You are on page 1of 18

Detangling consumer attitudes to better

explain co-branding success


Han-Chiang Ho
Department of Management and Marketing, Wenzhou-Kean University, Wenzhou, China
Nora Lado
Department of Business Administration, Universidad Carlos III de Madrid, Getafe, Spain, and
Pilar Rivera-Torres
Departamento de Dirección de Marketing e Investigación de Mercados, University of Zaragoza, Zaragoza, Spain

Abstract
Purpose – The purpose of this study is to examine consumer attitude toward a new type of co-branded products, which encompass attributes of
high-technology and luxury. The authors named these kinds of co-branded products as “high-tech luxury co-branded products” (HLCPs). Current
theoretical approaches used to study co-branding strategies cannot completely explain consumer attitude toward HLCPs. In this study, the authors
apply the ABC (affect-behavior-cognition) model of attitudes (as opposed to attitude as a whole) to explore how affect and cognition drive consumer
behavior toward HLCPs.
Design/methodology/approach – Questionnaires were used and the respondents totaled 483 in period 1 and 331 in period 2. Respondents were
collected using convenience sampling technique in one university in Spain and analyzed using structural equation modeling.
Findings – The authors finding revealed that consumers use both affect and cognition simultaneously when forming an attitude toward HLCPs.
Also, consumers’ perception of product fit represents a more relevant driver of consumer behavior with respect to brand fit. Appropriate theoretical
and managerial implications are derived from these results.
Originality/value – This study contributes to the understanding of consumers’ preferences toward high-tech luxury co-branded products.
Keywords Cognition, Affect, Co-branding, Brand fit, ABC model of attitude, Product fit
Paper type Research paper

1. Introduction 2007) – have not fully explained these products’ success and
the mechanisms that influence consumers’ decisions to prefer
Although many co-branding products are available in the such products. For example, Simonin and Ruth’s (1998)
market, this study examines the specific category of co-branded widely used theoretical model studies the effects of brand
products arising from the alliance between high-tech products alliances on consumer brand attitudes, but only addresses
and luxury brands, which we refer to as “high-tech luxury co- whether consumers might form either a positive or a negative
branded products” (HLCPs). Some examples of HLCPs are attitude toward a co-branded product. Further, their study
Samsung Armani cell phone, LG Prada cell phone, Asus does not explain which component of attitude drives
Lamborghini laptop and Segway Chanel personal transporter, consumers’ preferences. In the context of HLCPs, addressing
among others. Co-branded products are increasingly being which component influences consumers’ preferences is
introduced into the market. The success of these initiatives is highly relevant. Indeed, products of such a category
indirectly demonstrated by the release of new versions of the concurrently embed both functional attributes – in which
same HLCPs. For example, the first LG Prada cell phone was cognition is predominant – and symbolic attributes – in which
launched in 2006, and in 2012, LG released the newest affect is mostly important (Lim and Ang, 2008; Solomon
generation, LG Prada 3.0. Thus, HLCPs seem to attract et al., 2010; Voss et al., 2003). Consequently, this study
consumers’ attention and favor. assesses consumer preferences toward co-branded products
This study focuses on HLCPs because the existing (and HLCPs in particular) by considering three components
literature does not sufficiently explain the reasons for the of consumers’ attitudes, according to the ABC model of
success of these co-branding products. From the theoretical attitudes (Ray, 1973; Solomon, 1996). By doing this, it will
point of view, traditional co-branding models – such as be possible to disentangle the complex mechanism that brings
vertical and horizontal co-branding strategies (Helmig et al., consumers to form an attitude, by associating each attitude

The current issue and full text archive of this journal is available on
Emerald Insight at: www.emeraldinsight.com/1061-0421.htm Received 18 November 2015
Revised 11 May 2016
22 September 2016
Journal of Product & Brand Management
23 September 2016
26/7 (2017) 704–721 27 March 2017
© Emerald Publishing Limited [ISSN 1061-0421] 29 April 2017
[DOI 10.1108/JPBM-11-2015-1039] Accepted 30 April 2017

704
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

component to specific product attributes and by assessing Ueltschy and Laroche, 2004) and increasing the host brand’s
which component–attribute pair mainly drives consumer financial burden (Blackett and Boad, 1999). The alliance
preferences. between two brands may confuse consumers about the image
In addition, one of the major challenges to co-branding of constituent brands and consequently damage the brand
success is consumers’ perceptions of “product fit” and “brand equity of each brand (Park et al., 1996).
fit” of co-branded products, which are expected to play a
significant role in how consumers respond to co-branded 2.1 The formation of consumers’ attitudes toward
products (Simonin and Ruth, 1998). A high degree of fit is HLCPs
required to stimulate a desirable attitude and positive The marketing literature includes several theoretical models to
association, and eventually influence consumers’ behavioral address the study of consumer attitude. For example, the
intentions (Aaker and Keller, 1990; Helmig et al., 2007; theory of planned behavior contends that any social behavior
Simonin and Ruth, 1998). In the specific context of HLCPs, (or behavioral intention) could be interpreted by a few key
however, the concept of product fit and brand fit becomes antecedents such as attitude toward the behavior, subjective
puzzling. High-tech products and luxury brands have indeed norm and perceived behavioral control. When exploring
different attributes. While luxury brands are exclusive (Phau consumers’ attitudes toward luxury brands, affect and
and Prendergast, 2000) and costly (Keller, 2009), high-tech cognition are the prevailing attitudes used by researchers
products are functional and useful (Hawkins et al., 2000). In (Abbott et al., 2009; Riscinto-Kozub et al., 2013). Ray (1973)
addition, the concept of product fit and brand fit between high- and Solomon (1996) provide the ABC model of attitudes,
tech and luxury brand companies might also create disharmony which concludes that attitude is formed by three main
in consumers’ mind. Lastly, it is important to analyze components, namely, affect, behavior and cognition, with
consumers’ perceptions regarding product fit and brand fit and affect and cognition influencing consumer behavior (Laurent
how these perceptions influence consumers’ attitudes. Again, and Kapferer, 1985; Putrevu and Lord, 1994; Dubé et al.,
the adoption of a three-component model of attitude formation 2003). Subsequent studies have shown that, even though both
provides us support to disentangle this question. affect and cognition take part in consumers’ assessment of
The findings from this study would thus help marketers intrinsic and extrinsic product characteristics (Da Silva and
move toward better decision-making on the introduction and Syed Alwi, 2006), the affect seems more relevant in the
positioning of co-branded products. More specifically, this evaluation of hedonic products (Mano and Oliver, 1993; Kim
study provides firms with an improved understanding of the and Morris, 2007), while the cognition is associated with the
mechanisms by which consumers form an attitude toward evaluation of utilitarian products (Kim and Morris, 2007).
HLCPs. In turn, firms may apply this information to define the Given that HLCPs include attributes of both types of products,
content of advertising messages or shape communication both affect and cognition should influence consumer behavior.
strategies. Similarly, the predominance of one product attribute In general terms, cognition can be considered as the result of
over others in influencing consumers’ preferences might lead consumer’s thought process, which generates knowledge,
producers of HLCPs to define more effective positioning awareness, opinion and perception. The cognition is important
strategies. for prospective behavioral motivations to understand product
The rest of the paper is organized as follows: Section 2 attributes (Caro and Garcia, 2007). While most literature point
reviews the relevant literature and presents the main out that both affect and cognition are important for the
hypotheses, Section 3 discusses the empirical tests, Section 4 evaluation of brands, a few empirical studies show that the
presents the results and Section 5 concludes with a discussion rational part of a brand is assessed prior to its emotional part
of the findings and managerial implications of the study. (Bhat and Reddy, 1998).
Affect also has received intensive attention by several
marketing studies (Aaker and Shansby, 1982; Burke and Edell,
2. Literature review and hypotheses
1989). Affect is based on feelings toward specific stimuli
Co-branding is one of the possible forms of a brand alliance (Westbrook, 1987; Anand et al., 1988). In daily consumption
strategy (Simonin and Ruth, 1998; Rao et al., 1999; Geylani events, affect typically represents a crucial determinant of
et al., 2008; Cunha et al., 2015), or a specific arrangement consumer behavior (Anand et al., 1988). For instance,
within brand extension possibilities (Washburn et al., 2000). consumers often use affect to form a priori consumption
Geylani et al. (2008) define co-branding as a combination of experience in which they form their future behavioral decisions
two existing brand names to form a separate and single product (Cowley, 2007).
with a composite brand name. In the specific case of HLCPs, both affect and cognition are
Co-branding can provide both benefits and drawbacks to expected to drive consumer behavior. In this case, the hedonic
firms. In terms of benefits, co-branding offers established nature of luxury brands emphasizes the affective component of
brands an opportunity to add immediate credibility to existing attitude, while the functional nature of high-tech products
brands and increase sales of existing products (Mazodier and accentuates the cognitive dimension. For example, a well-
Merunka, 2014). The likelihood of the success of co-branded established study supports that utilitarian benefits are relevant
products increases with the degree of awareness, brand equity, in the customer decision process for the purchase of customized
familiarity and quality of the constituent brands. On the PCs (Tang et al., 2010). For the purchase of a smartphone,
contrary, co-branding also presents risks such as raising evidence demonstrates that occasional users primarily use
consumer mistrust, damaging the host brand’s image (Chang, smartphones to fulfill utilitarian needs and tend to perceive
2009), diluting the host brand’s equity (Washburn et al., 2000; these devices as functional tools (Dong-Hee, 2012).

