Examining the Association Between Personal and Cultural Values and Cross-National Customer Satisfaction

Although most marketing scholars agree that customer satisfaction is a critical determinant of post purchase attitude and product choice, research has remained largely confined to the US and few Western European countries. Given the increasing size of markets in developing countries, and that customer satisfaction is at the heart of all marketing activities, international marketers should have a deeper understanding of the determinants of satisfaction responses of the international customer. A review of present satisfaction literature reveals the domination of the expectation~disconfirmation paradigm. Proponents of this paradigm suggest that satisfaction following a product experience depends on the magnitude and direction of the perceived disparity between actual attribute performance and prior expectations. While much empirical evidence supports this linkage, relatively little is known about the extent to which elements of the cultural environment influence the satisfaction formation process especially in non-Western cultures. Research findings show that consumer values and performance expectations are associated. Building on these findings, the next research question becomes: "To what extent do personal and cultural values influence the formation of satisfaction judgments? " This study investigates the extent to which cultural values and personal values are associated, and whether these values systems relate to benefits sought in products, expectations of attribute performance, and the formation of overall satisfaction judgments. To investigate these relationships, a conceptual framework is advanced that explains satisfaction both from a cognitive perspective via disconfirmation, and from a cultural perspective through consumer values. This framework is then formalized and tested in two countries with different national cultures, namely Kuwait and the US. Six hypotheses, derived from logical reasoning and from previous literature, were tested by data collected from a computer notebook study involving participants from each country. Results suggest the presence of a linear relationship between cultural values and consumer personal values, and between consumer values and benefits sought in notebook computer. The results also show a better fit of the valued-benefit performance congruency model (VB-P) in explaining satisfaction responses of the Kuwaitis when compared to the disconfirmed-expectancy model. Both models perform equally well in explaining the satisfaction responses of the US customers. ' The conclusions of the study should be of a practical value to marketers seeking to sell their products in foreign markets and striving to maintain a high level of customer satisfaction relative to the competition. This goal can not be effectively obtained in the absence of a clear understanding of factors contributing to the generation of high satisfaction, and the particular process by which customer satisfaction evolves following a product consumption experience. Contrary to propositions of the disconfirmedexpectancy paradigm, it was shown that Kuwaiti consumers are more concerned with the kind of benefits the product is providing than what manufacturers are promising in terms of technical product attributes. The relative importance of these benefits was found to be different from one culture to another, depending on the type of dominant values held by the consumer. It is logical then to argue that value and benefit mapping should precede the introduction of products in every international market, and that competing solely on the basis of quality of technical attributes, which might be irrelevant to customer wants and preferences, could be the wrong approach in today' s highly competitive markets.


'
The conclusions of the study should be of a practical value to marketers seeking to sell their products in foreign markets and striving to maintain a high level of customer satisfaction relative to the competition. This goal can not be effectively obtained in the absence of a clear understanding of factors contributing to the generation of high satisfaction, and the particular process by which customer satisfaction evolves following a product consumption experience. Contrary to propositions of the disconfirmedexpectancy paradigm, it was shown that Kuwaiti consumers are more concerned with the kind of benefits the product is providing than what manufacturers are promising in terms of technical product attributes. The relative importance of these benefits was found to be different from one culture to another, depending on the type of dominant values held by the consumer. It is logical then to argue that value and benefit mapping should precede the introduction of products in every international market, and that competing solely on the basis of quality of technical attributes, which might be irrelevant to customer wants and preferences, could be the wrong approach in today' s highly competitive markets.

AKNOWLEDGEMENT
Th completion of this doctoral dissertation marks the end of doctoral program that started a bit over four years ago. While I am happy to complete this ambitious project, I feel that many parts of this work would have never been accomplished without the support and the help of many great individuals.
The starting point is to thank the Lord for providing me with such wonderful parents. Not only they brought me to life and took care of me when I couldn't sustain on my own, but they also planted in me that nothing is impossible with determination and ' hard work. It is very sad, though, that my dearest father did not live long enough to celebrate this achievement, may his soul rest in peace. But I am glad and thankful that my mother is still alive and has already shared my success with her tears of joy.
But closer than anybody else, my beloved wife was the best companion during this long endeavor. Bringing up two beautiful children and staying next to me away from her family and loved ones is indeed very graceful. Though words can hardly describe my great indebtedness, the next best thing is to say: thank you Hana for your genuine love and massive support! While the former have provided me with the emotional and moral support that I needed at home, this work was in part the result of association with many brilliant professionals in the Marketing Department at the University of Rhode Island.
Perhaps the main contributor to this work has been my firm and friendly mentor, Dr. Carol Surprenant. During the course of my program, she provided me with thoughtful advice, constructive comments, and used her deep knowledge of the vast IV literature of satisfaction and its related topics to guide and improve my work. Without her advice, I would have probably ended up with a complex dissertation project that would have taken many more years to complete and millions of dollars to finance. Thank you for believing in me and for supporting my projects throughout the program.
Many other individuals also deserve special thanks for their genuine help. Many thanks go to Dr. Lisa Harlow for her considerable help in teaching me much of the statistics used in this dissertation and for her great amicable teaching style. Thanks to Dr. Ven Venkatesan for being part of this long process and for his guidance on conducting international research. My thanks also go to my additional committee members, Dr. ' Jonathan Schroeder and Dr. Tim Tyrrell, for their insightful comments.
During the program I have met some wonderful people who have later become my good friends. My appreciation goes to Tim Greenlee for his companionship and support primarily in the first year of the program. Many thanks go to Jim Curran for his valuable assistance in running focus groups and student recruitment, and most importantly, for his humor. Thanks to Elizabeth Purinton for her kind help in student recruitment and thoughtful comments on this dissertation manuscript.

STATEMENT OF GOAL AND THE NEED FOR THE STUDY GLOBALIZATION AND CROSS-NATIONAL CUSTOMER SATISFACTION
Although most marketing scholars agree that customer satisfaction is a critical determinant of attitude and product choice, research has remained largely confined to the US and a few Western European countries. Given the increasing size of developing markets, an~ given that customer satisfaction is at the heart of all marketing activities, international marketers should have a deeper understanding of the determinants of the satisfaction responses of the international customer.
Customer satisfaction has been traditionally defined as the emotional outcome of a product use experience. Interest in studying customer satisfaction has grown as businesses strive to create sustainable competitive advantages in increasingly competitive markets. In response to mounting pressures from the consumer side, more marketers prefer utilizing satisfaction-focused strategies over price discounting strategies designed to encourage switching, or forced-loyalty strategies to inhibit exiting (through using, for example, tie-ins products and frequent flyer programs). Embracing the marketing concept as the modem marketing philosophy has also increased the interest in studying the satisfaction construct in consumer behavior. It is not surprising that Kotler (1991) considers customer satisfaction to be the best indicator of future profits.
The desirable effects of maintaining high level of customer satisfaction on postpurchase attitudes and behaviors are documented in the marketing literature. In his 1 review of the literature, Yi ( 1990) cites many articles showing a high level of satisfaction creates positive attitudes toward the brand relative to competing brands, motivates the dissemination of favorable information (e.g., positive word-of-mouth), and leads to a high level of repeat purchasing intentions (i.e., psychological loyalty). As a strategic focus, satisfaction is also a less expensive means to protect present customers from the competition. Dutka (1994) postulated that satisfaction-focused strategies minimize consumer attrition compared to price strategies designed to attract new customers and brand switchers through discounting. This means that by maintaining a stable customer base, markettng costs per customer should be lower since the cost of maintaining existing clients is three to five times cheaper than the cost of attracting new prospects (Lowenstein 1995, p. 2). Anderson and Sullivan (1993) have found empirical evidence linking high satisfaction to high quality perceptions, and hence, to stronger repurchase intentions. In acknowledging this pivotal role, many firms have tied executives' compensation to market measures of customer satisfaction.
Thus, the centrality of customer satisfaction to all marketing activities has made its study a vibrant area of marketing research. Investigation into this topic has been the focus of numerous marketing studies, and has produced a large, useful literature. Most of this literature, however, is derived from studies conducted in the US and few Western countries, a home for less than 6% of the world population. In today' s world economy, the trend toward globalization is likely to continue. International trade has been steadily growing in terms of the number of nations involved and the volume of traded products (estimated at $4.57 trillion of products annually, The World Bank Annual Report 1991 , p 36). This global interdependency not only brought along valuable economic growth, but 2 also new challenges to international marketers. It is unknown whether existing Western literature on satisfaction and its antecedents can be applied to other international markets, or should be considered culture-bound.
Examining the cross-national applicability and validity of present customer satisfaction knowledge, and other marketing concepts and models, requires conducting more cross-cultural research. In this regard, Durvasula et. al (1993), among many others, caJled for more rigorous examination of cross-national applicability of present consumer behavior models and constructs, and the validity of between-construct linkages. The general lack of cross-cultural consumer research is not limited to satisfaction but also ' extends to other business research as well. Adler (1983) warned that the growth of published cultural validation of western business theories has fallen far behind the fast growth of international commerce. It is quite possible that this gap in the international marketing literature might have led many international marketers to be less sensitive to the diverse norms, values, lifestyles, and needs of the international consumer in foreign markets. Consequently, one might expect that an inadequate understanding of the intricacies of international consumers could lead to declining product sales and to diminishing product popularity, even though the product performs well in home markets.
Among the factors contributing to the failure to understand host national cultures has been the general acceptance of the convergence thesis (Heenan 1988). This simply represents the expectation that the fast growth of international travel, telecommunication, and international media should slowly dissolve much of existing cultural differences, paving the way for the world to be dominated by one global culture. It has been argued that success in such a world culture depends entirely on the ability to deliver products that are technologically superior relative to the competition, focusing primarily on the wants and desires of the "typical" customer rather than on the specific preferences of the international customer (Levitt 1983). However, evidence from the existing consumer literature shows that, although people are becoming increasingly exposed to other cultures, their native culture still exerts a significant influence on many aspects of consumption behavior. (See for example, Arnould 1989, Engledow and Thorelli 1974, Maddox, Gronhaug, and May 1978, Holbrook and Schindler 1994, Walendrof and Arnould 1988, Ford, La Tour, and Henthorne 1995, Alden, Hoyer and Lee 1994, Graham, Kim, Lin, and Robinson 1988, Hernandez, Strahle, Garcia, and Sorensen 1991. Even ' within a single country, ethnographic studies have uncovered distinct consumption patterns among key ethnic groups (i.e., Blacks, Italian, W ASPs, and Jews in the US society, Hirschman l 985). This suggests that culture continues to be the driving force behind many consumer attitudes and behaviors, including those pertaining to post purchase satisfaction.
Interest in studying the meaning and the antecedents of cross-national customer satisfaction has been rising. Sustained income growth and greater accessibility to foreign products provide the international customer with the freedom to choose between competing brands with varying features and prices, thereby creating intense competition between global and local brands. Also, given the limited knowledge about the determinants of cross-national satisfaction, a rise in negative consumer sentiment toward the multinational company' s products can easily damage its market share in foreign markets, especially when local manufacturers are competing with imported products on the basis of patriotism. While a great many companies doing business internationally spend large amounts to measure satisfaction in their various national markets, these measures largely consist of simple adaptation/translations of measures developed in Western countries, notably the U. S. Studying cross-national customer satisfaction can be viewed as a natural extension of present satisfaction research. One should view culture not as icons and artifacts, but more as a process, which predisposes people to think and behave in a distinctive manner. This leads to the development of social norms, value systems, as well as cultural schematization of things, objects, and activities. Given that activities are culturally determined and that the same product might convey different meanings in two ' or more cultures, this culture-specific meaning of objects and activities might have a profound impact of what might drive the customer to be satisfied or dissatisfied following a product experience. Investigating these cross-cultural differences should rely on a sound conceptual framework. As many authors have argued (e.g. , Clark 1990), studying cross-cultural differences from a national character perspective can be a very useful approach to studying the pervasive effects of cultures on its members. Hofstede's empirical delineation of the dimensions of national character stands as the most notable example of this line of study, which seems to be a relevant framework of cultural studies.
The following chapter provides a contemporary literature review on: 1) defining national culture in terms of core cultural values; 2) establishing the conceptual linkages between cultural values and personal values; 3) examining existing customer satisfaction literature and the potential effects of values on customer satisfaction; and 4) advancing a number of hypotheses concerning the potential influence of cultural values on the various constructs involved in the satisfaction formation process at the consumer level.

REVIEW OF RELEVANT LITERATURE
There is a growing body of literature relating national culture, as an environmental research variable, to many aspects of consumer behavior. However, only a tiny fraction of this literature has focused on identifying observed differences in satisfaction following a product experience between consumers within two or more to confusion as to how cultural elements fit into the domain of consumer behavior.

NATIONAL CULTURE AND CONSUMER BEHAVIOR
Although there may be many levels at which culture exists and operates (e.g., family, work, or the social class), the highest level is the culture of a nation or a geographical region (Trompenaars 1994, p. 8). National culture has been variously conceptualized in different disciplines, and it seems that the nature of the field and the context of the study determine what this term represents. Anthropologists, for instance, view national culture broadly as the peculiar way of life shared by a group of humans, 6 including elements pertaining to economics and technology, social structure, and ideology (Ronen 1986, p. 19). Marketing and business researchers, on the other hand, view national culture more as a dynamic process affecting a wide range of behaviors, particularly those relating to buying and consumption. Under both conceptualizations, cultural imprints on consumer behavior are profound and can be hardly missed. From native culture, people learn how to feel, think, and act. As Kotler views it, "Culture is the most fundamental determinant of person's wants and behavior." (Kotler 1994, p.174).
Cultural influences on humans are unconscious and enduring. Individuals learn cultural meanings, concepts, and traditions without questioning early in their lives, and these ' elements become then the basis for solving life problems and for understanding the outside world. Core cultural elements, moreover, are relatively stable over time.
Attempts aimed at changing centrally held cultural beliefs and values have usually been met with strong resistance. Through socialization, culture has been successfully transferred from one generation to the next.
While studying cultural influences on consumer behavior is of interest to international marketers, attempts to operationalize this seemingly discernible, yet illusive, concept have been generally unsatisfactory. Henry (1976) mentions that, at his time, almost 164 definitions of cultures emerged since 1952. Consequently, the presence of such numerous definitions has confused the meaning of culture, leading some researchers to overlook linkages between cultural influences and specific consumer behaviors because they consider it too broad to be relevant. To overcome such definitional ambiguity, a more specific delineation of national culture is advanced.

National Culture as a System of Values
National culture can be described as a multilevel concept in which consumer thinking and actions are reflective of a deeper. more abstract level consisting of internal cultural values, beliefs, and norms. It is argued here that conceiving national culture as a set of values is more suitable for consumer research because this operationalization focuses less on actions and more on the roots and mechanisms determining these actions.
Therefore, Hofstede' s ( 1980b) definition of national culture as the "collective mental programming that conditions peoples' values and perceptions" is adopted in this study.
Viewed as conceptions of the desirable, learned from the social and physical ' environment, core cultural values are social patterning that prescribe and proscribe many aspects of consumer behavior. For example, Gatignon and Robertson (1985) reported a cross-national study investigating women's status in different cultures found that female involvement in the work force seems to determine future sales of some products such as dishwashers. Henry (1976) found a strong correlation between consumer cultural values and preference for a particular category of automobiles. In a later study, Ness and Stith (1984) found significant differences between cultural values of mid-class white and black consumers and proposed that designing customized advertising to each group is instrumental to the success of an advertising campaign when targeting a mass market.
Existing consumer motivation research indicates that cultural values can be considered as the broadest set among other value systems (e.g., personal values, family values, workplace values) whereas personal values are considered the most specific. The possible interaction between cultural values and personal values is of primary interest to studies investigating cross-national similarity and differences in consumer behavior.

