advertised brand to be highly correlated (γ = 0.56, p < .01). Box’s M-Statistics
was detected. for the variate was 42.87 (p = .17), suggesting that the observed
covariance matrices of the dependent variables were equal
Manipulation Checks across conditions. Multivariate results indicated significant
Chi-square tests were conducted to assess whether the visual main effects of product type (F = 8.06, p < .01) and NFC
representation was perceived as intended. The results were sig- (F = 15.67, p < .01). The interaction between ad type and
Downloaded by [University of La Rioja] at 04:34 13 February 2014

nificant in both product types (utilitarian product: χ 2 = 142.48, product type was significant (F = 12.62, p < .01). The three-
p < .001; hedonic product: χ 2 = 145.36, p < .001). As expected, way interaction among ad type, product type, and NFC was
most participants correctly identified what metaphor type the ad also significant (F = 3.85, p < .05). Additional examinations
was designed to convey. Detailed results are summarized in the of univariate results and mean comparisons were assessed. All
upper portion of Table 1. A manipulation check further showed univariate results are summarized in the upper portion of Table 2.
a significant effect of ad type on perceived metaphoric level for
both selected products (utilitarian product: F (2,203) = 3.22, Tests of Main Hypotheses
p < .05; hedonic product: F (2,199) = 6.30, p < .01) with an Hypothesis 1 predicted that the use of any metaphor, implicit
implicit metaphor being perceived to be more metaphoric than or explicit, would lead to higher positive attitude toward
an explicit metaphor (utilitarian product: M = 5.03 versus 4.55; the advertised brand and purchase intentions vis-à-vis the
hedonic product: M = 4.31 versus 4.03). An explicit metaphor nonmetaphor condition. The results showed that, compared
was deemed more metaphoric than a nonmetaphor (utilitarian with the nonmetaphor (M = 3.25), both the explicit metaphor
product: M = 4.55 versus 3.05; hedonic product: M = 4.03 ver- (M = 4.71) and implicit metaphor conditions (M = 4.58)
sus 2.91). No interaction effects among ad type, product type, significantly improved attitude toward the advertised brand
and NFC were found. Thus, the manipulation of ad type was (F (1,407) = 33.84, p < .001). Post hoc tests with Bonferroni
successful. adjustments were further performed. No difference was found
Consistent with pretest results, the hedonic product (M = between the explicit and implicit metaphors (F (1,265) = 0.51)
5.80) was perceived to be more hedonic than the utilitarian while a difference was found between the explicit metaphor
product (M = 3.51) (F (1,407) = 141.82, p < .001), and the ad and the nonmetaphor ad (F (1,274) = 36.21, p < .001). A
difference was also observed between the implicit metaphor ad
TABLE 1 and the nonmetaphor ad (F (1,274) = 31.13, p < .001). The
Summary of Manipulation Checks on Ad Type: patterns for purchase intentions are similar to those for attitude
Percentage of Participants in Each Condition Correctly (Mimplicit metaphor = 3.88, Mexplicit metaphor = 3.94, Mnonmetaphor =
Identified the Ad Type 3.22) (F (1,407) = 12.71, p < .001). Post hoc tests showed a
difference between the explicit metaphor and the nonmetaphor
Utilitarian product Hedonic product
ad (F (1,274) = 13.19, p < .001) and a difference between
Study 1 (shampoo) (chocolates)
the implicit and the nonmetaphor ads (F (1,274) = 11.45, p <
Nonmetaphor ad 94% 94% .001). Thus, hypothesis 1 was fully supported.
Explicit metaphor ad 83% 84% With regard to the two types of metaphor, participants’
Implicit metaphor ad 85% 91% responses were affected by the three-way interaction among
metaphor type, product type, and NFC (attitude: F (1,265) =
Utilitarian attribute Hedonic product 4.90, p < .05; purchase intentions: F (1,265) = 5.15, p < .05).
Study 2 (Internet searching) (movie watching) For participants with high NFC, we predicted that influences of
explicit and implicit metaphors on attitude toward the advertised
Nonmetaphor ad 85% 82%
brand and purchase intentions would be contingent on product
Explicit metaphor ad 91% 88%
types. As expected, the results were significant (attitude: F
Implicit metaphor ad 86% 80%
(1,149) = 7.57, p < .01; purchase intentions: F (1,149) = 7.71,
 MISSING INGREDIENTS IN METAPHOR ADVERTISING 85

TABLE 2
Univariate Results for Attitude Toward the Advertised Brand and Purchase Intentions
Attitude toward the brand Purchase intentions
Independent variables F value df η2 F value df η2
Study 1
Ad type (AT) 32.70∗∗ (2,203) 0.05 7.62∗∗ (2,203) 0.07
Product type (PT) 13.84∗∗ (1,203) 0.07 12.71∗∗ (1,203) 0.06
Need for cognition (NFC) 5.18∗ (1,203) 0.03 5.27∗ (1,203) 0.03
AT × PT 5.01∗ (2,203) 0.03 3.61∗ (2,203) 0.04
AT × NFC 1.76 (2,203) 0.02 2.18 (2,203) 0.01
PT × NFC 2.06 (1,203) 0.01 1.83 (1,203) 0.01
AT × PT × NFC 4.90∗ (2,203) 0.03 5.15∗ (2,203) 0.03
Study 2
Ad type (AT) 4.52∗ (2,311) 0.03 4.98∗ (2,311) 0.03
Product attribute (PA) 2.83 (1,311) 0.02 1.59 (1,311) 0.01
Downloaded by [University of La Rioja] at 04:34 13 February 2014

Need for cognition (NFC) 0.75 (1,311) 0.01 0.12 (1,311) 0.01
AT × PA 1.16 (2,311) 0.01 1.48 (2,311) 0.01
AT × NFC 1.49 (2,311) 0.01 0.95 (2,311) 0.01
PA × NFC 0.83 (1,311) 0.01 0.72 (1,311) 0.01
AT × PA × NFC 4.12∗ (2,311) 0.02 3.99∗ (2,311) 0.02
∗
p < .05; ∗∗ p < .01.

