Journal of Business Research 67 (2014) 278287

Contents lists available at ScienceDirect

Journal of Business Research

Understanding consumer's brand categorization across three

countries: Application of fuzzy rule-based classication
Shimi Naurin Ahmad , Marie-Odile Richard 1
Earl G. Graves School of Business and Management, Morgan State University, 1700 East Cold Spring Lane, Baltimore, MD 21251, USA

a r t i c l e i n f o a b s t r a c t

Article history: In multi-brand situations, people categorize all known brands into subsets called consideration, hold, foggy and
Received 1 November 2012 reject sets. This is the BrisouxLaroche model. Traditional brand categorization models including this, assume
Received in revised form 1 May 2013 that consumers can properly categorize each brand into these subsets. However, a brand with both perceived
Accepted 1 May 2013
positive and negative attributes increases the difculty for a consumer to decide about the placement into sub-
Available online 18 June 2013
sets. This study investigates consumers' brand categorization when a brand has both perceived positive and neg-
Keywords:
ative attributes. We propose that a brand may belong to more than one subset (decision fuzziness). Using
Fuzzy rule-based classication fuzzy-rule-based classication, this is investigated across three cultures (Chinese, Japanese and Kazakhstan)
Brand categorization and two product categories: Quick Service Restaurants (QSR) and beer. The ndings conrm that decision fuzz-
Consideration set iness varies across cultures. Chinese consumers have less decision fuzziness for foreign brands than for local
China brands in the QSR market. In general, the opposite is found to be true for Japanese and Kazakh consumers.
Japan 2013 Elsevier Inc. All rights reserved.
Kazakhstan

1. Introduction do consumers categorize a brand in this scenario? The traditional model

As the number of brands in a product category increases, so does the
difculty for consumers to make a choice. Among all brands available,
only some might be known to consumers. This subset is the consumer's
awareness set. While some brands within this subset are considered for
purchase, some are rejected. The rst set of brands belongs to the con-
sideration set and the rejected brands belong to the reject set. A con-
sumer might hold opinions about some brands which they will not
consider buying in the near future which constitute the hold set. How-
ever, there are some brands that are not even evaluated for purchase
and these are in the foggy set. In the marketing literature, this concep-
tualization is called brand categorization (Brisoux & Laroche, 1980).
This taxonomy assumes that consumers are able to properly catego-
rize each brand in these sets. However, an alternative with both per-
ceived positive and negative attributes increases trade-offs and in turn
decision difculty (Chatterjee & Heath, 1996) which may lead to ambig-
uous categorization, or deferral of decision (Dhar, 1997). For example,
during the Maple Leaf meat recall in 2009, consumers who were devoted
users of the brand may have rejected it at that time. However, when the
Maple Leaf CEO apologized and pledged consistent quality in the future,
those who rejected the brand may have reconsidered the situation. Con-
sumers hold some positive and negative opinions about the brand at the
same time. This conict makes the categorization decision difcult. How
Corresponding author. Tel.: +1 301 859 4404.
E-mail addresses: shiminahmad@gmail.com (S.N. Ahmad), odile10@hotmail.com
(M.-O. Richard).
1
Tel.: +1 514 738 3520.
of brand categorization cannot distinguish these cases.
Drawing on literature, our study tries to ll the void by proposing
that when a decision is not obvious in consumers' minds, a brand may
belong to more than one subset of the awareness set. This phenome-
non of one brand belonging to more than one set is termed decision
fuzziness. According to the BrisouxLaroche model of brand categori-
zation, the levels of cognition about a brand, attitude towards the
brands, condence in evaluation and purchase intention determine
a brand's categorization to either the consideration, hold, reject or
foggy set (Table 1). Therefore, these determine decision fuzziness.
The next question is, does decision fuzziness vary across brands in a
product category? Research showed that consumers' belief about
appropriateness to purchase foreign made products (ethnocentrism)
affects the attitude towards local and foreign brands and the purchase
intention within a category (Watson & Wright, 2000). Therefore, ethno-
centrism affects the determinants (attitude and purchase intention) of
the decision fuzziness and in turn affects decision fuzziness of local
and foreign products. Moreover, researchers showed that culture inu-
ences consumer ethnocentrism (Yoo & Donthu, 2005). Therefore, we
believe that decision fuzziness varies across foreign and local brands
and cultures, and it is important to examine the effect of culture on
decision fuzziness of foreign and local brands. Thus, another goal is to
investigate decision fuzziness across cultures (Chinese, Japanese and
Kazakhs). We investigate this phenomenon across two categories
(Quick Service Restaurant (QSR) and beer) for foreign and local brands.
We also examine how this pattern of decision fuzziness can be modied
by adding a managerially actionable positive attribute, a coupon, to the
offering.

