Journal of International Business Studies (2003) 34, 53–65

& 2003 Palgrave Macmillan Ltd. All rights reserved 0047-2506 $25.00
www.jibs.net

How perceived brand globalness creates brand
value
Jan-Benedict EM Steenkamp1
Rajeev Batra2
Dana L Alden3
1
Tilburg University, The Netherlands; 2University
of Michigan Business School, Ann Arbor, USA;
3
University of Hawaii at Honolulu

Correspondence:
J-B E.M. Steenkamp, Tilburg University,
P.O. Box 90153, 5000 LE, Tilburg,
The Netherlands. Tel: þ 31 134662916;
E-mail: J.B.Steenkamp@uvt.nl

Example - Global brand - Trend?

Received: 21 June 2001
Revised: 28 November 2001
Accepted: 3 March 2002
Online publication date: 28 November 2002

Abstract
In today's multinational marketplace, it is increasingly important to understand
why some consumers prefer global brands to local brands. We delineate three
pathways through which perceived brand globalness (PBG) influences the
likelihood of brand purchase. Using consumer data from the U.S.A. and Korea,
we find that PBG is positively related to both perceived brand quality and
prestige and, through them, to purchase likelihood. The effect through
perceived quality is strongest. PBG effects are weaker for more ethnocentric
consumers.
Journal of International Business Studies (2003) 34, 53–65. doi:10.1057/
palgrave.jibs.8400002

Introduction
Frito-Lay recently changed its leading potato chip brand in the
Netherlands from Smiths to Lay's in order to 'capture the affinity
that an international brand generates' (Anonymous, 2001a, 72).
This case is not unique. Many multinational corporations today are
altering their brand portfolios in favor of global brands. For
example, both Procter & Gamble (P&G) and Unilever have greatly
pruned the number of brands they market around the world, often
disposing of those with limited global potential (Pitcher, 1999).
The telecom giant Vodafone is replacing local brand names by the
global Vodafone name (Anonymous, 2001b). Although there is a
dearth of formal definitions of global brand in the literature, it is
commonly agreed that they are brands that consumers can find
under the same name in multiple countries with generally similar
and centrally coordinated marketing strategies (Yip, 1995; Branch,
2001).
Several reasons are offered for moves toward global brands. First,
globalization can yield economies of scale and scope in R&D,
manufacturing, and marketing (Yip, 1995). Second, its strategic
appeal increases as meaningful segments of consumers around the
world develop similar needs and tastes (Hassan and Katsanis, 1994).
Third, globalization speeds up a brand's time to market by reducing
time-consuming local modifications (Neff, 1999). Such arguments
have been present in the literature for many years and are now
widely accepted.
Recently, consumer preference for brands with 'global image'
over local competitors, even when quality and value are not
'objectively' superior, has been proposed as a fourth reason for
companies to move toward global brands (Shocker et al., 1994;

Perceived brand globalness creates brand value

J-B EM Steenkamp et al

54

Kapferer, 1997). Research indicates that corporations take advantage of such image-enhancing
effects by positioning brands as 'global' in their
communications, using message elements such as
brand name, logo, ad visuals and themes, etc.
(Alden et al., 1999). Although the belief that
perceived brand globalness (PBG) creates consumer
perceptions of brand superiority is widely asserted
(e.g. Shocker et al., 1994; Kapferer, 1997; Keller,
1998), it can be challenged. There is, for instance,
the phenomenon of consumer ethnocentrism (CET), a
well-established bias among many consumers in
favor of home-grown products (Shimp and Sharma,
1987). There is also evidence that many consumers
prefer brands with strong local connections
(Zambuni, 1993), and this leads some to argue that
consumers have no intrinsic preference for global
brands, and that corporate enthusiasm on this front
is misguided (De Mooij, 1998, 39).
Given this unresolved debate, there is clearly a
need to investigate whether consumers prefer
global brands and, if they do, the reasons (perceived
higher quality, higher prestige, etc.) that underlie
such a preference. In investigating the factors that
may predict a preference for global brands over
local brands, we hypothesize three pathways
through which PBG creates additional brand value.
Two of these pathways are posited to occur
indirectly through the quality and prestige associations of PBG. The third pathway involves the direct
effect of PBG on brand value. We focus on the
consumer's likelihood of purchasing the brand as
our measure of brand value or utility (Aaker, 1991).
We therefore test the relative influences of quality,
prestige and PBG per se on purchase likelihood.
We also test a potential 'offsetting' strategy
that local brands can pursue – positioning the
brand as a local 'icon' of the country in question.
In addition to examining these main effect
relationships, we also study the moderating role
of a key individual-difference variable: consumer
ethnocentrism.