p < .01). The implicit metaphor was more effective than the ex- condition than in the low-NFC condition. More thoughts were
plicit metaphor in promoting the hedonic product (attitude: Ms = expected to be generated by the implicit metaphor condition
5.35 and 4.99, F (1,65) = 5.13, p < .05; purchase intentions: Ms than the nonmetaphor condition, while the explicit metaphor
= 4.93 and 4.13, F (1,65) = 14.02, p < .01). The results were condition would fall somewhere in between. More favorable
consistent with the predictions in hypothesis 2. The explicit thoughts were expected to be generated by the utilitarian
metaphor was more effective than the implicit metaphor in product with the explicit metaphor condition than the utilitarian
promoting the utilitarian product (attitude: Ms = 5.23 and 4.06, product with the implicit metaphor condition. In contrast,
F (1,83) = 9.18, p < .01; purchase intentions: Ms = 4.44 more favorable thoughts were expected to be generated by the
and 3.55, F (1,83) = 4.07, p < .05). Thus, hypothesis 3 was hedonic product with the implicit metaphor condition than the
supported. hedonic product with the explicit metaphor condition. The dif-
For participants with low NFC, we predicted that influences ferences should occur only for people with higher NFC, not for
of explicit and implicit metaphors would not differ in utilitarian those with lower NFC. To test those expectations, two separate
and hedonic product ads. As expected, the difference was not ANOVA procedures were conducted, one with the total number
significant (attitude: F (1,119) = 0.03; purchase intentions: of thoughts and the other with the number of favorable thoughts
F (1,119) = 0.01). We observed only the main effect of product as the dependent variable. Ad type, product type, and NFC
type (attitude: Ms = 4.86 and 3.90, F (1,183) = 9.13, p < were the independent variables in the first procedure, and the
.01; purchase intentions: Ms = 3.98 and 3.13, F (1,183) = focus was on two metaphor conditions in the second procedure.
7.46, p < .01). Collectively, these findings provided support for First, the ANOVA on total thoughts showed significant main
hypothesis 4. effects of NFC (F = 16.89, p < .01; Mhigh = 3.98, Mlow =
2.87) and ad type (F = 18.12, p < .01; Mimplicit metaphor = 3.87,
Other Results on Cognitive Responses Mexplicit metaphor = 3.23, Mnonmetaphor = 2.55). No other effects
(Thought-Listing Analysis) were significant. Post hoc tests confirmed that the differences
In all, 369 participants answered the open-ended question. among three ad types were all significant (implicit versus ex-
Nonresponse rates were similar across conditions (F = 1.23). plicit metaphor: F = 6.12, p < .05; explicit versus nonmetaphor:
Two judges, who were blind to experimental conditions, F = 6.49, p < .05). Second, the ANOVA on favorable thoughts
independently coded the listed thoughts into favorable, neutral, showed an interaction effect among metaphor type, product
or unfavorable to the ad message. Interrater agreement was high type, and NFC (F = 5.12, p < .05). For the participants with
(γ = .90), and disagreements were resolved through discussion. higher NFC, the utilitarian product with the explicit metaphor
More thoughts were expected to be generated in the high-NFC (M = 3.01) was associated with more favorable thoughts
 86 CHANG AND YEN

than were the products with implicit metaphors (M = 2.35) To generalize the results of Study 1, a nontraditional student
(F = 5.57, p < .05). The hedonic product with the implicit sample was employed in Study 2.
metaphor (M = 3.17) was associated with more favor-
able thoughts than was the product with the explicit metaphor Study 2
(M = 2.55) (F = 5.92, p < .05). No such differences were found
Instead of manipulating product type, perceived product
in the participants with lower NFC. The results are consistent
characteristics were directly measured. It was thus important
with the findings of the previously reported hypothesis tests.
to choose a product containing both utilitarian and hedonistic
qualities; thus, a mobile phone was selected (Kim and Huang
Follow-Up for Study 1
2010). Utilitarian tendency refers to the user’s preference to use
With regard the chocolate executions in the experiment, the
mobile services that have more functional and economically
text in the nonmetaphor condition (“Rich and creamy”) is more
meaningful values, such as mobile shopping, mobile banking,
likely to promote notions of calories and weight gain than is
and searching for news. Hedonic tendency refers to the pref-
the text in the metaphoric condition (“You would like to make
erence to use mobile services that provide emotional or psy-
a sweet commitment”). In addition, the right-hand corner illus-
chological value, such as mobile chatting with friends, mobile
tration in the implicit metaphor condition reveals the product’s
games, sports, MP3 music playing, and watching movies.
decorative packaging. This product feature is not communicated
in the explicit metaphor condition. In addition, the similarity be-
Downloaded by [University of La Rioja] at 04:34 13 February 2014