0148-2963/$ see front matter 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jbusres.2013.05.014
 S.N. Ahmad, M.-O. Richard / Journal of Business Research 67 (2014) 278287
Table 1
Brand categorization (Laroche et al., 2005).
Consideration set Hold set Foggy set Reject set
Cognition High Average to Low Average Low
Attitude High Average Low Lower than Average
Condence High Average to Low Average Low
Intention High Average to Low Low Low
To test the hypotheses, we applied the fuzzy rule based (FRB) clas-
sication on the BrisouxLaroche brand categorization model. Unlike
the probability based model, FRB provides degrees of membership or
belongingness for more than one set at the same time and this is the
central aspect to be investigated.
Next, we briey review the literature on brand categorization and
fuzzy rule based methodology. Then, in three studies, we examine de-
cision fuzziness. The rst explores decision fuzziness of the Chinese
QSR market compared with the Japanese QSR market. In the second,
the Japanese QSR market is compared with the Japanese beer market.
In the last study, the Japanese and Kazakhstan beer markets are com-
pared. We conclude with managerial implications and limitations.
2. Conceptual background
When faced with several brands in a category, consumers categorize
the available brands into different sets. This brand categorization
concept was rst put forward by Howard (1963). In 1980, Brisoux and
Laroche presented a complete brand categorization paradigm (Fig. 1).
Accordingly, consumers classify all the brands in a category into two
sets, namely the foggy (attributes are not processed) and the processed
sets. The brands in the processed set are evaluated and put into
either consideration, reject or hold sets. There are four determinants
(cognitions, attitudes towards the brand, condence in evaluation and
purchase intentions) of each of the sets and these vary from low to
high (Table 1).
The cognitions, condence in evaluation and purchase inten-
tions are average to low for the hold set, and average for the reject
set. Attitudes are average for the hold set and low for the reject
set. For example, consumers may have average cognitions and con-
dence towards well-known brands such as the Maple Leaf brand in
Canada. However, consumers may not have clear attitudes and purchase
intentions towards the brand because of a particular attribute (e.g., con-
fusion from reconsidering the quality of the Maple Leaf meat after the
CEO's assurances). Consequently, attitudes and purchase intentions to-
wards the brand may be between average and low. Now, if we classify
the brand into one of these sets based on consumers' cognitions, con-
dence in evaluation, brand attitudes and purchase intentions (Table 1),
it is likely that the brand would belong to both the hold and reject sets
Fig. 1. The BrisouxLaroche model of brand recognition.
279
to similar degrees because of the overlapping of the determinant vari-
ables of the hold and reject sets.
Moreover, there is an intuition in the literature supporting this prop-
osition. As mentioned by Yoon, Thompson, and Parsa (2009), these sub-
sets of awareness set may not be mutually exclusive and thus, there
exists the possibility of one brand belonging to more than one set.
Therefore,
H1. A brand may belong to more than one subset of awareness set in
the consumer mind.
This phenomenon of a brand's belongingness to more than one sub-
set of awareness in consumers' minds is termed pattern of decision
fuzziness. This pattern increases when multiple inuences (positive
and negative) act on consumers' decision making processes.
2.1. Culture and decision fuzziness
As mentioned, studying decision fuzziness across cultures needs
careful examination of the effects of culture on four variables (brand
attitude, cognition, condence in evaluation and purchase intention).
In some cultures consumers have favorable attitudes and high pur-
chase intentions towards foreign brands and in others they do not
(Yoo & Donthu, 2005). Researchers suggest that culture affects con-
sumer ethnocentrism which in turn affects the attitudes and purchase
intentions towards buying foreign brands (Watson & Wright, 2000;
Yoo & Donthu, 2005). Consumer ethnocentrism is the beliefs held
by consumers about appropriateness, indeed morality, of purchasing
foreign made products (Shimp & Sharma, 1987, p 280). Highly
ethnocentric consumers have a less favorable attitude towards
foreign brands. Researchers link the collectivism dimension of culture
to consumer ethnocentrism (Yoo & Donthu, 2005). Studies also suggest
that less favorable attitudes towards foreign brands are manifested
more frequently among consumers with high collectivism (Balabanis
& Diamantopoulos, 2004).
2.2. Choice of context
We examine decision fuzziness across three countries (China, Japan
and Kazakhstan) with both foreign and local brands in each market.
These countries are predominantly collectivist (Hofstede, 1991). Apart
from geographical proximity, the cultural dimension scores (Table 5)
are similar. By choosing these countries, we see a range in scores and
thus differences within the collectivist umbrella. China portrays a classic
example of collectivism, Japan is less so, with Kazakhstan somewhere
between the two.
We chose foods as our product category since food taste preferences
are shaped by culture (Wright, Nancarrow, & Kwok, 2001). They noted
that marketers need to recognize the tension between the local
280 S.N. Ahmad, M.-O. Richard / Journal of Business Research 67 (2014) 278287

Table 2a Table 2c
Sample demographics of Chinese fast food survey. Sample demographics of Kazakhstan survey.

Variables Percentage (%) Variables Percentage (%)

Gender Male 49.3 Gender Male 41.9

Female 50.7 Female 58.1
Age Under 20 15.7 Age Under 20 29.1
2029 27.3 2029 34.1
3039 25.0 3039 12.8
4049 19.7 4049 15.6
5059 10.3 5059 7.3
Over 60 2.0 Over 60 1.1
Education Some schools 56.0 Education Some schools 16.2
College/Bachelor 43.7 College/Bachelor 62.6
Masters 0.3 Masters 21.2
Family income Under 999 0.7 Family income Under T2000 0.6
10001999 15.3 T2000T8000 5.0
(In Chinese RMB) 20002999 38.3 T8000T20000 30.3
30003999 24.3 T20000T50000 35.2
40004999 11.3 T50000T1000000 15.3
50006999 7.0 Over 100000 14.0
70008999 2.3
9000 or over 0.7

brands. They correspond to different segments of the Japanese QSR in-

cultural values and new choices from other countries and this fact
makes it suitable to study decision fuzziness. We chose Quick Service
Restaurants (QSR) because of the popularity and availability of
foreign/local brands.
We compare the Chinese and Japanese QSR markets. However, re-
spondents in both markets indicated that environment and service
quality are very important when choosing a brand. To rule out possible
alternative explanations that observed phenomena might have been
caused by a good restaurant environment of western QSR brands, we
examine the beer category. Preferences towards beer brands do not
depend on the establishment since it is not sold in a specic place and
this allows directly comparing foreign and local brands.
2.2.1. The Chinese market
The Chinese QSR market is coexisting with Chinese-style and
western-style fast foods. Chinese-style fast food restaurants emerged
after western companies such as KFC and McDonald's entered China
in 1987 and 1990 respectively. These Chinese outlets have a fast food
mode that combines traditional Chinese diet (noodle or dumpling)
with the modern fast food facilities. Thus, we included KFC, McDonald's
and Pizza Hut as western QSRs and Ma Lan Hand Pulled Noodles, Beijing
Quan Du Roast Duck and Tian Jian Steam Stuffed Buns as Chinese QSRs.
These brands have been the focus of studies on QSR brands in China
(Teng, 2009). Although local brands are learning modern management
techniques, they are sometimes considered to have lower service qual-
ity (Gao, Pan, Tse, & Yim, 2006).
dustry (chicken, sandwich, hamburger, and beef bowl). MOS Burger is
a giant chain in Japan which has expanded to other countries. Unlike
other local QSR brands, MOS Burger provides burgers with distinct
Japanese taste, green tea dessert with sweet red beans and is directly
comparable to any other big western fast food chain (Lee, 2006). In con-
trast, Lotteria is a copied version of McDonald's; and Matsuya offers a
local fast food version (beef bowl).
When comparing the Japan QSR market with the Japan beer
market, ve leading brands: three Japanese (Kirin lager, Asahi super
Dry and Kirin Ichiban Shibori) and two foreign brands Budweiser
(North American) and Heineken (European) are chosen. Since there
is no evidence that the Japanese prefer North American brands to
European brands, we expect this inclusion to enhance the generaliz-
ability of our results.
2.2.3. The Kazakhstan market
We examine one foreign brand, Baltika, and ve local brands (Bavaria,
Zhigulevskoe, Tian Shan, Karagandinskoe and Irbis) in Kazakhstan. The
reason for choosing one foreign brand is that the Kazakhstan beer market
consists mainly of foreign subsidiaries. Therefore, it is hard to classify
them as foreign or local. Baltika is a Russian imported beer but adapted
to local taste. The locals are produced in Kazakhstan to make these brands
Kazakh. For example, the name Irbis represents a national symbol in
Kazakhstan. Among these, Zhigulevskoe, Tian Shan, and Karagandinskoe
are positioned in the economy segment; Baltika, Irbis and Bavaria are
premium quality.