705
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Provided that affect is mainly associated with products’ A high degree of fit results in a positive assessment of co-
hedonic attributes (luxury attributes, in our case) and cognition branded products (Aaker and Keller, 1990; Boush and Loken,
to utilitarian attributes (high-tech attributes, in our case) (Lim 1991; Bhat and Reddy, 1998; Ramanathan and Purani, 2014;
and Ang, 2008; Solomon et al., 2010; Voss et al., 2003), it is not Xiao and Lee, 2014) and increases the likelihood of success of
clear whether one attitude component will prevail over the co-branded products (Park et al., 1996; Simonin and Ruth,
other in driving consumer behavior. We thus prefer to consider 1998; Menon and Kahn, 2003; Helmig et al., 2007; Monga and
that both components may influence consumer behavior. Lau-Gesk, 2007). By contrast, if product fit and brand fit of
Consequently, we hypothesize: two constituent brands are inconsistent, the co-branded
product might create incompatible beliefs and judgments in
H1a. Affect positively influences consumer behavior toward consumers’ minds.
HLCPs. In the case of HLCPs, we expect product fit and brand fit to
H1b. Cognition positively influences consumer behavior positively influence consumer behavior. However, it may be
toward HLCPs. challenging for consumers to evaluate both in circumstances in
which constituent brands originate from very different product
Many authors conclude that affect dominates over cognition and brand categories. Based on these assumptions, we draw the
when predicting consumer behavior (Allen et al., 1992; Morris following hypothesis:
et al., 2002; Keer et al., 2014). Other authors (Anand et al.,
1988; Bhat and Reddy, 2001; Saunders and Buehner, 2013) H2. Consumer behavior toward HLCPs is positively
have shown that cognition is expressed by consumers before influenced by the degree of product fit of constituent
affect, and this process leads to behavioral intention. Lastly, in brands.
an attempt to overcome such contrasting views, other H3. Consumer behavior toward HLCPs is positively
researchers (Dubé et al., 2003) have introduced a more influenced by the degree of brand fit of constituent
complex hierarchical structure of consumer attitudes that both brands.
affect and cognition influence each other (Fink and Yolles,
2015) and neither prevails over the other. In this sense, Shiv
and Fedorikhin (1999) argue that the direction of causality 2.3 Acceptance of high-tech products is a basic
between affect and cognition remains largely unexplored. It is determinant of consumer behavior toward HLCPs
thus unclear which component dominates over the other or High-tech products are associated with innovation and
whether both components influence each other. Sometimes the characterized by their research and development intensity
affect and cognition toward a product might occur individually (Hansen and Serin, 1997). Sahadev and Jayachandran
or exist both at the same time. Consequently, we derive the (2004) also provide a definition of high-technology
following hypothesis: products. They define high-technology products as product
categories characterized by certain attributes such as short
H1c. Affect is associated positively with cognition.
product life cycle, high level of risk in the users’ behavior
changing, indispensability of infrastructure, lack of clearly
defined industry standards and uncertainty on product
2.2 The impact of product fit and brand fit on consumer functionality.
behavior toward HLCPs Because HLCPs involve the congruence of a high-tech
The marketing literature emphasizes that partner selection is the product with luxury attributes, it is essential to consider
most important decision for a company in pursuing a co- consumers’ acceptance of high-tech products in general.
branding strategy (Rao and Ruekert, 1994; Simonin and Ruth, Consumers’ acceptance of high-tech products (i.e. the ability to
1998; Rao et al., 1999; Sénéchal et al., 2014). Regarding the appraise the characteristics) plays a critical role in influencing
concept of brand extension, consumers’ perception of the parent favorable attitudes about HLCPs. Therefore, the acceptance of
brand and the extended product itself are the key points when high-tech products should represent a key determinant of
evaluating extended products (Shen, 2014; Zheng et al., 2015). consumer behavior toward HLCPs. In this sense, Davis (1989)
Simonin and Ruth (1998) distinguish two main types of fit, points out that the perception of products’ usefulness and ease-
namely, “product fit” and “brand fit”. Product fit represents of-use are essential components that influence consumers’
the consumers’ perception of the extent of compatibility of two acceptance of technological products. With the development of
(or more) product categories, whereas brand fit is the degree of science and technology, high-tech products have become more
consistency of brand images of each partner (Park et al., 1991). delicate and complex, integrating many advanced functions
If product categories or brand images of two constituent into one device. Because a significant amount of techno-babble
brands are somehow incompatible, consumers might have terminology has been created by high-tech companies and the
difficulty in understanding why the two constituent brands are technical jargon used by trained salespeople may confuse
associated and therefore question the competence of consumers, the evaluation of products’ attributes is not an easy
constituent brands to produce high-quality co-branded task. As a consequence, shopping for high-tech products may
products. For instance, adding a hedonic feature to a utilitarian involve a big challenge.
product would cause an incongruent addition (Azhari and In addition, a common perception of high-tech products is
Afiff, 2015). In this vein, Xiao and Lee (2014) suggest that that they provide both conveniences and problems for users
consumers must understand the fit between and among the co- (Hawkins et al., 2000). For example, a user may experience
branding products and the constituent brands. pleasure in using computers to analyze data. However, the

706
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

same user may also feel frustrated by the inability to fully use all affect-behavior sequence (Solomon, 1996). In contrast, based
of the functions provided by the software. This paradox causes on the experiential hierarchy, consumers initially act on the basis
anxiety and frustration in users of high-tech products. of emotional reactions (“feeling”), usually associated with
Therefore, we hypothesize: intangible product attributes (e.g. aesthetic and brand). Thus,
consumer behavior typically follows an affect-beliefs-behavior
H4. Consumer behavior toward HLCPs is positively sequence (Solomon, 1996).
influenced by the acceptance of high-tech products. In the case of HLCPs, it is likely that both emotional
sequences in consumer behavior may occur. On the one hand,
the luxury attribute is expected to provide prestige,
2.4 The effect of attitudes toward luxury brands on
conspicuousness, materialism, hedonism, self-identity,
consumer behavior toward HLCPs
uniqueness and quality to consumers. All of these attributes
In brand alliances, existing attitudes toward constituent brands
primarily impact the consumers’ affect and exceed the values
are highly associated with the attitudes toward partner brands
that cognition would provide (Wiedmann et al., 2007). As such,
(Simonin and Ruth, 1998), as a favorable attitude could create
we can expect affect to play a mediating role between existing
positive spillover to influence consumers’ post-attitude toward
attitudes toward luxury brands and consumer behavior toward
the brand. Consequently, it is appropriate to determine the role
HLCPs. On the other hand, the cognition is expected to
of attitudes toward luxury brands (i.e. the motivations that
mediate the relationship between acceptance of high-tech
bring consumers to possess luxury brands) in explaining
products and consumer behavior toward HLCPs. Hence, affect
consumers’ attitudes toward the co-branded products,
and cognition are expected to play different roles in the case of
provided that luxury is one of the two constituent attributes of
HLCPs, which include attributes associated to both “knowing”
HLCPs.
and “feeling” components. In this study, affect and cognition
A luxury brand is defined as “those whose ratio of
are not necessarily opposite sides of the same coin; the
functionality to price is low, while the ratio of intangible and
characteristics of a product to associate with consumers’ affect
situational utility to price is high” (Nueno and Quelch, 1998,
attitude do not imply that consumers dissociate from cognition.
p. 61). Some motivations behind the possession of luxury
Affect and cognition exist both simultaneously and are relative
brands are owing to social recognition (Park et al., 2008),
levels in consumers’ mind (Shavitt, 1989). As a consequence,
parental influence (Prendergast and Wong, 2003) and social
we propose the following hypotheses:
acceptance (Summers et al., 2006). Wilcox et al. (2009) also
pointed out that luxury brands serve two relevant functions for H6a. Affect positively mediates the effect of consumers’
consumers, namely, the social-adjustive function and the value- attitude toward luxury brands on consumer behavior
expressive function. The social-adjustive function posits that toward HLCPs.
consumers are attracted to luxury brands for image-relevant
reasons. Luxury brands also help people maintain relationships H6b. Cognition positively mediates the effect of consumers’
and obtain recognition in the social context. On the other hand, acceptance of high-tech products on consumer
the value-expressive function refers to intrinsic aspects related behavior toward HLCPs.
to the functionality of products or their quality. In this sense,
Furthermore, as mentioned earlier, product-brand fit
consumers are motivated to possess a luxury brand to form self-
influences consumer behavior. However, it is reasonable to
expression. The two functions may occur separately or
presume that the fit also influences consumer behavior in
simultaneously (Shavitt, 1989).
different ways, such as by generating distinct affect,
Sometimes consumers may have a negative reaction
behavior and cognition. In a study by Bhat and Reddy
toward luxury brands to the extent that their value
(1998), the impact of product fit on extension evaluation is
consciousness, personal attitude or social influence results
found to be independent of extensions of symbolic and
in a not-to-buy recommendation (Phau and Teah, 2009;
functional brands. Brand fit is critical to the extension of
Phau et al., 2009).
symbolic brand only with durable brand extensions (e.g.
All of these effects might have an impact on the success of
brands of wrist watches). For non-durable brand extensions
HLCPs. Regardless of the function that prevails in consumers’
(e.g. brands of ice creams), brand fit is equally critical to
minds, a positive attitude toward luxury brands should
functional and symbolic brands. When extensions are
represent a condition for the development of a positive behavior
symbolic (versus functional), the affect seems more
toward HLCPs. Based on these considerations, we draw the
influential than the cognition, even though the latter is more
following hypothesis:
important in shaping product fit of extensions.
H5. Consumer behavior toward HLCPs is positively In the context of this study, it is reasonable to presume that
influenced by their existing attitude toward luxury consumer behavior is influenced by product fit and brand fit
brands. both directly and indirectly. Apart from the direct effect, the
perceived fit between the constituent brands and product
categories should impact consumers’ affect and cognition of the
2.5 The mediating role of affect and cognition co-branded product, and in turn, on behavior. Thus, the affect
According to the standard learning hierarchy of the ABC model and cognition partially mediate the relationships between
of attitudes, when consumers emphasize a product’s consumer behavior toward HLCPs and product fit and brand
functionality, price and availability of high-involvement in fit. Based on these considerations, we draw the following
“knowing” products, their behavior typically follows a beliefs- hypotheses (Figure 1):

707
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Figure 1 Theoretical model I Luxury brands used for this survey have been selected among
those ranking the highest in the Digital IQ Score. Furthermore,
with respect to the first survey, we preferred to select high-tech
brands from more than one country (Sony is from Japan and
Philips is from The Netherlands). A similar approach has
guided our choice of luxury brands (Burberry is from the UK
and cK is from the USA). Lastly, we chose headphones as our
tested products because headphones are very common
electronic products among young consumers (e.g. university
students). As for the first survey, respondents were randomly
assigned to one of the two versions of the survey.
3.1.2 Samples
We chose university students in the community of Madrid, Spain
(undergraduate level), as our respondents to increase
homogeneity of the sample and to minimize random error caused
by selecting the general public (Calder et al., 1981). The student
sample was selected based on its internal validity for consumer
behavior analysis recognized by previous related academic work
H7a. Affect positively mediates the effect of product fit on
on attitude formation (Backhaus et al., 2001; Sicilia et al., 2006).
consumer behavior toward HLCPs.
According to the research from Stevens (2011), most objections
H7b. Affect positively mediates the effect of brand fit on to the use of student samples focus on the lack of generalizability.
consumer behavior toward HLCPs. When there is no theoretical reason to expect that one specific
sample would react in a different way than another, the research
H7c. Cognition positively mediates the effect of product fit priority is accurate to establish causal relationships. The internal
on consumer behavior toward HLCPs. validity justifies the use of student sample, because, as Peterson’s
(2001) meta-analysis found, college students tend to be more
H7d. Cognition positively mediates the effect of brand fit on homogeneous than nonstudents. In addition, because business
consumer behavior toward HLCPs. students of a prestigious university are expected to attain their
professional career goals or obtain high managerial positions, this
demographic is likely an important potential target group for
3. Methodology high-tech and luxury brands.
Students completed the surveys as a partial fulfillment of
3.1 Data collection process and measures their course in marketing management. In an attempt to
3.1.1 Research design enhance consumers’ understanding of the terms, luxury brands
The empirical test of the theoretical model uses data collected and high-tech products, definitions of these two terms were
through a survey submitted in 2010 to a sample of university listed at the beginning of the survey. The first survey was
students. The survey has been designed using real-world performed in 2010. We obtained a total of 566 responses. To
examples. Specifically, the first version of the survey used the ensure the validity of the responses, respondents with low levels
“Samsung Giorgio Armani” cell phone, while in a second of brand familiarity were eliminated from the database. Results
version, we used the “LG Prada” cell phone. Such co-branded also show that (Table I), on average, brand familiarities of these
products have been considered suitable for our study because four brands are high and similar among them:
they combine high-tech and luxury attributes. Respondents 1 Samsung (M = 5.95);
were randomly assigned to one of the two versions of the 2 LG (M = 5.90);
questionnaire. 3 Giorgio Armani (M = 5.60); and
To generate robust estimates and increase the 4 Prada (M = 5.46).
generalizability of results, we replicated the survey three years After excluding invalid and low brand familiarity toward the
later in 2013. In the second survey, real HLCPs have been products, we were left with 483 valid components (87 per cent)
replaced with fictitious co-branded products. We adopted of which 39 per cent of the respondents were males. The
products that consumers have never seen and used and, hence, demographics of respondents are presented in Table II.
do not possess well-formed preferences. The second survey In 2013, surveys containing the same questions were
requires consumers to focus on the evaluation of products distributed to students at the same university but with different
themselves (Zhao et al., 2011). Therefore, consumers’ products and respondents. The selection of valid components
preferences are not subject to discreet evaluation, nor are they followed the same procedure as the first survey. Again, similar
sensitive to the appraisal context. In this case, our fictitious to the results of our first survey, on average, brand familiarities
HLCPs are: among these four brands are also high and close:
 Sony and Calvin Klein (cK); 1 Sony (M = 5.78);
 Sony and Burberry; 2 Philips (M = 5.77);
 Philips and Calvin Klein (cK); and 3 Burberry (M = 5.77); and
 Philips and Burberry headphones. 4 cK (M = 5.80).