tween the wrapped product and a ring in the explicit metaphor Test Attributes and Brand Name Selection
condition may suggest there is a hardness to the product. To To determine the manipulation of product attributes, a pretest
rule out these alternative explanations for the relatively unfa- was conducted. After reviewing marketing documents (e.g.,
vorable advertising effects that resulted from these conditions, catalogs, and webpages) and previous studies about mobile
a follow-up study was conducted with an independent sample phone shoppers’ perceptions (e.g., Chitturi, Raghunathan, and
of undergraduate students (N = 62, 30 males and 32 females). Mahajan 2007; Riquelme 2001), we selected the eight salient
Respondents viewed one of the three ads used in hedonic attributes for use in the pretest. A total of 50 undergraduate
product (chocolates) and then answered questions on product students (23 males and 27 females) at a university in southern
perceptions associated with weight (the product in the ad make Taiwan participated. They were asked to evaluate these attributes
me think of calories and weight gain), perceived product hard- as to their perceived value on 7-point scales from Not important
ness (the ad suggests a hardness to the product), and decorative to Very important. For each attribute, the participants were also
quality (the ad is associated with product’s decorative package) asked to indicate the degree of utilitarian/hedonic value on a
on 7-point Likert scales (1 = Disagree, 7 = Agree). There were 7-point scale from Totally utilitarian to Totally hedonic. Par-
no statistically significant differences on distribution of demo- ticipants rated Internet search (M = 5.51) and video watching
graphics. The three measures were observed to be insignificant (M = 5.32) as most utilitarian and hedonic, respectively. There
among the three ad conditions (F = 1.55, 0.98, and 1.23 re- was no significant difference in perceived importance (T = 0.99)
spectively). Those measures were confirmed not to confound and perceived value (T = 1.93) of the two selected attributes.
the experimental results. Detailed results are summarized in the A two-dimensional scale from Voss, Spangenberg, and
upper portion of Table 4. Grohmann (2003) was used to measure utilitarian/hedonic per-
ceptions. Participants were asked to rate the extent to which
Study 1 Discussion they felt the use of a mobile phone was perceived as ineffective/
The experimental ads worked as expected, with the effects effective, unhelpful/helpful, not functional/functional, unneces-
of explicit and implicit metaphors being heterogeneous, subject sary/necessary, impractical/practical, not enjoyable/enjoyable,
to product type and NFC. Nevertheless, a major limitation of not delightful/delightful, dull/exciting, not fun/fun, and not
Study 1 is that different products were selected to represent util- thrilling/thrilling in 7-point semantic scales. A factor analy-
itarian and hedonic products. Even with extra care in choosing sis with varimax rotation confirmed the two dimensions with
test products with similar degrees of product familiarity and the first five items grouped as the utilitarian perceptions (α =
favorability, the influence of product selection itself cannot be 0.92) and the later five grouped as the hedonic perceptions (α =
ruled out. In addition, the manipulations of metaphor explicit- 0.87). Indices of utilitarian/hedonic perceptions were created by
ness between utilitarian and hedonic products were not consis- calculating the mean of each of the five items.
tent. While the visuals used in the hedonic product conditions The same participants were given a list containing 10 brand
were similar, the visual differences between the nonmetaphor names of mobile phones and were asked to circle the three brand
ad and the two metaphor ads of the utilitarian product condition names with which they were most familiar. Nokia, Samsung, and
could also have been confounding. Therefore, it is important LG were identified as the top three, with scores of 95%, 89%,
that the advertising effects be compared when the same product and 82%, respectively. The results are similar to the worldwide
is tested. A pretest was also conducted to minimize differences cell phone market share statistics and consumer preferences
in perceived product benefits among experimental conditions. released by ABI Research (2008). Nokia was thus chosen as the
 MISSING INGREDIENTS IN METAPHOR ADVERTISING 87

test brand to make our experimental ads realistic. Respondents and a two-dimensional scale to measure utilitarian/hedonic per-
in the pretest did not participate in the main study. ceptions. Once the computer detected the completeness of these
questions, it activated the “continue” button.
Stimulus Materials Depending on the net value of utilitarian/hedonic percep-
For the nonmetaphoric condition, “Now you can use Inter- tions, the respondents were then exposed to one of two condi-
net any time, everywhere” and “Now you can watch movies tions: a mobile phone framed as utilitarian (Internet search) or
any time, everywhere” were the statements used for utilitar- hedonic (video watching). If the net value of a participant’s per-
ian and hedonic versions, respectively. “Control the world us- ceptions was higher than zero (i.e., he or she viewed a mobile
ing your hands” and “Enjoy your own luxury movie theater” phone as more utilitarian), that participant would be directed
were metaphoric headlines describing the utilitarian and hedonic to the ad depicting the utilitarian attribute. If the net value was
value of the mobile phone, respectively. The same headline was lower than zero (i.e., he or she viewed a mobile phone as more
used in the implicit and explicit metaphor ads (see Appendix 1). hedonic), then the individual would be assigned to view the
A computer mouse linked to the Internet with the Google home- ad portraying the hedonic attribute. If the net value was zero,
page on the screen provided a metaphoric image for framing the the system chose randomly whether the participant would be ex-
mobile phone as utilitarian. The only difference between the ex- posed to the utilitarian or the hedonic condition. After determin-
plicit and implicit metaphor ads was whether or not the mobile ing which condition the respondent was assigned, one of three ad
phone was used to replace the mouse. To frame the mobile phone types (implicit metaphor versus explicit metaphor versus non-
Downloaded by [University of La Rioja] at 04:34 13 February 2014

as hedonic, a single person alone in a luxury movie theater, rais- metaphor) was randomly selected to be the stimulus ad. After
ing his arms, and sitting in the middle provided the feeling that viewing the designed ad, the respondents completed a ques-
the entire movie theater was owned by one person. The movie tionnaire containing a thought-listing task, dependent measures,
Avatar was chosen to be on the screen because it was the most and manipulation checks on metaphoric presentation. Similar to
popular movie at the time this study was conducted. The differ- Study 1, two sets of questions were used to assess the manipula-
ence between the explicit and implicit ads was whether or not tion check of metaphoric type.3 Finally, they clicked the Submit
the screen on the theater was replaced with the phone. All the button. The participants were debriefed and thanked. The study
ads were A4 size and were inspired by existing branded ads. took most participants approximately 20 minutes.