2.2.2. The Japanese market 2.3. Hypotheses development

In the Japanese QSR market, we include three foreign (McDonald's,
Subway and KFC) and three local (MOS Burger, Lotteria and Matsuya)
Table 2b
Sample demographics of Japanese study.
2.3.1. China
Previous scholars speculated that consumer nationalism could be-
come a barrier to multinational companies in China (Wang & Wang,

Variables Percentage (%) Table 3a

Japanese fast food Gender Male 66.9 International and local brands' proles across the four sets in Chinese and Japanese fast
Female 33.1 food and Japanese and Kazakhstan beer markets.
Age 1820 0.8
Consideration Hold Foggy Reject
2124 82.0
2529 17.2 Chinese fast food Foreign 83% 8% 6% 3%
Occupation Student 100 Local 44% 25% 7% 23%
Japanese beer Gender Male 69.8 Japanese fast food Foreign 56% 26% 11% 7%
Female 30.2 Local 52% 24% 10% 14%
Age 1820 42.0 Japanese beer Foreign 22% 20% 32% 26%
2124 55.9 Local 59% 20% 12% 9%
2529 2.1 Kazakhstan beer Foreign 71% 16% 8% 5%
Occupation Student 100 Local 51% 18% 13% 18%
 S.N. Ahmad, M.-O. Richard / Journal of Business Research 67 (2014) 278287 281

Table 3b Table 4
Local brands' proles across the four sets in Chinese fast food market. Average coefcient alpha values of the measurements.

Consideration Hold Foggy Reject Cognition Attitude toward Condence in Purchase

the brand evaluation intention
Ma Lan Hand Pulled Noodles 40% 23% 25% 12%
Tian Jian Steam Stuffed Buns 30% 23% 30% 17% Chinese fast food 0.76 0.81 0.84 0.68
Beijing Quan Du Roast Duck 54% 35% 7% 4% Japanese fast food 0.89 0.88 0.91 0.90
Japanese beer 0.86 0.83 0.79 0.88
Kazakhstan beer 0.87 0.94 0.82 0.93

2007). However, the evidence proved the opposite. The Chinese

consumer's attitude towards foreign brands, especially western
brands, is quite favorable (Wang, 2000) although there is evidence
that the Chinese are ethnocentric (Hsu & Nien, 2008). To reconcile
this, Dong and Tian (2009) argued that Chinese nationalism is con-
ceived through the consumption of western brands to erase China's
past humiliation and create a progressive future. Chinese consumers
take foreign brands, especially western brands, as the symbol of democ-
racy and wholeheartedly adopt them. These are seen as instruments of
democratization and economic progress (Dong & Tian, 2009). For
advanced social standing, the Chinese imitate the consumption practice
of westerners (Wang, 2000). Consumers like global products and feel
that elements of foreign, or western, culture are appropriate (Zhou &
Hui, 2003).
By comparing foreign and domestic brands in China, Gao et al. (2006)
nd that foreign brands are associated with superior core advantages
such as the universal traits of a strong brand such as superior quality, in-
novation, value pricing, and services, as well as intangible aspects. A
brand with core advantages gives consumers a positive emotional asso-
ciation regarding brand choice (Schmitt & Simonson, 1997). Western
QSR brands have a competitive advantage of having superior core advan-
tages. However, local QSR brands have the advantage of offering foods
suitable to local taste (Schuiling & Kapferer, 2004) but often lack modern
management techniques and early mover advantage (Gao et al., 2006).
Altogether, foreign QSR brands in China have fewer negative inu-
ences compared to their positive attributes and thus create fewer
conicts in consumers' minds than their domestic counterparts.
We examine another product category, beer, in the same market
which is unlikely to be affected by the restaurant environment or
the service related variable. Moreover, unlike the QSR market, the
Japanese have a long history of beer making dating back to 1870
and the beer is also exported to many countries. Consequently, unlike
western QSR brands, the foreign beer brands in Japan do not have the
early mover advantage. In this market foreign brands have a negative
attribute of being foreign.
These brands do not have the competitive advantages regarding
restaurant environment or service quality as in the QSR market.
Therefore, we expect a more pronounced pattern than in the Japan
QSR market.
H3b. In the Japanese beer market, foreign brands have higher decision
fuzziness than the local beer brands.
2.3.3. Kazakhstan market
We compare the Japanese beer market with that of Kazakhstan.
Beer consumption in Kazakhstan is growing due to double digit
shifting from hard liquor to pub culture (Jun, 2006). There is growing
evidence that in Kazakhstan, consumer ethnocentrism leads to high
consumption of local brands and to lower levels of consumption of
foreign brands (Rybina, Reardon, & Humphrey, 2010).
Therefore, we expect foreign brands to have higher decision fuzz-
iness than the local beer brands.

H2. In the Chinese QSR market, foreign brands have a lower pattern
of decision fuzziness than their domestic counterparts.
2.3.2. Japan
In adopting foreign brands, one difference between China and Japan
is that Japanese consumers evaluate domestic brands more favorably
than foreign ones (Gurhan-Canli & Maheswaran, 2000; Nagashima,
1970) and they have a preference towards local brands (Hsieh, 2004).
Researchers showed that Japanese consumers are signicantly more
ethnocentric than consumers in the United States (Neuliep, Chaudoir,
& McKroskey, 2001). Therefore, foreign or western brands have the
disadvantages of being foreign, although they are high in product and
service quality. Moreover, some local brands (such as MOS Burger)
have outlets overseas signifying high quality with local taste.
Thus we expect foreign brands to have high decision fuzziness be-
cause they are foreign but with high quality.
H3a. In the Japanese QSR market, foreign brands have a higher pattern
of decision fuzziness than domestic brands.
H4. In Kazakhstan beer market, foreign brands have higher decision
fuzziness than the local beer brands.
Decision fuzziness can be desirable or undesirable. Any multiple ex-
istences with the consideration set are undesirable. However, a brand
which belongs to the reject set and hold set simultaneously is more like-
ly to move to the hold set than a brand which belongs to the reject set
only. Here, the fuzziness is comparatively desirable. Next, we investi-
gate how decision fuzziness might be changed by a marketing action
such couponing. Research showed that price discounts move a brand
from the hold set to the consideration set (Teng, 2009). We applied cou-
pon to local brands in the Chinese QSR market and examined how this
changes the pattern of decision fuzziness.
It is imperative that the application of coupons adds more positive
attributes to the brand. Thus,
H5. By offering coupons the brand moves consumers more towards
considering the brand. Decision fuzziness may increase or decrease
depending on current state.