Research hypotheses
Manufacturers have products that are internally
considered to be 'global' because they share similar
technical specifications (Branch, 2001). However,
the issue here is whether a brand benefits from
consumer perceptions that it is 'global' – a perception
that can be formed only if consumers believe the
brand is marketed in multiple countries and is
generally recognized as global in these countries.
Such a perception can be formed in one of two

Journal of International Business Studies

ways. First, consumers may learn that the same
brand is found in other countries, through
media exposure (for example seeing the brand
name in coverage of an overseas sports or
concert event), word of mouth (friends or relatives
returning from abroad passing on the news), or
their own travel overseas. Second, a brand may
assert or imply its 'globalness' even if it is not
available worldwide, through marketing communications that use brand names, endorsers, advertising themes, packaging and other symbols widely
associated with a 'modern', urban lifestyle (Alden
et al., 1999). The questions of interest here are
whether consumer perceptions of brand globalness
affect purchase likelihood, why (i.e., through which
pathways), and for whom (i.e., are there moderating
factors?).

Pathways through which perceived brand
globalness affects purchase likelihood
We suggest that the appeal of global brands arises
from three different but not mutually exclusive
sources: higher perceived quality, higher prestige,
and the psychological benefits of PBG per se. These
sources of global brand appeal provide three
pathways through which PBG can affect purchase
likelihood: a direct effect, and indirect effects via
brand quality and brand prestige (Han, 1990).
Hypotheses about these pathways and other
influencing factors follow. Figure 1 presents these
relationships and hypotheses graphically.
Some authors have asserted that consumers may
prefer global brands because of associations of
higher prestige (Kapferer, 1997). As Kochan (1996,

Consumer
Ethnocentrism
H6b
H6a
H3 (Belongingness)

Perceived
Brand
Globalness

H2,H4
H1

Perceived
Brand
Quality
Brand
Purchase
Likelihood

H5b

Brand
Local Icon
Value

H5a

Brand
Prestige

H5c

Figure 1

Conceptual model.