Pre- and Postmanipulation Measures Results of Study 2

Premanipulation measures included brand familiarity, brand With regard to the distribution of background demographics
favorability, and utilitarian/hedonic perceptions. Two questions and NFC among the experimental groups, no significant differ-
assessed participants’ perceptions of favorability and familiar- ences were detected. The scales for NFC, purchase intentions,
ity with the Nokia brand based on 7-point semantic differen- attitude toward the advertised brand, and utilitarian and hedo-
tial scales. The two-dimensional scale (Voss, Spangenberg, and nic evaluation exhibited high reliability (α values of 0.87, 0.88,
Grohmann 2003) measured utilitarian/hedonic perceptions. The 0.89, 0.91, and 0.89, respectively). No statistically significant
net value was calculated by taking the difference between the in- difference was shown in the measures of brand favorability (F =
dices of utilitarian and hedonic perceptions. Measures in Study 1.01) or familiarity (F = 0.56) across experimental conditions.
2 were similar to those in Study 1. No moderating effects of the demographic variables on attitude
toward the advertised brand or purchase intentions were found.
Participants and Procedure As in Study 1, the sample was divided into high and low groups
A nontraditional student sample was chosen because mobile based on a median split on the NFC scores (Mhigh NFC = 4.89,
phones with the selected attributes are generally not cheap. A Mlow NFC = 3.69) (F = 272.89, p < .001).
total of 312 part-time undergraduate and graduate students (170
males and 142 females) enrolled at a university in southern Manipulation Checks
Taiwan participated in the experiment. For their participation, As expected, a manipulation check showed a significant
the students were either offered extra credit or were given credit effect of ad type on perceived metaphoric level for both selected
to be counted as a part of their course work. Respondents ranged attributes (utilitarian: F (2,131) = 4.82, p < .05; hedonic:
in age from 20 to 68 (M = 32.7, SD = 8.41). F (2,179) = 5.12, p < .05). The implicit metaphor was
Prior to the experiment, researchers first contacted instructors perceived to be more metaphoric than the explicit metaphor
who held their classes in a computer lab. Arrangements were (utilitarian: M = 4.79 versus 4.29; hedonic: M = 5.45 versus
made with course instructors for the experiment to take place. 4.83). The explicit metaphor was viewed as more metaphoric
The experiments were conducted during the regular class by an than the nonmetaphor (utilitarian: M = 4.29 versus 3.51;
administrator who was blind to the research hypotheses. The test hedonic: M = 4.83 versus 4.11). Chi-square tests were then
instrument was accessed via a website. Participants were told to performed to evaluate whether the visual representation was
evaluate an ad that would be launched for a forthcoming adver- perceived as intended. The results were both significant in
tising campaign. They first filled out a premanipulation ques- both product attributes (utilitarian: χ 2 = 24.85, p < .001;
tionnaire, including perceived brand familiarity and favorability hedonic: χ 2 = 34.94, p < .001). Most participants correctly
 88 CHANG AND YEN

identified what metaphor type the ad was designed to convey. are similar to those for attitude toward the advertised brand
Results are summarized in the lower portion of Table 1. None (Mimplicit metaphor = 3.89, Mexplicit metaphor = 3.85, Mnonmetaphor =
of interactions among ad type, product attribute, and NFC were 2.85) (F (1,311) = 8.91, p < .01). Post hoc tests indicated that
found. Thus, the manipulation of ad type was successful. the differences were both significant (explicit metaphor versus
nonmetaphor: F (1,205) = 10.23, p < .01; implicit metaphor
Results of Multivariate Analysis of Variance versus nonmetaphor: F (1,205) = 9.01, p < .01). These results
and Univariate Results gave support to hypothesis 1.
MANOVA was conducted after confirming that attitude to- With regard to the two types of metaphor, participants’
ward the advertised brand and purchase intentions were highly attitude toward the advertised brand was affected by the
correlated (γ = 0.51, p < .01). A significant main effect of three-way interaction among metaphor type, product type, and
product attribute (F = 3.05, p < .05) and a three-way interac- NFC (attitude: F (1,206) = 4.12, p < .05; purchase intentions:
tion among ad type, product attribute, and NFC were observed F (1,206) = 3.99, p < .05) (see Table 3 for the means across
(F = 3.21, p < .05). Further examinations of univariate results experimental groups). For participants with high NFC, an inter-
and mean comparisons were then performed. The univariate action between product attribute and metaphor type was found
results are summarized in the lower portion of Table 2. on attitude (F (1,101) = 5.34, p < .05) and purchase intentions
(F (1,101) = 5.16, p < .05). When the mobile phone was framed
Tests of Main Hypotheses as hedonic, the implicit metaphor was more effective than the
Downloaded by [University of La Rioja] at 04:34 13 February 2014

The main effect of ad type was observed (F (1,311) = 9.43, explicit metaphor (attitude: Ms = 4.87 and 4.45, F (1,43) =
p < .01). Both the explicit metaphor (M = 4.99) and the implicit 4.65, p < .05; purchase intentions: Ms = 4.10 and 3.51, F (1,43)
metaphor (M = 4.83) led to more positive attitude toward = 4.72, p < .01). Hypothesis 2 was thus supported. When the
the advertised brand than the nonmetaphor (M = 3.83). Post mobile phone was framed as utilitarian, the explicit metaphor
hoc tests with Bonferroni adjustments were further performed. was more effective than the implicit metaphor (attitude: Ms =
No difference was found between the explicit and implicit 5.30 and 4.88, F (1,57) = 4.95, p < .01; purchase intentions:
metaphors (F (1,205) = 0.38), while a difference was found Ms = 4.23 and 3.65, F (1,57) = 5.39, p < .05). Hypothesis
between the explicit metaphor ad and the nonmetaphor ad 3 was thus supported. For low-NFC participants (N = 160),
(F (1,205) = 10.23, p < .01). A difference was also observed influences of explicit and implicit metaphors on attitude
between the implicit metaphor ad and the nonmetaphor ad toward the advertised brand and purchase intentions were not
(F (1,205) = 9.01, p < .01). The patterns for purchase intentions significantly different under either the utilitarian and hedonic

TABLE 3
Advertising Effectiveness as a Function of Ad Type, Product Type, and NFC
Utilitarian product Hedonic product
High NFC Low NFC High NFC Low NFC
Study 1
Attitude toward the advertised brand as the dependent variable
Nonmetaphor ad 3.25 (0.99) 2.88 (1.26) 3.33 (1.27) 3.54 (1.00)
Explicit metaphor ad 5.23 (1.10) 3.84 (1.16) 4.99 (0.64) 4.78 (1.21)
Implicit etaphor ad 4.06 (1.27) 3.96 (1.70) 5.35 (1.38) 4.94 (0.73)
Purchase intentions as the dependent variable
Nonmetaphor ad 3.07 (1.23) 2.94 (1.10) 3.13 (0.99) 3.75 (0.98)
Explicit metaphor ad 4.44 (1.39) 3.18 (1.09) 4.13 (0.93) 4.02 (1.28)
Implicit metaphor ad 3.55 (1.38) 3.08 (1.49) 4.93 (1.28) 3.94 (0.79)
Study 2
Attitude toward the advertised brand as the dependent variable
Nonmetaphor ad 3.53 (1.73) 4.24 (1.43) 3.46 (1.52) 4.14 (1.09)
Explicit metaphor ad 5.30 (1.03) 4.88 (0.87) 4.45 (1.26) 5.33 (0.92)
Implicit metaphor ad 4.88 (1.17) 4.56 (1.15) 4.87 (0.74) 5.00 (0.87)
Purchase intentions as the dependent variable
Nonmetaphor ad 2.96 (1.10) 2.72 (1.33) 3.10 (1.52) 2.76 (0.90)
Explicit metaphor ad 4.23 (1.02) 3.75 (0.83) 3.51 (1.15) 4.10 (1.13)
Implicit metaphor ad 3.65 (1.43) 3.70 (1.14) 4.10 (1.09) 3.81 (1.40)
Note: Numbers in parentheses are standard deviations of treatment cells.
 MISSING INGREDIENTS IN METAPHOR ADVERTISING 89