708
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Table I Mean values and standard deviations of items brand familiarity


Year 2010 Year 2013
Samsung-Armani N: 178 LG-Prada N: 153 Sony-cK N: 226 Philips-Burberry N: 257
Brand Familiarity Mean SD Mean SD Mean SD Mean SD
High-tech brand* 5.95 0.94 5.90 0.90 5.78 0.97 5.77 1.01
Luxury brand* 5.60 1.05 5.46 1.35 5.80 1.38 5.77 1.27
Note: *Three items (Appendix)

We obtained 331 valid components (81 per cent) of a total of respondents recognized SAMSUNG and LG as high-tech
407 components of which 49 per cent of the respondents were brands, and that ARMANI and GUCCI as luxury brands.
males. The ratios of the sample size to survey items (30 items) Before testing the model fit of the structural model, it is
for the sample satisfies the minimum requirements specified by recommended to perform a rigorous testing of measurement
Gorsuch (1983) and Thompson (2000). reliability and validity (Close et al., 2006). To measure
affect and cognition, we used the items from Ratchford
3.1.3 Measures
(1987). Acceptance of high-tech products has been
For both sets of surveys, participants completed an online
measured by using the items by Roehm and Sternthal
questionnaire, which provided a product information sheet
(2001). To measure attitude toward luxury brands, we
consisting of three parts – product name, a picture of the followed the scale developed by Wilcox et al. (2009). To
product and a set of product attributes including price. All measure the product fit and brand fit, we used the scale of
measures (Appendix 1 and Table III) used in the two sets of Aaker and Keller (1990). Lastly, consumer behavior has
surveys were evaluated on a 7-point Likert scale, ranging been measured by means of a scale related to the
from 1 (strongly disagree) to 7 (strongly agree). We recommendation to buy, which we derived from Baker and
translated our questionnaires into Spanish using a back- Churchill (1977).
translation to ensure comparability and equivalence in the
meaning of questionnaires (Brislin, 1970; Hult et al., 2008).
3.2 Data analysis
We adopted these items from previous studies and
Given the objective of this study, we conduct our analysis in
performed a pretest (participants N = 40) to refine the
two steps. First, we analyze the dimensional structure of the
measurement instruments. In the pretest, all of the
theoretical constructs. We subsequently analyze the respective
measurement models in terms of reliability and construct
Table II Demographics of respondents validity by multi-group analysis. Lastly, to test our hypothesis,
Characteristics Year 2010 (%) Year 2013 (%) we evaluate the results of the structural equation models with
latent variables.
Gender In this research, we used structural equation modeling
Male 46 44 (SEM). SEM enables us to simultaneously analyze our
Female 54 56 observed and latent variables from a perspective multi-
Age sample (Bollen, 1989; Jöreskog, 1993) and to obtain test
Under 20 years 37 28 and estimate measurement and/or structural models based
21-30 year 63 72 on robust statistics with multivariate non-normality
(Bentler, 2006; Muthén and Muthén, 1998-2012). To
Annual household income estimate our models, we selected two programs – MPLUS
Under e6,000 8 7 7.11 (Muthén and Muthén, 1998-2012) and EQS 6.2
e6,001-18,000 28 24 (Bentler, 2006). Thus, for assessing the overall suitability of
e18,001-30,000 49 35 the proposed models, we analyzed:
e30,001-60,000 12 26  the robust chi-square, the root mean square error of
e60,001 or over 3 8 approximation (RMSEA);
Part-time job  the standardized root mean square residual (SRMR, the
Yes 16 35 smaller the coefficient is, the better the model fit is); and
No 84 65  the comparative fit index (CFI).

Electronic device possession* We also conduct the robust Lagrange multiplier test (Bentler,
EBook/PDA 12 18 2006) to assess the multi-group equality constraints. Here, a
Music player 42 35 RMSEA of 0.08 to 0.10 is considered a mediocre fit and below
Mini PC/Netbook 17 40 0.08 provides a good model fit (MacCallum et al., 1996).
Navigation system 18 28 Recommendations for SRMR are less than 0.08 and a cutoff
Digital camera 42 36 criterion of CFI is larger or equal to 0.9 (Hu and Bentler, 1999).
To assess the measurement model, we follow the approach
Notes: N2010 = 483 and N2013 = 331; *Allow to choose more than one
suggested by Bagozzi (1980, 2010). After assessing their fit, we
device
analyze the reliability of the constructs, together with their

709
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Table III Mean and standard deviations of observed variables


Year 2010 Year 2013
Code Construct and items Mean SD Mean SD
ATTLB Attitude toward luxury brands
ATTLB_1 Luxury brands reflect the kind of person I see myself to be 3.03 1.76 3.00 1.63
ATTLB_2 Luxury brands help me communicate my self-identity 2.98 1.75 2.92 1.64
ATTLB_3 Luxury brands help me express myself 2.54 1.60 2.57 1.52
ATTLB_4 Luxury brands help me define myself 2.72 1.62 2.73 1.56
ATTLB_5 Luxury brands are a symbol of social status 5.07 1.62 5.23 1.53
ATTLB_6 Luxury brands help me fit into important social situations 3.84 1.89 3.94 1.92
ATTLB_7 I like to be seen wearing luxury brands 3.13 1.84 3.44 1.84
ATTLB_8 I enjoy it when people know I am wearing a luxury brand 2.81 1.85 3.03 1.60
PF Product fit
PF_1 I think LG’s products and PRADA’s products are a complementary product combination 3.21 1.43 3.24 1.43
PF_2 I think LG’s products and PRADA’s products can be used together in a natural manner 3.72 1.50 3.83 1.43
PF_3 I think LG’s products and PRADA’s products are an appropriate product combination 3.60 1.49 3.52 1.36
BF Brand fit
BF_1 The images or associations that you might have for the brand of LG and PRADA are consistent 3.81 1.53 3.86 1.35
BF_2 The images or associations that you might have for the brand of LG and PRADA are complementary 3.46 1.60 3.52 1.50
BF_3 The images or associations that you might have for the brand of LG and PRADA are expected 2.97 1.54 3.07 1.56
ACCHT Acceptance of high-tech products
ACCHT_1 I like high-tech products 5.20 1.47 5.19 1.44
ACCHT_2 I think high-tech products are useful 5.57 1.25 5.60 1.16
ACCHT_3 I think high-tech products are good 5.32 1.23 5.43 1.14
ACCHT_4 I find that high-tech products are practical 5.42 1.26 5.44 1.25
ACCHT_5 High-tech products are worth owning 5.16 1.34 5.18 1.26
ACCHT_6 High-tech products are impressive 5.49 1.42 5.71 1.26
ACCHT_7 High-tech products are valuable 5.28 1.30 5.38 1.17
ACCHT_8 High-tech products are advanced 6.01 1.12 6.17 0.95
AFF Affect
AFF_1 Owning a LG PRADA cell phone expresses my personality 3.38 1.89 3.44 1.86
AFF_2 Owning a LG PRADA cell phone is based on a lot of feeling 2.82 1.66 3.03 1.59
COG Cognition
COG_1 Owning a LG PRADA cell phone is mainly logical or objective 2.25 1.29 2.23 1.23
COG_2 Owning a LG PRADA cell phone is based mainly on functional facts 2.49 1.45 2.43 1.42
REC Recommendation to Buy
REC_1 Would you like to try the LG PRADA cell phone? 2.21 1.51 2.12 1.38
REC_2 I would patronize this LG PRADA cell phone 2.48 1.53 2.35 1.43
Notes: 2010: LG; 2013: SONY; 2010: SAMSUNG; 2013: PHILIPS; 2010: PRADA; 2013: BURBERRY; 2010: ARMANI: 2013: cK

convergent and discriminant validity. The reliability of the significance of the estimated parameters and the percentages of
observed variables is evaluated with the standardized factor explained variance.
loadings of the indicators for each construct and their reliability
coefficients. Furthermore, to evaluate the precision for each of 4. Analysis of results
the latent variables, we used the composite reliability coefficient
(CRC) (McDonald, 1985) and the average variance extracted 4.1 Measurement model
(AVE) (Fornell and Larcker, 1981). With a given model of To test the hypotheses in the empirical model of the year
measurement, the parameters of interest for the evaluation of 2010 and 2013 samples concurrently, it is necessary to
discriminant validity are the AVE. The estimation of the assess the factor structure. We applied a confirmatory factor
squared correlations among the latent variables, in parallel, will analysis of the observed variables. The complete theoretical
be analyzed to determine whether 1 lies inside all the M. model is a first-order seven-factor model in which each
confidence intervals around the correlation estimate between scale indicated a distinct latent factor. Tables IV, V and VI
any two factors (Anderson and Gerbing, 1988). Lastly, to test show the results of the measurement models. The measure
our theoretical framework, we consider the statistical of goodness-of-fitness indicates a poor fit of the data to the
model in both studies (Study 2010: x 2[329] = 1287.39,