Perceived brand globalness creates brand value

J-B EM Steenkamp et al

55

xii) notes, 'the brands most admiredyare global
brands.' If global brands have higher prestige, it
could be because of their relative scarcity and
higher price compared with local brands (Bearden
and Etzel, 1982). Despite exceptions (Coca-Cola, for
example), evidence indicates that global brands are
typically more scarce and more expensive than
local brands (Batra et al., 2000). It is well established
that higher price and greater scarcity create greater
aspirational, prestige appeal (e.g. Bearden and Etzel,
1982). Global brands may also connote cosmopolitanism (Thompson and Tambyah, 1999). Certain
consumers are said to buy global brands to enhance
their self-image as being cosmopolitan, sophisticated, and modern (Friedman, 1990). The worldwide scale of these brands also allows them to be
associated with globally recognized events (such as
the Olympic Games and the World Soccer Cup)
and celebrities (Steffi Graf, Tiger Woods, Harrison
Ford, or Gerard Depardieu, for example). Through
a process of meaning transfer, the prestige attached
to these events and celebrities may be transferred
to the sponsoring global brand (McCracken,
1986). Although it is mentioned anecdotally
that PBG enhances brand prestige (e.g. Batra et al.,
2000, 84), empirical evidence is not available.
To test this likelihood, we offer the following
hypothesis:
H1: PBG is positively associated with the brand's
perceived prestige.
The second potential association of brand globalness concerns perceived quality. Brand name is a key
indicator of quality (Rao and Monroe, 1989), and a
global image can arguably enhance the brand's
perceived quality. If a brand is viewed as globally
available, consumers may attribute higher quality
to the brand because such quality is likely to be
thought of as critical to global acceptance (Kapferer,
1997; Keller, 1998). As noted by Alden et al. (1999),
global brands often advertise their worldwide
availability and acceptance. For example, ads for
Nivea Visage and P&G's major international
detergent brand, Ariel, featured brand-quality
testimonials from women in different countries.
Vodafone's push toward global branding stems
from wanting consumers everywhere to view it as
a signal of high-quality, dependable service (Anonymous, 2001b, 20). However, empirical tests of the
claim that brand globalness contributes to perceptions of higher quality are rare in the published
literature. Even rarer are studies that control, as we

shall, for other pathways (such as brand familiarity)
Thus we propose:
H2: PBG is positively related to consumer perceptions of brand quality.
A third reason for a global brand preference may
be the globalness per se of such brands, independent
of any effects via prestige and quality. We refer to
this as the belongingness pathway, because global
brands offer purchasers the opportunity to acquire
and demonstrate participation in an aspired-to
global consumer culture (GCC; Alden et al., 1999).
This is possible because such brands often appeal to
human universals and are purchased to signal
membership in worldwide consumer segments
(Dawar and Parker, 1994). Several authors
(Appadurai, 1990; Hannerz, 1990) note that media
flows, increased travel, and other factors are
creating widely understood symbols and meanings
reflected in global brands that, in turn, communicate membership in the global consumer community
with all its positive connotations (McCracken,
1986). PBG per se may therefore also be an added
value for consumers, distinct from and above any
incremental quality or prestige connotations associated with the brand. Hence we posit the existence
of a belongingness pathway:
H3: After controlling for brand quality and
prestige, PBG is positively associated with
consumers' purchase likelihood.
We further expect that the pathway through
brand quality will have the strongest influence on
purchase likelihood. Perceived quality is the
primary driver of purchase likelihood, irrespective
of product category, consumer segment or time
frame (Jacoby and Olson, 1985). Furthermore, the
importance of brand name as a quality cue and the
quality associations of global brands have been
widely noted in the literature (Yip, 1995). In
contrast, the prestige pathway may be more
limited to:
(1) Specific consumer markets/segments (young
urban consumers, for example).
(2) Some product categories (such as those that are
socially visible).
(3) Certain stages of market evolution (e.g., where
global brands are a novelty, extremely expensive, or scarce: Ger et al., 1993).

Journal of International Business Studies

Perceived brand globalness creates brand value

J-B EM Steenkamp et al

56

Finally, the belongingness pathway should be the
weakest, since according to authors such as Samli
(1995) GCC is still in its infancy, and local culture
has a stronger influence. Therefore we hypothesize:
H4: Of the three pathways through which PBG is
related to purchase likelihood, the pathway
via perceived quality is the most important
overall (across all markets and consumer
segments).