Research on Consumer Personal Values
"Personal Values" are an organized system of centrally held beliefs based on what individuals consider as important at various stages of their lives. Essentially, human values evolve very early in life, mainly in the first five years or so, and are stored in long-term memory in a hierarchical cognitive structure. Humans acquire values because they serve important cognitive functions. As social adaptation theory suggests, values are another way that humans can become efficient when dealing with their social and physical environments (Kahle 1996). Rather than remembering the details of each experience ~ith a particular object or class of objects, people relieve the stress on their long-term memories by summarizing their experiences in the form of attitudes and values. The key difference between the two concepts, however, is that an attitude is a belief specific to a certain object while values are global beliefs that transcend both objects and object class.
Within consumer literature, consumer values were consistently shown to be a key determinant of consumer behavior (Rokeach 1973;Kahle 1983Kahle , 1986Kahle , and 1996. As Clawson and Vinson (1977, p. 400) put it, " ... Values may prove to be one of the more powerful explanations of, and influences on, consumer behavior." Once formed, these values influence consumer behavior in two major ways. First, values become important goals that need to be attained and fulfilled which then influence consumer consumption patterns and product/brand preferences, especially in situations where individuals must choose from several alternatives. Empirical research has shown that consumers with a high value for excitement preferred outdoor activities involving high level of energy, whereas people who ranked security highly preferred reading and watching more news 9 and exhibited less preference for outdoor activities. Second, values serve as moral standards upon which objects and events can be considered to be "just," "right," "fair," and "appropriate" in life (Posner and Munson 1979). For example, the influence of values on personal judgment is apparent in views expressed about social issues (e.g., prolife and pro-choice debate over abortion), in logic used to reason out arguments (economic benefits of child slavery), and in deciding whether or not to engage in specific behaviors when the situation seems conflicting (Should I gamble even though I consider it to be immoral?).  (Beatty et. al. 1985, Kahle 1983. Post hoc refinement of the scale resulted in accepting eight personal values commonly shared by individuals with different rank order (Table 2. 1 ). 10 The use of values in market segmentation is prevalent in modem marketing management (e.g., Kamakura and Mazzan 1991 ). Target marketing has been quite successful in optimizing business resources by eliminating the focus on the wrong customer. Several marketers use modem value scales such as V ALS II (SRI International) and LOY to identify the best fit between marketing mix variables (product attributes, price positioning, brand image, themes of advertising campaigns, and selected distribution outlets) and the value structure of potential consumers. Of the three value scales (LOY, RVS, and V ALS II), LOY seems the most appropriate for many consumer researchers. Because the V ALS II scale is only available for commercial use, all data ' pertaining to factor loadings and scale reliability are proprietary information.

National Culture and Consumer Personal Values
Because humans are born into, and raised in, an existing culture, researchers posit that personal values are derived from, and modified by, several cultural institutions such as family, work, school, religious institutions, and social organizations (Clawson and Vinson 1977, p. 401 ). Consider, for example, how respect for the elderly is first learned and then reinforced in many Eastern and Middle-Eastern cultures. A child who observes parents and significant others treating older people respectfully tends to behave similarly, especially when that behavior is encouraged by rewards (verbal complements and positive facial gestures) and deviation is curbed with punishment (frowning, deprivation of valued possessions or privileges).
While both are called "values", personal values and cultural values differ on a number of characteristics. One key difference is that personal values represent what is important and desirable to an individual, whereas cultural values are viewed as what society promotes as preferred and useful to the common good. These values has been accepted and carried from one generation to the next. Individuals are expected to abide by cultural values since they are realities that must be respected (i.e., cultural norms), yet personal values are desired modes of behavior and an end-estate of existence that reflect the needs and objectives in life for an individual. Another difference is that, unlike personal values, cultural values generally lack a specific rank order of importance, and are shared by members of a culture with varying intensity.
The hierarchy of consumer personal values in different regions in the U. S. and in a few other countries was investigated in a series of cross-national studies. Khale ( 1986) ' studied the rank-order of personal values possessed by a probability sample composed of 2,000 residents living in nine geographical locations as suggested by the Census Regions in the United States. Findings showed that the ranking of individual values differed significantly from one region to another. In a series of studies conducted on measuring cross-national differences in the ranking of personal values, several authors utilized the LOV scale to measure the ranking of eight personal values in five Western countries, namely Great Britain, Germany, Denmark, Norway, United States, and the former Soviet Union, and in one Eastern country, Japan (cf. Beatty et al 1993, Grunert and Scherhorn 1990, Kahle, Beatty and Homer 1989, Kahle, Poulus, and Sukhdial 1988. As shown in  Conducting cross-cultural and cross-national research has been generally avoided, not only because it requires a much more complex methodology than what indigenous research usually requires, but also due to the general lack of clear research paradigm that would guide the selection of countries and the explication of findings. When countries (as surrogates of cultures) are selected on the basis of a single dimension or even without a priori criteria, observed cross-country similarities and differences may be attributed to a variety of cultural and individual sources that can be extremely difficult to disentangle a posteriori. To avoid these conceptual shortcomings, it has been frequently emphasized that one should select an appropriate framework prior to proceeding with any crossnational research.

13
A particularly useful approach that has gained wide prominence in organizational behavior is studying the impact of culture on individuals' values from a National Character perspective (Clark 1990, Inkeles andLevinson 1969). Briefly defined, national character reflects the adult modal personality that is prevalent across social groups living in one nation. The characteristics of this common personality are the direct outcome of the interplay between the social environment and its members; group members consistently exhibit synchronous patterns of behaviors. Clearly, the aim of the national character is not to generate superficial stereotypes about how groups typically behave, but rather to characterize intramural values and subtle behavior traits. It is believed, ' therefore, that dimensions of national character be of "higher-order abstractions that refer to stable, generalized dispositions or modes of functioning . .. " (Inkeles and Livinson 1969, p. 426).
The literature contains many methods that can be helpful in describing the traits of national character and can be generally classified into two main approaches: culturecentered or a personality-centered (Clark 1990). The culture-centered approach is a deductive framework that starts with critical observation of types of social structures, artifacts, and collective behavior, and identifies common elements that help sketch the modal personality that best represents the culture. Obviously, the underlying rationale behind this approach is that the institutions which make up a culture must have a psychological unity and an individual personality that reflects the national unity.
The personality-centered approach starts with the observation and measurement of representative random samples to generate evaluations and identify common traits of the group. It uses a modal personality to represent a culture where individual personality traits are observed, enumerated, tabulated, and aggregated. Within the personalitycentered approach, actual dimensioning of predominant cultural values may take any of two methods, purely conceptualized dimensions (Kluckhohn and Stradtbeck 1961) or empirical-based dimensions (Hofstede l 980b ).
The first empirical effort to validate the concept of national character was started by Hofstede (1980bHofstede ( , 1983 and has revealed significant new findings. Methodologically, Hofstede surveyed workplace values possessed by 116,000 IBM workers of various occupations in 40 countries in two-wave studies. In the first study, a factor-analysis of the data revealed that nations can be relatively positioned on a bi-polar continuum in ' terms of ( 1) how members perceive their own power relative to others', and how they relate to authority (high/low Power-Distance), (2) the relative tolerance of ambiguity, lack oflife structure, and risk-taking in life (high/low Uncertainty-Avoidance), (3) the degree to which society believes that members consider gender to be a determinant of specific roles and to what extent members endorse gender-specific values (Feminine cultures vs. Masculine cultures), and (4) how strongly members relate and show concern for others (Individualist cultures Vs. Collectivist cultures). A summary of these dimensions appears in Appendix (1).
The validity of these dimensions was demonstrated in two subsequent studies. In a follow-up study that included more countries, Hofstede conducted a factor analysis of worker responses to the same scale and managed to extract the same dimensions thereby lending more support to the validity and reliability of these dimensions. In addition, a subsequent qualitative analysis of data drawn from completely unrelated sources in these nations (e.g., data from surveys of managers values systems training at !MEDE in Lausanne, Switzerland, and via content analysis of children's books) was found to be correlated with one or more of these dimensions (Hofstede 1980, p.11, 22).
Perception of the four dimensions of national character in organizational literature has been generally positive. Many researchers emphasized the logical appeal of these dimensions and their implications on questioning the applicability of theories of international management to foreign companies. For example, Triandis (1982) states: The dimensions identified by Hofstede certainly make sense. One has a "deja vu" feelings about some of them ... One can also bring empirical support for such dimensions from one's own experience. For example, the Power-Distance dimension seems to make sense in terms of such matters as the use'Ofnonreciprocal forms of address (e.g., Usted-Tu) in some countries, in contrast to reciprocal forms (e.g., Du-Du/Site-Site) widely used in other countries. One can also think of the protection of the leader (King or President) in high power distance countries as opposed to the mingling of leaders with population in low power distance countries. Thus, there is something that the dimensions picks on that makes intuitive as well as empirical sense ... [Hofstede study] will stand as one of the major landmarks of cross-cultural research for many years to come. " Triandis (1982, p 87-90).
Despite such enthusiasm, a few researchers have expressed their reservations about the reliability of these dimensions, especially when studies employing these dimensions are conducted at the individual level. For example, Dorfman and Howell (1988) pointed out that Hofstede focused on between-countries similarities and differences rather within-countries variations. It follows that these dimensions have been constructed based on mean scores of values at the country level, therefore, these dimensions represent an "ecological" level of analysis unrelated to individual traits. In response to the doubt in the validity of these findings at the individual level, Hofstede and Bond (1984) conducted a follow-up study on data collected by Ng. et al. (1982) who surveyed student personal values in six Asian and Pacific countries. Their findings show that the importance ofRokeach's 18 terminal and instrumental values corresponds largely 16 to the positioning of each country on the four dimensions of national character.
Another criticism was directed at the validity of the some of the dimensions. Robinson (1983) points out that while the measurement of power distance and uncertainty avoidance would have improved if more items were added, both Individualism and Masculinity dimensions have been operationalized based on items that have shown strong empirical significance rather than solid theoretical relationship.
Specifically, he points out that: [items composing individuality] seems to be a hodgepodge of items, few of which are directly related to individualism/collectivism ... Moreover, labeling the index a "masculinity" index tells us nothing about the specific value ' being measured. One could construct a number of indices based on current Western stereotypical polarizations of the sexes (e.g., emotional/rational, realistic/scientific, extrinsic/intrinsic) and label them "masculinity" index. " Robinson 1983, p. 114).
Taking these concerns into account, it seems that the perceived power distance and relative tolerance of uncertainty are among the best conceptualized and measured dimensions in Hofstedes' cross-cultural study of national character.
In sum, using the concept of national character as theoretical framework for crosscultural studies seems to be an appropriate approach, with some recognized limitations.
Hosfstede' s 1980 multinational study stands as the most notable empirical research that focused on identifying a few dimensions of national character, despite some of the shortcomings in the measurement of collectivism and masculinity. The differences between cultures in terms of power distance and uncertainty avoidance seem to be among the best-measured and more relevant to consumer behavior. The linkages between these cultural tendencies and the kind of values possessed by consumers might also be interesting to consumer researchers.

RESEARCH ON CROSS-NATIONAL CUSTOMER SATISFACTION
Prior to discussing recent developments in cross national satisfaction research, it seems appropriate to start with a formal definition of this concept. Customer satisfaction is the summary psychological state resulting from a product consumption experience, which encompass emotional elements (pleasure/displeasure), attitudinal consequences (positive/negative), and co native tendencies (complaining/promoting). Obviously, a poor product evaluation generates negative feelings of dissatisfaction that may trigger consumer voice, exit, or reduction in present consumption under monopolistic market conditions. A high appraisal of the product, however, improves future buying intention ' and presumably, brand loyalty.
Published research attempting to find differences in cross-national satisfaction and complaining behavior has reached some interesting yet inconsistent results. In a study designed to measure the cultural differences in complaining behavior between US and Puerto Rican consumers, Hernandez et al. (1991) found dissatisfied Puerto Rican consumers to be less likely to complain following products failures as opposed to their American counterparts. In a similar vein, Yaun's (1988) cultural analysis of customer satisfaction concept in China revealed that Chinese consumers don't blame their product when it fails but rather attribute it to their fate and destiny. While the preceding studies found behavioral differences, other comparative studies have found some similarities in postpurchase behaviors among international consumers. Alden, Stayman, and Hoyer (1994) conducted an experiment to examine the effects of the congruency between supplied product information and expectations of product performance on product evaluation strategies for American and Thai consumers. Participants from both countries 18 have shown a similar increase in evaluation efforts as information did not match up with actual attributes, especially under high risk conditions. Recent developments in international marketing literature include two research streams that can be useful in studying cross-national satisfaction for international marketers and for social policy makers. The first stream is focused on improving the calibration of satisfaction data collected in a multinational context. Specifically, some researchers question whether satisfaction ratings of the same product provided by two different international consumers can be meaningfully used for between-country comparisons pefore adjusting for country-specific measurement biases. Agrawal and Desmet (1994) provided theoretical support for constructing the Transnational Index of Customer Satisfaction before companies can use consumers' feedback as a monitoring mechanism for assessing its present competitive standing. Unfortunately, they stopped short of providing the procedures required to construct such an index. Crosby ( 1992) has suggested that using standardized method and scales to collect satisfaction responses in different countries would probably minimize random measurement error. Data must be readjusted using a "Correction Factor" drawn from the baseline or "reference country" (i.e., the headquarters) prior to interpretation or identification of performance gaps. He concluded that companies should be attentive not only to the reliability via standardization, but also to insure validity through comparative scaling techniques.
Another stream of research has emphasized the need to evaluate consumer welfare by measuring overall customer satisfaction feelings toward leading brands within each industry over equal periods of time. The basic objective is to provide legislatures and business decision-makers with an adjunct measure of the economic activities that can supplement the predominantly financial measures of economic performance reported periodically by government agencies . Fornell's (1992) original work on the Swedish Customer Satisfaction Barometer (SCSB) and the American Customer Satisfaction Index (Fornell et al. 1996) allowed for the first time a tracking of annual consumer satisfaction feedback in key industries to measure its improvement and to validate the much conceptualized links between satisfaction and other marketing variables such as homogeneity of market demand, brand loyalty, and profitability.
While each stream has made its unique contribution to enhance construct calibration for managerial and social policy use, a number of research gaps still exist in ' the literature. Obviously, enhancing cross-national satisfaction measures has been a useful diagnostic tool in judging whether or not customers were adequately satisfied following product consumption, yet the ability to identify performance inadequacies as an input for designing effective customer-focused strategies requires more grounded understanding of the primary determinants of satisfaction of the international consumer.
Although much progress was made in delineating the antecedents of satisfaction at the individual level, published satisfaction studies have been exclusively focused, in their own right, on studying how the typical Western consumer views product satisfaction by defining its cognitive antecedents and emotional/conative consequences. While this has clearly advanced understanding of how satisfaction feelings interact with other key consumer behaviors, one could mistakenly generalize these definitions and linkages to the behavior of international consumer before cross-cultural exploratory and theorytesting research is conducted. In the next section, a brief review of Western satisfaction models is presented and evaluated.