710
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Table IV Results of the complete measurements models: theoretical and empirical across year
Complete theoretical M. model Complete empirical M. model
Year 2010 Year 2013 Year 2010 Year 2013
Code l* R2 Rel. l* R2 Rel. l* R2 Rel. l* R2 Rel.
Attitude toward luxury brands ATTLB
ATTLB_1 0.84* 0.71 0.77 0.84 0.71 0.75 0.84* 0.71 0.77 0.84 0.71 0.75
ATTLB_2 0.92* 0.85 0.61 0.93 0.86 0.59 0.92* 0.85 0.61 0.93 0.86 0.59
ATTLB_3 0.89* 0.79 0.92 0.85 0.89* 0.79 0.92 0.85
ATTLB_4 0.90* 0.81 0.92 0.85 0.90* 0.81 0.92 0.85
ATTLB_5 0.45* 0.20 0.37 0.14 0.45* 0.20 0.37 0.14
ATTLB_6 0.70* 0.49 0.64 0.41 0.70* 0.49 0.64 0.41
ATTLB_7 0.72* 0.52 0.67 0.45 0.72* 0.52 0.67 0.45
ATTLB_8 0.72* 0.52 0.70 0.49 0.72* 0.52 0.68 0.46
Product fit PF
PF_1 0.84* 0.71 0.78 0.84 0.71 0.80 0.84* 0.71 0.73 0.84 0.71 0.73
PF_2 0.72* 0.52 0.62 0.74 0.55 0.64 0.72* 0.52 0.62 0.75 0.56 0.61
PF_3 0.79* 0.62 0.82 0.67 0.79* 0.62 0.82 0.67
Brand fit BF
BF_1 0.74* 0.55 0.68 0.65 0.42 0.65 0.74* 0.55 0.68 0.46
BF_2 0.84* 0.71 0.49 0.82 0.67 0.45 0.83* 0.69 0.81 0.66
BF_3 0.46* 0.21 0.49 0.24 0.46* 0.21 0.50 0.25
Acceptance of high-tech products ACCHT
ACCHT_1 0.76* 0.58 0.77 0.81 0.66 0.75 0.76* 0.58 0.77 0.81 0.66 0.75
ACCHT_2 0.88* 0.77 0.59 0.90 0.81 0.57 0.88* 0.77 0.59 0.90 0.81 0.57
ACCHT_3 0.83* 0.69 0.85 0.72 0.83* 0.69 0.85 0.72
ACCHT_4 0.86* 0.74 0.83 0.69 0.86* 0.74 0.83 0.69
ACCHT_5 0.81* 0.66 0.74 0.55 0.81* 0.66 0.74 0.55
ACCHT_6 0.57* 0.32 0.57 0.32 0.57* 0.32 0.57 0.32
ACCHT_7 0.72* 0.52 0.68 0.46 0.72* 0.52 0.68 0.46
ACCHT_8 0.69* 0.48 0.59 0.35 0.69* 0.48 0.59 0.35
AffectAFF
AFF_1 0.76* 0.58 0.78 0.74 0.55 0.80 0.76* 0.58 0.78 0.74 0.55 0.80
AFF_2 0.79* 0.62 0.60 0.86 0.74 0.64 0.79* 0.62 0.60 0.85 0.72 0.64
Cognition COG
COG_1 0.89* 0.79 0.86 0.85 0.72 0.83 0.89* 0.79 0.86 0.85 0.72 0.83
COG_2 0.83* 0.69 0.74 0.80 0.64 0.68 0.83* 0.69 0.74 0.80 0.64 0.68
Recommendation to buy REC
REC_1 0.87* 0.76 0.87 0.77 0.59 0.82 0.87* 0.76 0.87 0.77 0.59 0.82
REC_2 0.87* 0.76 0.76 0.86 0.74 0.67 0.87* 0.76 0.76 0.86 0.74 0.67
Goodness x 2 [329] = 1287.39 RMSEA = x 2 [329] = 1595.36 RMSEA = x 2 [335] = 1285.75 RMSEA = x 2 [335] = 1612.08 RMSEA =
of Fit 0.08 SRMR = 0.05 CFI = 0.88 0.11 SRMR = 0.08 CFI = 0.78 0.08 SRMR = 0.05 CFI = 0.88 0.11 SRMR = 0.08 CFI = 0.78
Notes: Standardized path coefficients are reported and all l are significant at *p < 0.00; reliability of the latent variables: CRC/AVE. Complete Theoretical
M. Model: Complete Theoretical Measurement Model. Complete Empirical M. Model: Complete Empirical Measurement Model: with PRODUCT-BRAND
FIT (P-BF)

RMSEA = 0.08, SRMR = 0.05 and CFI = 0.88; Study 2013: Thus, we estimated a first-order six-factor model (complete
x 2[329] = 1595.36, RMSEA = 0.11, SRMR = 0.08 and empirical measurement model). The measures of goodness-of-
CFI = 0.78). The standardized parameter estimates fitness show reasonable fit of the data to the model in both
obtained do not provide clear evidence of the reliability and studies (Study 2010: x 2[335] = 1285.75, RMSEA = 0.08,
convergent discriminate of our seven dimensions across the SRMR = 0.05 and CFI = 0.88; Study 2013: x 2[335] =
samples. The most important issue is the high correlation 1612.08, RMSEA = 0.11, SRMR = 0.08 and CFI = 0.78). Our
between product fit and brand fit that shows the existence of results provided no problems with the discriminate validity, but
one unique latent variable (U2010 = 0.98 [0.88, 1.08] and the complete empirical measurement model yielded significant
U2013 = 0.99 [0.88, 1.08]). We decided to make one unique problems of reliability and convergent validity of the observed
latent variable product-brand fit. variables (l < 0.70, R2 < 0.50). Thus, we eliminated the

711
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Table V Results of the reduced empirical measurements model and invariance factorial across years
Reduced empirical M. model: Reduced empirical M. model:
without constraints with constraints
Year 2010 Year 2013 Year 2010 Year 2013
Code l* R2 Rel. l* R2 Rel. l* R2 Rel. Constraints x 2[1] Constraints x 2[1]
Attitude toward luxury brands ATTLB
ATTLB_1 0.86* 0.74 0.90 0.85 0.72 0.91 0.85* 0.72 0.90 kATTLB_1 0.74 r2ATTLB_1 1.06
ATTLB_2 0.94* 0.88 0.81 0.94 0.88 0.83 0.94* 0.88 0.81 kATTLB_2 1.29 r2ATTLB_2 1.69
ATTLB_3 0.89* 0.79 0.93 0.86 0.90* 0.81 kATTLB_3 0.29 r2ATTLB_3 4.58
ATTLB_4 0.90* 0.81 0.92 0.85 0.91* 0.83 kATTLB_4 0.46 r2ATTLB_4 2.81
Product-brand fit p-BF
PF_1 0.85* 0.72 0.80 0.83 0.69 0.81 0.84* 0.71 0.80 kPF_1 1.17 r2PF_1 0.00
PF_2 0.72* 0.52 0.63 0.76 0.58 0.65 0.74* 0.55 0.64 kPF_2 0.03 r2PF_2 1.64
PF_3 0.78* 0.61 0.82 0.67 0.79* 0.62 kPF_3 0.95 r2PF_3 4.21
BF_2 0.83* 0.69 0.81 0.66 0.82* 0.67 kBF_2 1.00 r2BF_2 1.50
Acceptance of high-tech products ACCHT
ACCHT_1 0.76* 0.58 0.83* 0.81 0.66 0.83 0.76* 0.58 0.83 kACCHT_1 2.64 r2ACCHT_1 1.89
ACCHT_2 0.90* 0.81 0.69 0.91 0.83 0.69 0.91* 0.83 0.70 kACCHT_2 1.80 r2ACCHT_2 0.95
ACCHT_3 0.82* 0.67 0.84 0.71 0.83* 0.69 kACCHT_3 0.26 r2ACCHT_3 4.20
ACCHT_4 0.87* 0.76 0.86 0.74 0.87* 0.76 kACCHT_4 1.47 r2ACCHT_4 0.17
ACCHT_5 0.81* 0.66 0.73 0.53 0.79* 0.62 kACCHT_5 4.14 r2ACCHT_5 0.71
Affect AFF
AFF_1 0.76* 0.58 0.78 0.75 0.56 0.80 0.76* 0.58 0.79 kAFF_1 0.44 r2AFF_1 0.08
AFF_2 0.79* 0.62 0.60 0.85 0.72 0.64 0.81* 0.66 0.62 kAFF_2 0.01 r2AFF_2 1.39
Cognition COG
COG_1 0.89* 0.79 0.86 0.86 0.74 0.83 0.88* 0.77 0.85 kCOG_1 0.10 r2COG_1 0.29
COG_2 0.83* 0.69 0.74 0.80 0.64 0.69 0.82* 0.67 0.72 kCOG_2 0.00 r2COG_2 0.51
Recommendation to buy REC
REC_1 0.87* 0.76 0.87 0.77 0.59 0.82 0.83* 0.69 0.85 kREC_1 1.89 r2REC_1 0.64
REC_2 0.87* 0.76 0.76 0.86 0.74 0.67 0.87* 0.76 0.72 kREC_2 0.11 r2REC_2 0.25
Goodness of Fit x 2 [137] = 311.73 x 2 [137] = 464.23 x 2 [312] = 832.21 RMSEA = 0.06 SRMR = 0.06 CFI = 0.95
RMSEA = 0.05 SRMR = RMSEA = 0.08 SRMR =
0.04 CFI = 0.97 0.05 CFI = 0.90
x 2 [274] = 781.83 RMSEA = 0.07 SRMR = 0.05 CFI
= 0.95
Notes: Standardized path coefficients are reported and all l are significant at *p < 0.00; Reliability of the latent variables: CRC/AVE. Reduced Empirical
M. Model: Reduced Empirical Measurement Model: with PRODUCT-BRAND FIT (P-BF) and without ATTLB_5-ATTLB_6-ATTLB_7 ATTLB_8- BF_1- BF_3-
ACCHT_6- ACCHT_7- ACCHT_8

observed variables with reliability problems in both samples; model is deemed to demonstrate adequate reliability,
ATTLB_5, ATTLB_6, ATTLB_7, ATTLB_8, BF_1, BF_3, convergent and discriminant validity.
ACCHT_6, ACCHT_7 and ACCHT_8. At this point, we evaluated the factorial invariance between
The new measurement model has six latent variables and the two different samples (year 2010 and year 2013) applied to
19 observed variables. The indexes of goodness-of-fit index, the covariance structure; this is the measurement equivalence
of the reduced empirical measurement model, show that this or metric invariance across the factor loadings and the
model fits the data reasonably well (Study 2010: x 2[137] = measurement error variances. First, we estimated the reduced
311.73, RMSEA = 0.05, SRMR = 0.04 and CFI = 0.97; empirical measurement model simultaneously in both samples
Study 2013: x 2[137] = 464.23, RMSEA = 0.08, SRMR = without constraints. The goodness-of-fitness of this model is:
0.05 and CFI = 0.90). Table V shows that the standardized x 2[274] = 781.83, RMSEA = 0.07, SRMR = 0.05 and CFI =
parameter estimates obtained are the evidence of the 0.95. Second, we estimated the same model with 38 equality
reliability and convergent validity of six factors (l > 0.70, constraints on parameter – 19 factor loadings and 19 error
R2 > 0.50, AVE > 0.50 and CRC > 0.70). The AVEs are variances (Table V). The goodness-of-fitness reveals a
above the squared correlation among the latent variables reasonable fit of the data: x 2[312] = 832.21, RMSEA = 0.06,
(AVE > U1/2) and of course, 1 is outside the confidence SRMR = 0.06 and CFI = 0.95. Table V shows the x 2[1] for
intervals around the correlation estimate between any two each factor loading and for each measurement error variance:
latent variables (Table VI). As such, this measurement 38 individual contrasts of equal the parameters across the two