Brands as icons of local culture
The emergence of global brands does not suggest
that PBG is the only route to success. Despite the
advent of global culture, local culture remains a
central influence on consumer behavior and
individual identity (Samli, 1995). An alternative
route is to become an icon of the local culture.
Indeed, various scholars have expressed skepticism
about the existence of a standardized GCC and its
consequent strengthening of global brands, arguing
instead that we are much more likely to find
localized cultural variants that are not standardized
across all markets (Ger et al., 1993). These authors
also suggest that localized aspects of consumption
return to the forefront once the 'novelty' of
Western and/or global offers wears off. For example,
Belk (2000, 14) reports that, although elite consumers in Zimbabwe measure their consumption
success by adopting the symbols of the developed
world, they have also retained a strong local culture
orientation. Similarly, in a study of ethnicity and
consumption in Romania, Belk and Paun (1993,
193) find that Romanian consumers desire 'celebrations of disappearing folk culture' that bind them
to a mythical past.
Building on the frequent consumer desire for
local culture, Ger (1999) develops a conceptual
framework for competitiveness of local companies
and brands in global markets. Within the context of
global production and consumption, local brand
managers can achieve competitive success by using
local cultural capital and targeting and positioning
based on a deeper understanding of local culture,
tastes, and needs. In her framework, these two key
components of competitive success for the local
brand create 'a sustainable unique value and offer
the symbolism of authenticity and prestige' (Ger,
1999, 70). Because the interpretation of 'high
quality' may also differ across markets, consumers
may believe that a local brand is more in line with
local quality needs. Thus Ger (1999) proposes that

Journal of International Business Studies

local brands should 'out-localize' global brands by
presenting the 'local-as-an-alternative.' In sum, we
hypothesize:
H5a: The extent to which a brand is perceived to be
an icon of the local culture is positively related
to consumer perceptions of brand prestige.
H5b: The extent to which a brand is perceived to be
an icon of the local culture is positively related
to consumer perceptions of brand quality.
H5c: After controlling for brand quality and
prestige, the extent to which a brand is
perceived to be an icon of the local culture is
positively associated with consumers' purchase likelihood.
Note that H1–3 and H5 are not mutually
incompatible. A brand can rate high or low on both
the local and the global dimension. Heineken and
Coca-Cola are brands that are both perceived to be
global as well as strong icons of Dutch and U.S.
culture, respectively. The peanut butter brand Calve´
is a strong icon of Dutch culture but is not
perceived to be a global brand by Dutch consumers.
Sony is perceived to be a global brand but not an
icon of Japanese culture. Finally, consumers may
perceive certain brands to be neither a strong icon
of their local culture nor high on PBG (e.g. Dodge,
in the U.S.).

Moderating role of consumer ethnocentrism
Previous research has documented the enhanced
appeal of global brands among certain segments,
such as teenagers and businesspeople (Hassan and
Katsanis, 1994; Walker, 1996). In this study we
examine the moderating role of CET. CET is defined
as 'the beliefs held by consumers about the
appropriateness, indeed morality, of purchasing
foreign-made products' (Shimp and Sharma, 1987,
280). Ethnocentric consumers take pride in their
country's brands, symbols, and culture. They are
less open to foreign cultures, and are less cosmopolitan. CET is closely linked to economic nationalism (Baughn and Yaprak, 1996).
In the thinking underlying economic nationalism, local brands provide a link between the
national economy and individual well-being.
Global brands may be viewed suspiciously as a
threat to national economic prosperity. In the mind
of ethnocentric consumers, global brands pose not
only an economic but also a cultural threat. Lower
levels of cosmopolitanism and openness to foreign
cultures further contribute to negative evaluations

Perceived brand globalness creates brand value

J-B EM Steenkamp et al

57

of global brands by more ethnocentric consumers.
Ethnocentric consumers may even be willing to
sacrifice 'objective' gain (higher quality, lower
price, etc.) to enjoy the psychological benefit of
avoiding contact with the outgroup (i.e., the global
culture) by purchasing local brands (Baughn and
Yaprak, 1996). On the other hand, consumers low
on CET are more cosmopolitan in outlook and have
a higher degree of cultural openness (Baughn and
Yaprak, 1996). They do not attribute to local brands
an intermediate role between the national economy and individual well-being, and, if anything,
derive psychological gain from having contact with
the GCC, of which global brands are an important
component. Based on this literature, we hypothesize:
H6a: The positive association of perceived brand
globalness with purchase likelihood will be
weaker for more ethnocentric consumers.
H6b: The positive relationship between perceptions of the brand as an icon of the local
culture and purchase likelihood will be
stronger for more ethnocentric consumers.