conditions (attitude: F (1,105) = 0.89; purchase intentions: of bipolar adjectives (easy/difficult to understand, and straight-
F (1,105) = 1.21). Only the main effect of product attribute on forward/confusing) (Gkiouzepas and Hogg 2011; correlation =
attitude was found (Mhedonic versus Mutilitarian = 5.15 and 4.75, 0.90) were used. In the utilitarian conditions, to check whether
F (1,105) = 3.98, p < .05). Hypothesis 4 was supported. the use of words “control the world” in the headlines of the
metaphor or whether the illustration of the connection between
the phone and a computer may render the product more attrac-
Other Results on Cognitive Responses tive, perceived benefit of control was assessed with a Likert scale
(Thought-Listing Analysis) (ad is associated with power of control). In the hedonic condi-
In all, 281 participants answered the open-ended question. tions, to check whether watching a movie through touch-screen
Nonresponse rates were similar across conditions (F = 1.42). No phones confers a product advantage, perceived attribute (watch-
participants mentioned Avatar, the movie shown on the Nokia ing movie through touch-screen phones is a unlikable/likable at-
phone’s screen in the ad, suggesting little influence of movie tribute) was assessed. All the items were measured using 7-point
affinity. Similar to Study 1, two judges coded the thoughts scales.
independently. Interrater agreement was high (γ = .86). The Among experimental groups, there were no statistically sig-
patterns of results were similar to those in Study 1, and the ex- nificant differences on distribution of the selected demographic
pectations were supported by the analysis. First, the ANOVA variables. Picture realism (F = 0.91) was found insignificant
on total thoughts showed significant main effects of NFC (F = across six experimental conditions. Participants’ comprehen-
Downloaded by [University of La Rioja] at 04:34 13 February 2014

13.12, p < .01; Mhigh = 3.67, Mlow = 2.91) and ad type (F = sion (F = 1.05) and fit level (F = 0.80) were confirmed to be
10.31, p < .01; Mimplicit metaphor = 3.63, Mexplicit metaphor = 3.24, similar between explicit and implicit metaphors. Comprehen-
Mnonmetaphor = 2.88). No other effects were significant. Post hoc sion (F = 0.93) and fit level (F = 0.56) were confirmed to
tests confirmed that differences among the three ad types were be similar between utilitarian and hedonic versions. Perceived
all significant (implicit versus explicit metaphor: F = 5.56, p < benefit of control was insignificant in utilitarian conditions (F =
.05; explicit versus nonmetaphor: F = 5.12, p < .05). Second, 1.78). Perceived attribute was also insignificant in the hedonic
the ANOVA on favorable thoughts showed an interaction effect conditions (F = 1.51). Detailed results are summarized in the
among metaphor type, product type, and NFC (F = 5.89, p < lower portion of Table 4. The results provided evidence that the
.05). For the participants with higher NFC, the utilitarian prod-
uct with the explicit metaphor (M = 3.21) was associated with
more favorable thoughts than that with the implicit metaphor TABLE 4
(M = 2.56) (F = 5.87, p < .05). The hedonic product with the Results of Follow-Ups
implicit metaphor (M = 3.49) was associated with more favor- Explicit Implicit
able thoughts than that with the explicit metaphor (M = 2.89) Nonmetaphor metaphor metaphor
(F = 5.73, p < .05). No such differences were found in the
participants with lower NFC. The results are consistent with the Study 1 Follow-Up (Hedonic Products Only)
findings of the previously reported hypothesis tests. Product perceptions 5.02 (1.23) 4.92 (1.32) 4.88 (1.19)
associated with
weight
Follow-Up for Study 2 Perceived product 4.59 (1.11) 4.72 (1.21) 4.81 (1.25)
Although the thought-listing results provided participants’ hardness
cognitive responses to support the hypotheses, it is difficult Decorative quality 4.88 (0.94) 4.99 (1.05) 5.06 (1.12)
to rule out the alternative explanation that the results might Study 2 Follow-Up
be explained by picture realism, fit level, comprehension, and Utilitarian attribute:
product perception across conditions.4 To address these issues, Picture realism 6.09 (0.98) 6.01 (1.05) 5.97 (0.91)
a follow-up study was undertaken with an independent sample Fit level 6.09 (0.85) 5.88 (0.92) 5.95 (1.06)
of part-time undergraduate students (N = 145, 70 males and 75 Comprehension 6.14 (0.87) 5.33 (0.92) 5.45 (0.98)
females). Participants saw one of the six ads used in Study 2 Perceived benefit of 5.56 (1.10) 5.76 (1.12) 5.80 (1.04)
and answered a battery of questions. control
Picture realism was measured by two semantic differential
Hedonic attribute:
items (realistic/unrealistic and real/fictitious) (Gkiouzepas and
Picture realism 6.21 (1.09) 6.13 (1.17) 6.19 (1.03)
Hogg 2011; correlation = 0.88). Fit level was measured using
Fit level 6.02 (1.03) 5.89 (1.11) 5.70 (1.04)
a three-item Likert scale (ad is appropriate for the product; ad
Comprehension 6.08 (0.88) 5.21 (0.92) 5.14 (0.98)
highlights product features; ad is good match with the product
Perceived attribute 5.38 (1.11) 5.24 (1.04) 5.30 (1.09)
depicted) (Putrevu 2008; α = 0.86). To check whether or not the
predicted advantages of metaphor incur a cost in terms of com- Note: Numbers in parentheses are standard deviations of treatment
prehension, a semantic differential scale consisting of two pairs cells.
 90 CHANG AND YEN