Table 5
Table 3c Hofstede's (1991) cultural dimension score.
Local brands' proles across the four sets in Kazakhstan market.
Dimension China Kazakhstana Japan
Consideration Hold Foggy Reject
Collectivism/individualism 20 21 46
Bavaria 50% 19% 17% 14% Power distance 80 68 54
Zhigulevskoe 43% 15% 7% 35% Masculinity 66 41 95
Tian Shan 63% 12% 11% 14% Uncertainty avoidance 30 55 92
Karagandinskoe 37% 25% 18% 20% Long term orientation 118 unavailable 80
Irbis 55% 24% 17% 4% a
Source: http://internationalbusiness.wikia.com/wiki/Kazakhstan_Uncertainty_avoidance.
282 S.N. Ahmad, M.-O. Richard / Journal of Business Research 67 (2014) 278287
Fig. 2. Mathematical formulation of membership function.
2.4. Choice of method
Probability theory based classication presupposes that an object
can only be in one class and provides likelihoods for each of the clas-
ses. These likelihoods describe the uncertainty of event occurrence. In
contrast, the fuzzy rule based classication system provides a degree
of membership which is not a probability and unlike probability theory,
it may be assigned with complete certainty (Lewis-Beck, Bryman, &
Liao, 2004). Viswanathan and Childers (1999) put forward a compre-
hensive fuzzy set approach which distinguishes between degree of
membership and probability. In their study, products have different de-
grees of membership for each product attribute. For example, they as-
sert, a Ford Contour may possess membership of m = 0.3 in the
economy car and m = 0.8 in the family car meaning that it has both
inherent attributes of being an economy and family car in varying
degrees.
The fact that fuzzy techniques allow partial membership to more
than one category is crucial here. To examine decision fuzziness, FRB
is needed since, unlike probability theory, it allows for the simultaneous
existence of a brand in more than one set. FRB uses a set of rules
extracted from the BrisouxLaroche model (Table 1) which has been
tested and validated in many contexts and across product categories
(Laroche & Parsa, 2000; Laroche, Takahashi, Kalamas, & Teng, 2005).
Using these rules, degree of membership in each set is calculated for a
brand and a consumer. The degree of membership is the degree to
which a brand is a member of a set. If the membership values of any
Fig. 3. Membership functions.
of the two sets are very close, the brand exists in both sets with similar
strengths and decision fuzziness occurs.
For example, if McDonald's has the degree of membership of 0.79 for
hold and 0.75 for consideration set, it can be argued that the t be-
tween the brand and the consumer hold set, as well as consideration
set, is almost the same. Now, if the consumer explicitly states that the
brand is in his hold set, we can infer that although the consumer is
not actively considering purchasing the brand, he has an inclination to-
wards buying it. The same cannot be argued given the probability of the
two sets. Exact knowledge removes uncertainty. If the probability of
being in the hold set is 0.79, and a consumer states his preference
(hold set) explicitly, the probability becomes 1 for the hold set and 0
for the other sets. Moreover, it is not possible to have two probability
values of 0.79 and 0.75 mathematically.
2.4.1. Fuzzy rule-based classication
We modeled the awareness set as a fuzzy rule based classication
system (FRB) and the concordance between the FRB and the
self-stated classications are analyzed. For correctly classied cases,
we compare the two highest memberships among the subsets to see
the closeness and to decipher the consumer's hidden inclination
based on the measures of determinants.
Zadeh (1965) introduced the fuzzy set theory. Fuzzy systems are
mathematically based in linguistics (high, low, etc.) rather than with
numerical variables. A fuzzy set is a class of objects with a continuum
of grades of membership and is characterized by a membership func-
tion. Fuzzy classication rules classify an object into a set with a degree
of membership ranging from 0 to 1. This classication system is adapted
from Hellmann and Jager (2002) where they used the technique to clas-
sify images to different classes such as forest, etc.
The linguistic terms like high, average and low are dealt with
fuzzy sets and are depicted by membership functions. The mathemat-
ical formulation and the graphical representation of membership
function are presented in Figs. 2 and 3. This can be seen as a function
: X [0,1]. Here, X can only take a value in between 1 and 7 as cog-
nitions, condence, attitudes and purchase intentions are measured
with a 7-point scale. The value (x) at x X is the grade of member-
ship of x to (Zadeh, 1965). For each x X fuzzy set measures the
extent to which the linguistic term is fullled for x.
In mathematical term, classication rules are written as follows:
N N N N
RN : IF x1 is A1 AND x2 is A2 AND AND xn is An THEN class is C :
Each variable xk is dened on a set Xk and is described by linguistic
terms Aik (high, average, low). Each input n-tuple (x1 xn) of values of
the variables are mapped on a certain class c {c1, c2 cc}. The set of
rules {Ri|i = 1, N} may be considered as a piecewise dened function
= {1 n} with Ci if 1 AI1 and n AIn (Hellmann & Jager,
 S.N. Ahmad, M.-O. Richard / Journal of Business Research 67 (2014) 278287
2002). In the rst step, the grade of fulllment of each rule antecedent is
calculated as
i i i
i 1 xi 2 x2 n xn i 1 N: several QSR brands in China (Japan). If a consumer is not aware of a
Here linguistic AND is modeled by the minimum operator . The
values i are the degrees to which a rule Ri hold for (x1, , xn).
For each class ci {c1, c2, , cc} is modeled as a so-called singleton
1ci : {c1, c2, , cc} [0,1]. Thus 1ci y 1 for y = ci and 0 otherwise.
Each rule provides a fuzzy output
out
x1 ;;xn y i 1ci y
where y ranges in the set of classes. Each rule contributes to the nal
fuzzy output of the rules to an extent of i. To get the nal output, we
aggregate the singletons to one fuzzy set.
out N out N
 in evaluation, and purchase intentions are measured. The alpha
x1 ;;xn y i;x1 ;;xn y i 1ci y
i1 i1
denotes maximum and is modeling the linguistic OR. This gives for
each class y {c1, c2, cc} a grade of being compatible with the param-
eters (x1 xn). The decision is made in favor of the class Ci which has
maximum out
ix1 ;;xn :. This is called defuzzication. If the algorithm
yields for a brand 0.9 for set i and 0.85 for set j, then a decision in
favor of set i is taken although the two values are hardly distinguishable.
We take this advantage of observing the membership for each class and
thus the brands with indistinguishable membership values for more
than one set.
For validation, we compare the classication accuracy of FRB to
Support Vector Machine (SVM) which is equivalent to fuzzy IFTHEN
rules (Pitiranggon, Benjathepanun, Banditvilai, & Boonjing, 2010). It out-
performs discrete choice models in marketing (Cui & Curry, 2005).
2.4.2. Support Vector Machine
The SVM performs classication between two classes by constructing
a decision surface between samples of two classes, maximizing the
margin between them. The classication function (Vapnik, 1998) is
! chase intentions towards McDonald's are measured.
X
N
f x sign i yi K xi ; x b
i1
where x Rd is an input vector, xi is a training sample vector along with
its label yi (+1,1)N (class labels for binary situation). b is a bias and
K is a kernel function which maps the vectors into a higher dimensional
space by the non-linear mapping function f: Rd Rl, where l > d or l
could even be innite. We utilize a multi-class classication method.
3. Study 1
We examine the Chinese and Japanese QSR markets and compare
them on local and foreign brands on decision fuzziness. In the Chinese
market, three hundred and one consumers were surveyed about three
foreign brands (McDonald's, KFC and Pizza Hut) and three local brands
(Ma Lan Hand Pulled Noodles, Tian Jian Steam Stuffed Buns and Beijing
Quan Du Roast Duck). The sample demographics are shown in Table 2a
(49.3% male and 50.7% female aged between 18 and 59). In the Japanese
QSR market, three hundred and thirty nine consumers were surveyed
about three foreign brands (McDonald's, KFC and Subway) and three
local brands (MOS Burger, Matsuya and Lotteria). The sample demo-
graphics are shown in Table 2b (66.9% male and 33.1% female aged be-
tween 18 and 29). Choosing brands from a set may bias the result.
Among the local brands we chose the brands with similar proles
which also fairly match their foreign counterparts (Tables 3a and 3b).
283
3.1. Questionnaire
A screening question asked respondents whether they had heard of
brand, it falls under unawareness set and the respondent does not cate-
gorize the brands further. If known, respondents then categorized QSR