Covariates
Although this study focuses on the pathways
through which PBG influences purchase likelihood,
other exogenous influences are likely. Three sets of
covariates are included in our analyses. First, brand
familiarity is included because previous research
suggests that it may have an important impact on
perceived brand quality, brand prestige, and
purchase likelihood, whether or not a brand is
perceived as global (e.g. Keller, 1998). Second,
country-of-origin (CO) image is included to control
for the possibility that a certain global brand may
attain higher prestige, quality, and/or purchase
likelihood because it comes from a particular
foreign country, rather than because it is global.
Previous work has documented the importance of
CO in consumer evaluations (Peterson and Jolibert,
1995). Finally, we add brand dummies to the
analyses to control for unobserved, brand-specific
effects (such as objective quality, distribution
coverage, and market share). Since product category effects are also captured by brand dummies
(each category was represented by two brands), we
also control for product category effects (such as
differences in perceived risk) by using these brand
dummies. Controlling for all these variables
provides a stronger test of our hypotheses, and

produces more accurate estimates of the true effects
of PBG.

Method
Procedure
Our hypotheses are tested using consumers from
Korea and the U.S.A. There are substantial cultural
differences between these countries. Korea belongs
to the top half of the countries surveyed on power
distance, uncertainty avoidance, and long-term
orientation, and to the bottom half on individualism and masculinity. The position of the U.S.A. is
exactly the reverse (Hofstede, 1991). Given these
large differences, our research setting provides a
stringent test of the generalizability of our hypotheses (Van de Vijver and Leung, 1997). The
hypotheses are tested across four product categories
including durables and nondurables, involving
eight brands per country, further increasing the
generalizability of the results.
Product categories and brands were selected from
a larger set developed through focus groups in both
countries. It was important for categories to vary
across the nondurable–durable continuum, to
increase the generalizability of our results. The
categories finally chosen included foods/beverages,
personal care products, and consumer durables. It
was also important for the brands within each
category to vary on globalness while being reasonably familiar. Two brands in each category were
chosen deliberately to represent the global and
local ends of the continuum in order to increase the
probability of variance on perceived brand globalness. Categories differed somewhat between countries, as well-known brands that were positioned as
more global in both the U.S.A. and Korea were not
available. Product categories (brands) selected for
the U.S.A. were cola drinks (Coca-Cola, Royal
Crown Cola), facial cream (Nivea, Pond's), color
TV sets (Sony, RCA), and wristwatches (Seiko,
Timex). For the Korean study, cola drinks (CocaCola, Chilsung), toothpaste (Colgate, Lucky),
color TV sets (Sony, Samsung), and refrigerators
(Whirlpool, Goldstar) were selected.
In both countries, respondents were resident
women responsible for at least half of the shopping
in the household. In the U.S., a survey was
mailed to a random sample of households from a
leading sampling firm's list. The response rate was
12% (247/2093). This response rate is typical
for mail surveys (Dillon et al., 1994, 144). Given
reported experiences in other Asian countries

Journal of International Business Studies

Perceived brand globalness creates brand value

J-B EM Steenkamp et al

58

(Kumar, 2000, 313), surveys were administered in
Korea using personal interviews to ensure sufficient
response. Employing randomized cluster sampling,
approximately 2000 households in Seoul were
selected and, of these, 370 agreed to participate
(18.5%).
The English version of the questionnaire was
double back-translated into Korean. To avoid
respondent fatigue in both countries, product
categories were rotated across questionnaires, in
sets of two categories for any one questionnaire.
Target brands were presented through their names
(i.e., logos were not used). For each product
category the brands were also rotated across
questionnaires. Hence each respondent answered
questions for two brands in each of two product
categories, a total of four observations per respondent. Though these multiple observations are not
independent, several authors have established that
bias due to non-independence of observations is
small when samples are large (Hunter and Schmidt,
1990, 452).