manipulations of ad type confounded with the communication aging information efficiently while the pleasure of metaphoric
of product benefits could be minimized. processing provides sufficient interest to process information.
An implicit metaphor may undermine persuasion by directing
attention away from the advertising claim because implicitness
GENERAL DISCUSSION may be too complex and cause cognitive overload for the one
We worked on metaphor advertising by identifying bound- processing the ad message.
ary conditions associated with the influence of two types of Third, implicit metaphors had an advantage when promoting
metaphor on consumers’ evaluation of an advertised prod- hedonic products. Ads using implicit metaphors seemed to pro-
uct/brand. Our taxonomy extends that proposed by Gkiouzepas vide more opportunities for imagination, befitting the nature of
and Hogg (2011). Specifically, Gkiouzepas and Hogg found that hedonic products. Imagination could provide a beneficial impact
juxtapositioning or synthesizing the advertised product in a vi- by communicating sensory information that reinforces advertis-
sual had an effect on the ad’s effectiveness when the objects in ing claims, given that in hedonic consumption an individual is
the visual comparison were conceptually similar (e.g., an or- prone to adopt sensory processing. Aesthetics generated by im-
ange and marmalade) but had no effect when the objects in the plicit metaphors may provide relevant information out of sen-
comparison were conceptually dissimilar (e.g., a car and tan- sory cues. Therefore, compared with explicit metaphors, ads
ning bed). Study 1 in the current research shows that metaphor using implicit metaphors were more likely to arouse favorable
type (explicit versus implicit) influences the ad’s effectiveness attitude toward the advertised brand and raise purchase inten-
Downloaded by [University of La Rioja] at 04:34 13 February 2014

for high-NFC consumers, even when dissimilar objects form tions when the product or value is perceived as hedonic. The
the basis of the visual metaphor (an eraser and a bottle of sham- interaction between product type and type of metaphor reflected
poo, or a chocolate and a diamond). The results indicated that resource-matching theory: persuasion is optimized when the de-
focusing on the comparison between two types of metaphor mands of an advertising execution match the cognitive resources
without considering other factors might be overly simplistic. of the ad viewer. The thought-listing analysis also provided ev-
Four specific observations are noteworthy. idence of attention and information processing, and the results
First, the effects of metaphor type were contingent on an could be used to support the hypotheses.
individual’s NFC. Metaphors engender deeper processing of Fourth, metaphor type did not matter to low-NFC individuals,
messages and enhance attitude and purchase intentions when and only the superior effects of product hedonic value over util-
NFC is high. Some metaphor types worked better on partici- itarian value were found. This finding was consistent with past
pants with high NFC. However, these differences did not emerge research suggesting that low-NFC individuals are more prone to
for participants low in NFC. The findings suggest that visual the influences of symbolic cues and avoid elaborative processing
metaphors are more likely to be successfully interpreted by (Martin, Sherrard, and Wentzel 2005). Because of elaboration
people with higher NFC. The results were relevant to previous avoidance, people with low NFC tend to be influenced by ad-
research (e.g., Haugtvedt, Petty, and Cacioppo, 1992; Petty and vertising’s symbolic aspects, like sensory cues. Hedonic needs
Cacioppo 1986) that low-NFC individuals’ response to a persua- reflect sensory pleasure and sensory cues are highlighted in ads
sive message are less influenced by the message-relevant cues. of hedonic products. Hedonic information was thus preferred
The findings also echo Jeong (2008) in that visual metaphors de- by low-NFC individuals.
liver arguments central to the message content, which influences The findings have important implications for practice. In
cognitive elaboration and advertising persuasion. terms of message strategy, an ad message should, of course,
Second, an interaction between product type and metaphor convey product value. The current investigation provides guid-
type on high-NFC people was observed. This study showed ance for practitioners on how to frame the value of the product
systematic effects on consumer responses by comparing two by choosing the right type of metaphor and frame for the ad. Per-
different product types. Study 1 used fictitious brand names for ceived utilitarian or hedonic value determines which metaphor
high-purchase frequency and low-value products. Study 2 fo- type should be used. Altering the metaphor type is a relatively
cused on a high-price product with a real brand name. The results straightforward task. An appropriately presented visual should
provided converging evidence that consumers with high NFC be incorporated into an existing marketing message or media
show a more positive attitude and higher purchase intentions campaign to increase its effectiveness when facing consumers
in response to an explicit metaphor than an implicit metaphor with high NFC.
when utilitarian value is highlighted. The explicit execution of The de facto market segmentation based on NFC may offer
the product image fits into the rational appeal. We used resource- potential payoffs to advertisers who strive to discover and take
matching theory (Anand and Sternthal 1989; Martin, Sherrard, advantage of the diversity of audience characteristics. Although
and Wentzel 2005; Peracchio and Meyers-Levy 1997) to explain advertisers who would like to employ metaphors in their ads
the superiority of explicitness under such circumstances. When cannot administer NFC scales to members of their target audi-
facing a product with utilitarian value, an individual is prone to ence, such difficulty does not diminish the importance of under-
evaluate the advertising claim directly. The explicit execution standing how NFC affects metaphor processing. One can argue
of the product image in the metaphor assists him or her in man- that differences in NFC among individuals are tantamount to
 MISSING INGREDIENTS IN METAPHOR ADVERTISING 91