brands into different sets. The consideration set was measured using
these two questions: (1) Indicate your rst choice from all the brands
you are aware of from the list below and (2) Indicate which other
brands (from the same product category) you would consider selecting
when your rst choice is not available. The hold set was measured
by asking of those brands, are there any which you have formed an
opinion of but cannot say whether or not you would be willing to select
it?. The reject set was measured by asking the respondents to indicate
the brands they would denitely not consider buying, and nally the
foggy set was measured by asking them to indicate among the brands
that they know, those that they have not yet formed an opinion of.
In the next part, brand cognitions, Brand attitudes, condence
measures are provided in Table 4.
In the Chinese market, scenario based coupons (worth RMB 10) for
each of the three local QSR brands (30 for Tian Jian Steam Stuffed Buns,
29 for Malan Hand Pulled Noodles, 28 for Beijing Quan Du Roast Duck)
are provided randomly at the end of the main questionnaire. Their
brand attitudes and purchase intentions are measured again. 77 re-
sponses are analyzed as the rest did not know the local brands. This re-
sult is compared with the result prior to the couponing application.
3.2. Procedure
The following procedure was performed in MATLAB.
Step 1 If a consumer is aware of a brand, he categorizes the brand in
either the consideration, hold, foggy or reject set (for example,
respondent 1 answered that McDonald's is in his consider-
ation).
Step 2 Each respondent expresses his cognitions, brand attitudes, con-
dence, and purchase intentions towards a brand if it is in
either his consideration, hold, foggy or reject set. In this exam-
ple, respondent 1's cognitions, attitudes, condence, and pur-
The degree of memberships for the consideration, hold, foggy and
reject sets are obtained according to the rules based on the Brisoux
Laroche model.
Rule 1 IF cognitions are High AND attitudes are High AND condence
is High AND Intentions are High THEN class = Consideration
set.
Rule 2 IF cognitions are (Average or Low) AND attitudes are Average
AND condence is (Average or Low) AND Intentions are
(Average or Low) THEN class = Hold set.
Rule 3 IF cognitions are Average AND attitudes are Low AND condence
is Average AND Intentions are Lowest THEN class = Reject set.
Rule 4 IF cognitions are Lowest AND attitudes are Lower than Average
AND condence is Lowest AND Intentions are Low THEN
class = Foggy set.
There are four membership functions for four determinants and
each has High, Average and Low curves in it. The mathematical
formulation (shown in Fig. 2) of mf1 is for Low, mf2 for Average
and mf3 for High. The same function is used for all four determi-
nants. Fig. 3 is the example for cognitions; it is the same for attitude,
condence or purchase intention and Y-axis represents degree of
membership to a High, Average and Low fuzzy set. The Gaussian
shape has been shown to be the most adequate choice of membership
function (Kreinovich, Quintana, & Reznik, 1992).
284 S.N. Ahmad, M.-O. Richard / Journal of Business Research 67 (2014) 278287
Continuing with this example, consider the following estimated
membership values which were obtained by using fuzzy classication
rules: Consideration = 0.791 (from Rule 1), Hold = 0.766 (from
Rule 2), Foggy = 0.026 (from Rule 3), and Reject = 0.042 (from Rule 4).
Step 3 The fuzzy classication concludes that brand 1 is in the consid-
eration set in respondent's 1 mind since its membership value
is the highest (0.791). This is a correct classication since the re-
spondent mentioned that brand 1 was in the consideration set
(Step 1). Otherwise it is considered as a misclassication.
Step 4 Next, the largest value (0.791) is compared with the second
largest (0.766) to see if the difference is very small. These two
values are very close. Thus the brand belongs to the consider-
ation set and the hold set in respondent 1's mind with almost
the same strengths.
Step 5 This procedure is repeated for every participant who expressed
that brand 1 is in one of the four sets. The whole procedure is
repeated for the remaining brands.
Researchers generally designate two degrees of memberships as
very close when these values are apart by 0.1 or less (Pendharkar,
2003). We counted the number of cases where these membership
values (two highest) are apart by 0.03, 0.05, 0.10 and 0.15 (Fig. 4).
The rst portion of each plot is almost horizontal meaning that the
number of fuzzy cases remains almost unchanged in this range. To per-
form a conservative test, the rest of the calculations uses a 0.05 differ-
ence for statistical testing.
3.3. Results and discussion
Following this procedure, the whole awareness set was modeled as
a fuzzy rule based classication system (Table 6). The second column
represents the number of respondents who were aware of the brand.
The hit rate ranges from 64.2% for Pizza Hut to 97.7% for KFC, which is
high. The classication accuracy of this FRB is comparable and even
sometimes slightly better than that of SVM. Thus, the fuzzy rule based
classication system can model the data well and is a valid model of
brand categorization.
Among these correctly classied cases of KFC consumers, 3 cases
showed close membership values for another set yielding 1.1% of deci-
sion fuzziness. McDonald's and Pizza Hut have 1.4% and 3.7% decision
fuzziness respectively. Decision fuzziness is very small for western
brands; however the percentage becomes larger for the local brands. A
two-tailed test nds a signicant difference in decision fuzziness
Fig. 4. Number of fuzzy cases (line represents brands; red, green, blue and yellow lines represent China, Japan FF, Japan beer, and Kazakhstan market, respectively.
between these two types of QSR brands (z = 9.64, p b 0.01). As
expected, in the Chinese market western brands have a lower decision
fuzziness since these brands have very few negative attributes compared
to their positive attributes and thus have less decision conict. Overall,
H1 and H2 are supported.
Using the same procedure, we compare these results with that of the
Japanese QSR market (Table 6). The hit rate is excellent (the average is
around 70%). As expected, western brands have higher decision fuzzi-
ness than the local ones. In this market, MOS Burger has the lowest de-
cision fuzziness. McDonald's and Matsuya have similar fuzziness and
nally KFC, Subway and Lotteria have higher percentages than the
other brands. Although the differences in decision fuzziness between
foreign and local brands are in the hypothesized direction, it is not sta-
tistically signicant. The differences in fuzziness between MOS Burger
and the foreign brands are statistically signicant (z = 3.15,
p b 0.01). Thus, H3a is partially supported. Lotteria is a copied version
of McDonald's and is perceived more like McDonald's with a somewhat
lower quality. Therefore, it has high decision fuzziness and contributed
towards the overall result. The result for Lotteria supports our reasoning
that when there are both perceived positive and negative attributes
(domestic brand, copied version of foreign brand with lower quality),
decision fuzziness occurs. MOS Burger is a Japanese brand with local
taste and excellent quality. Therefore, consumers have little conicting
inuence on the decision process and have low decision fuzziness.
However, as a whole the difference between local and foreign brands'
decision fuzziness was not statistically signicant.
For more insight about the fuzzy cases, we compare the purchase in-
tentions of fuzzy and non-fuzzy cases in this market. The rst group of
cases includes the cases where McDonald's is in consideration set and
the second group consists of cases where McDonald's belongs to both
the consideration and hold sets. Third, fourth and fth groups are for
hold set (non-fuzzy cases), hold and reject set (fuzzy cases), and reject
set (non-fuzzy cases) only. Fig. 5 shows that the purchase intentions of
the consideration set (non-fuzzy cases) are signicantly higher than
that of fuzzy cases. The reverse is true for the reject set (non-fuzzy
cases) and the hold and reject sets (fuzzy cases). It is important to
note that the fuzzy cases were classied by the FRB system only and
were not stated by consumers. Thus FRB actually models the awareness
set better and represents the hidden inclinations of consumers.
In summary, in the Chinese market, the foreign brands have lower
decision fuzziness than the local ones and in general, the reverse in
the Japan market. However, one of the local brands Lotteria had
high decision fuzziness. Therefore, to get a clearer picture in this
 S.N. Ahmad, M.-O. Richard / Journal of Business Research 67 (2014) 278287 285