712
Table VI Correlation matrix and confidence intervals of the latent variables

Model Year 2010 Year 2013


Complete theoretical
M. Model ATTLB PF BF ACCHT AFF COG ATTLB PF BF ACCHT AFF COG

0.21*** 0.07
Product fit PF (0.13, 0.29) (0.03, 0.17)
0.23*** 0.98*** 0.10 0.99***
Brand fit BF (0.14, 0.31) (0.88, 1.08) (0.02, 0.22) (0.89, 1.09)
Acceptance of
high-tech products 0.14*** 0.19*** 0.24** 0.14** 0.21*** 0.24**
ACCHT (0.06, 0.22) (0.11, 0.27) (0.16, 0.32) (0.06, 0.22) (0.13, 0.29) (0.16, 0.32)
0.66*** 0.41*** 0.40*** 0.14** 0.67*** 0.26*** 0.15*** 0.11*
Affect AFF (0.58, 0.74) (0.33, 0.49) (0.32, 0.48) (0.06, 0.22) (0.59, 0.75) (0.18, 0.34) (0.05, 0.25) (0.01, 0.19)
0.21*** 0.49*** 0.48*** 0.11** 0.41*** 0.25*** 0.38*** 0.31*** 0.03 0.42***
Detangling consumer attitudes

Cognition COG (0.13, 0.29) (0.41, 0.57) (0.42, 0.56) (0.03, 0.19) (0.33, 0.49) (0.17, 0.33) (0.30, 0.46) (0.25, 0.39) (0.11, 0.05) (0.34, 0.50)
Recommendation 0.33*** 0.47*** 0.47*** 0.10* 0.56*** 0.61*** 0.37*** 0.37*** 0.32*** 0.03 0.53*** 0.64***
to buy REC (0.25, 0.41) (0.39, 0.55) (0.41, 0.55) (0.02, 0.18) (0.48, 0.64) (0.53, 0.69) (0.29, 0.45) (0.29, 0.45) (0.26, 0.40) (0.11, 0.05) (0.45, 0.61) (0.56, 0.72)
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres

Complete
Empirical M. Model ATTLB P-BF ACCHT AFF COG ATTLB P-BF ACCHT AFF COG
Product-brand fit 0.22*** 0.08
P-BF (0.14, 0.30) (0.04, 0.18)
Acceptance of
high-tech products 0.14*** 0.21*** 0.14** 0.22***

713
ACCHT (0.06, 0.22) (0.13, 0.29) (0.06, 0.22) (0.14, 0.30)
0.66*** 0.41*** 0.14** 0.68*** 0.22*** 0.11*
Affect AFF (0.58, 0.74) (0.33, 0.98) (0.06, 0.22) (0.60, 0.76) (0.14, 0.30) (0.03, 0.19)
0.21*** 0.48*** 0.11** 0.41*** 0.25*** 0.36*** 0.03 0.42***
Cognition COG (0.13, 0.29) (0.40, 0.56) (0.03, 0.19) (0.33, 0.49) (0.17, 0.33) (0.28, 0.44) (0.11, 0.05) (0.34, 0.50)
Recommendation 0.33*** 0.47*** 0.10* 0.56*** 0.61*** 0.37*** 0.35*** 0.03 0.53*** 0.64***
to buy REC (0.25, 0.41) (0.39, 0.55) (0.04, 0.18) (0.48, 0.64) (0.53, 0.69) (0.29, 0.45) (0.27, 0.43) (0.09, 0.05) (0.45, 0.61) (056, 0.72)

Reduced Empirical
M. Model ATTLB P-BF ACCHT AFF COG ATTLB P-BF ACCHT AFF COG
0.22*** 0.09
Product-brand fit (0.14, 0.30) (0.03, 0.19)
Acceptance of
high-tech products 0.10** 0.20*** 0.12** 0.24***
ACCHT (0.02, 0.18) (0.12, 0.28) (0.04, 0.20) (0.16, 0.32)
0.64*** 0.40*** 0.11** 0.67*** 0.20*** 0.10*
Affect (0.56, 0.72) (0.32, 0.48) (0.03, 0.19) (0.59, 0.75) (0.12, 0.28) (0.02, 0.18)
0.21*** 0.47*** 0.10* 0.41*** 0.27*** 0.33*** 0.04 0.42***
Cognition (0.13, 0.29) (0.39, 0.55) (0.02, 0.18) (0.33, 0.49) (0.19, 0.35) (0.25, 0.41) (0.12, 0.04) (0.34, 0.50)
Recommendation 0.32*** 0.46*** 0.08* 0.56*** 0.61*** 0.37*** 0.33*** 0.03 0.53*** 0.64***
Volume 26 · Number 7 · 2017 · 704–721
Journal of Product & Brand Management

to buy (0.24, 0.40) (0.38, 0.54) (0.02, 0.16) (0.48, 0.64) (0.53, 0.69) (0.29, 0.45) (0.26, 0.40) (0.09, 0.05) (0.45, 0.61) (056, 0.72)
Notes: ATTLB: Attitude toward Luxury Brands. In brackets: Confidence intervals at 95%; and ***p < 0.00; **p < 0.05; *p < 0.10; Complete Theoretical M. Model: Complete Theoretical Measurement
Model. Complete Empirical M. Model: Complete Empirical Measurement Model: with PRODUCT-BRAND FIT (P-BF). Reduced Empirical M. Model: Reduced Empirical Measurement Model: with PRODUCT-
BRAND FIT (P-BF) and without ATTLB_5-ATTLB_6-ATTLB_7 ATTLB_8- BF_1- BF_3- ACCHT_6- ACCHT_7- ACCHT_8
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

samples (Rivera and Satorra, 2002). For the factor loadings, Figure 2 Theoretical model II
only one hypothesis of equal the parameters is rejected
(l ACCHT_5: 2010 = l ACCHT_5: 2013; x 2[1] = 4.14, p > 0.04).
Regarding the error variance, only three of the 19 tests conducted
lead us to reject the null hypothesis of equality the error variance
(s 2ATTLB_3: 2010 = s 2ATTLB_3: 2013, s 2PF_3: 2010 = s 2PF_3: 2013
and s 2ACCHT_3: 2010 = s 2ACCHT_3: 2013 with p > 0.04). Lastly, we
evaluated simultaneously the 38 equality constraints on the
parameter. We used SBDIFF.EXE to compute significance
test on the difference between Satorra–Bentler scaled chi-
square statistics (Crawford and Henry, 2003; Satorra, 2000;
Satorra and Bentler, 2001). It is used for the calculation of
these statistics:
 the Satorra–Bentler x 2 for the more constrained model
 x 2 [312] = 832.21;
 the Normal x 2 for the more constrained model  x 2
[312] = 1013.95; product-brand fit on consumer behavior toward HLCPs”.
 the Satorra–Bentler x 2 for the less constrained model  x 2 Figure 2 shows the new conceptual model.
[274] = 781.83; and We estimated our structural model. The goodness-of-
 the Normal x 2 for the less constrained model  x 2[274] = fitness reveals reasonable fit of the data: Study 2010:
943.62. x 2[139] = 367.98, RMSEA = 0.05, SRMR = 0.04 and
CFI = 0.96; Study 2013: x 2[139] = 483.76, RMSEA =
The Satorra–Bentler Scaled Difference reach 54.06, with 38 0.08, SRMR = 0.06 and CFI = 0.90. Table VIII shows the
degrees of freedom and the p > 0.04 (Table VII). relationship among all constructs. It describes the effects of
Accordingly, we cannot reject the hypothesis of equal the affect and cognition on recommendation to buy. Both affect
factor loadings and the measurement error variances in the (AffR (2013) = 0.26, p < 0.01; AffR (2010) = 0.35, p < 0.01)
Reduced Empirical Measurement Model, but we can reject and cognition (CogR (2013) = 0.46, p < 0.01; CogR (2010) =
the hypothesis of equal the correlations between the latent 0.41, p < 0.01) positively influence recommendation to buy in
variables (D x 2 = 1.98, df = 53, p < 0.00). As such, this both periods. Thus, H1a and H1b fail to reject. In addition,
measurement model is deemed to demonstrate strict H1c is no rejection because there exists a correlation
invariance in the covariance structure about loading significant and positive between affect and cognition
invariance and homogeneity measurement error but not in (CorreAC (2013) = 0.32, p < 0.01; CorreAC (2010) = 0.25, p <
the correlation matrix (Meredith, 1993). Consequently, the 0.01) (Figure 3).
covariance structures across the time periods are not Table VIII represents the relationship between product-brand
invariants, and we cannot combine the two samples for fit and recommendation to buy that is significant in both periods
estimation of our structural model. (PBFR (2013) = 0.15, p < 0.01; PBFR (2010) = 0.13, p < 0.01).
“H2-H3” is no rejection. Acceptance of high-tech products has no
4.2 Structural model impact on recommendation to buy in both periods (AccR (2013) =
Having tested the dimensional structure of the variables, 0.09, p > 0.10; AccR (2010) = 0.02, p > 0.10). Therefore, H4
first owing to combine product fit and brand fit into one is rejected. Attitude toward luxury brands does not influence
construct, product-brand fit, this study uses reduced recommendation to buy (AttR (2013) = 0.10, p > 0.10; AttR (2010) =
empirical measurement model to get the empirical results. 0.02, p > 0.10). It concludes that H5 is rejected.
In addition, apart from H1a, H1b, H1c, H4, H5, H6a and With regard to the results of affect and cognition as mediators,
H6b, maintain the same hypotheses as the theoretical in both periods, attitude toward luxury brands positively
measurement model. The new hypotheses are as follows: influences affect (AttA (2013) = 0.62, p < 0.01; AttA (2010) = 0.57,
“H2-H3: Consumer behavior toward HLCPs is positively p < 0.01), affect positively influences recommendation to buy
influenced by the degree of product-brand fit of constituent (AffR (2013) = 0.26, p < 0.01; AffR (2010) = 0.35, p < 0.01) and
brands”; “H7a-H7b: Affect positively mediates the effect of the indirect effect is positive and significant (AttA ! AffR (2013) =
product-brand fit on consumer behavior toward HLCPs” and 0.16, p < 0.01; AttA ! AffR (2010) = 0.20, p < 0.01), failing to
“H7c-H7d: Cognition positively mediates the effect of reject H6a.