Measures
All scales used, along with their sources and their
coefficients of reliability, are detailed in the
Appendix. Only two of the scales are new. As
mentioned earlier, consumers can form the perception that a brand is 'global' if they come to believe
that the brand is also available in other countries
through media coverage, word of mouth, or travel
and/or if they see brand communications stressing
'globalness' through associations with GCC symbols (Alden et al., 1999). Expanding on the scale
used by Batra et al. (2000), we measured PBG with
three items indicating the degree to which consumers thought the same brand was marketed in
countries beyond their own. Note that PBG is not
an either/or construct but a continuous variable, as
some brands (Coca-Cola, for example) are higher
on PBG than other brands (Philips, for example).
Perceptions of the brand as an icon of local culture
were measured using three items indicating the
degree to which it was associated with, and
symbolized, the local country or culture.
CO perceptions were measured for the countries
in which the mother company of the brand was
located, using four items derived from previous
country-of-origin research (Hunter and Nebenzahl,
1984). CE was measured with four of the highestloading items from the original CETSCALE study
(Shimp and Sharma, 1987). Steenkamp et al. (1999)
reported correlations above 0.95 between this four-

Journal of International Business Studies

item scale and the 10-item CETSCALE for Great
Britain, Belgium, and Greece.

Results
Cross-national measurement validation
Before comparing results from the U.S.A. and
Korea, configural and metric invariance were tested
using multigroup confirmatory factor analysis
(CFA; Mullen, 1995; Steenkamp and Baumgartner,
1998). Brand-level measures were analyzed simultaneously to rigorously test their psychometric
properties. Configural invariance of the six-factor
model was supported as the CFA model fit was
good: w2(152) ¼ 1218.69 (Po0.001), CFI ¼ 0.933,
TLI ¼ 0.907, CAIC ¼ 1990.10. All factor loadings
were significant at Po0.001, and the (within-group
standardized) factor loadings were all above 0.5. All
factor correlations were significantly below unity
(Po0.0001), supporting discriminant validity between the constructs (Anderson and Gerbing,
1988). Equality of factor loadings was also supported: w2(161) ¼ 1266.20 (Po0.001), CFI ¼ 0.931,
TLI ¼ 0.909, CAIC ¼ 1958.71. CFI decreased an
insubstantial 0.002, whereas CAIC and TLI,
indicators of goodness of fit and model parsimony,
improved. These findings provide strong evidence
to support the metric invariance of brand-level
measures (Steenkamp and Baumgartner, 1998).
Configural invariance was also obtained for CO
perceptions. The fit of the (one-dimensional) CFA
model was good: w2(4) ¼ 39.83 (Po0.001),
CFI ¼ 0.990, TLI ¼ 0.969, CAIC ¼ 176.21. All factor
loadings were significant at Po0.001 and above
0.6. Metric invariance was also supported:
w2(7) ¼ 59.83 (Po0.001), CFI ¼ 0.985, TLI ¼ 0.974,
CAIC ¼ 170.63. Finally, CET exhibited configural
invariance. The fit of the (one-dimensional) CFA
(Po0.01),
model
was
good:
w2(4) ¼ 13.85
CFI ¼ 0.992, TLI ¼ 0.975, CAIC ¼ 132.65. All factor
loadings were significant at Po0.001 and above
0.7. Metric invariance was also supported:
w2(7) ¼ 16.37 (P ¼ 0.022), CFI ¼ 0.992, TLI ¼ 0.986,
CAIC ¼ 112.89.
In sum, cross-national invariance of the measures
used was supported. The good model fit and the
significant and high factor loadings further support
the unidimensionality and convergent validity of
the constructs (Anderson and Gerbing, 1988). Items
were averaged for each scale to obtain composite
scores for the various constructs. Mean values of
key constructs are provided in Table 1. Given that
metric invariance is established, we can now validly

Perceived brand globalness creates brand value

J-B EM Steenkamp et al

59

Table 1

Means and standard deviations of key constructs

Purchase intention
Perceived global brand
Perceived brand quality
Brand icon vale
Brand familiarity

U.S.A. (n ¼ 898)

U.S.A.