Table 6
Results from Chinese, Japanese and Kazakhstan QSRs and beer markets.

Country/category Brand names Number of Correct Fuzzy pattern among correctly 95% Condence Correct classication
respondent classication classied (0.05 differences) interval (SVM)
(FRB)

China/fast food KFC 266 260 (97.7%) 3 (1.1%) (0%2.45%) 255 (95.8%)
McDonald's 250 244 (97.6%) 2 (1.4%) (0%3.3%) 244 (97.6%)
Pizza Hut 212 136 (64.2%) 5 (3.7%) (0.5%6.8%) Used to train SVM
Ma Lan Hand Pulled Noodles 180 120 (66.7%) 25 (20.1%) (13.5%28.1%) Used to train SVM
Beijing Quan Du Roast Duck 116 84 (72.4%) 9 (10.7%) (5.1%16.3%) 82 (70.7%)
Tian Jian Steam Stuffed Buns 155 104 (66.7%) 25 (24.0%) (15.8%32.2%) 102 (65.6%)
Japan/fast food KFC 212 141 (66.5%) 18 (12.7%) (7.3%18.3%) 139 (65.6%)
McDonald's 266 215 (80.8%) 22 (10.2%) (6.2%14.3%) 226 (84.9%)
Subway 224 118 (53%) 16 (13.6%) (7.4%19.7%) Used to train SVM
MOS Burger 244 191 (78.3%) 9 (4.7%) (1.7%7.7%) 200 (81.9%)
Matsuya 219 150 (68.5%) 14 (9.3%) (4.7%13.9%) 152 (69.4%)
Lotteria 251 141 (56.2%) 26 (18.4%) (12.1%24.9%) Used to train SVM
Japan/beer Budweiser 230 98 (42.6%) 22 (22.5%) (14.2%30.7%) 122 (53.0%)
Heineken 207 76 (32.7%) 15 (19.8%) (10.8%28.7%) Used to train SVM
Kirin Lager 208 121 (58.2%) 19 (15.7%) (9.2%22.2%) Used to train SVM
Kirin Ichiban Shibori 233 132 (56.7%) 14 (10.6%) (5.4%15.9%) 152 (65.2%)
Asahi Super Dry 231 133 (57.6%) 8 (6.02%) (1.9%10.0%) 156 (67.5%)
Kazakhstan/beer Bavaria 101 65 (64.5%) 7 (10.7%) (3.2%18.3%) 59 (58.4%)
Zhigulevskoe 116 83 (71.6%) 27 (32.5%) (22.5%42.6%) Used to train SVM
Tian Shan 105 77 (73.3%) 10 (12.3%) (5.5%20.5%) 73 (69.5%)
Karagandinskoe 81 48 (59.3%) 16 (19.7%) (11.1%28.4%) 47 (58.0%)
Irbis 80 51 (63.8%) 2 (3.9%) (0%9.25%) Used to train SVM
Baltika 98 74 (75.5%) 2 (2.7%) (0%6.4%) 69 (70.4%)

market we examine another food category, beer, where restaurant
quality and service are irrelevant in decision making.
4. Study 2
We examine the Japanese beer market and compare it to Japan's
QSR market.
4.1. Data and methodology
Replicating the previous study, 291 consumers were surveyed. The
sample demographics are shown in Table 2b (69.8% male and 30.2% fe-
male aged 1829). The brand proles of foreign and local brands are
provided in Table 3a.
brewers need to make a unique beer to compete with the local ones.
Therefore, these brands are showing high decision fuzziness. This result
supports our proposition.
5. Study 3
We examine and compare the decision fuzziness of the Kazakhstan
and Japan beer markets.
5.1. Data and methodology
179 consumers were surveyed following Study 1. The sample de-
mographics are provided in Table 2c (41.9% male and 58.1% female
aged 1859). The brand proles are shown in Table 3a and Table 3c.