Table VII Fit indices for the nested sequence in the multiple group confirmatory factor analysis
Model Normal x 2 S-B x 2 df Dx 2 Ddf p
Configural invariance 943.62 781.83 274 – – –
Loading invariance 964.65 812.32 293 23.15 19 >0.233
Homogeneity of measurement errors 1013.95 832.21 312 54.06 38 >0.042
Equality of correlations 1063.49 851.56 327 71.98 53 <0.001
Note: Each nested model contains its constraints, plus the constraints of all previous, tenable models

714
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Table VIII Results of the structural model


Attitude Acceptance of
toward luxury Product-brand high-tech
Year 2010 brands fit products Affect Cognition R2
Direct effects
Affect 0.57*** 0.28*** 0.47
Cognition 0.48*** 0.01 0.23
Recommendation to buy 0.02 0.13** 0.02 0.35*** 0.41*** 0.50
Indirect effects
Affect fi Recommendation to buy 0.20*** 0.10**
Cognition fi Recommendation to buy 0.20*** 0.01
Year 2013
Direct effects
Affect 0.62*** 0.15** 0.43
Cognition 0.36*** 0.10 0.12
Recommendation to buy 0.10 0.15** 0.09 0.26*** 0.46*** 0.48
Indirect effect
Affect fi Recommendation to buy 0.16** 0.04
Cognition fi Recommendation to buy 0.16*** 0.06
Notes: Standardized path coefficients are reported. ***p < 0.00; **p < 0.05; *p < 0.10. Goodness of Fit of the Structural Model 2010: x 2 [139] =
367.98, RMSEA = 0.05, SRMR = 0.04 and CFI = 0.96. Goodness of Fit of the Structural Model 2013: x 2 [139] = 483.76, RMSEA = 0.08, SRMR = 0.06 and
CFI = 0.90

Figure 3 Estimated model (2010/2013)

In 2013, the relationship between acceptance of high-tech H7c-H7d states that cognition plays a mediator role in the
products and cognition is insignificant (AccC (2013) = 0.10, p > relationship between product-brand fit and consumers’
0.10). In addition, the relationship is insignificant in 2010 recommendation to buy. This hypothesis is not rejected as
(AccC (2010) = 0.01, p > 0.10) either. Hence, H6b is not product-brand fit positively influences cognition (ProC (2013) =
supported. 0.36, p < 0.01; ProC (2010) = 0.48, p < 0.01), cognition
Product-brand fit positively influences affect (ProA (2013) = positively impacts recommendation to buy in both periods
0.15, p < 0.01; ProA (2010) = 0.28, p < 0.01), and affect (CogR (2013) = 0.46, p < 0.01; CogR (2010) = 0.41, p < 0.01)
positively influences recommendation to buy in both periods and the indirect effects are positive and significant in both
(AffR (2013) = 0.26, p < 0.01; AffR (2010) = 0.35, p < 0.01). In periods (ProC (2013) ! CogR (2013) = 0.16, p < 0.00;
this case, the indirect effect is positive and significant in 2010, ProC (2010) ! CogR (2013) = 0.20, p < 0.01).
but not in 2013 (ProA (2013) ! AffR (2013) = 0.04, p > 0.01; Furthermore, concerning the entire model, it demonstrates
ProA (2010) ! AffR (2010) = 0.10, p < 0.01). H7a-H7b is thus a similar percentage of variance of recommendation to buy in
partially supported. both periods (R22013 = 0.48; R22010 = 0.50), and the

715
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

significant results shown in Table VII reveal a similar model (ACCR (2013) = 0.26 > ACCR (2010) = 0.20). It implies that,
for two periods. For both periods, attitude toward luxury even though consumers may suffer a serious economic crisis
brands and acceptance of high-tech products do not have any and have passive expectation for future economic growth,
direct effect on consumers’ recommendation to buy HLCPs. their desires to have luxury brands are not stopped by the
The effects of attitude toward luxury brands and product-brand financial crisis. In this sense, these results seem to contrast
fit on recommendation to buy are mediated by affect and prior studies by Kim et al. (2009), which point out that a
cognition. financial crisis changes consumers’ perceptions, and
consumers become more conservative leading them to prefer
5. Conclusions more security and trust. Also, the results are inconsistent with
prior studies by Reyneke et al. (2012), which show that
5.1 Discussion economic and financial crisis influence the perception of
The central aim of this study is to identify the success drivers of customers.
co-branded products, especially for HLCPs, by detangling Sixth, the effects of acceptance of high-tech products on
consumer attitudes. This study provides new insights into cognition are negative in the two periods. The reason for this
important issues regarding product preference development and phenomenon might be caused by the features of HLCPs.
the exact direction in which to strengthen product evaluation. Because HLCPs combine the features of high-tech technology
Much of the previous research on co-branding primarily analyzed and luxury, the price of HLCPs is much higher than the cell
the impact of behavioral intention, without untangling the phones with similar specifications. It also suggests that
different paths through which the various successful drivers of co- consumers perceive HLCPs as luxury brands, not only high-
branding may impact consumer behavior. By detangling tech products.
consumers’ attitudes into their three main components, this Taken together, our findings thus confirm that both affect
paper sheds light on the subtle routes underlying consumer and cognition, and product-brand fit play significant roles in
behavior in the case of co-branded products. driving consumer behavior toward HLCPs.
Our results from both 2010 and 2013 offer significant
theoretical and managerial contributions to co-branding 5.2 Managerial implications
literature in several ways. First, consumers’ affect and cognition Co-branding may offer a solid strategy to help companies
do influence their consumer behavior directly. This is most likely survive the current financial crisis (Lee et al., 2006), by
caused by the two specific attributes of HLCPs that arouse both contrasting sales decline and worsening business performance.
consumers’ affect and cognition. As the study of Kempf (1999) For example, because HLCPs provide consumers with both
suggests, while the evaluations of functional products are mainly high-tech and luxury attributes, they may represent a relatively
impacted by affect and cognition, the evaluations of hedonic easy possibility for constituent companies to extend their
products are mainly influenced by affect. product lines and increase consumers’ loyalty. The high-tech
Second, concerning brand fit and product fit, the results attribute may help luxury brands enhance exclusivity and high-
demonstrate that the convergence of these two constructs also end brand image; for high-tech products, a luxury attribute
impact consumer behavior regarding HLCPs. Our results may provide added value to more practical functions.
suggest that potential clients should consider both constructs as Therefore, our results have critical implications for
intrinsically joined. They should evaluate the fit between the marketing managers. First, although HLCPs combine two
products and the fit between the brands in a very close way, not diverse components within one product, consumer behavior is
discriminating between either construct. In this case, the still symmetrically influenced by the two types of attributes:
complementarity among both brands is the key driver concept. high-tech and luxury attributes. Second, because luxury brands
Product-brand fit is a very important component driving could extend to more product categories than high-tech brands
consumer behavior. It is the only construct with both direct and (Park et al., 1991), it is recommended that high-tech producers
indirect effects on consumers’ recommendation to buy align with well-known luxury partners to take advantage of
HLCPs. Its indirect impact is mediated by both the affect and luxury attributes.
cognition. This implies that marketers could emphasize the luxury
Third, in terms of the total impact of the two periods, the attributes of these HLCPs with advertising and promoting.
total effect of product-brand fit (PBFTC (2013) = 0.36; This may also increase the unique feature of HLCPs and
PBFTC (2010) = 0.48) on cognition is larger than those of positively impact consumers’ purchase motivation (Mazodier
affect (PBFTA (2013) = 0.15; PBFTA (2010) = 0.28). and Merunka, 2014). Third, product-brand fit has a direct and
Fourth, our results show that the correlations between affect positive impact on consumer behavior and also an indirect and
and cognition in the two periods are positive. This is in line with positive impact on it via the affect and cognition. This result
the ABC model of attitudes formation, which demonstrates suggests that firms should meaningfully focus on finding the
that consumers perceive both affect and cognition as essential most harmonious product categories when forming a co-
to recommendation intention when assessing HLCPs. branding alliance. Therefore, in selecting a partner with whom
Fifth, according to the annual report of the management to align, firms should strongly consider how their respective
consulting firm Bain & Co.[1], the luxury brands sales in product categories fit in consumers’ minds. Last, acceptance of
2012 have shown a 10 per cent increase with respect to 2010, high-tech products plays no role in 2010, but has a negative role
notwithstanding the financial crisis. The total effect of in 2013. This finding implies that the attribute of high-
attitudes toward luxury brands on recommendation to buy technology does not promote favorable preferences in
are significant, and the effect in 2013 is larger than 2010 consumers. We can deduce that, when advertising HLCPs, an

716
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

over-emphasis on the high-tech attribute might lower (rather Anand, P., Holbrook, M.B. and Stephens, D. (1988), “The
than enhance) consumer attitudes. formation of affective judgments: the cognitive-affective
model versus the independence hypothesis”, Journal of
5.3 Limitations and further research Consumer Research, Vol. 15 No. 3, pp. 386-391.
This study suffers from several limitations that deserve new Anderson, J.C. and Gerbing, D.W. (1988), “Structural
avenues for future research. First, all participants in this study equation modeling in practice: a review and recommended
were university students. Although student samples reduce the two-step approach”, Psychological Bulletin, Vol. 103 No. 3,
problems of heterogeneity, caution is needed when generalizing pp. 411-423.
these results to the general public. Future research should aim Azhari, M.Z. and Afiff, A.Z. (2015), “The coherence and
to extend respondents to other groups (e.g. office workers). congruence of convergence in consumer electronics”, Journal
Second, this only considered consumers of one country to test of Product & Brand Management, Vol. 24 No. 4, pp. 377-385.
consumer attitude. This may result in the lack of generality. Backhaus, K., Muhlfeld, K. and Van Doorn, J. (2001),
Cross-cultural differences are also an important issue when “Consumer perspectives on standardization in international
discussing consumer behavior. Hofstede (2001) suggests that advertising: a student sample”, Journal of Advertising
Eastern and Western consumers have varying perspectives Research, Vol. 41 No. 5, pp. 53-61.
when faced with the same event. Future research that Bagozzi, R.P. (1980), Causal Models in Marketing, John Wiley &
incorporates cross-cultural differences would be valuable in this Sons, New York, NY.
respect. Third, this study exclusively relies on the ABC model Bagozzi, R.P. (2010), “Structural equation models are modelling
of theoretical approach to analyze attitudes formation. tools with many ambiguities: comments acknowledging the
However, other theories of consumer attitude and behavior also need for caution and humility in their use”, Journal of Consumer
exist. Future research could apply theories such as the Psychology, Vol. 20 No. 2, pp. 208-214.
functional theory of attitudes by Katz (1960) and multi- Baker, M.J. and Churchill, G.A. (1977), “The impact of
attribute attitude model by Wilkie and Pessemier (1973) to physically attractive models on advertising evaluations”,
assess the robustness of our results. Journal of Marketing Research, Vol. 14 No. 4, pp. 538-555.
Fourth, gender differences have been considered an Bentler, P.M. (2006), EQS 6 Structural Equations Program
important variable of difference in consumers’ product use and Manual, Multivariate Software, Encino, CA.
brand choice (Gould and Stern, 1989). Further research that Bhat, S. and Reddy, S.K. (1998), “Symbolic and functional
encompasses attitudinal differences between males and females positioning of brands”, Journal of Consumer Marketing,
would be fruitful toward the development of a full Vol. 15 No. 1, pp. 32-43.
comprehension of HLCPs. Lastly, in our theoretical approach, Bhat, S. and Reddy, S.K. (2001), “The impact of parent brand
we have considered consumer behavior as the main dependent attribute associations and affect on brand extension evaluation”,
variable. However, from the viewpoint of firms, other variables Journal of Business Research, Vol. 53 No. 3, pp. 111-122.
might play a more relevant role, such as brand equity Blackett, T. and Boad, B. (1999), Co-branding: The Science of
(Washburn et al., 2000). Consequently, future research could Alliance, Palgrave Macmillan, London.
explore whether the co-branding weakens or strengthens the Bollen, K.A. (1989), Structural Equations with Latent Variables,
brand equity of constituent brands before and after Wiley, New York, NY.
implementing the co-branding strategy. Boush, D.M. and Loken, B. (1991), “A process-tracing study
of brand extension evaluation”, Journal of Marketing
Research, Vol. 28 No. 1, pp. 16-28.
Note Brislin, R.W. (1970), “Back-translation for cross-cultural
research”, Journal of Cross-Cultural Psychology, Vol. 1 No. 3,
1 www.ipmark.com/pdf/lujo_2012.pdf [Last accessed: June
pp. 185-216.
10, 2014].
Burke, M.C. and Edell, J.A. (1989), “The impact of feeling on
ad-based affect and cognition”, Journal of Marketing
Research, Vol. 26 No. 1, pp. 69-83.
Reference
Calder, B.J., Phillips, L.W. and Tybout, A.M. (1981),
Aaker, D.A. and Shansby, J.G. (1982), “Positioning your “Designing research for application”, Journal of Consumer
product”, Business Horizons, Vol. 25 No. 3, pp. 56-62. Research, Vol. 8 No. 2, pp. 197-207.
Aaker, D.A. and Keller, K.L. (1990), “Consumer evaluations Caro, L.M. and Garcia, J.A.M. (2007), “Cognitive–affective
of brand extensions”, Journal of Marketing, Vol. 54 No. 1, model of consumer satisfaction: an exploratory study within
pp. 27-41. the framework of a sporting event”, Journal of Business
Abbott, M., Holland, R., Giacomin, J. and Shackleton, J. Research, Vol. 60 No. 2, pp. 108-114.
(2009), “Changing affective content in brand and product Chang, W.L. (2009), “Roadmap of co-branding positions and
attributes”, Journal of Product & Brand Management, Vol. 18 strategies”, The Journal of American academy of Business,
No. 1, pp. 17-26. Vol. 15 No. 1, pp. 77-84.
Allen, C.T., Machleit, K.A. and Kleine, S.S. (1992), “A Close, A.G., Finney, R.Z., Lacey, R.Z. and Sneath, J.Z.
comparison of attitudes and emotions as predictors of (2006), “Engaging the consumer through event marketing:
behavior at diverse levels of behavioral experience”, Journal linking attendees with the sponsor, community, and brand”,
of Consumer Research, Vol. 18 No. 4, pp. 493-504. Journal of Advertising Research, Vol. 46 No. 4, pp. 420-433.