Korea (n ¼ 1460)

Korea

Mean

s.d.

Mean

s.d.

4.82
4.93
5.26
4.07
5.53

1.90
1.50
1.34
1.53
1.28

4.78
4.72
5.04
5.19
5.33

1.45
1.58
1.13
1.61
2.12

estimate the structural relations between the
constructs and test the hypotheses in a crossnational setting (Steenkamp and Baumgartner,
1998).

Test of hypotheses
To test the hypotheses, structural equation modeling was used. Each latent construct was measured
by a single indicator scale, formed by averaging
multiple items and fixing error variance at a level
appropriate to its coefficient alpha reliability
(Anderson and Gerbing, 1988). The use of summated indicators is common when the model is
highly complex. Furthermore, this approach typically leads to more precise structural estimates,
provided the measures are unidimensional
(Bandalos, 2002), as is the case in this study. Brand
familiarity, country-of-origin image, and brand
dummies were added to each structural equation
as covariates to control for possible confounds. To
avoid overfitting, and to increase degrees of freedom, only brand dummies with Po0.10 or better
were retained to estimate the 'trimmed' model. In
the tables below, control variable data are omitted
as they are not the study's focus. The multigroup
model with the U.S.A. and Korea as two groups
yielded a good fit: w2 (21) ¼ 48.187 (Po0.001),
CFI ¼ 0.997, TLI ¼ 0.976, CAIC ¼ 1704.80. The unstandardized structural coefficients, appropriate in
cross-national research (Singh, 1995), are reported
in Table 2.
Rival models Before we present estimates for the
hypothesized model, it is important to establish
that this model fits better than other plausible rival
models. We estimated two plausible rival models.
The first rival model specifies the reverse causal
direction between prestige/quality and PBG. It
could be argued that a brand is considered to have
'global or world-class levels' of prestige and
perceived quality if its prestige and perceived

Table 2 Unstandardized structural coefficients (with standard
errors), total sample

N

Total U.S.A.
sample
897

Total Korea
sample
1460

Global brand-Prestige
Local icon-Prestige
Global brand-Quality
Local icon-Quality
Global brand-PI (direct)
Local icon-PI (direct)
Prestige-PI
Quality-PI
Global brand-PI (total)
Local icon-PI (total)

0.361
0.309
0.152
0.049
0.112
0.116
0.042
1.244
0.317
0.191

0.434 (0.056)*
0.104 (0.048)***
0.573 (0.051)*
0.061 (0.033)
0.032 (0.057)
0.117 (0.038)**
0.048 (0.019)***
0.732 (0.056)*
0.367 (0.080)*
0.157 (0.051)**

(0.103)*
(0.060)*
(0.063)***
(0.042)
(0.075)
(0.051)***
(0.028)
(0.074)*
(0.091)*
(0.061)**

*Po0.001, *Po0.01, *Po0.05.

quality are very high. The fit of this model
(w2(21) ¼ 83.705, Po0.001; CFI ¼ 0.994; TLI ¼ 0.946;
0.946; CAIC ¼ 1740.32) is significantly worse than
our hypothesized model. In addition, since the two
models have the same d.f., their w2 statistics can be
compared directly. This comparison reveals that the
rival model's fit is 74% weaker.
Second, a model was estimated specifying bidirectional relations between prestige/quality and
PBG. This model posits that prestige and quality
affect PBG and vice versa. Our theoretical model is a
nested version of this mode...