Present Customer Satisfaction Models
The pervasiveness of the marketing concept as the main school of thought in academic and practitioner literatures has elevated the interest in studying the antecedents and consequences of customer satisfaction. Modem marketing approach prescribes satisfaction of consumer needs and desires to be the chief goal of modem marketing management. Research evidence has clearly shown that while discount strategies using price cuts to motivate brand switching succeed in increasing market share, manufacturers should strive to maintain the lowest price position in order to prevent price conscious customers from switching to other discounted brands. Strategies emphasizing customer ' satisfaction, on the other hand, were found to be strongly associated with a favorable postpurchase attitude about the brand, increased level of reported intent to repurchase, and strong consumer brand loyalty.
The early stream of research exploring the determinants of satisfaction has focused almost exclusively on identifying demographic/psychographic variables that describe highly satisfied customers as opposed to customers who are least satisfied.
Results from these studies show that highly satisfied customers tend to be older, less educated, have more satisfaction with life in general while individuals with high income are less likely to be satisfied. While these studies might have identified some possible covariates, this stream does little to advance a fuller conceptual understanding of the satisfaction process.
To address more fundamental questions regarding the source of pre-consumption quality beliefs, the nature of the performance evaluation process and the set of intervening variables that moderate or enhance satisfaction, two theoretical structures 21 have been suggested in the consumer literature. Each structure has taken a different approach in defining the key constructs involved and the relationships among these constructs. While both theories seem to agree that satisfaction judgments are the main result of an evaluation process, each has a different basis against which performance evaluations are compared and comprise a different process of making evaluations.

Disconfirmed-Expectancy Paradigm
The primary satisfaction framework that has received strong empirical support is the disconfirmed-expectancy paradigm. According to this framework, satisfaction results from disconfirmation of expectations, which is the gap between expected attribute performance and the realized attribute performance. According to Howard and Sheth (1969) and Oliver (1980), a product performance that exceeds (lags) prepurchase expectations was found to evoke positive (negative) surprise emotions, which then creates an increased (decreased) level of reported satisfaction. The main constructs in this framework are described below.
Expectations of Attribute Performance. Studies attempting to model satisfaction at the individual level have operationalized expectations as a measure of consumer subjective anticipation of how the product will perform on a set of product attributes considered to be relevant. These subjective probabilities, then, become the standards against which post-use perceived attribute performances are evaluated.
To measure the role of expectations on performance evaluations, Cardozo (1965) designed an experiment in which subjects were presented with catalogues that featured high-quality and low-quality pens followed by a request to evaluate actual pens of different quality using catalogue pens as a frame of reference. Since it was found that 22 does objective performance, even though one might expect to find a strong (but not perfect) correlation between the two.
The impact of expectations on the evaluation of perceived performance was tested in many performance evaluation studies (e. g., Anderson 1973;Olson and Dover 1979;Olshavsky and Miller 1972). Some studies have found consumers to bias (assimilate) their performance evaluations toward their initial expectations when their expectations remained within their latitude of acceptance. Larger expectations-performance disparities caused consumers to shift their evaluations away from prior expectations.

Disconfirmation of Expectations.
The theoretical support for including disconfirmation as an immediate antecedent of satisfaction originates from the axioms of Helson's (1964) adaptation level theory. This theory posits that a given object is always evaluated in relation to some chosen standard. Consistent with the traditional view that cognition precedes emotions, Helson suggests that a positive deviation from the standard can create pleasant feelings of surprise, which then enhance the final evaluation. A negative deviation from the chosen frame of reference will lead to disappointment and to a less favorable evaluation.
In the context of product-use experience, the studies that examined the relative impact of product performance on overall satisfaction have reached inconsistent findings.
While Oliver (1980) demonstrated in a structural model of consumer satisfaction with an automobile that disconfirmation actually had an immediate and direct influence on satisfaction judgments, other studies found performance to be the only antecedent of satisfaction with a video disk player (Churchill and Surprenant 1982), phone services (Bolton and Drew 1991), and a compact disk player (Tse and Wilton 1988).

Assessment of Disconfirmed-Expectancy Paradigm
In the satisfaction literature, it is still controversial whether or not disconfirmed expectations and perceived performance are the only determinants of satisfaction responses to product experience. Proponents of the disconfirmed-expectancy model admit that the role of expectations in determining performance evaluation and satisfaction is not as clear as it had been originally suggested (cf . Oliver 1980). Some researchers are skeptical about whether "performance expectations" are in fact the main calibrations used by customers to evaluate postpurchase product performance (e. g., LaTour and Peat 1979, Churchill and Surprenant 1982, Cadotte, Woodruff, and Jenkins 1987 ' Reilly 1983). Miller (1977) argues that performance expectations is a broad concept that includes several types of expectations including ( 1) ideal performance which reflects the desired level of performance, (2) expected (i.e., predictive) performance that is based on the subjective probability of how the product will perform, (3) deserved (i.e., normative) performance, or what customers think they should be getting given the monetary and non-monetary sacrifices expended to acquire the product, and (4) minimum-tolerable, representing the least amount of performance acceptable from the standpoint of the customer. Among these four types of expectations, normative expectations play a stronger role in influencing satisfaction judgments than do predictive expectations, partly because normative expectations are contrived from previous experiences whereas predictive expectations are built around other consumer experiences and product claims supplied by manufacturers and retailers. In their model of satisfaction with restaurant experience, Cadotte, Woodruff, and Jenkins (1987) found that satisfaction arises mainly by a comparison between product normative expectations and perceived performance rather than from predictive expectations alone.
Results from several product evaluation studies have shown that the credibility of the expectation source (experience, word of mouth, and manufacturers' claims, La Tour and Peat 1979) and the magnitude of initial performance expectations (high/low) moderate the direct effects of expectations on satisfaction. For example, Anderson (1973) suggests that a small variance between expectations and performance generates favorable evaluation of attribute performance, whereas large disparities between expectations iilld performance (above some intrinsic threshold called latitude of acceptance) trigger negative product evaluations and diminish satisfaction. Studies investigating whether initial expectations have an independent direct influence on satisfaction concluded that satisfaction judgments were found to be displaced toward expectations. This functional relationship has not been stable across various product classes. Churchill and Surprenant (1982) found an independent direct effect of expectations on satisfaction and an indirect effect of expectations through disconfirmation. Yet, when a non-durable product was used (i.e., plant), performance emerged as the only determinant of overall satisfaction. When examined at the conceptual level, the inconsistent empirical relationship between expectations and satisfaction are expected to occur because performance expectation represents consumers' predictions of attribute performance and consumers are primarily concerned with receiving valued benefits which may not correspond totally to existing product attributes. Hence, there are some situations where a pleasant surprise resulting from a better-than-expected performance may enhance satisfaction responses, while in other situations, positive disconfirmation appears to be totally unrelated to satisfaction. For example, regardless of whether this has been expected or not, mechanical failures while using a product will always create feelings of dissatisfaction (La Tour and Peat 1979), whereas one might expect that a superior performance of a less valued attribute (e.g., low calorie snacks) would not affect the satisfaction response of the indifferent customer (e.g., less health-conscious customers).
Furthermore, studies attempting to validate the propositions of the disconfirmedexpectancy paradigm have operationalized prepurchase expectations as if consumers are always capabl~ of transforming their wants and desires from a general state to an attribute-specific level. While this transformation is relatively easy in products with predominantly experience attributes (i.e., attributes that can be easily evaluated through product use; sweetness of a beverage), the task is complicated by the absence of salient product attributes required to evaluate product performance, particularly when products lacks a strong performance component upon which product quality can be unambiguously evaluated (i.e., credence products). For example, the inherent difficulty of evaluating the quality of a home air filter is much greater than the picture quality assessment of a television set.
Perhaps one of the more convincing arguments against the role of disconfirmedexpectancy on satisfaction and related postpurchase feelings and behaviors stems from the theoretical definition. Disconfirmation is basically the effect of surprise on evaluation, which might have an intense but highly perishable impact on satisfaction. As time passes, the impact of surprise on performance evaluation is less likely to impact attitude or repurchase intentions. This weakening in the linkage between the affective and conative aspects of satisfaction is responsible for the recent suspicion among marketing practitioners about the usefulness of maintaining high satisfaction as opposed to strategies designed to increase consumer retention (Lowenstein 1995).

Valued Benefits as Determinants of Satisfaction
As consumers see it, products represent a bundle of benefits rather than a mere set of attributes. These benefits may or may not be isomorphic to the utilitarian features of the product. Desired benefits are needs that originate both from physiological and psychological/sociological needs. Because most products provide essentially the same level of satistaction of the physiological need, products are actually competing on the basis of their ability to fulfill the psychological/sociological needs of the customer.
Vehicles, for example, are perceived by consumers as not only a medium of transportation with specific mechanical features but also as a display of personal wealth or social status, as an entertainment instrument, and as a symbol about the personal life style, to name a few. If strongly held values play an important role in defining the psychological/sociological needs, it may be reasonable to expect that satisfaction with a chosen product is related to what is valued.
To introduce the notion of values to satisfaction, Westbrook and Reilly (1983) proposed an alternative satisfaction model labeled the Value-Precept Disparity Paradigm (VPDP). This model reconceptualized satisfaction as the difference between what is valued by the customer and what has been experienced during product use. According to the authors, the main contribution is that, unlike the disconfirmation model which is descriptive of the process, the VPDP model offers a more parsimonious explanation of the satisfaction process, and has much theoretical support in the marketing literature. d d th is model seems to be in agreement with the theoretical proposition of Gutman' s In ee ' (lgSZ) Means-End model, which postulates that consumer desires are extracted from high-end values, and that consumers are actively seeking products with attributes that can satisfy these desires. However, when the VPDP model was empirically tested against the disconfirmation paradigm, neither models fit the data properly.
A later study by Parakash (1984) attempted to investigate the relationship between personal values and customer expectation of product performance. Parakash factor-analyzed 18 terminal values and 18 instrumental values possessed by white and black U. S. consumers, and then correlated these factors to consumer normative ' expectations of product attributes (how the product should perform). His findings revealed that consumers basically derive their normative expectations from their personal values, which in tum exert more influence on satisfaction than did either predictive expectations (how the product will perform) or comparative expectations (how well the brand will perform compared to similar brands). These findings are important in light of the fact that normative expectations were found to be more directly related to satisfaction than were predictive expectations.
In a more recent study, Spreng, MacKenzie and Olshavsky ( 1996) used structural equation modeling to examine the relationships between consumer desires, expectations, and perceived performance, and their relative influence on overall satisfaction responses.
In controlled experimental setting, participants were assigned to six experimental groups (high/low expectations, high/low desires, and high/low performance) and were asked to use and evaluate their satisfaction with a Camcorder. Their findings indicated that the congruency between desires and product performance strongly influenced attribute 29 satisfaction (structural path coefficient= .63) as opposed to disconfirmation of expectation (structural path coefficient= .34).
Within the context of identifying the key determinants of cross-national customer satisfaction, incorporating consumer personal values as the main explanatory variable of their satisfaction response to product use experiences appears to be appropriate. There is a strong conceptual linkage between consumer values and their needs, wants, and desires which subsequently determine what is wanted from the product. Furthermore, there is promising empirical evidence that supports the relationship between the hierarchy of consumer personal values and the favorableness of satisfaction responses. In light of the ' fact that the personal values of consumers are expected to be shaped by consumer culture and vary significantly from one nation to another, the extent to which heterogeneity of cross-cultural values impact the cross-national satisfaction with product and services process becomes an interesting research question. This question that had little or no been attention in consumer literature.

SYNOPSIS
In the world economy, the rapid proliferation of international business has increased the need to understand consumer behavior from a cross-national perspective.
As competition increases in international markets, management places more emphasis on protecting their customer base from switching to competitors (i.e., customer retention) by making consumers more happy and satisfied relative to competing products. Although a number of satisfaction models have been accepted by researchers and practitioners, these models generally explain this phenomenon at the individual level, independent of the cultural environment of the consumers. Clearly, applying these models cross-nationally  ( 1980,1983) shows that cultural values differ significantly among nations along his four dimensions of national character.
In the satisfaction literature, some empirical support was found for the impact of satisfaction, but two of his dimensions, namely power distance and uncertainty avoidance, seem to be more directly related to normative expectations, product evaluations, and satisfaction responses. Furthermore, it was argued that the specific meaning of individualism and masculinity seems to be culturally dependent, rendering scientific measurement controversial. It may also be true that cultures differ along many dimensions other than what Hofstede has studied and delineated, however, including all possible dimensions will undoubtedly jeopardize the scientific parsimony of study.
The second purpose is to compare the relative ability of the disconfirmation paradigm and the value-precept-disparity paradigm to explain observable variations in cross-national satisfaction judgments. The disconfirmation paradigm represents an individual model of satisfaction that excludes cultural values as a key construct, whereas the valued-benefits perfom1ance congruency paradigm explicitly includes consumer values as a determinant of satisfaction judgment. A path model representing the major constructs and their relationships appears in Figure 1. A summary of all hypotheses of the study is provided in table 3 .2 at the end of this chapter.  power and status enjoy advantages such as social respect, greater access to economic opportunities, and an ability to control others' behavior. Over time, a social equilibrium process occurs in which individuals with low power strive to eliminate the gap in status by employing one or more power expanding strategies. These strategies include enhancement of physical strengths (i.e., athletic), intellectual abilities (e.g., education, skills), accum\llation of wealth (i.e., saving, investment), and possession of statusenhancing products. While each of the preceding strategies increases perceived social power, the last method is particularly popular since power enhancing effects are almost instantaneous compared to other methods. To counterbalance the effect of newly gained power by weaker individuals, high-power individuals will either increase their power or reduce another persons' attempt to gain power.
To test the relative effect of the variability of culture-wide power distance on the ordering of personal values within each culture, a correlation analysis was performed between scores given to different countries on the power distance index as measured by According to the definition of personal values included in LOY, individuals who attach high importance to self-respect tend to maintain inner peace with themselves, like to be accepted by others the way they are, and generally feel more confident in themselves and their abilities irrespective of other's opinions. In comparison, selffulfillers favor consuming self-enhancing products and, as accomplishers, they prefer symbols of achievement and success (Kahle 1996). Taking these preliminary findings into consideration, it is reasonable to expect that: HI: In general, consumers living in nations with large social power distance tend to value products that provide benefits which enhance status and social image more than consumers living in nations with low power distance. • RankinUSS g of eight personal values m Great Britain, Germany, Derunark, Norway, United States, former R, and Japan. • Significant at the a. < .05 l!Jlcertaintv Avoidance. As the label of this dimension suggests, uncertainty avoidance reflects the degree to which members of a nation tolerate role ambiguity and lack of strocture in life. This can be seen as a relative measure of fear of what the future might bring, a concept Hofstede labeled "National Anxiety". Attempting to cope with future uncertainties produced several strategies that differ in terms of scope. On the individual level, guarding against the future requires continually enhancing personal skills, talents, and abilities. On the social level, individuals will try to reduce future uncertainties by pursuing strong interpersonal relationships with others and firm interfamilial ties.
Results from the preceding preliminary correlation analysis between country scores on Uncertainty-Avoidance Index (UAI) and personal values are promising. Of the eight values, self-fulfillment and the need for warm relationships were positively associated with UAI though one of these correlations was not significant due to low power (r = .95, p < .01 level; and r = .8, p < .07 respectively). While uncertainty-minimization strategies can be effective, they tend to produce negative consequences to individuals. Striving to enhance one' s abilities will more likely create personal dissatisfaction. This theoretical relationship is supported by significant negative correlation between self-respect and uncertainty avoidance (r = -.86, p <. 05). Hence: H2a: In general, consumers living in nations with low tolerance for uncertainty (high uncertainty avoidance) place more importance on personal abilities and interpersonal relationships than do individuals living in low uncertainty avoidance nations.
In the context of cross-national satisfaction, it was found that cultures do indeed differ in terms of their tolerance for future ambiguity. Within cultures, intolerance for risk can take one of many forms, including reducing risk in purchasing. In general, a major reason why consumers prefer known brand names products to generic products is the level of confidence in the quality of performance. Well-known brands establish considerable trust and awareness with extensive promotional campaigns and standing behind their claims of product quality. Taking this linkage into consideration, it is expected that: H 2 b: Members of cultures with high uncertainty avoidance generally prefer familiar brands with known levels of attribute quality to unfamiliar brands.