unplanned market segmentation, and that advertisers may im- the hedonic product condition (chocolates), three choices were (1) a
prove advertising results through judicious ad placement be- model wearing a chocolate on a ring, (2) a model wearing a diamond
cause audiences with different degrees of NFC may have dif- ring, and (3) a model’s hand without anything worn on the fingers.
ferent media habits. For example, certain technical publications 3. Participants first had to identify the visual image from three descrip-
(e.g., Wall Street Journal) may attract audience groups with tions. In the utilitarian attribute condition, the three options were
(1) a Nokia phone displaying the Google homepage on its screen,
higher NFC more than less cognitively demanding magazines
(2) a Nokia phone connected to a computer, and (3) a computer
(e.g., People magazine or tabloid-type magazines). mouse connected to a computer. In the hedonic attribute condition,
the three descriptions were (1) a Nokia phone used to watch a movie,
DIRECTIONS FOR FUTURE RESEARCH (2) a person alone watching a movie in a luxury movie theater, and
This research raised a number of issues that should be ex- (3) a person alone watching a movie from a Nokia screen.
amined in more depth in future research. One avenue is to work 4. The authors thank the anonymous reviewer who raised these issues.
on the combination of visual and verbal metaphors. This re-
search focused only on the visual–verbal congruence. Future
research may compare their separate and additive effects based REFERENCES
ABI Research (2008), “Mobile Phone Market Shrugging Off Economic
on product type and NFC. In addition, artful deviations from Downturn; 2007 Shipments Increased by 15.8% to 1.15 Billion,” available at
metaphors could be studied. Some metaphors are less familiar http://www.thefreelibrary.com/Mobile+Phone+Market+Shrugging+Off+
Downloaded by [University of La Rioja] at 04:34 13 February 2014

to consumers and require more effort before they can be compre- Economic+Downturn%3B+2007+Shipments. . .-a0173759229
hended (Phillips and McQuarrie 2004). Some metaphors deviate Anand, Punam, and Brian Sternthal (1989), “Strategies for Designing Persua-
too much and create confusion that may backfire in attitude and sive Messages: Deductions from the Resource Matching Hypothesis,” in
Cognitive and Affective Responses to Advertising, Patricia Cafferata and
intentions. Given the differences noted, an interesting avenue Alice M. Tybout, eds., Lexington, MA: Lexington, 135–59.
for future research would be to consider the potential effects Ang, Swee Hoon (2002), “Effects of Metaphoric Advertising Among Mainland
of complexity and richness of metaphor itself with variables Chinese Consumers,” Journal of Marketing Communications, 8 (3), 178–88.
studied in this research. Ang, Swee Hoon, and Elison A.C. Lim (2006), “The Influence of Metaphors
What other boundary conditions of individual differences and Product Type on Brand Personality Perceptions and Attitudes,” Journal
of Advertising, 35 (2), 39–53.
may come into play? Metaphoric thinking ability is found to Brennan, Ian, and Kenneth D. Bahn (2006), “Literal versus Extended Symbolic
affect creativity (Burroughs and Mick 2004), which implies that Messages and Advertising Effectiveness: The Moderating Role of Need for
individuals with higher metaphoric thinking ability are more Cognition,” Psychology and Marketing, 23 (4), 273–95.
likely to appreciate metaphor ads and be better able to dif- Brislin, Richard W. (1987), “The Wording of Translation of Research Instru-
ferentiate different types of metaphors. In addition, message- ments,” in Field Methods in Cross-Cultural Research, Walter J. Lonner and
John W. Berry, eds., Beverly Hills, CA: Sage, 137–64.
processing motivation can be elicited from product-category Burroughs, James E., and David Glen Mick (2004), “Exploring Antecedents
involvement (Peracchio and Meyers-Levy 1997). In contrast to and Consequences of Consumer Creativity in a Problem-Solving Context,”
NFC (i.e., a message recipient’s internal processing motivation), Journal of Consumer Research, 31 (2), 402–11.
product-category involvement is an externally induced interest Cacioppo, John T., and Richard E. Petty (1982), “The Need of Cognition,”
in a message. Future research may examine whether individual Journal of Personality and Social Psychology, 42 (January), 116–31.
Cacioppo, John T., Richard E. Petty, and C.F. Kao (1984), “The Efficient As-
differences in metaphoric thinking ability or product-category sessment of Need for Cognition,” Journal of Personality Assessment, 48 (3),
involvement play a role similar to that played by NFC in this 306–17.
research. Callow, Michael A., and Leon G. Schiffman (1999), “A Visual Esperanto?
This research suggests that not all metaphors are created The Pictorial Metaphor in Global Advertising,” in European Advances in
equal, at least in terms of metaphoric type. When seeking the Consumer Research, Bernard Dubois, Tina M. Lowrey, L.J. Shrum, and
Marc Vanhuele, eds., vol. 4, Provo, UT: Association for Consumer Research,
advantages that explicit and implicit pictorial metaphors pro- 17–20.
vide in obtaining and sustaining attention, advertisers should Chitturi, Ravindra, Rajagopal Raghunathan, and Vijay Mahajan (2007), “Form
consider product type and individual differences in NFC. Such versus Function: How the Intensities of Specific Emotions Evoked in Func-
knowledge is useful to advertisers who are interested in maxi- tional versus Hedonic Trade-Offs Mediate Product Preferences,” Journal of
mizing the impact of their advertising dollars. Marketing Research, 44 (4), 702–14.
Dhar, Ravi, and Klaus Wertenbroch (2000), “Consumer Choice Between Hedo-
nic and Utilitarian Goods,” Journal of Marketing Research, 37 (1), 60–71.
NOTES Frey, Kurt P., and Alice H. Eagly (1993), “Vividness Can Undermine the Per-
1. Dandruff is known as the most common hair problem in Taiwan suasiveness of Messages,” Journal of Personality and Social Psychology, 65
(1), 32–44.
(Landseed Hospital Newsletter, 2012). In a pretest, 40 undergradu-
Gkiouzepas, Lampros, and Margaret K. Hogg (2011), “Articulating a New
ates were asked to identify a hair problem after viewing either the
Framework for Visual Metaphors in Advertising,” Journal of Advertising,
explicit or the implicit ad. All of them associated dandruff with the 40 (1), 103–20.
advertised shampoo. Haugtvedt, Curt, Richard E. Petty, and John T. Cacioppo (1992), “Need for
2. In the utilitarian product condition (shampoo), three alternatives Cognition and Advertising: Understanding the Role of Personality Vari-
were (1) a couple smiling at each other, (2) a blackboard eraser in ables in Consumer Behavior,” Journal of Consumer Psychology, 1 (3),
a model’s hand, and (3) a bottle of shampoo in a model’s hand. In 239–60.
 92 CHANG AND YEN