4.2. Results and discussion 5.2. Result and discussion

In the Japanese beer market, the local brands create much less deci-
sion fuzziness than the foreign brands (z = 2.85, p b 0.01). Here again,
the hit rate is good (Table 6). H3b is supported. We found that
Budweiser and Heineken have high decision fuzziness. All local brands
have low decision fuzziness. The Japanese are very loyal to their nation-
al brands. Japan produces excellent beers and thus this has always been
a challenge for the foreign brands. Therefore, the local beer brands have
less decision fuzziness. Although Budweiser and Heineken have good
quality as evidenced by the fact that they survived in this tough market,
they are foreign brands. Knight (1996) pointed out that in Japan, foreign
Fig. 5. Purchase intention of fuzzy and non-fuzzy cases.
Unlike Japan, little evidence of the preferences of the Kazakh people
is available. Kazakhstan has a wide variety of beers and most of the for-
eign brewers in Kazakhstan have subsidiaries. The brands we classify as
local are mostly produced with foreign technology and adapted to
the local taste (Tian Shan is produced with Dutch technology,
Karagandinskoe is produced by Efes, a Turkish brewer). Therefore, it is
not clear if consumers would differentiate these beers in terms of local
versus foreign brands. Moreover, although Baltika is classied as a for-
eign beer in Kazakh market, it is based in Russia to which Kazakhstan
is historically related. Thus, these brewers are not exactly local or
foreign to them.
The results (Table 6) show that the hit rate is excellent (ranging
from 63.8% to 75.5%). Baltika and Irbis have the least decision fuzziness.
Irbis is a niche product positioned as a high quality brand. Zhigulevskoe
and Karagandinskoe have higher decision fuzziness. Both of these
brands are positioned in the economy segment and are produced by
Efes Breweries International. Tian Shan is a Heineken product produced
with Dutch technology and is also positioned in the economy segment.
Bavaria is a premium beer from The Netherlands. Therefore, in this
market, the brands which claim to have premium quality have lower
decision fuzziness. H4 is not fully supported. Contrary to expectations,
foreign brand Baltika has low decision fuzziness. This premium beer
from Russia has been around long before Kazakhstan earned its
286 S.N. Ahmad, M.-O. Richard / Journal of Business Research 67 (2014) 278287
Table 7
Results of coupon in the Chinese market.
Belongs to Moves to
Consideration Consideration & Hold Hold &
Hold Reject
Consideration 37
Consideration & Hold 6 1 1 of sets provided by respondents. This information can be very useful to
Hold 9 9 8 managers and this methodology has an enormous potential to be used
Hold & Reject 1 1 2 in other contexts where t between an object and class is required.
Reject 2 From a managerial perspective, the ndings are important. Our re-
sovereignty and has adapted to the local taste. Consequently,
consumers' perception of its foreignness is not clear and might have
driven the results. Because of foreign subsidiaries, other brands may
not be purely local. Therefore, this interplay might have caused the
higher degree of decision fuzziness of the locally produced brands.
In summary, our results show that in the Kazakh beer market,
brands which have adapted to local taste, have high quality and are
attached to Kazakhstan nationalism have lower decision fuzziness
than the brands positioned in the economic segment.
6. Effects of couponing in the Chinese QSR market
We examined the effects of couponing on decision fuzziness in the
Chinese market. We follow the same procedure and found ve patterns
in the data (Table 7). Among the 77 cases, 28 (36.4%) move in a more
positive direction. There are cases where some brands moved from
the hold set to the consideration and hold sets and some brands from
the hold and reject sets to the hold set only. So, in the rst scenario
the fuzziness decreased, whereas it increased in the second. H5 receives
support. This is an important result from a manager's perspective as the
decision fuzziness can be changed by applying appropriate incentives.
7. Conclusions and managerial implications
The BrisouxLaroche brand categorization model divides con-
sumers' awareness set into four sets with a presumption that each
brand will be included in one of these four sets. However, we nd
evidence that a brand may belong simultaneously to more than one
set in consumers' minds, especially when a brand has both positive
and negative attributes associated to it. Consumers may not absolute-
ly consider or reject that brand. We investigate this phenomenon
across three cultures (China, Japan and Kazakhstan), with two types
of products (QSR and beer), including foreign and local brands.
The contributions are twofold. First, by using fuzzy theory which al-
lows the brands to have (if it exists) a degree of membership in each set
simultaneously, we found empirical evidence that a brand may exist in
more than one set of awareness. This phenomenon in the whole aware-
ness enables us to better understand the brand selection process. This
research contributes to the literature by demonstrating a unique phe-
nomenon hidden in consumers' minds. We also found evidence of vary-
ing degrees of decision fuzziness across products and cultures. Chinese
consumers have low decision fuzziness for foreign brands as opposed to
the local ones, and vice versa for the Japanese. The ndings are revealing
and have great managerial importance. Fuzzy cases have signicantly
different purchase intentions than the non-fuzzy cases. Therefore,
these cases are able to capture the hidden inclinations in consumers'
minds beyond the self-stated measure. Also, the application of coupon-
ing may change decision fuzziness.
Secondly, we contribute methodologically as follows: the classica-
tion accuracy is comparable to well-known SVM. Moreover, using FRB,
we can see consumers' hidden inclinations. For example, the degree of
membership and probabilities (from SVM) for four sets for a specic re-
spondent are: SVM: C = 0.0401, H = 0.1188, F = 0.2955, R = 0.5456;
FRB: C = 0.0126, H = 0.5892, F = 0, R = 0.6016. It shows that
although both methods correctly classied the case (reject), fuzzy clas-
sication picked up the hidden inclination of considering the brand with
a strength of 0.58 (very close to 0.60) true value to the hold set. The same
cannot be argued for probabilities. This inclination towards considering
Reject
(or rejecting) the brand hidden in the measures can only be known
through fuzzy classication, and it goes beyond the self-stated measure
search shows that a managerially actionable task, couponing, may
change decision fuzziness. If a brand belongs to both the reject and
the hold sets, the manager has to take action by eliminating some of
the negative attributes. These consumers are more likely to buy that
brand in the future than those who only rejected the brand. More-
over, if a brand belongs to the foggy and the reject sets, managers
should increase their advertising efforts since it is more likely that