Personal Values and Product Satisfaction
The other goal of this study is to examine the relationship between cross-national consumer valties and customer satisfaction with products. Basically, consumers have needs that can be fulfilled by many alternatives available in the market, each of which has attributes that provide different benefits. After evaluating the quality of relevant attributes for each alternative, consumers choose the brand that is more likely to produce those desired consequences and that is least likely to lead to unwanted consequences.
Satisfaction, then, becomes an emotional response that results from a comparison process between what consumers consider desirable consequences and actual postpurchase attribute performance.
Desired consequences, according to Gutman (1982), can be categorized based on the need fulfilled. These needs may be physiological, geared toward maintaining the survival of the person (e.g., hunger, thirst, and need for shelter and medicine). These needs could be psychological, satisfying innate emotional and cognitive needs of the individual (e.g., maintaining a high level of respect, motivation and strong selfconfidence, and adopting healthy lifestyles), or these needs could be sociological, reflecting how an individual would like others to react to the product and to the self (e.g., the need to belong, association with a larger group, achieving social respect). While most researchers agree that there are few variations among individuals in terms of the number and hierarchy of physiological needs, they admit that consumers differ widely in terms of their psychological and sociological needs, which are partly determined by their cultural settings and personal values.

Valued-Benefit Performance {VB-P) Congruency Model o(Satisfaction.
Unlike the prediction of the disconfirmation paradigm which portrays satisfaction as a function of expectation-minus-performance, the (VB-P congruency) paradigm conceptualizes satisfaction as the level of congruency between valued benefits and the perceived quality of attribute performance. As the product is perceived to bring the customer closer to attaining his or her valued benefits, it becomes more likely that s/he will be satisfied.
Valued Benefits. Products can be viewed as providing benefits that have consequences for the customer. In order for these consequences to be relevant to the final evaluation of product, they should first be consequential to the customer. We know that what is desired in a product is derived primarily from what is valued by the individual, hence values as a favored end-state of existence and preferred mode of conduct can be considered as a primary source of consumers' needs, wants, and desires. Therefore, it is expected that: H3: Desired benefits expected to be fulfilled through product consumption are directly related to highly important consumer values.
Perceived Performance. As discussed earlier, it is expected that a high level of perceived performance is capable of overwhelming the emotional and cognitive state of 38 the customer. Satisfaction judgments will be swayed particularly in products with attributes that have a salient performance component. Evidence supporting the effect of performance on satisfaction can be found in several studies that have shown that the level of perceived performance has a direct and positive influence on satisfaction and an indirect and positive impact through the cognitive comparison between what was initially wanted and what was actually experienced. In a controlled experimental setting, Olshavsky and Miller (1972) examined the effects of high and low performance and expectations on product evaluations of tape recorders. They concluded that high levels of expectations ~d performances lead to a favorable evaluation of product performance, yet they did not measure their impact on satisfaction. In a study explicitly measuring satisfaction, Churchill and Surprenant ( 1982) have found that while both performance and disconfirmation had a positive and direct effect on satisfaction with the plant, only the perceived performance had a statistically significant impact on the reported satisfaction with a video disk player. It is expected that: H 4 a: Perceived performance will have a direct and positive influence on satisfaction.
~ b: A high level of perceived performance will have an indirect positive impact on the value-performance congruency (or in other words, a negative impact on value-performance disparity).

Benefits-Performance Congruency. Much evidence from early research on
product evaluation has demonstrated that customers are unable to evaluate the quality of product performance unless compared to some standard of choice. According to the adaptation level theory (Helson 1964), consumers evaluate product performance using a standard of comparison. An observed disparity between the performance and the chosen standard of comparison represents a conflict between internally held beliefs and ·ved reality This often leads consumers to a negative emotional state. This percet · cognitive conflict can be successfully removed by eliminating the disparity, which can be accomplished by shifting performance evaluation either toward or against the standard.
Support for these relationships can be found in the Assimilation-Contrast theory, where Sherif and Hovland ( 1961) postulated that the magnitude of the disparity between actual performance and the chosen standard distorts performance ratings. They have shown that when the performance-standard disparity lies within the latitude of acceptance, individuals often assimilate their performance ratings toward their standard, but when the disparity falls,outside the acceptance range, the consumer's evaluation of performance quality deteriorates, and consumers shift their performance away from their standard.
Utilizing valued benefits as a research variable has been evident in some disconfirmation studies. While Oliver's (1980) cognitive model of satisfaction advances expectations as the primary antecedent of reported satisfaction, subjects were explicitly asked to provide their subjective disconfirmation of whether the benefits and problems of the flu vaccine did or did not exceed original expectations.
In the context of cross-national satisfaction, valued benefits can be viewed as the desired level of product performance that justifies the monetary and non-monetary costs the consumer has undergone to acquire the product. With perceived performance held constant, an increase in the magnitude of valued benefits will be more likely to create a negative gap in the valued benefits-perceived performance construct. Hence: Hsa: As the level of wanted benefits increases, the congruency between performance and value congruency decreases, ceteris paribus.
The level of the incongruity between acquired benefits (perceived performance) and desired benefits is expected to generate a negative effect on satisfaction. Evaluation 40 f d ucts starts with the assessment of the perceived product benefits. These benefits 0 pro will be evaluated relative to what the customer perceives to be valuable and important.
As discussed earlier, customers' perception of the valued and desired stems from consumer personal values. Therefore, a superior performance on attributes that brings the customer closer to attaining their highly ranked personal values will lead them to greater satisfaction with products. Inferior performances of valued attributes will make customers less satisfied. Hence: Hs b: Satisfaction with the product is expected to be inversely related to the disparity between valued benefits and perceived attribute performance.

Disconfirmed-Expectancv.
Results from past research provide evidence of the direct effect of disconfirmation on the satisfaction outcome of consumption. As conventionally described, better-than-expected performance creates pleasant feelings of surprise while sub-expectation performance leads to feelings of disappointment and anger. When a customer finds out that the total savings from converting to a new heating system is 10 percent rather than the 35 percent that was initially promised, the outcome of purchasing the new heating system is more likely to be frustration and strong dissatisfaction.
The relative explanatory power of the disconfirmation paradigm vis-a-vis the value-congruency paradigm of cross-national customer satisfaction is tested.
Specifically, the disparity between initial performance expectations and perceived level of performance should create feelings of pleasures, while a disappointingly poor performance should make the customer dissatisfied. Therefore, it is expected that: H6: Product satisfaction is expected to be a function of the perceived difference between expected attribute performance and realized attribute performance. Applicability of the satisfaction responses of citizens of the performance higher than two countries is related to the magnitude what they initially expected, disconfirmation model of of disconfirrnation observed between they become pleasantly satisfaction expected attribute performance and surprised and hence, more perceived quality of performance.

METHODOLOGY
There are three basic objectives for this study. The first objective is to explore a cultural interaction between cultural values and personal values. Although the intention here is not to propose a direct causal relationship between the two sets of values, it is believed that, based solely on the temporal order of existence, formulation of individuallevel values could be influenced, in part, by dominant cultural values, and thus, one expects to fincl an empirical linkage between the two constructs. The second objective is to examine the relationship between the two kinds of values and consumers preference for specific product benefits. The final objective is to evaluate the relative explanatory power of the disconfirmed-expectancy model of customer satisfaction in comparison to the value-congruency model in determining the correlates of cross-national customer satisfaction. A depiction of the two models including its dependent and independent constructs appears in figure 2.

Design.
The research design specified for this study is a 2 x 2 between-group, quasi-experimental design with participants from two countries (U. S.A./Kuwait), and two levels of expectations (high/low). In each country, half of the sample was randomly assigned to each of the two levels of expectations. As illustrated in figure 2, expected performance and valued benefits are the two exogenous independent variables in the valued-benefit performance congruency model while expected performance is only exogenous variable in the disconfirmed-expectancy model. Attempting to manipulate expectations was deemed necessary to avoid the inadvertently grouping subjects with to the procedures employed by Spreng, MacKenzie, and Olshavsky (1996).
In the context of this study, performance was treated as a measured variable and not as a manipulated variable. This decision was made in part because the performance of the computer is inherently multidimensional and complex, involving a large number of the evaluative criteria, or attributes. In addition, a true manipulation of computer performance is questionable because evaluation of perceived performance is tam inated by attribution tendencies, such as attributing good or poor performance to con manY plausible sources including quality of hardware, software, perceived lack of necessary computing skills, or a combination of these contaminants.

Countries.
The objectives of this research include the assessment of the relative influence of cultural values on consumer values and to examine the impact of nonproduct factors (personal values) and product-related factors (valued-benefits, attribute expectations, performance, valued-benefits congruency, and disconformation) on post use satisfaction. To test these underlying research hypotheses, two countries with heterogeneous' cultural value systems were needed. Although the accessibility to subjects has contributed to the selection of these two nations, the aim was to select two countries that: (1) Table 4.1 reveals the similarity between the socioeconomic conditions of the residents of both countries. Furthermore, Kuwait, as a part of a cluster of Arab countries, scored high on both power distance and uncertainty avoidance, low on individualism, and average on masculinity. The United States, in comparison, scored low on power distance between individuals, low on uncertainty avoidance, high on both masculinity and individualism.

Stimuli Product.
A new computer notebook (Dell Latitude XPi CD) was used as the stimulus product for both groups in this study. This product category was specifically chosen for the following reasons. First, Nelson ( 1970) and Zeithaml ( 1988) provided literature showing that evaluation of products with attributes that can be evaluated prior to purchase (i.e., search attributes) or during actual use (experience attributes) are more likely to trigger a cognitive evaluation process than those composed of attributes that can never be evaluated (i.e., credence attributes). In this case, notebooks have an adequate number of search attributes to allow creation of expectations and predictions about the likelihood of obtaining desired benefits, and enough salient performance attributes to allow an evaluation of actual performance. Second, there has been a rapid diffusion of personal computer machines for business and household use in both countries, and the college student population has been one of the target markets for notebook makers. As such, participants are expected to possess ample prior knowledge to identify what kind of benefits and attributes are wanted in a computer and to judge its performance. Finally, an important objective of our model is to examine whether greater power distance leads consumers to adopt status-enhancing consumption. Because notebook computers are considered as high-end products with prices far exceeding their desktop counterparts, and because they are portable and can be carried and used in public, notebook computers qualify as a status symbol, very similar to owning a cellular phone.

_MEASURES.
As shown in the structural model (refer to Figure 1 ), satisfaction experience with a product is related to two levels: 1) no~-product related factors including cultural (i.e., country) and individual level factors (valued benefits, attributes, and performance), and 2) product-related factors specific to the purchase-consumption experience. values. The approach to measuring each of these two dimensions is described below.

POWER DISTANCE SCALE (PDS) .
This construct was traditionally measured using 3 items mentioned in the Hofstede original scale. All of these items dealt with employee perception of the behavior of their management (i.e., "To what extent employees are afraid to disagree with their manager?", "What is the present decision-making style of I anager?" and "what is the preferred decision-making style?"). It was feared that yourm · ' measuring such a broad societal tendency using only three items related to workplace values might be too restrictive and may not capture the true presence of power distance.
In order to improve the measurement of this cultural dimension, a broader scale has been constructed including two of the three original items and 9 additional questions related to behaviors and perceptions closely associated with these dimensions. These additional items were all derived from Hofstede's observation of the broad consequences of Power Distances in society. All questions are scored using an 11-point scale, ranging from strongly,disagree (1) to strongly agree (11). This scale is represented by items 10 to 20 in the final questionnaire (appendix 2A).

49
The Satisfaction Constructs Set.
This set consisted of one dependent variable and three independent variables.

Satisfaction (dependent variable). The dependent measure used in this study
is customer satisfaction. The satisfaction scale used in this study was designed to measure the emotional outcome of the notebook experience after completing the handson evaluation. Satisfaction with the computer was assessed using an overall satisfaction measure with bi-polar items anchored by "Delighted-Terrible'', "Very Dissatisfied-Very Satisfied", and "Very Displeased-Very Pleased". These items coincided with the widely accepted portfayal of satisfaction as an emotional construct following product experience (Westbrook, 1980). This scale is represented by items 83 to 85 in the final questionnaire.
It is important to mention that some cross-cultural studies implicitly assume the equivalence of constructs without any attempts to support this assumption. In this study, however, special attention was given to explore the nature of the satisfaction construct and to examine its conceptual equivalency in both cultures before empirical comparisons were conducted. Details of this effort are outlined in the procedures section.

Valued Benefits (independent variable). This measure was constructed
specifically for this study. Indeed, identifying the relevant benefits is one of the objectives in this study. The conceptual distinction between attributes and benefits is straightforward; attributes reflect what the manufacturer claims a product can do for the consumer, while benefits represent the actual utility consumers receive from the product as they perceive it from their side. If a product performs in a manner identical to the engineering specifications but fails to provide what the consumer is willing to receive, the product may not succeed in creating high satisfaction. The approach taken to identify these benefits in both countries is described in the "procedures" section. It was found that the common benefits valued in both cultures were four: portability, multimedia, accessibility to information, and social perception. The scale measuring valued benefits is represented by items 3 7 to 40 in the final questionnaire.

Expectation of Attribute Performance (independent variable). Measures
of expected attribute performance were taken on 10 key attributes frequently mentioned in advertisements, review articles and frequently mentioned by computer salespeople.
These attributes were technical in nature including the following items: lightness, size and keyboard'comfort, CPU speed, screen quality and size, RAM size, modem speed, ease of upgrading, perceived quality of brand name, length of warranty period and after sale service, and being reasonably priced given expected performance (items 41-50).
Manipulation. Expectations of attribute performance were manipulated by providing two versions of ads for the notebook. The first version contained messages highlighting portability of the computer as well as technical specifications about the weight and the size of the notebook and the durability of the battery. The second version of the ad contained messages stressing the multimedia capabilities of the computer and its related technical features (screen size and quality, speakers, and game features). Refer to Appendix 3A-3D for a copy of both ads.
The response format used for this scale is an 11 point rank scale ranging from (1) representing highly unexpected to (11) highly expected. After the first pretest, many Participants complained about the lack of identifying categories for the numbers in the scale. A typical question was "which number to choose for generally expected when the scale ranges from 7 to 10?". Subsequently, 7 categories were constructed to provide a d fi nition of the representation of each number. The second pretest was successful clear e in the sense that respondents understood the scale and no complaints or questions were received about this uncertainty. 4. Perceived Attribute Performance. fhis is a 12-items scale designed to measure how participants evaluated attribute performance after using the product using an I I-point scale ranging from ( 1) bad performance to (I I) exceptional performance.
This I2-items scale is represented by questions 51-62. Initial pretest revealed that none of the 20 respondents selected a number below good performance, which eventually led to restricting answers to a smaller subset of the original scale. In response, the anchors were changed to I -4 points reflecting ascending levels of good performance, 5-8 indicating excellent performance, and 9-1 I representing exceptional performance. To examine the presence of boundary effects, a frequency table was constructed. Results did not show any sign of clumping at the lower side of the scale. (See appendix 4).