Homer, Pamela, and Sun-Gil Yoon (1992), “Message Framing and the Interre- Morgan, Susan E., and Tom Reichert (1999), “The Message Is in the Metaphor:
lationships Among Ad-Based Feelings, Affect, and Cognition,” Journal of Assessing the Comprehension of Metaphors in Advertisements,” Journal of
Advertising, 21 (1), 19–33. Advertising, 28 (4), 1–12.
Jeong, Se-Hoon (2008), “Visual Metaphor in Advertising: Is the Persuasive Okada, Erica M. (2005), “Justification Effects on Consumer Choice of He-
Effect Attributable to Visual Argumentation or Metaphorical Rhetoric?,” donic and Utilitarian Goods,” Journal of Marketing Research, 42 (1), 43–
Journal of Marketing Communications, 14 (1), 59–73. 53.
Kahn, Uzma, Ravi Dhar, and Klaus Wertenbroch (2005), “A Behavioral De- Peracchio, Laura A., and J. Meyers-Levy (1997), “Evaluating Persuasion-
cision Theory Perspective on Hedonic and Utilitarian Choice,” in Inside Enhancing Techniques from a Resource-Matching Perspective,” Journal
Consumption: Frontiers of Research on Consumer Motives, Goals, and De- of Consumer Research, 24 (2), 178–91.
sires, S. Ratneshwar and David Glen Mick, eds., New York: Routledge, Petty, Richard E., and John T. Cacioppo (1986), Communication and Per-
144–65. suasion: Central and Peripheral Routes to Attitudes Change, New York:
Kassin, Saul M., Marisa E. Reddy, and William F. Tulloch (1990), “Juror In- Spring-Verlag.
terpretation of Ambiguous Evidence: The Need for Cognition, Presentation Petty, Richard E., John T. Cacioppo, and David T. Schumann (1983), “Central
Order, and Persuasion,” Law and Human Behavior, 14, 43–55. and Peripheral Routes to Advertising Effectiveness: The Moderating Role
Kim, Dan J., and Yujong Y. Huang (2010), “A Study of Mobile Internet User’s of Involvement,” Journal of Consumer Research, 10 (2), 135–46.
Service Quality Perceptions from a User’s Utilitarian and Hedonic Value Phillips, Barbara J., and Edward F. McQuarrie (2004), “Beyond Visual
Tendency Perspectives,” Information Systems Frontiers, 4, 63–84. Metaphor: A New Typology of Visual Rhetoric in Advertising,” Market-
Lagerwerf, Luuk, and Anoe Meijers (2008), “Openness in Metaphorical and ing Theory, 4(1/2), 113–36.
Straightforward Advertisements,” Journal of Advertising, 37 (2), 19–30. Putrevu, Sanjay (2008), “Consumer Responses Toward Sexual and Nonsexual
Landseed Hospital Newsletter (2012), “Dandruff Problems and Care,” Appeals: The Influence of Involvement, Need for Cognition (NFC), and
Downloaded by [University of La Rioja] at 04:34 13 February 2014

available at http://www.landseedhospital.com.tw/lishin/news/news inner. Gender,” Journal of Advertising, 37 (2), 57–69.

php?id=647 Putrevu, Sanjay, Joni Tan, and Kenneth R. Lord (2004), “Consumer Responses
Maheswaran, Durairaj, and Joan Meyers-Levy (1990), “The Influence of Mes- to Complex Advertisements: The Moderating Role of Need for Cognition,
sage Framing and Issue Involvement,” Journal of Marketing Research, 27 Knowledge, and Gender,” Journal of Current Issues and Research in Adver-
(3), 361–67. tising, 26 (1), 9–24.
Martin, Brett A., Michael J. Sherrard, and Daniel Wentzel (2005), “The Role Riquelme, Hernan (2001), “Do Consumers Know What They Want?” Journal
of Sensation Seeking and Need for Cognition on Web-Site Evaluations: of Consumer Marketing, 18 (5), 437–48.
A Resource-Matching Perspective,” Psychology and Marketing, 22 (2), Singer, Jerome L. (1966), Daydreaming: An Introduction to the Experimental
109–26. Study of Inner Experience, New York: Random House.
McQuarrie, Edward F., and David G. Mick (1999), “Visual Rhetoric on Adver- Toncar, Mark, and James Munch (2001), “Consumer Responses to Tropes in
tising: Text-Interpretive, Experimental, and Reader-Response Analyses,” Print Advertising,” Journal of Advertising, 30 (1), 55–65.
Journal of Consumer Research, 26 (1), 37–54. Voss, Kevin E., Eric R. Spangenberg, and Bianca Grohmann (2003), “Measuring
McQuarrie, Edward F., and David G. Mick (2003), “Visual and Verbal Rhetorical the Hedonic and Utilitarian Dimensions of Consumer Attitude,” Journal of
Figures under Directed Processing versus Incidental Exposure to Advertis- Marketing Research, 40 (3), 310–20.
ing,” Journal of Consumer Research, 29 (4), 579–87. Wood, Walter (1960), “Psychological Dimensions of Consumer Decision,” Jour-
McQuarrie, Edward F., and Barbara J. Phillips (2005), “Indirect Persuasion in nal of Marketing, 24 (1), 15–19.
Advertising: How Consumer Process Metaphor Presented in Pictures and Zhang, Yong (1996), “Responses to Humorous Advertising: The Moderating
Words,” Journal of Advertising, 34 (2), 7–20. Effect of Need for Cognition,” Journal of Advertising, 25 (1), 15–32.
 MISSING INGREDIENTS IN METAPHOR ADVERTISING 93

APPENDIX 1: EXPERIMENTAL MATERIALS USED IN STUDY 1 AND 2 (TRANSLATED INTO ENGLISH)

Downloaded by [University of La Rioja] at 04:34 13 February 2014
 94 CHANG AND YEN
Downloaded by [University of La Rioja] at 04:34 13 February 2014

Note: We hired a doctoral student—who is bilingual in English and Chinese and had work experience in an international advertising
agency—to translate the materials into English. Second, we used help from a bilingual English professor (living in Taiwan for more than
20 years) to translate them back to Chinese. After the back-translation, the original and back-translated instruments were compared and points
of divergence were noted. The translation was then corrected to more accurately reflect the intent of the wording in the original language. We
attempted to make sure that the only difference between the stimuli and the ones in this appendix is the language. (Color figures available
online).