717
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Cowley, E. (2007), “How enjoyable was it? Remembering an versus new alternatives”, Structural Equation Modeling, Vol. 6
affective reaction to a previous consumption experience”, No. 1, pp. 1-55.
Journal of Consumer Research, Vol. 34 No. 4, pp. 494-505. Hult, G.T.M., Ketchen, D.J. Jr., Griffith, D.A., Gonzalez-
Crawford, J.R. and Henry, J.D. (2003), “The depression Padron, T., Harmancioglu, N., Hung, Y., Talay, M.B. and
anxiety stress scales: normative data and latent structure in a Cavusgil, S.T. (2008), “Data equivalence in cross-cultural
large non-clinical sample”, British Journal of Clinical international business research: assessment and guidelines”,
Psychology, Vol. 42, pp. 111-131. Journal of International Business Studies, Vol. 39 No. 6,
Cunha, M. Jr., Forehand, M.R. and Angle, J.W. (2015), pp. 1027-1044.
“Riding Coattails: when co-branding helps versus hurts less- Jöreskog, K.G. (1993), “Testing structural equation
known brands”, Journal of Consumer Research, Vol. 41 No. 5, models”, in Bollen, K. and Long, J.S. (Eds), Testing
pp. 1284-1300. Structural Equation Models, Sage Publications, Newbury
Da Silva, R.V. and Syed Alwi, S.F. (2006), “Cognitive, Park, CA, pp. 294-316.
affective attributes and conative, behavioural responses in Katz, D. (1960), “The functional approach to the study of
retail corporate branding”, Journal of Product & Brand attitudes”, Public Opinion Quarterly, Vol. 24 No. 2,
Management, Vol. 15 No. 5, pp. 293-305. pp. 163-204.
Davis, F.D. (1989), “Perceived usefulness, perceived ease of Keller, K.L. (2009), “Managing the growth tradeoff:
use, and user acceptance of information technology”, MIS challenges and opportunities in luxury branding”, Brand
Quarterly, Vol. 13 No. 3, pp. 319-340. Management, Vol. 16 Nos 5/6, pp. 290-301.
Dong-Hee, S. (2012), “Cross-analysis of usability and Kempf, D.S. (1999), “Attitude formation from product trial:
aesthetic in smart devices: what influences users’ distinct roles of cognition and affect for hedonic and
preferences?”, Cross Cultural Management, Vol. 19 No. 4, functional products”, Psychology & Marketing, Vol. 16 No. 1,
pp. 563-587. pp. 35-50.
Dubé, L., Cervellon, M.C. and Jingyuan, H. (2003), “Should Keer, M., Conner, M., Van den Putte, B. and Neijens, P.
consumer attitudes be reduced to their affective and (2014), “The temporal stability and predictive validity of
cognitive bases? Validation of a hierarchical model”, affect-based and cognition-based intentions”, British Journal
International Journal of Research in Marketing, Vol. 20 No. 3, of Social Psychology, Vol. 53 No. 2, pp. 315-327.
pp. 259-272. Kim, J. and Morris, J.D. (2007), “The power of affective
Fink, G. and Yolles, M. (2015), “Collective emotion regulation response and cognitive structure in product-trial attitude
in an organisation – a plural agency with cognition and formation”, Journal of Advertising, Vol. 36 No. 1,
affect”, Journal of Organizational Change Management, pp. 95-106.
Vol. 28 No. 5, pp. 832-871. Kim, M., Lado, N. and Torres, A. (2009), “Evolutionary
Fornell, C. and Larcker, D.F. (1981), “Evaluating structural changes in services attribute importance in a crisis scenario”,
equation models with unobservable variables and Journal of Service Research, Vol. 11 No. 4, pp. 429-440.
measurement error”, Journal of Marketing Research, Vol. 18 Laurent, G. and Kapferer, J.N. (1985), “Measuring consumer
No. 1, pp. 39-50. involvement profiles”, Journal of Marketing Research, Vol. 22
Geylani, T., Inman, J.J. and Hofstede, F.T. (2008), “Image No. 1, pp. 41-53.
reinforcement or impairment: the effects of co-branding on Lee, S.J., Kim, W.G. and Kim, H.J. (2006), “The impact of co-
attribute uncertainty”, Marketing Science, Vol. 27 No. 4, branding on post-purchase behaviors in family restaurants”,
pp. 730-744. International Journal of Hospitality Management, Vol. 25
Gorsuch, R.L. (1983), Factor Analysis, Lawrence Erlbaum, No. 2, pp. 245-261.
Gillsdale, NJ. Lim, E.A.C. and Ang, S.H. (2008), “Hedonic vs. utilitarian
Gould, S.J. and Stern, B.B. (1989), “Gender schema and consumption: a cross-cultural perspective based on cultural
fashion consciousness”, Psychology and Marketing, Vol. 6 conditioning”, Journal of Business Research, Vol. 61 No. 3,
No. 2, pp. 129-145. pp. 225-232.
Hansen, P.A. and Serin, G. (1997), “Will low technology McDonald, R.P. (1985), Factor analysis and related methods,
products disappear? The hidden innovation processes in low Lawrence Erlbaum Associates Publishers, Hillsdale, NJ.
technology industries”, Technological Forecasting and Social MacCallum, R.C., Browne, M.W. and Sugawara, H.M.
Change, Vol. 55 No. 2, pp. 179-191. (1996), “Power analysis and determination of sample size for
Hawkins, D.I., Best, R.J. and Coney, K.A. (2000), Consumer covariance structure modeling”, Psychological Methods, Vol. 1
Behavior, Building Marketing Strategy, 8th ed., Irwin, No. 2, pp. 130-149.
McGraw-Hill, Homewood, IL, New York, NY. Mano, H. and Oliver, R.L. (1993), “Assessing the
Helmig, B., Huber, J.A. and Leeflang, P.S.H. (2007), dimensionality and structure of the consumption experience:
“Explaining behavioural intentions toward co-branded evaluation, feeling and satisfaction”, Journal of Consumer
products”, Journal of Marketing Management, Vol. 23 Research, Vol. 20 No. 3, pp. 451-466.
Nos 3/4, pp. 285-304. Mazodier, M. and Merunka, D. (2014), “Beyond brand
Hofstede, G. (2001), Culture’s Consequences: Comparing Values, attitude: individual drivers of purchase for symbolic
Behaviors, Institutions and Organizations Across Nations, 2nd cobranded products”, Journal of Business Research, Vol. 67
ed., Sage, Thousand Oaks, CA. No. 7, pp. 1552-1558.
Hu, L.T. and Bentler, P.M. (1999), “Cutoff criteria for fit Menon, S. and Kahn, B.E. (2003), “Corporate sponsorships of
indexes in covariance structure analysis: conventional criteria philanthropic activities: when do they impact perception of