Desired
Value and Performance Congruency. This four-item scale was constructed to measure respondents' judgement about attaining their desired benefits given the level of performance already experienced. More specifically, respondents were asked to evaluate the disparity, as they perceive it, between the four main benefits (portability, multimedia, accessibility to information, and social perception) and the perceived level of attribute performance. The four items have an I I -point response format, categorized into three groups as the following: 1-4 (less than what I expected), 5-7 (similar to what I expected), and 8-11 (more than what I expected). This scale is represented by questions 63-66 in the final questionnaire. 52 6 . Disconfirmation. This subscale, consisting of 12 items (67-78), was truc ted to measure the subjective difference between consumer initial expectations cons and perceived attribute performance. This widely accepted method measures the valence of the subjective expectation-performance discrepancy as opposed to simply using difference scores between expectations and performance. The response format chosen for this scale was an 11-point evaluation scale falling into three distinct groups (1 -4 = less than expected, 5-7 = as expected, 8-11 = better than expected).

C. Possible Covariates
Altho~gh most of the research that identifies the determinants of customer satisfaction has traditionally focused on product-oriented factors, research on purchasing as a social exchange process led to the identification of two intervening (moderating) variables of satisfaction responses. Because computer evaluation is knowledge-based, familiarity with computers was measured.
Equity. This concept comprises fairness; individuals demand this in most dyadic relationships, particularly commercial transactions. A relationship is judged to be unequal when the proportion of outcomes to sacrifices of the individual parties varies beyond some idiosyncratic level of tolerance.
Within the context of customer satisfaction, consumers recall and compare their realized benefits to their monetary and non-monetary costs then they judge their relative benefits in relation to other customers or the seller. Sacrifices disproportionately higher than benefits (deficit) decrease satisfaction while large rewards (surplus) will enhance product satisfaction ratings. Given that perceived equity varies from one culture to th r it seems necessary to measure the presence of (in)equity perceptions in both ano e' countries and to assess their relative association with satisfaction.
Attribution. The evaluation of a product can be influenced by the attributions consumers make about who is responsible for observed superior or inferior quality of product performance. Results from research on attribution show that because consumers are protective of their self-concept, they tend to attribute product failures to others while attributing successful performance to themselves (Oliver and DeSarbo1988). Research on attribution effects uncovered these attribution tendencies. Valle and Wallendorf (1977) have shown that when a product fails to perform as anticipated, 75 percent of consumers blame others but when product performs well, nearly the same proportion attribute the superior performance to internal causes. As such, this tendency should be measured and controlled for, if necessary.
Prior Experience. In the context of this study, one could argue that knowledge about the category (computers) and the specific product (notebooks) may be different between the two countries. Single item measures of all three variables were included in the questionnaire (questions 86-88).

Participants.
The aggregate sample used in this study consisted of two samples recruited from two presumably different student populations, namely Kuwait University and the University of Rhode Island (representing the Northeast region). These populations were chosen with many considerations in mind. First, within the local economy, the potential purchasing power of college students after graduation is greater than their less-educated counterparts. Second, given that a college education is fully subsidized at Kuwait University and partially subsidized by the Rhode Island state mm ent for resident students at the University of Rhode Island, diversity among gove college students tends to be larger than what can be found in private colleges. This probably provides a better representation of the state population. Third, the product used in this study is widely used by both student groups on both campuses.
A total of 126 participants were recruited from business classes taught at the Faculty of Administrative Sciences at Kuwait University. Recruitment was restricted to subjects who ~e Kuwaiti citizens educated in Kuwaiti schools. Students were asked to participate in a study on notebook computers by responding to a number of questions about themselves and their perception of their society followed by questions about the computer. Enrolment in the study was voluntary and no financial or course credits were offered. Six surveys were excluded from the analyses due to missing responses, reducing the number of usable surveys to 120.
The initial US sample consisted of 125 participants recruited from business classes taught at college of business at the University of Rhode Island. Only American students who completed all their formal education in the US were recruited. At the beginning, participation was voluntary, however, attendance was far less than anticipated.
To encourage participation, students were offered a little extra credit in their class.
Participants were asked to sign up for individual appointments.  Translation. Because of the concern that translation of Western scales into a different language may actually have altered the individual meaning of the items, a system of translation and b'ack-translation procedures was adopted for this study. The original English questionnaire was given to four Kuwaiti bilingual individuals for translation into Arabic. Three of these translators were students graduated from English schools in Kuwait while the fourth was a Kuwaiti female student with an advanced English degree.
The four copies of translated questionnaires were then compiled to produce one Arabic questionnaire. Three research assistants with graduate business degrees from the US translated these questionnaires back to English. Both the original versions and the backtranslated versions were examined for textual differences, however no differences were observed and no necessary refinements were necessary. Copies of the questionnaire, ads, and consumer reports in both languages appear in appendix 2A-2D and 3A-3D.
Equivalency of C onstructs.
The obvious linguistic and cognitive differences between cultures poses a unique problem to cross-cultural researchers--whether the constructs established in one culture mean the same thing in another culture. Several articles on cross-cultural methodology recommended establishing functional equivalence of main constructs rather than simply ·ng its existence without empirical validation. In the present research, the meaning assunu of satisfaction and the determination of relevant valued benefits seem to be the two constructs with meanings that are entirely determined by the individual. Hence, it is possible that the definition of these two constructs might be culture-dependent.
Satisfaction Construct. While there has been extensive research on satisfaction in the US, no prior work exploring the concept of satisfaction from the perspective of the Kuwaiti consumer was found. One could argue that, although termed similarly, satisfaction with product and services in the US may represent a concept that can be ' different to Kuwaitis. Approach to explore Satisfaction. The approach taken to explore satisfaction was the thought elicitation procedure using the Critical Incidence Method (e.g., Bitner et al. 1990). Briefly described, the method is a face-to-face, structured open-ended focus group interviews in which the interviewer asks the participants about memorable incidents that were particularly satisfying and dissatisfying. This method is especially effective in exploratory research because firsthand experiences are inherently much more accurate and reliable compared to archival complaining data or manager's perceptions of what might be satisfying or dissatisfying to the customer.
The effort to outline the satisfaction concept involved one focus group in the US and two focus groups in Kuwait. The US focus group consisted of 20 juniors (equal mix of males and females) taking a consumer behavior marketing course. Participants were asked to remember particularly satisfying or dissatisfying experience(s) and to describe, as clearly as possible, what caused them to feel that way. A total of seven incidents, four dissatisfying and three satisfying, were revealed at the meeting.
Two Kuwaiti focus groups were conducted. The first focus group consisted of 15 college students (3 males and 12 females) recruited from a business class taught in Kuwait University. Roughly half of the participants were business majors while the other half were not. The second focus group was conducted using 7 college students ( 6 males and one female) taking a different business class in a manner identical to the first group.
The two focus groups provided a total of 9 satisfying experiences and 12 dissatisfying experiences. All of the experiences were then transcribed for further analysis. Refer to appendix 5 and appendix 6 for a summary of these transcribed experiences.
The cov-tent of the 25 experiences collected from all participants in both countries were first content analyzed to insure that (1) experiences were not vague, general to product class rather to a specific product, or were related to other people's experiences, and (2) experiences included direct causes of (dis)satisfaction as perceived by the participants. Consequently, two incidents (Incident # 13 and # 15) were dropped from the Kuwaiti sample due to generality and perception of the category rather than a single brand, whereas one vague experience (Incident # 6) was dropped from the US responses.
The remaining 22 incidents were then enumerated, sorted, and coded. Two judges (research assistants), working independently, were asked to read the coded experiences carefully and check for pattern similarity and differences in order to construct categories for the satisfying and unsatisfying responses. The resultant categories in each country were similar, both for satisfying and unsatisfying experiences (See Table 4.3). On the basis of these results, one could conclude that satisfaction in both cultures is an emotional outcome of a consumption experience triggered by product and non-product factors, and leads to attitudinal and behavioral consequences.  .. , etc.). Another key benefit in notebooks was the quality of multimedia audio-video output (opening ad statements such as "Visual Feast .." and "the power of multimedia notebook". It was also found that branded computers were desirable because they usually come with excellent technical service and longer warranty (i.e., the best of customer service and tech support). Unfortunately, review articles were directed at the American consumer and it would be a leap of faith to assume that other intemational consumers tend to value the same benefits.

58
59 The second source of information to define the main benefits was computer notebook advertisements found in a popular computer magazine (Computer Shopper) and in a local computer exhibition in Kuwait (INF097). Ads collected were analyzed to identify the most important benefit highlighted in the ad, typically in boldface print and background image. After analyzing ads directed at personal users (business users were excluded), roughly the same benefits were highlighted (portability, power and multimedia, connectivity and information accessibility, and, finally, brand image  The second US focus group consisted of20 students (the same students who participated in the second focus group exploring satisfaction were reused here). Similar to the first group, the interviewer used the same free format and open-ended questions.
Students were asked, among other things, to name the most important things to consider when buying a new computer. Results from this group are described in Table 4.5. Measures of satisfaction, its antecedents, and covariates. 3.

. Demographic variables.
Each participant was asked to respond to items in section 2. They were then given a one-page mock consumer review article that described notebook computers and The number of variables with missing observations was low ( 4 7 variables in both sets, less than 25% of the total number), their frequency was low (29 variables have one missing response, 8 variables have 2, 7 variables have 3, and 2 variables have 4), and the pattemoftheir presence appears to be random. However, one variable, evaluation of modem performance, was found to have a suspiciously large number of missing values in each (6 and 8 missing observations each set respectively). As a standard procedure, replacement with the mean was employed to correct this flaw in all variables, but in the case of modem performance, pre and post replacement comparisons were conducted and then later found to be minimal (pre-replacement standard deviations were 3.23 and 2.77 whereas post-replacement standard deviations were 3.14 and 2.67). This variable was, therefore, retained for further analyses.
Normality. Most statisticians recommend screening raw data prior to any ' advanced multivariate analysis to assess the extent to which each continuous variable departs from normality (Tabachnick and Fidell 1989). The third and fourth moments were used for this purpose, where a skewness range of-2 to +2 was acceptable, while for kurtosis, the permissible range was -1.5 and 6. Descriptive analysis generated by the SPSS computer package showed that, except for three variables in the Kuwaiti data set, all skewness and kurtosis statistics appear to be within the acceptable range. These three variables were: one personal value statement (to be well-respected by others) and the first two measures of overall satisfaction with skewness equal to -3.66, -2.45, and -2.25 and kurtosis equal to 14.94, 7.94, and 7.33 respectively. The reflect-then-square root transformation method was used to smooth out high kurtosis in these three variables Reliability Analysis. There are several criteria to help the researcher assess the reliability of each scale including the test-retest method, alternative form method, split-halves method, and internal consistency method. From a conceptual standpoint, the validity of the test-retest method is prone to subject learning effects while results from split-halves method depends on how items were divided into halves. Given the absence of alternative forms of present scales, the reliability of each scale was examined by calculating Cronbach' s alpha coefficient of internal consistency (see table 5.1), which is based on the average of inter-item correlations. According to Novick and Lewis (1967), this reliability measure is a conservative estimate of lower bound reliability of unweighted items in a scale. Even though there was no agreed-upon minimum value to judge reliability"a reasonable threshold frequently mentioned in the business literature has been set at .70, although some researchers considered alpha values below .70 to be acceptable in research of exploratory purposes (e. g., Hair et. al. 1995, p. 641 ).
Scale Dimensionality. Factor analysis can be a useful tool for examining the underlying structure of scales to assess its factorial validity, assuming that the scales used are theoretically sound. This procedure allows the researcher to identify the number of dimensions in each scale and the items that contribute to the formation of each dimension. The procedure becomes even more useful when some of the scales will be used as exogenous or endogenous constructs in structural model.

Exploratory Factor Analyses with Principal Component extraction procedures
were applied to all items in each scale for each sample. In the statistical literature, many different criteria have been suggested to determine the number of factors to be retained.
To improve the quality of this delicate decision, a combination of criteria was used in the following order. First, factors retained should explain at least 50% of the total variance.
Second, scree plots were employed as a graphical tool to define a cutoff point separating . 85 . 91 .87 • Satisfaction was measured both at the overall level and at the benefit level. To make tested models equivalent in terms of their dependent construct, only the first scale was utilized.  were used as criteria to examine the relative similarity of internal factor structure. It was concluded, based on the following discussion, that these five factors exhibited much more similarity than distinctiveness across the two samples.
The first component contained two variables, namely X9 " wanting child obedience" anQ XI I "wanting employee respect'', and was, consequently, catalogued as respectively), and therefore were classified as "Consequences of Uncertainty Avoidance" factor. The third factor consisted of two variables, namely X23 "wanting detailed law" and X24 "work stability" (LK = .76 and .65 and Lus = .57 and .47 respectively), as well three variables unique to the US sample which were eliminated. Thus, this factor can be looked upon as a "Social Stability" component. The fourth retained factor composed mainly of one unique sample-common variable, X20 "managers' freedom in making risky decisions" (LK = .71 and Lus = .85 respectively) along with sample-specific variables which were later deleted. This factor was, therefore, labeled "Attitude Toward Risk". The last factor extracted from each sample was extremely difficult to interpret =--because completely different indicators loaded on this factor in each sample, and therefore was considered as a complex component.
Because of the inherently subjective nature of a visual inspection method of rotated loadings, it was felt that a statistical test with more discriminatory power should be performed to examine the validity of these findings. Cattell's Salient Similarity Index, s, (Cattell and Baggaley 1960) was selected for this study (see Cattell et. al 1969 for a full explanation of ¢.is method and significance tables). To calculate the value of the s index, a frequency table was constructed for each factor extracted (refer to Appendix 8A and SB). Factor loadings below .4 were considered "Hyperplane", whereas a value over the minimum was classified as either "Positive Salient" or "Negative Salient" depending on the sign. The obtained s value was then compared to a table value based on the Hyperplane percentage and number of variables in the factor. A p-value less than .05 was considered significant. Of the five factors extracted from PDS, the first four factors were found to be significantly similar for the two groups (p-value for these factors were in the following order: .008, .015, .028, and< .001) while factor 5 was not (p-value = .105). The similarity testing for the first two factors extracted from UAS were found to be significant (p-value for both factors< .001) while the third factor was not (p-value = .07). Testing the significance for the last two factors was indeterminate due to the absence of table values for Hyperplane proportions below 60% and the inapplicability of interpolation methods for this case. Out of the 10 factors examined in both sets, 7 were found to be similar, and therefore, were retained for further analyses and discussion.
't the failing of some items to emerge at the desired level of significance. In the desp1 e following sections, measurement findings from each scale in each country are reported.
Results of (M)ANOVA on (PDS). Multivariate Analysis of Variance (MANOV A) was run using all items in PDS as continuous dependent variables and using country as the dichotomous independent variable (Table 5.3). Special attention was paid to insure that MANOV A assumptions were met including normality, independence among observations, and multivariate normality. Only results from the Box test showed an inequality of the variance-covariance matrix, however, this inequality is not problematic be~use the cell size is approximately equal (Hair et. al. 1995, p. 275).   . 000 .271 . 000 .271 . 000 .271 ::.:: ·::::::· :::: .. . 007 .239

EFFECT
. 057 .966 . 001 .087 . 005 .189 . 041 .886 .000 .052 .120 1.000 . 005 .196 . 044 .911 . 071 . 990 . 015 .479 To measure general direction of this omnibus between-sample difference, all . s on the scale were summed up to form a composite score (negatively framed item questions were reversed) and then analyzed using one-way ANOV A. Results clearly showed that Kuwaitis score higher on power distance scale relative to Americans,    Mean scores and significance of the remaining four items were inconsistem with the prior predictions either due to insignificant between-country differences or to significant results but in the wrong order. Two items have means that were found to be significantly different from each other in the wrong direction (detailed laws are required in organizations and workers are perceived to be under more pressure than what they should be). Covntry means of one item were found to be in the correct order but at an insignificant level (attitude toward conflicting ideas in society), whereas another item generated two means with insignificant differences and in the wrong order (the need to adhere to social rules irrespective of agreement). To allow the full measurement of all aspects of uncertainty avoidance regardless of between country significance, all items were retained for further analyses.