the brand is being rejected because of lack of knowledge about the
brand.
8. Limitations and future research
Like most studies, this one has limitations. This research does not
compare collectivist countries with individualistic ones. Based on the
literature review, it can be inferred that individualistic countries will
show less decision fuzziness towards foreign brands. A direct test
might have been a stronger test of our proposition. Future research
may explore this aspect.
Acknowledgments
The authors gratefully acknowledge the nancial support of the
Fonds Qubcois de la recherche sur la socit et la culture.
References
Balabanis, G., & Diamantopoulos, A. (2004). Domestic country bias, country-of-origin
effects, and consumer ethnocentrism: A multidimensional unfolding approach.
Journal of the Academy of Marketing Science, 32, 8095.
Brisoux, J. E., & Laroche, M. (1980). A proposed consumer strategy of simplication for cat-
egorizing brands. In John H. Summey, & Ronald D. Taylor (Eds.), Evolving Marketing
Thought for 1980 (pp. 112114). Carbondale, IL: Southern Marketing Association.
Chatterjee, S., & Heath, T. (1996). Conict and loss aversion in multi-attribute choice:
The effects of trade-off size and reference dependence on decision difculty.
Organization Behavior and Human Decision Processes, 67(2), 144155.
Cui, D., & Curry, D. (2005). Prediction in marketing using the support vector machine.
Marketing Science, 24(4), 595615.
Dhar, R. (1997). Context and task effects on choice deferral. Marketing Letters, 8(1),
119130.
Dong, L., & Tian, K. (2009). The use of Western brands in asserting Chinese national
identity. Journal of Consumer Research, 36(3), 504523.
Gao, G. Y., Pan, Y., Tse, D. K., & Yim, C. (2006). Market share performance of foreign and
domestic brands in China. Journal of International Marketing, 14(2), 3251.
Gurhan-Canli, Z., & Maheswaran, D. (2000). Determinants of country-of-origin evaluations.
Journal of Consumer Research, 27(1), 96108.
Hellmann, M., & Jager, G. (2002). Fuzzy rule based classication of polarimetric SAR
data. Aerospace Science and Technology, 6, 217232.
Hofstede, G. (1991). Cultures and Organizations: Software of the Mind. New York: McGrawHill.
Howard, J. A. (1963). Consumer Behavior: Application of Theory. New York: McGraw-Hill.
Hsieh, M. (2004). An investigation of country-of-origin effects using correspondence
analysis: A cross-national context. International Journal of Market Research, 46(3),
267295.
Hsu, J., & Nien, H. (2008). Who are ethnocentric? Examining consumer ethnocentrism
in Chinese societies. Journal of Consumer Behavior, 7(1), 436447.
Jun, J. (2006). Central Asia: Culture shift sends beer popularity soaring. http://www.
rferl.org/content/article/1067941.html (Retrieved from).
Knight, G. (1996). Educator insights: International marketing blunders by American
rms in JapanSome lessons for management. Journal of International Marketing,
3(4), 107129.
Kreinovich, V., Quintana, C., & Reznik, L. (1992). Gaussian membership functions are
most adequate in representing uncertainty in measurement. Proceedings of NAFIPS:
North American Fuzzy Information Processing Society Conference, Puerto Vallarta.
Laroche, M., & Parsa, H. (2000). Brand management in hospitality: An empirical test of
the BrisouxLaroche model. Journal of Hospitality and Tourism Research, 24(2),
199222.
 S.N. Ahmad, M.-O. Richard / Journal of Business Research 67 (2014) 278287
Laroche, M., Takahashi, I., Kalamas, M., & Teng, L. (2005). Modeling the selection of
fast-food franchises among Japanese consumers. Journal of Business Research,
58(8), 11211131.
Lee, D. (2006). Popularity of Japanese convenience food (fast food). http://uniorb.com/
ATREND/Japanwatch/fastfood.htm (Retrieved June 19, 2010, from UNIROB Asian
trend: website).
Lewis-Beck, M. S., Bryman, A., & Liao, T. F. (2004). The Sage Encyclopedia of Social Science
Research Methods. Thousand Oaks, CA: Sage publication.
Nagashima, A. (1970). A comparison of Japanese and US attitudes towards foreign
products. Journal of Marketing, 34, 6874.
Neuliep, J., Chaudoir, M., & McKroskey, J. (2001). A cross cultural comparison of ethnocen-
trism among Japanese and United States college students. Communication Research
Reports, 18(2), 137146.
Pendharkar, P. C. (2003). Characterization of aggregate fuzzy membership functions using
Saaty's eigenvalue approach. Computers and Operations Research, 30, 199212.
Pitiranggon, P., Benjathepanun, N., Banditvilai, S., & Boonjing, V. (2010). Fuzzy rules gen-
eration and extraction from support vector machine based on kernel function ring
signals. International Journal of Engineering and Applied Sciences, 6(4), 244251.
Rybina, L., Reardon, J., & Humphrey, J. (2010). Patriotism, cosmopolitanism, consumer
ethnocentrism and purchase behavior in Kazakhstan. Organizations and Markets in
Emerging Economies, 1(2), 92107.
Schmitt, B., & Simonson, A. (1997). Marketing Aesthetics: Title Strategic Management of
Brands, Identity, and Image. New York: The Free Press.
Schuiling, I., & Kapferer, J. (2004). Executive insights: Real Differences between local and
international brands: Strategic implications for international marketers. Journal of
International Marketing, 12(4), 97112.
287
Shimp, T., & Sharma, S. (1987). Consumer ethnocentrism: Construction and validation
of the CETSCALE. Journal of Marketing Research, 24(1), 280289.
Teng, L. (2009). A comparison of two types of price discounts in shifting consumer's
attitudes and purchase intentions. Journal of Business Research, 62(1), 1421.
Vapnik, V. (1998). Statistical Learning Theory. New York: Willey.
Viswanathan, M., & Childers, T. L. (1999). Understanding how product attributes inuence
product categorization: Development and validation of fuzzy set based measures of
gradeness in product categories. Journal of Marketing Research, 36(1), 7594.
Wang, J. (2000). Foreign Advertising in China: Becoming Global, Becoming Local. Ames:
Iowa University Press.
Wang, J., & Wang, Z. (2007). The political symbolism of business: Exploring consumer
nationalism and its implications for corporate reputation management in China.
Journal of Communication Management, 11(2), 134139.
Watson, J., & Wright, K. (2000). Consumer ethnocentrism and attitudes toward domestic
and foreign products. European Journal of Marketing, 34(9&10), 11491166.
Wright, L. T., Nancarrow, C., & Kwok, P. (2001). Food taste preferences and cultural
inuences on consumption. British Food Journal, 103(5), 348357.
Yoo, B., & Donthu, N. (2005). The effect of personal cultural orientation on consumer
ethnocentrism: Evaluations and behaviors of U.S. consumers toward Japanese
products. Journal of International Consumer Marketing, 18(1/2), 744.
Yoon, H. J., Thompson, S. S., & Parsa, H. G. (2009). Bayesian approach to assess
consumer's brand selection process and identication of brand attributes in a
service context. International Journal of Hospitality Management, 28(1), 3341.
Zadeh, L. (1965). Fuzzy sets. Information and Control, 8(3), 338353.
Zhou, L., & Hui, M. (2003). Symbolic value of foreign products in the People's Republic
of China. Journal of International Marketing, 11(2), 3658.