718
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

sponsor brand?”, Journal of Consumer Research, Vol. 13 Ray, M. (1973), “Marketing communications and the
No. 3, pp. 316-327. hierarch-of-effects”, in Clarke, P. (Ed.), New Models for Mass
Meredith, W. (1993), “MI, factor analysis and factorial Communications, Sage, Beverly Hills, pp. 147-176.
invariance”, Psychometrika, Vol. 58, pp. 525-543. Reyneke, M., Sorokácová, A. and Pitt, L. (2012), “Managing
Monga, A.B. and Lau-Gesk, L. (2007), “Blending cobrand brands in times of economic downturn: how do luxury
personalities: an examination of the complex self”, Journal of brands fare?”, Journal of Brand Management, Vol. 19 No. 6,
Marketing Research, Vol. 44 No. 3, pp. 389-400. pp. 457-466.
Morris, J.D., Woo, C., Geason, J.A. and Kim, J. (2002), “The Riscinto-Kozub, K., Anthony O’Neill, M. and Palmer, A.A.
power of affect: predicting intention”, Journal of Advertising (2013), “Emotional antecedents and outcomes of service
Research, Vol. 43 No. 3, pp. 7-17. recovery: an exploratory study in the luxury hotel industry”,
Muthén, L.K. and Muthén, B.O. (1998-2012), Mplus User’s Journal of Services Marketing, Vol. 28 No. 3, pp. 233-243.
Guide, 5th ed., Muthén and Muthén, Los Angeles, CA. Rivera, P. and Satorra, A. (2002), “Analysing group
Nueno, J.L. and Quelch, J.A. (1998), “The mass marketing of differences: a comparison of SEM approaches”, in
luxury”, Business Horizons, Vol. 41 No. 6, pp. 61-68. Marcoulides, G. and Moustaki, I. (Eds.), Latent Variable and
Park, C.W., Jun, S.Y. and Shocker, A.D. (1996), “Composite Latent Structure Models, Lawrence Erlbaum Associates, New
branding alliances: an investigation of extension and Jersey, pp. 86-104.
feedback effect”, Journal of Marketing Research, Vol. 33 Roehm, M.L. and Sternthal, B. (2001), “The moderating
No. 4, pp. 453-466. effect of knowledge and resources on the persuasive impact
Park, C.W., Milberg, S. and Lawson, R. (1991), “Evaluation of of analogies”, Journal of Consumer Research, Vol. 28 No. 2,
brand extensions: the role of product feature similarity and pp. 257-272.
brand concept consistency”, Journal of Consumer Research, Sahadev, S. and Jayachandran, S. (2004), “Managing the
Vol. 18 No. 2, pp. 185-193. distribution channels for high-technology products”,
Park, H.J., Rabolt, N.J. and Jeon, K.S. (2008), “Purchasing European Journal of Marketing, Vol. 38 Nos 1/2, pp. 121-149.
global luxury brands among young Korean consumers”, Satorra, A. (2000), “Scaled and adjusted restricted tests in
Journal of Fashion Marketing and Management, Vol. 12 No. 2, multi-sample analysis of moment structures”, in
pp. 244-259. Heijmans, R.D.H., Pollock, D.S.G. and Satorra, A.
Peterson, R.A. (2001), “On the use of college students in social (Eds), Innovations in Multivariate Statistical Analysis: A
science research insights from a second-order meta- Festschrift for Heinz Neudecker, Kluwer Academic
analysis”, Journal of Consumer Research, Vol. 28 No. 3, Publishers, London, pp. 233-247.
pp. 450-461. Satorra, A. and Bentler, P.M. (2001), “A scaled difference chi-
Phau, I. and Prendergast, G. (2000), “Consuming luxury square test statistic for moment structure analysis”,
brands: the relevance of the ‘rarity principle”, Journal of Psychometrika, Vol. 66 No. 4, pp. 507-514.
Brand Management, Vol. 8 No. 2, pp. 122-138. Saunders, T.S. and Buehner, M.J. (2013), “The gut chooses
Phau, I. and Teah, M. (2009), “Devil wears (counterfeit) faster than the mind: a latency advantage of affective over
prada: a study of antecedents and outcomes of attitudes cognitive decisions”, The Quarterly Journal of Experimental
toward counterfeits of luxury brands”, Journal of Consumer Psychology, Vol. 66 No. 2, pp. 381-388.
Marketing, Vol. 26 No. 1, pp. 15-27. Sénéchal, S., Georges, L. and Pernin, J.L. (2014), “Alliances
Phau, I., Teah, M. and Lee, A. (2009), “Targeting buyers of between corporate and fair trade brands: examining the
counterfeits of luxury brands: a study on attitudes of antecedents of overall evaluation of the co-branded product”,
Singaporean consumers”, Journal of Targeting, Measurement Journal of Business Ethics, Vol. 124 No. 3, pp. 365-381.
and Analysis for Marketing, Vol. 17 No. 1, pp. 3-15. Shavitt, S. (1989), “Products, personalities and situations in
Prendergast, G. and Wong, C. (2003), “Parental influence on attitude functions: implications for consumer behavior”,
the purchase of luxury brands of infant apparel: an Advances in Consumer Research, Vol. 16 No. 1, pp. 300-305.
exploratory study in Hong Kong”, Journal of Consumer Shen, F. (2014), “Perceived fit and deal framing: the
Marketing, Vol. 20 No. 2, pp. 157-169. moderating effect of perceived fit on sales promotions in line
Putrevu, S. and Lord, K.R. (1994), “Comparative and and brand extensions”, Journal of Product & Brand
noncomparative advertising: attitudinal effects under Management, Vol. 23 Nos 4/5, pp. 295-303.
cognitive and affective involvement conditions”, Journal of Shiv, B. and Fedorikhin, A. (1999), “Heart and mind in
Advertising, Vol. 23 No. 2, pp. 77-91. conflict: the interplay of affect and cognition in consumer
Ramanathan, J. and Purani, K. (2014), “Brand extension decision making”, Journal of Consumer Research, Vol. 26
evaluation: real world and virtual world”, Journal of Product No.3, pp. 278-292.
& Brand Management, Vol. 23 No. 7, pp. 504-515. Sicilia, M., Ruiz, S. and Reynolds, N. (2006), “Attitude
Rao, A.R. and Ruekert, R.W. (1994), “Branding alliances as formation online: how the consumer’s need for cognition
signals of product quality”, Sloan Management Review, affects the relationship between attitude towards the website
Vol. 31 No. 1, pp. 87-97. and attitude towards the brand”, International Journal of
Rao, A.R., Qu, L. and Ruekert, R.W. (1999), “Signaling Market Research, Vol. 48 No. 2, pp. 139-154.
unobservable product quality through a brand ally”, Journal Simonin, B.L. and Ruth, J.A. (1998), “Is a company known by
of Marketing Research, Vol. 36 No. 2, pp. 258-268. the company it keeps? Assessing the spillover effects of brand
Ratchford, B.T. (1987), “New insights about the FCB grid”, alliances on consumer brand attitudes”, Journal of Marketing
Journal of Advertising Research, Vol. 27 No. 4, pp. 24-38. Research, Vol. 35 No. 1, pp. 30-42.

719
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Solomon, M.R. (1996), Consumer Behaviour, Prentice Hall, Westbrook, R.A. (1987), “Product/consumption-based
Englewood Cliffs, NJ. affective responses and postpurchase processes”, Journal of
Solomon, M.R., Bamossy, G., Askegaard, S. and Hogg, M.K. Marketing Research, Vol. 24 No. 3, pp. 258-270.
(2010), Consumer Behaviour: A European Perspective, 4th ed., Wiedmann, K.P., Hennigs, N. and Siebels, A. (2007),
Prentice Hall, London. “Measuring consumers’ luxury value perception: a cross-
Stevens, C.K. (2011), “Questions to consider when selecting cultural framework”, Academy of Marketing Science Review,
student samples”, Journal of Supply Chain Management, Vol. 7 No. 7, pp. 1-21.
Vol. 47 No. 3, pp. 19-21. Wilcox, K., Kim, H.M. and Sen, S. (2009), “Why do
Summers, T.A., Belleau, B.D. and Xu, Y. (2006), “Predicting consumers buy counterfeit luxury brands?”, Journal of
purchase intention of a controversial luxury apparel Marketing Research, Vol. 46 No. 2, pp. 247-259.
product”, Journal of Fashion Marketing and Management, Wilkie, W.L. and Pessemier, E. (1973), “Issues in marketing’s
Vol. 10 No. 4, pp. 405-419. use of multi-attribute attitude models”, Journal of Marketing
Tang, Z., Chen, X. and Xiao, J. (2010), “Using the classic Research, Vol. 10 No. 4, pp. 428-441.
grounded theory approach to understand consumer Xiao, N. and Lee, S.H. (2014), “Brand identity fit in co-
purchase decision in relation to the first customized branding: the moderating role of C-B identification and
products”, The Journal of Product and Brand Management, consumer coping”, European Journal of Marketing, Vol. 48
Vol. 19 No. 3, pp. 181-197. Nos 7/8, pp. 1239-1254.
Thompson, B. (2000), “Ten commandments of structural Zhao, M., Hoeffler, S. and ZauberMan, G. (2011), “Mental
equation modeling”, in Grimm, L.G. and Yarnold, P.R. simulation and product evaluation: the affective and
(Eds), Reading and Understanding More Multivariate Statistics, cognitive dimensions of process versus outcome simulation”,
American Psychological Association, Washington, DC, Journal of Marketing Research, Vol. 48 No. 5, pp. 827-839.
pp 261-283. Zheng, C., Ma, K., Duan, Q. and Wang, H. (2015),
Ueltschy, L.C. and Laroche, M. (2004), “Co-branding “Ineffective brand extensions and the Sisyphus effect”,
internationally: everyone wins?”, Journal of Applied Business Journal of Product & Brand Management, Vol. 24 No. 2,
Research, Vol. 20 No. 3, pp. 91-102. pp. 165-171.
Voss, K.E., Spangenberg, E.R. and Grohmann, B. (2003),
“Measuring the hedonic and utilitarian dimensions of
Further reading
consumer attitude”, Journal of Marketing Research, Vol. 40
No. 3, pp. 310-320. Conway, J.M. and Lance, C.E. (2010), “What reviewers
Washburn, J.H., Till, B.D. and Priluck, R. (2000), “Co- should expect from authors regarding common method bias
branding: brand equity and trial effects”, Journal of Consumer in organizational research”, Journal of Business & Psychology,
Marketing, Vol. 17 No. 7, pp. 591-604. Vol. 25 No. 3, pp. 325-334.

720
Detangling consumer attitudes Journal of Product & Brand Management
Han-Chiang Ho, Nora Lado and Pilar Rivera-Torres Volume 26 · Number 7 · 2017 · 704–721

Appendix

Table AI Measures
No. Questions
Attitude toward Luxury brands (Wilcox et al., 2009)
1 Luxury brands reflect the kind of person I see myself to be
2 Luxury brands help me communicate my self-identity
3 Luxury brands help me express myself
4 Luxury brands help me define myself
5 Luxury brands are a symbol of social status
6 Luxury brands help me fit into important social situations
7 I like to be seen wearing luxury brands
8 I enjoy it when people know I am wearing a luxury brand
Product fit (Aaker and Keller, 1990)
1 I think LG’s products and PRADA’s products are a complementary product combination
2 I think LG’s products and PRADA’s products can be used together in a natural manner
3 I think LG’s products and PRADA’s products are an appropriate product combination
Brand fit (Aaker and Keller, 1990)
1 The images or associations that you might have for the brand of LG and PRADA are consistent
2 The images or associations that you might have for the brand of LG and PRADA are complementary
3 The images or associations that you might have for the brand of LG and PRADA are expected
Acceptance of High-tech products (Roehm and Sternthal, 2001)
1 I like High-tech products
2 I think high-tech products are useful
3 I think high-tech products are good
4 I find that high-tech products are practical
5 High-tech products are worth owning
6 High-tech products are impressive
7 High-tech products are valuable
8 High-tech products are advanced
Affect (Ratchford, 1987)
1 Owning a LG PRADA cell phone expresses my personality
2 Owning a LG PRADA cell phone is based on a lot of feeling
Cognition (Ratchford, 1987)
1 Owning a LG PRADA cell phone is mainly logical or objective
2 Owning a LG PRADA cell phone is based mainly on functional facts
Recommendation to buy (Baker and Churchill, 1977)
1 Would you like to try the LG PRADA cell phone?
2 I would patronize this LG PRADA cell phone
Brand familiarity (Simonin and Ruth (1998)
1 I am extremely familiar with the LG brand name
2 I definitely recognize the LG brand name
3 I definitely have heard of the LG brand name before

Corresponding author
Han-Chiang Ho can be contacted at: sausare@yahoo.com.tw

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

721

You might also like