Association between PV and Cultural Dimensions
One of the main objectives ofthis exploratory study was to examine the possible observations for each variable required for performing this procedure was also met, since the variable sets contained 19 total variables while the sample size in each set is 120 and 121 respectively. All of the analyses were done using the SAS program (refer to Appendix l lA, l lB for bivariate correlation matrix between the two PDS, UAS and LOY scale).
Assessing the relative strength of association between the criterion variable set and the predictor variable set was performed by examining both the magnitude and the statistical significance of the canonical function (using F distribution tests) as well as the 79 redundancy index for each canonical variate derived from the data set. While inspection of the canonical coefficient provides a useful approximate measure of the strength of relationship between the variates formed in the two sets, this measure does not reveal the amount of the variance shared in one set with variables in the opposite set. Calculation of a redundancy index was done in the following order. First, the amount of variance explained in the criterion variable is calculated by averaging squared canonical loading (canonical loadings represent the correlation between the criterion variable and their canonical variate). Second, the canonical correlation coefficient between the criterion and the predictor ~ariable was first squared and then multiplied by the average squared canonical loadings derived from the first step (Hair et. al., 1995).

CANONICAL CORRELATION ON LOV AND PDS--Pooled Sample
Performing canonical analysis between the 8-item LOV scale (criterion variables) and the 11-item PDS (predictor variables) generated a total of eight canonical functions along with four multivariate tests of statistical significance (see Table 5.5). Assessment of the overall model fit showed that, taken collectively, the full set of eight canonical functions was significantly different from zero, yet, of these 8 functions, only the first was significantly different from zero and seemed worthy of further investigation (Canonical correlation coefficient = 0.50, Eigenvalue = .34, F value = l .79, p <.0001).
To obtain a better understanding of structure of this retained function, canonical loadings relating the variables in each set with their variate were obtained, using canonical loadings of .3 as a minimum acceptable criterion (see Table 5.6). Results show that, on the criterion side, sense of belonging (.72), need to be well-respected (.59), sense of security (.35), and sense of accomplishment (.31) formed most of this variate. The

CANONICAL ANALYSIS BETWEEN PDS AND LOV--Kuwait and US
A separate canonical analysis was performed on data collected from each country to assess whether the same results will hold in the two samples. Two canonical analyses Were performed independently, one using the Kuwaiti data set while the other was run on 82 the US data (refer back to Out of the 11 items, 5 were similar in magnitude and direction (XlO, Xl, X13, X18, and X 19), whereas 2 differed in terms ofloading size (X9 and X 12), two loadings in terms of sign (X 14 and X 17), and two variables in terms of size and direction (X 15 and X 16).
Results from redundancy analysis showed that 2% of the variance in PDS was accounted for by variance in LOY for the Kuwaiti group, while the percentage was twice as high for the US sample (4%). When viewed from the other side, the proportion of variance in variate derived from the predictor set (PDS) accounted for by the variation in the criterion set was 3.6% and 3.2%, for Kuwait and US, respectively. . d for further analyses, yet the composition of the canonical variates was not always re tame ·stent across the two samples. For the Kuwaiti group, an examination of the matrix cons1 of canonical loadings showed that the canonical variate consisted of sense of belonging, security, and less preference for an exciting life, in descending order. Viewed from the opposite side, the largest loadings observed between PDS variables and the variate were between close employee supervision, child obedience, and student fear of disagreement.

Interpretation
In a general sense, these variables can be tied together as behavioral consequences of the presence of high power distances.
Oisaggrt4gate canonical analysis performed on US sample showed the same two variables observed in the Kuwaitis sample to be the main determinants of the retained variate, namely sense of belonging and more life excitement in descending order. On the PDS side, the composition of the variate was different. It was found that employee respect, low level of employee supervision and positive attitude toward power and wealth were the main determinants of this variate.

CANONICAL CORRELATION BETWEEN LOV AND UAS--Pooled Sample
In this part, examining the magnitude and significance of association between personal values and uncertainty avoidance was accomplished by running a canonical correlation where LOV served as variables in the criterion set while the 10-item UAS used as variables in the predictor set (Table 5.7). To evaluate the strength of association between the two sets, eight canonical variates were extracted, of which two were found to be significant and were retained for further analyses (Canonical Correlation = .64 and .40, F == 2.83, 1.50, while p< .0001 and .008 respectively). A closer examination of the two functions at the country level reveals that, although the second canonical function was statistically significant in the pooled sample, it failed to emerge as a significant canonical function at the country level and hence, was deleted from subsequent analyses. Followup canonical procedures aimed at uncovering the factor structure of the first retained variate revealed two LOY items to be the main contributors (see table 5  . ce in canonical root derived from the UAS set account for by its variables was van an 16 .5%, whereas the amount of variance in UAS explained by LOV set was 6.8%.
Although the absence of minimum cut-off criteria to evaluate the practical significance of the obtained makes evaluation somewhat subjective. these results compare favorably with the previous results, hence interpreted as a moderate association.

CANONICAL ANALYSIS BETWEEN LOV AND UAS--Kuwait and the US
Part of the results reported in Table ( Results obtained from performing redundancy analysis on the US sample were relatively stronger. Shared variance between the criterion set and its own variables was 24%, while the shared variance between this variate and canonical variables on the predictor set was 9 JS%. Taken from the other side, the canonical root derived from the predictor set explained 19.7% of the variance its own variables while it accounted for 7.6% of the variance in the criterion set. Being well respected and accomplishment were transformed using the reflect-then-inverse and the inverse methods respectively.   such as clothing quality, vehicle luxury, and house elegance) showed that Kuwaitis scored higher (MK =8.88 and Mus =8.12, main effect for country difference Fp ,239) = 9.548,p < .002). To assess the relative strength between perceived power distance and the tendency to consume status-enhancing products, a bivariate correlation was computed using individuals' composite score and scores obtained on item 9. Results revealed a fair correlation coefficient (r = .279) which is significant at .001 level (see Table 5.11).   rtance on strengthening family ties to guard against future hardships (social level impo strategy) and making the best use of talents and abilities (individual level strategy) than others living in low uncertainty avoidance cultures (i.e., Americans). Results obtained from one-way ANO VA and correlation procedures were mixed (see table 5.10 and table S.11). On the one hand, Kuwaitis expressed more emphasis on establishing strong familial ties to guard against future hardships than did Americans ( MK=9.18 and Mus= 7.87,country main effect F[l ,239] = 103.288, p < .001) while a moderate size correlation was found between UAS composite score and strengthening familial ties (bivariate r = .39, p < .01 ). Qn the other hand, no statistically significant differences were observed between the level of importance Kuwaitis and Americans assign to making the best use of their talents and abilities (MK=9.37 and Mus = 9.34, country main effect Fp ,239J = 0.047, p < .91). In the second part, it was proposed that members of the high uncertainty avoidance culture prefer purchasing familiar brands to generics. Results obtained from between-countries ANOVA were statistically insignificant ( MK=6.98 and Mus= 6.74, country main effect Fp , 2391 = 3.455, p < .44) and no further follow-up analysis was performed. It was, therefore, concluded that no support was found for this part.

Hypothesis Three.
This hypothesis was intended to measure the relationship between consumer personal values with what consumers seek as benefits in products. The approach taken to test this hypothesis was to measure the practical and statistical significance of association between LOV with valued benefits in computer notebook, namely portability, multimedia, information, and reputation, using canonical correlation procedures. Table 5.12 illustrates the overall model fit.   the second and the third functions were statistically significant, the magnitude of variance explained was low (R 2 = .08 for each functions) and neither of the two functions were significant at the country-level analysis. Therefore, they were deleted from subsequent analyses.   to be the marker variable followed by information seeking and multimedia features.

Identification of a general association between personal values and valued
notebook benefits was further examined by running Pearson bivariate correlational procedures. As shown in table 5 .14, the same pattern of results reemerged, where a need for enhanced social image was found to be associated with high regard for being accepted by others and for warm relationships with others. For the "information" benefit, the largest correlation coefficients were between need for others' respect and warm relationships with others, while for the multimedia benefit, the highest correlation was with belonging.

HYPOillESES SET 2
The three hypotheses included in this set were formulated to test the directional relationship between the antecedents of customer satisfaction derived from previous ' literature. The first step in testing these hypotheses was to specify two structural equation models, one representing the disconfirmed-expectancy paradigm of customer satisfaction while the other tests the Valued-Benefit-Performance Congruency model (VB-P Congruency) of satisfaction. Each of the two specified models was then estimated using data collected from samples drawn from each country. All of the structural equation models were estimated with the AMOS program (Arbuckle 1997).  257°0 . 174 .. .195 .. -.021 . 285°0 . 101 .174 .. .068 .080 -.018 . 145° .126  observed and predicted matrices. It is hard to achieve an insignificant chi-square, and so other indices (e.g. , CFI) are assessed instead (Bentler and Bonett 1980 1995). An examination of obtained results from the two samples reveals a generally g r eliability for all exogenous constructs with the exception of desired benefits for stron Americans. This construct fell short of the minimum established reliability value of .7, while the calculated validity was below the recommended minimum of .5. . 89 . 33 . 74 . 84 . 68 .78 Variable 2
In summary, overall fit indices, good reliability and validity of the measurement model, and the statistically significant parameter estimates obtained from fitting the VB-P Congruency model to the Kuwaiti sample lend strong support for its validity explanatory power. In comparison, the slightly less appealing validity results (i.e., the validity of the independent exogenous construct, desired benefits) obtained from applying the model to the US sample may cast some doubts about the validity of the measurement model and hence, the ability of the structural model to explain satisfaction responses.

Hypothesis Four.
Hypothesis four predicted that the perceived performance would exert two kinds of effects in the VB-P Congruency model; a direct antecedent to satisfaction and a direct effect on the congruency between valued benefits and perceived product performance (See figure 3 ). The first part, H4 a, predicts perceived performance between desired benefits and perceived performance (VB-P Congruency). Support derived from the data was also mixed. Performance was found to have a direct positive impact on the DV-P congruency in the Kuwaiti sample (p = .77, t-value = 5.60) while the same path obtained from the US sample was insignificant (p = .36, t-value = 1. 77).

Hypothesis Five.
As argued earlier, there is reason to believe that the perceived level ofVB-P Congruency is preceded by both the level of performance (tested earlier) and the initial level of valued benefits the consumer wants to receive from the product. The first part of this hypothesis was formulated to test the relationship between the level of valued benefits and the VB-P Congruency, predicting a negative relationship between the two constructs. Accordingly, an estimation of the path coefficients linking the two constructs was estimated for each sample. The obtained path coefficient was positive and significant in the Kuwaiti sample (p = .19, t-value = 2.01) but was insignificant in the US sample (p = .53, t-value = 1.75) due to large standard error.
Taken together, these findings suggest that, contrary to what was hypothesized, initial level of valued benefits has shown a positive association with VB-P Congruency, leading to the rejection of Hsa· The second part of the fifth hypothesis examines the relationship between VB-P Congruency and reported satisfaction. It has been argued that a positive congruency between the valued benefits and performance, assuming the performance is conducive to obtaining the desired benefits, will positively influence satisfaction rating of the products (Spreng, MacKenzie and Olshavsky 1996). Findings tend to support this contention in both samples (fQr Kuwait respecified model generated a ~ = .55 with t-value = 4.05; for US sample, ~ = .48 with t-value = 2.44).

Disconfirmed Expectancy Model-Kuwait and US
The disconfirmed-expectancy model, as specified for this study, consisted of four main constructs divided into one exogenous construct (expectation of attribute performance), and three endogenous constructs, namely, perceived performance of attributes (performance), perceived gap between expectations and performance Pooled sample but not to the US data). Encountering a negative variance can be attributed to many potential problems including non-normality of data, small sample size, and specification problems. Because the normality of data was fairly similar across the two samples and that sample size is practical restriction, it might have been a problem related to the specification of the model. Fortunately, this inadequacy was corrected by placing an equality restriction between the disturbance associated with performance and disconfirmation. Because these two constructs have roughly the same dimensions of the notebook, placing an equality constraint seemed logical yet admittedly, less than ideal.
The ove\all fit of disconfirmed-expectancy model for the Kuwaiti sample was less than acceptable (CFI = .85, RMSEA = .14), hence, the model was rejected (though further analyses will be performed to identify model imperfections). In comparison, results obtained from fitting the model to the US data was better (CFI=.92, RMSEA = .11 ), and that model was accepted.
Measurement Model. Two separate CF As were applied to the exogenous variable in each sample to examine the empirical reliability and validity of the exogenous construct. Refer to appendix 12B for a depiction of the model and to table 5 .17 for a summary of results. The calculated Cronbach reliability measure (alpha) for the Expectation construct was slightly lower than the acceptable minimum for the Kuwaiti sample, whereas the measure was strong for the US sample. Similarly, validity measures showed the variance extracted from each observed indicator was lower than 50%, for the Kuwaiti sample, but higher than that for the US sample (R 2 was .38 and .67 respectively).
Structural Model. For the purpose of assessing the strength of relationships between the model constructs, an assessment of the structural coefficients was undertaken for each sample. Across the two samples, some estimates of path coefficients were found to be significant at the .05 level while others were not significant (table 5.18).
Interestingly, the pattern of parameter significance was consistent in both samples, although the model fitted to the Kuwaiti sample has already been rejected. Attempts to respecify the model by eliminating paths with insignificant t-values did not produce any marked improvement in model fit or in any t-values associated with other insignificant path coefficients. Consequently, only the US model was retained for further analyses. Hypothesis Six. This hypothesis pertains to the effects of disconfirmation on overall satisfaction judgements. It has been postulated that a positive disparity between initial attribute expectations and perceived performance should lead to positive satisfaction judgements and vice versa. In essence, the expected path coefficient reflecting this pattern of relationship should be positive. An examination of the findings obtained fyom the pooled and US samples supports this hypothesis, (for pooled sample, 13 = .78 with tvalue = 4.47; for US, 13 = .45 with t-value = 2.27). In the Kuwaiti sample, however, this predicted relationship did not hold because the path coefficient was high but was not significant due to higher standard error. But since this model was rejected due to poor fit, the insignificant path observed in the Kuwaiti data is not reliable and should not affect these findings. Therefore, this hypothesis was supported.  54 . 34 . 68 .45 CHAPTER6

DISCUSSION
The results of this study were found to be consistent with predictions derived from previous research on cultural values, and provided some, though not complete, evidence supporting the effects of values on some important aspects of consumer behavior. This research has also shown that, as predicted by Hofstede (1980), Kuwaitis have a higher level of power distance and uncertainty avoidance compared to their ' American counterparts. Moreover, the general findings of this study tend to support the existence of a relationship between personal values and cultural values, though some of the results were not as strong as was predicted. As demonstrated by results obtained from canonical correlation analysis, it was found that there is, at least, one reliable form of linear association between each of the two dimensions of national character with personal values. The somewhat small size of redundancy variance explained by the canonical root did not lend equivocal support to practical significance of this root.
Another important objective of this study was to delineate the correlates of crossnational satisfaction at the individual level. Results obtained from notebook computer study have shown that while VB-P congruency model explained the satisfaction judgements both for Kuwaitis and for the US sample equally well, the disconfirmedexpectancy model emerged as a suitable paradigm to explain the satisfaction judgement of the US sample but not for the Kuwaiti sample. For the purpose of simplicity and clarity, the following discussion chapter has been divided into three sections to discuss the findings pertaining to each of these distinct but interrelated parts.

~rsonal and Cultural Value Systems
Findings obtained from applying canonical correlation procedures uncovered at least one reliable canonical root linking relative power distance and strength of uncertainty avoidance, separately, to personal values. The absence of rotational techniques in present statistical software hindered, to some extent, the interpretation of pairs of canonical variates, so the matrix of canonical loadings was examined to aid the decision as to identify the variables forming the extracted variates in each set.
For PDS, findings from the pooled sample reveal that a higher need for social acceptance and for others' respect is positively correlated with both social and managerial power-expressive behaviors. Follow up analysis performed on the Kuwaiti sample suggests that personal values related to the desire for more social acceptance, more security, and a less exciting life were all found to be associated with power-expressive social behaviors. In contrast, the same personal values reemerged in the US sample and were associated with power-expressive behaviors, but limited to managerial context.
Taken together, the pattern of these results strongly suggests that power expressive behaviors, either social or managerial, are associated with outward personal values. Yet, the strength of these results should be cautiously interpreted, because the amount of variability LOV explained by the variability of PDS was on the low side (from 2% to 4%). One explanation for observing such a small redundancy is related to the way the index is built. Because the procedure involves averaging squared correlation coefficients between a variate and the variables composing the other variate using equal weights, the large number of variables with insignificant coefficients have diluted the effects of the smaller number of variables with significant coefficients.
The canonical association between LOV and UAS was relatively stronger than the relationship observed between LOV and PDS, both in terms of magnitude of canonical roots (.52 to .64) and redundancy (4% to 9.4%). For Kuwaitis, preference for outerdirected values (belonging, warm relationship and being well respected) was associated with the magnitude and sources of uncertainty (level of worry about the future and relative intolerance of socially conflicting opinions). For the US group, mostly innerdirected values (security, accomplishment, and belonging) were found to be related to future anxiety and preference for homogenous social opinions.
These forms of association do not, however, ascertain whether people with a ' . stronger need for social acceptance and others' respect are driven to practice more powerexpressive behaviors or conversely, the presence oflarge power distance (and its related behaviors) leads its members to favor this subset of personal values (the same argument applies to observed relationship between LOV and UAS). Yet, it is intuitive that the latter is more likely. Through socialization, individuals learn the values, norms, and attitudes of the community which then become a primary influence on the way individuals define themselves, their roles, and their interests at the personal level.
Admittedly, answering this question requires a longitudinal analysis rather than the crosssectional approach taken in this study.

PDS, UAS and Relative Importance of Personal Values.
In the hypotheses section, a table was constructed to test the correlation between the ranking of personal values in six countries and the positioning of these countries on power distance and uncertainty avoidance (Table 2.1 ). According to the correlational analysis results shown in that table, it was expected that self-fulfillment would show a high and positive correlation with both power distance and uncertainty avoidance, while self-respect would show a negative significant correlation. The table also shows a negative correlation between sense of belonging and power distance. Because the number of countries involved in this study was not adequate to construct meaningful correlation analysis, the only possible analysis was to see whether between-country mean scores on these values would vary according to patterns of correlation reported in the table. More specifically, one would expect to find individuals living in the country with the higher power distance and uncertainty avoidance (i.e., Kuwait) to have a larger mean score on self-fulfillment and lower mean ~cores on self-respect and sense of belonging.
Main effects of country on the importance of personal values were examined using one-way ANO VA, and yielded mixed results. Citizens of both countries equally valued self-fulfillment, Kuwaitis were found to value self-respect less than Americans (consistent with earlier prediction), while Kuwaitis expressed higher desire for sense of belonging than did Americans (contrary to earlier predictions).
One plausible explanation for these discrepancies is the incompatible measurement used for personal values in both cases. Specifically, the approach taken in the present research is to measure values in terms of importance scores, while the correlation table was formed on basis of ranking of personal values. An individual may provide a score of 10 on two values separately, however, when he or she is asked to rank these two values, one will be ranked higher, even though both were equivalent on the first measurement. It follows that a table of correlation coefficients should be reconstructed on the basis of importance scores and not on ranking scores.
Complexity of Cultural Values. The multidimensionality of factor structures of the pDS and UAS scales was viewed, paradoxically, as a point of interest and a cause for concern. Hofstedes' measurement of workplace values involved three items for each of the two chosen cultural dimensions, but clearly, the relevance (and hence the validity) of such indicators become questionable when applied to cultural contexts other than the organizational culture. In response, the scale was expanded by adding items measuring the social and managerial consequences of power distance and uncertainty avoidance, items that Hofstede uncovered using qualitative analyses. Given the level of association that one might expect to find between indicators and consequences of a phenomenon, it was expected that two factors would emerge, the first measuring the magnitude of each dimension and the other measuring its consequences. However, results obtained from factor analyses suggested the presence of five reliable factors in each scale, with somewhat similar, but not identical, factor structures between the two countries. The retention of a 5-factor solution in each scale was immediately followed by an extensive factor labeling process that relied heavily on both the researcher' s conceptual judgement of marker item contents and on the use of more objective similarity indices to ensure factor interpretability and cross-group validity.
It could be said that the elaborate factor structure uncovered in this study indicates both the richness and the complexity of these two dimensions of national character.
These findings show that the extent to which power distances exist in society should be measured at many different levels including beliefs about sources of power, attitudes toward high power, and prevalence of power expressing behaviors in society. Likewise, the emergence of a 5-factor solution in the uncertainty avoidance scale suggests a complex structure of this cultural dimension, more than what Hofstede initially had proposed and measured in his famous study. It is the contention of the present author that the measurement of relative tolerance for future risk and lack of structure in life would be greatly enhanced if a more coherent set of items can be generated and tested in future research.

Gender of Participants.
There were some concern that observed differences in values scales, especially PDS and UAS, might have been gender-dependent. Because the Hofstede study was conducted in the late seventies where females in the workforce, especially in developing countries, might have been underrepresented, there was a chance that his scales might have been gender biased. The gender mix in the this study was fairly good, with almost 50% split between the two sexes in each sample. The application of one-way ANOV A procedures did not show any significant differences between males and females within each sample or between the two genders in the pooled sample.

112
The effects of gender on the preference for particular personal values over others were, unlike gender effects on cultural values, expected. According to the author of LOV (Kahle 1983), the level of importance assigned to some of the personal values in the LOV scale was expected to be different between males and females. For example, warm relationship with others and sense of belonging were expected to be valued more by women compared to men. One-way ANOV A results measuring main effects of LOV by gender in each sample were not generally supportive of Kahle's predictions. Of all values, American females reported higher value for warm relationships with others more than men did, bllt they unexpectedly placed higher importance on personal safety than did men. In comparison, Kuwaiti men valued fun and excitement more than Kuwaiti females, while they scored about the same on all other values. These results could be attributed to the specific nature of the sample in each group. Increased preference for security in life by US females may be attributed to the kind of threats present in living on campus while female students in Kuwait University live off-campus, and the crime rate is much lower than in the US.

Values and Wanted Product Benefits
Hypothesis one predicted that consumers in the country with high level of power distances (i.e., Kuwait) would exhibit more preference for products that provide enhanced social image compared to individuals living in countries with more power equality. The genesis of this hypothesis originates primarily from the theory of conspicuous consumption, which suggests that consumers may not consume products merely to satisfy themselves, but rather to satisfy others in their communities. But for ostentatious consumption to be prevalent in society, individuals should believe that evaluation of others based on the luxury of possession is a common practice, and more importantly, individuals should also believe that ostentatious consumption should contribute to improving their social status. As such, question 9 in the questionnaire was coined to measure the presence of this practice in both countries at an absolute level, while measuring the relative importance of creating a favorable reputation when using a mobile computer measured this phenomenon at the product level. Results generated from running one-way ANOV A procedures on these two items provided ample support for this prediction, although estimates of effect size measures (TJ 2 ) were relatively small (ri 2 = .04 and .07 respectively).
Theories of conspicuous consumption behavior were divided on identifying the main motive for pursuing this overt display of wealth (Mason 1981 ). In his book, he explained that the old school of thought attributes leisure consumption to innate personality traits that motivate the individual to favor distinctiveness from others, while the contemporaneous school considers this form of consumption as an effort on the part of the individual to improve social status. If one subscribes to the later view, overt display of power and wealth, and hence, ostentatious consumption, would then be more likely to occur in societies with large power disparities as opposed to societies with relatively even power structure. Fortunately, because the two cultures chosen for this study do not have family or birth-restricted social classes, social mobility of individuals can be achieved by means of conspicuous spending.
Because self-reporting was the approach taken in measuring the presence of conspicuous consumption, demand artifacts and social desirability may have inhibited the full measurement of this behavior. For Kuwaitis, admitting the presence of, let alone engaging in, the practices of conspicuous consumption is susceptible to the moderating effects of religion. According to Islamic beliefs, lavish lifestyle and wasteful spending are highly discouraged, which, under the pressure of social desirability, some respondents might have been less open and candid about this behavior. This concern becomes more real in light of the fact that Kuwaitis have reported higher adherence to religion than did their Americans counterparts at a statistically significant level.
The second hypothesis contained two parts with two distinct predictions. In the first part, it was stated that nationals of cultures with a high level of uncertainty avoidance attempt to reduce the undesirable effects of uncertainty by enhancing personal ' talents and abilities and by building stronger ties with their families. In the second part, it was stated that, given that uncertainty avoidance is, in fact, a culture-wide measure of risk aversion in general, it would be expected that avoiding ambiguity in general might also affect personal preferences for familiar well-known brands over generic products with uncertain, and highly variant, quality.
Results generated from one-way ANOVA and bivariate correlational analyses (see "Results" section) failed to provide firm support for all the predictions made in the first part of hypothesis 2. Respondents in the Kuwaiti group were found to be more concerned with building stronger ties with their families to cope with the unpredictable future, however, both Kuwaitis and Americans were equally concerned with making the best use of their talents and abilities. Similarly, no support was found for predictions made in second part of hypothesis 2.
One possible explanation for this lack of association lies with the level at which uncertainty-avoidance tendency may be operating. Anxieties generated from a high level of uncertainty are not manifested in individual level behaviors such as those related to consumption, but might be evidenced in other forms of social behaviors, such as more emphasis on familial unity.
Hypothesis three predicted that consumer needs, wants, and desires are directly related to consumers' personal values as measured by the LOV scale. Overall results obtained from canonical correlation analysis uncovered at least one significant canonical root. In depth analysis aimed at determining the nature of the retained root showed that outer-directed values were positively correlated with preference for social image, information, multimedia, and portability benefits in a descending order.

'
Generally, results generated from canonical correlation were supportive of the linkages between general human values and benefits wanted at the product level. In the consumer literature, it has been theorized that values guide many aspects of human behavior, including those pertaining to buying and consumption. Gutmans' (1982) means-end chain model conceptualizes values as "end states of existence" and products as "means" employed to satisfy an end. Thus, it is anticipated that in a competitive market with a variety of differentiated products, product choice, and ultimately the satisfaction with that choice, are dependent on the degree of match between initial personal values and the kind of benefits provided by the product.
From this study, and from other studies reported elsewhere, it was found that the cultural background of the consumer leads to subtle differences in the degree of importance assigned to some personal values. If one accepts the argument that values indeed guide consumer behavior, it might be reasonable to expect that cross-cultural 116 variability of importance assigned to values might very well lead to the display of different consumer behaviors, including product choice and evaluations.

y_alue Benefits vs. Expectations
Jn the previous discussion, it was shown that the kind of personal values consumers hold is directly related to the kind of benefits consumers look for in products.
Taking this relationship a step further, this association was tested using a real product (notebook computer) to examine its effects not only on what consumers want in a product, but also if attainment of these benefits would, in fact, lead to higher levels of satisfaction. In this regard, Prakash ( 1984) stated that " ... Consumer research has shown that expectations and confirmation of expectations are important determinants of satisfaction (Oliver 1980;Parakash 1981 ). The causal sequence is: Expectations lead to confirmation of expectations, which leads to satisfaction. With the introduction of personal values in this relationship, the sequence can be modified: personal values leads to expectations, which leads to satisfaction . . . " Parakash, 1984, P. 14 7) The approach of this study differs from Prakash in the sense that congruency between valued-benefits and perceived performance, and not the disparity between attributes expectations and perceived performance, should emerge as the determinant of satisfaction with the notebook.
Examining the suitability of disconfirmed-expectancy paradigm vs. the valuebenefit congruency paradigm in explaining the satisfaction responses of the international consumer has shown an interesting finding. To Kuwaitis, satisfaction responses to the notebook computer seems to be determined by the degree to which wanted benefits matched perceived performance, whereas for the U. S. sample, the disconfirmation of expectations emerged as the prime antecedent of satisfaction.
Hypothesis 4 predicted that perceived notebook performance should exert two kinds of effects; a direct influence on expressed satisfaction judgements about the product (H 4 a) and an indirect effect on satisfaction through the level of perceived consistency between wanted benefits and perceived performance (H4b). Results generated from applying structural equation modeling on the pooled sample showed no direct effects of performance on satisfaction, while the conceptualized indirect effects of performance on satisfaction through disconfirmation were supported.
The absence of direct effects of attribute performance on satisfaction comes as a surprise becaus~ the level of performance saliency is quite high when compared to other product categories with ambiguous performance dimensions (coefficients obtained from pooled samples were p = .16 for the VB-P model and p = -.01 for disconfirmation model respectively). The effects of performance on satisfaction were completely mediated by the VB-P disparity in the better-fitting VB-P congruency model, or by the disconfirmation of attribute performance in the disconfirmed-expectancy model. From a conceptual standpoint, these findings suggest that while performance is an important construct in understanding the satisfaction formation process, its absolute level may be irrelevant with a valence that results from some form of comparative process.
Hypothesis five consisted of two parts. In the first part, it was predicted that a high level of valued benefits is expected to be inversely related to the observed congruency between valued benefits and perceived performance. Surprisingly, the regression path coefficient between valued benefits and VB-P congruency was positive, which fails to support the hypothesized prediction. The second part of hypothesis 5 was supported. In this part a prediction was made that satisfaction with the notebook would be dependent on how closely product performance brought the consumer to attaining the benefits initially wanted.
Failing to support the linear relationship between valued benefits and VD-P congruency constructs is counterintuitive because as the level of benefits increases (as determined by the type of values), the ability to deliver a matching quality of performance becomes more difficult. This mediation of effects is consistent with the proposition of the contrast theory in the sense that consumers' satisfaction judgements seem to be shifting away from initial level of valued benefits.
Results &enerated from fitting the disconfimed-expectancy paradigm to the data, despite its inadequate fit, supported the prediction that satisfaction is a function of subjective disconfirmation between initial expectation level and perceived level satisfaction. Similar to the VB-P congruency, the path coefficient observed linking disconfirmation to satisfaction was strong and positive, indicating a positive linear relationship.
The empirical support found for the association between disconfirmation and satisfaction is in line with the prediction of the adaptation level theory (Helson 1964), which suggests that consumers tend exaggerate the gap between expectation and performance in ways that affects satisfaction. More specifically, it has been suggested elsewhere (e.g. , Yi 1990) that the surprise effects from resulting from performance surpassing (falling short of) initial expectations generates positive (negative) feelings that would then affect satisfaction.

CONCLUSIONS
As expansion to foreign market continues to be the trend, reexamination of existing consumer theories, or perhaps the development of new ones, might be necessary to understand the intricacies of cross-national consumer behavior. Whether the consumer is local or international, it appears that the most basic motivation for buying products and services is to satisfy needs, wants, or desires, which can be physiological and/or psychosocial. The introduction of Bentham's law of diminishing utility marked the first ' theorization of the critical interplay between satisfaction and consumer decision making, however, the identification of critical antecedents of satisfaction really progressed as the focus shifted toward the incorporation of psychological antecedents. While this progress evidently shows that we probably now know more about the antecedents of customer satisfaction than ever before, studying the interaction between consumer culture and consumer psychology has been excluded from present satisfaction research. This study is one of the first attempts to address this issue and to fill some of these research gaps. This has been accomplished by studying cross-cultural differences in terms of values, and traced the effects of these differences to choice of valued benefits in products, and measured how these benefits affects overall satisfaction judgement.
This study uncovered some important differences in the types of values held by individuals in Kuwait and the US. Kuwaitis, in general, choose outer-directed personal values, such as the need to belong and need to be well-respected more than Americans did, whereas Americans preferred inner-directed values, such as self-respect and excitement more than the Kuwaitis did. In addition, there is a large perception of power inequality between Kuwaitis, and there is also a greater tendency to engage in power expressive behaviors when compared to Americans. Kuwaitis were also found to express a general lack of tolerance toward uncertain situations and toward ambiguity in general, and consequently, were found to have stronger interfamilial ties to guard against future threats and unforeseen difficulties.
The specific links between values and consumers' desired benefits and overall satisfaction was formalized and tested in each country using a notebook computer as to-face administration of the study instead of using fictitious products or non-personal method of data collection. However, as explained in the next section, the study is not free of limitations.

Limitations of Study
As is the case with all studies, the generalizability and the validity of the study findings is constrained by the conceptual framework and design, characteristics of selected samples, and the exclusion of some potentially relevant variables. The identification of these limitations is instrumental for the proper interpretation of findings, 121 and should provide direction for future research.
This study investigates the impact of values on key benefits desired in products and on the satisfaction response of the customer. As someone might argue, portraying purchase events as goal-directed behaviors (products are considered as means to achieve desired consequences and valued end-states) is mainly derived from the expectancy-value approach in attitude and consumer research. Proponents of this approach consider the consumption experience as a purposeful behavior in which consumers create some form of subjective probability that the purchased product will eventually lead to the desired consequences. ~ile this portrayal might be evident in big-ticket items, it is expected that such conceptualization may not include habitual and serendipitous buying, or "lowinvolvement" purchase decisions with little information search and limited cognitive evaluation. Even for highly involving purchase decisions, it is also possible that the relative importance of benefits as choice criteria could be moderated by subjecting the customer to immense advertising and point-of-purchase promotional pressures. It is also important to note that satisfaction is measured after exposing the 122 participants to the stimulus product for a limited period of time. While using product trials as a surrogate for a true consumption experiences studies has been generally accepted and adopted in consumer satisfaction research (for notable examples see Oliver I 980, Churchill andSurprenant 1982, Spreng, MacKenzie andOlshavsky 1996), it is possible that a more prolonged consumption experience might create different results.
For example, it is possible that product knowledge increases with time (learning effects) and uses that seemed to be unimportant at first impression might become consequential benefits with time. In light of these time and learning effects, this study can be considered as a pilot research effort in the area of cross-national satisfaction. Future studies might improve their findings by adopting longitudinal research designs with purchased and consumed products, not just product trials.

Masculinity Femininity
Cultural Dimensions Proposed By Hofstede (1980Hofstede ( , 1983  Reflects the strength of the relationship between an individual and others 7.

8.
Stayi ng with one company is undesirable It is better to work for Individualism: Relationship is very loose. Caring for oneself or his/her immediate family member is concern. There is great individual freedom and less social intervention in one's life , Collectivism: People are born into ; collectivities or in-groups where everyone , looks after the interest of his/her in-group in exchange for protection when this person is in trouble or in need.

To make the best use of my talents and abilities
To be well respected by others To be safe and protected from harm and attack (feeling secure) To have self respect and to be proud of myself as I am To be successful C. This following set of questions describes how someone feels about his/her own culture. Please circle the number that represents your opinion.

C.
After you tried the notebook, how would you evaluate the following attributes?
I found the . ................ to    H. Imagine that you have bought the computer notebook that you have already tried and that you have paid the price mentioned in the commercial ad. Imagine further that you will be using the notebook in the same way that you have used the computer in the trial. The followin g statements reflect some of the thoughts that cross the minds of consumers after purchase. Please read each statement carefully and specify the extent to which you agree or disagree by circling the number that best represents your opinion.

I PART THREE
The following is a list of questions that will help us classify our data into groups. Please read the questions and circle the appropriate number.

Notebook Computers
In the past few years portable computing has managed to draw a large chunk of the personal computer market away from traditional desktop systems. This trend is due, in part, to the light weight and small size of notebooks that results from combining the three main components of computers (monitor, CPU, and keyboard) into a single unit.
This feature allows business people and educators to fit the notebooks into a carry-on travel case and run presentations and office work from a single piece of equipment. Although consumers want to buy the finest notebook that provides all the above benefits, doing so seems to be extremely difficult. Results from our research has shown inverse relationships among these benefits such as: Multimedia capabilities versus portability. Multimedia notebooks provides high quality audio video especially when using 3-D games yet these devices are typically larger and heavier than non-multimedia notebooks. It was also found the battery drains much faster when using multimedia features thereby limiting the ability to work on the run.
Price vs. brand names. It was also found that notebooks with well-known brands have better design and stronger outer shell as well as better after-sale services, however, they are priced higher than generic brands.
We concluded that it is important to decide ahead oftime what you want from a notebook before buying one. Those who are looking for the best sound and graphic  ,,-; __ ..i..,; .<. Ull · ~ n:, 1 11.. ..1.u16..ll   He was particularly liked the amazing flexibility of the frame where it could be stretched 180 degrees easily without breaking. He was happy with the extremely lightweight (3 grams) because he has no longer feel it on his nose, or see any marks on his upper nose. This price was high but he believed it worth it.

Incident # 2.
Although drinking salted yogurt drink was popular in Kuwait, this female student did not like any of these products since she was young. At one time, she tasted a new yogurt product called (Nadek) imported from a neighboring country (Saudi Arabia) and she liked it ever since. She specifically liked the taste and the thickness of the drink. After discovenng this brand, she was ever since a loyal customer, even if she has to pay a higher price than other cheaper competing brands.
Incident# 3. The subject of this unsatisfactory experience was a female student shopping in one of the most elegant shopping centers in Kuwait. The student was shopping with her brother in one of the shops with a beautiful interior design and decoration when she found nice selection of well-known brand of shoes with really good prices. They bought two pairs of shoes and using them at home, they found the two pairs to be uncomfortable and the label was counterfeited and was used to hide the name of the real country of origin, which is famous for counterfeiting popular brands. When they attempted to return the product, the saleswoman met them with a very angry face and rough treatment saying that one of the conditions posted throughout the shop and on the sales invoice was that "sold goods can not be returned or exchanged".

Incident # 4.
This experience was related to the use of mobile phones. This female student has an unlimited loyalty to one particular model of phones sold by the only cell Phone Company in Kuwait. She said that although the cell phone she likes a bit heavy and larger in size compared to the other phones, she likes the crystal clear connection and the quality of the speakers when calling anywhere in the country. She jokingly said that her father tried to offer her another phone and she turned his offer down.
Incident # 5. This unsatisfactory experience was related a brand new car the student bought "Dodge Intrepid" which had so many unexpected problems. Few days after purchasing the car, he noticed that the car shuts off few minutes after starting. He took it to the dealer to fix it. After many attempts to fix it, he was told that the car has no mechanical problems and the real problem is in the kind of fuel used. The chief of the maintenance department claimed the car requires unleaded gasoline and what is available in Kuwait is leaded, therefore, he should live with the problem and should not annoy them again. Few months later, he sold the car in the secondary market for great loss just to get rid of it. He vowed never to buy a Chrysler car again! Incident # 6. This female student cooks for her family and she admits that she is very picky about the ingredients. She frequently uses tomato pastes in her cooking and she believes that the cartoon-packaged Kuwait-Danish Diary (KDD) brand to be the best in the market. She said that she mistakenly bought an identically cartoon-packaged tomato paste produced in a neighboring country (Bahrain) and she immediately recognized the thin texture and fade red coloring when she mixed it with the rest of the ingredients. She said that she missed the tomato taste she usually feels when eating the food. She said the price difference between the two brands were negligible. One of the participants asked if she ever used canned tomato paste, she said that she did not like them because the container swells because of the intense heat which makes opening them quite messy. The paste also changes color and become more acidic.
Incident # 7. This student was happy about how she was treated in the US when bought things and complained about the quality. She traveled to the US last year and while shopping in one of the malls, she found what she though to be a nice pair of shoes. She bought it, but upon using it, she found the leather behind her ankle to be very irritating and lead to many sores. When she visited the mall again, she went to the shop, without the pair of shoes, only to complain about her experience so they can inform the producer about it. She was pleasantly surprised when the salesperson offered to get her money back or to exchange it with any other pairs in the shop. The student stressed the fact that she has already used it many times but she even surprised that the salesperson did not mind at all. She did not return it back because she was uncomfortable thinking that she is taking advantage of the store if every thing bought and used could be returned for full money back.

155
APPENDIX 5 (cont.) Incident # 8. A water pump was the focus of this unsatisfactory incident. Four years ago, the student wanted to buy a water pump to improve the water pressure in his house, and since then the pump ran very well without a problem. A few days ago, the pump broke down because of heavy winter rain soaked the pump and caused an electrical short circuit. A technician from another company was called to inspect it. He suggested that the pump was old and it is a good time to replace it with new pump that operates automatically whenever the water pressure drops when opening a water fountain. Eager to use this technology of convenience, he bought the new pump. Two days later the pump unexpectedly stopped and he contacted the company' s technician. Few days later, he had the same trouble, and made an appointment this afternoon to discuss the problem with the company. He knows that he won't get his money back and regretted disposing the old pump for a failing technology.
Incident# 9. The product reported in this unsatisfying incident was a "teeth cleaning machine" promoted in an infomercial aired by a satellite channel. The product claimed to ' clean all deposits in the teeth when used according to instructions. The student read the instructions carefully and used the machine. She experienced gum bleedings and tooth was drilled, not cleaned. To her dismay, the painful experience did not lead to cleaner teeth as promised by the ad. She could not return the product because the local agent refused to do so.

Incident# 10. This incident involves a student who is very happy about how the "Sanyo
Memory" day planner worked. He was particularly pleased with the large storing capacity of the machine, it's small size, its ability to be programmed using any regular computer, and the possibility of adding more memory. He said that it made his life easier.
Incident# 11. A female student was totally dissatisfied about the quality of a " Walk Master" exercise machine she has recently bought. The machine she used caused a sever pain in a group of muscles located at the lower part of the feet which caused her to stop exercising. The machine is electricity powered, and that, according to her, may have shifted the workout pressure to the wrong group of muscles.

Incident # 12.
The satisfying product in this experience was a set of reusable contact lenses. His tried to use optical glasses but found them be uncomfortable, and especially heavy on the nose. He does not mind the tedious routine of daily washing and disinfecting, and he does not prefer wearing the disposable lenses because they are lesser in terms of quality.
Incident # 13. The student was very dissatisfied with the quality of high priced pens. He said that these pens are beautifully designed (golden engraved and platted) and cost so much money but fail to provide nice lines. He was very upset with these pens when they stop working while he takes notes in the lecture.

Incident# 14.
This student has been delighted with the concentrated milk he drinks every morning. His experience with his favorite brand "Carnation" started when he was young. The taste is very delicious and he can' t get this kind of satisfaction from any other brand. He said that when the product is in short supply, he usually drives long distances, even traveling, to get it.
Incident# 15. This male student has been frustrated about many children toys he bought last time. He is particularly angry at how easy these new toys break in his children's hands in very short period of time. In his opinion, this is not a coincidence. He explained that toy companies sell inferior toys so parents has to repurchase new ones, especially when many brands are produced by the same company.
Incident# 16. For the second time, the cellular phone was the subject of unsatisfying incident. This male student said that his new phone has a very loud speaker because the magnet used in the speaker was far more powerful than speakers of other competing brands. He r~ad few articles in the papers about the grave health ramifications of exposing to such fields. To demonstrate to the group how powerful the magnet was, he took one of his keys out and was able to left the whole phone by attaching the keys to the speaker magnet.

Incident # 17.
A male student has complained about the quality of frozen food product he bought for himself and his family. The delicious photos displayed on the outer package lured him to buy the product, but was disappointed when he prepared them at home. " ... the food was partially thawed in the center, and the taste was really awful", he said. He couldn' t return tern back because the package is opened. One participant suggested using the microwave oven instead of the traditional oven may make the food cook evenly, but the student expressed grave concerns over the radiation emitted from these types of ovens.

Incident # 18.
A cellular phone was also the subject of another incident but in negative sense. The male student purchased a small-size mobile phone with a car installation kit. Rather than scheduling an appointment with the company to install the kit, he took it to a company-certified technician. When he used the phone on the freeway, the quality of sound is far less than if the whole unit is detached and used separately. He is bothered by this problem but he is less motivated to go to the technician to fix it because he doubts his competence.
APPENDIX 6 (cont.) Incident # 5. Another student related that she had recently purchased a printer to use with her computer. After doing some research on printers she decided to buy a Hewlett Packard HP820 because of the reputation of the manufacturer and several reviews of the printer that she read. She was looking for a computer, which would print with very crisp black and has found that this printer offers the quality of black printing, which was sought. Much to her pleasant surprise, she has found the color printing to also be very sharp and the speed of the printer to be much faster than she had imagined it would be. She is quite happy with the purchase.
Incident# 6. Number six was a male student who had purchased a stereo set which included all of the components required for playing music except a turntable. The brand he chose was Sony because of its reputation for electronics. The compact disk player is now broken. It will spin around and not allow any of the disks to play. The problem may have been caused by abuse from one of this student' s housemates. He lives in a house with 30 other men. He is frustrated because he cannot reach the store to try to get the CD player fixed, but he says that he would buy Sony again because he does not think that the breakdown was the machine' s fault.