Forthcoming at the Journal of"Personality & Sodal Psychology (2001) CONSUMPTION SYMBOLS AS CARRIERS OF CULTURE: A STUDY OF JAPANESE AND SPANISH BRAND PERSONALITY CONSTRUCTS Jennifer Aaker University of Stanford Veronica Benet-Martinez University of Michigan Jordi Garolera Universität Pompeu Fabra This research argues that the meaning embedded in consumption symbols, such as commercial brands, can serve to represent and institutionalize the values and beliefs of a culture. Relying on a combined emic-etic approach, we conducted four studies to examine how the symbolic and expressive attributes associated with commercial brands are structured, and how this structure varies across three cultures. Studies 1 and 2 revealed a set of 'brand personality' dimensions common to both Japan and the United States (Sincerity, Excitement, Competence, and Sophistication), as well as culture-specific Japanese (Peacefulness) and American (Ruggedness) dimensions. Studies 3 and 4, which extended this set of findings to Spain, yielded brand personality dimensions common to both Spain and the United States (Sincerity, Excitement, and Sophistication), plus non-shared Spanish (Passion) and American (Competence and Ruggedness) dimensions. The meaning of the culturally-common and -specific brand personality dimensions is discussed in the context of cross-cultural research on values and affect, globalization issues, and cultural frame shifting. "The Marlboro Man is an egoistic ideal; at home in his universe, master of his destiny. Thus, the Marlboro Man has come to symbolize individualism and independence. " (Vacker, 1992) Traditional research in both cultural and cross-cultural psychology has focused on culture-based effects by identifying the influence of culture on the individual (culture-affects-psyche; see Cooper & Denner, 1998). However, the reverse relationship also exists; individuals influence culture (psyche-affects-culture) by the creation of institutions, symbols, and practices that carry and validate particular cultural meaning systems (DiMaggio, 1997; Kitayama et al., 1997; Shore, 1996). In this research, we rely on this bi-directional conceptualization of culture to examine how cultural meaning is represented in the minds of individuals. We argue that, similar to cultural icons (e.g., Hong et al., 2000), reasons (e.g., Briley, Morris, & Simonson, 2000), and public messages (e.g., Kim & Markus, 1999), consumption symbols such as commercial brands (e.g., Marlboro cigarettes) can serve as carriers of culture. That is, the meaning embedded in brands can serve to represent and institutionalize the values and beliefs of a culture. To test this premise we raise the following question: To what degree are the symbolic and expressive attributes that people perceive in socially-constructed entities, such as commercial brands, organized similarly or differently across cultures? More specifically, to what degree do basic dimensions of brand personality, defined as a set of human-like attributes associated with a particular brand (Aaker, 1997), carry universal or specific cultural meaning? Insight into this question will shed more light on the degree to which culture and psyche are mutually constituted and how culture-specific and universal human needs are carried through the creation, perception, and use of non-human symbolic objects such as brands. Further, from a more applied perspective, the role that culture may play in people's perception of consumer goods needs to be examined against the assumption that market globalization makes all of us psychologically more similar (Hermans & Kempen, 1998). The Dynamic Role of Culture and the Meaning of Commercial Brands Much of the research in cross-cultural psychology has conceptualized culture as a broad, domain-general, and stable set of value tendencies (e.g., individualism-collectivism, power distance; Hofstede, 1980). In this light, the portrayal of culture Jennifer I „ Aaker, Stanford Business School, Stanford University; Veronica Benet-Martine^, Department of Psychology, University of Michigan; jordi Garolera, Department of I economics and Business, Universität Pompeu tabra. 7 'he ßrst two authors contributed equally and are listed in alphabetical order. I 'he authors would like to thank the following individuals for their insight and intellectual support regarding this work: Alexander Biel, Montserrat Bonamusa, Steve Cole, Aimee D rolet, Asako Hoshino, Hee/ung Kim, Uiona I jee, Michael Adorns, and Ravi Pillai. We also thank Research International, Marketing Intelligence Corporation and Marketing Science Institute for their financial support. Correspondence concerning this manuscript should be addressed to Jennifer I ^ Aaker, Stanford Business School, 518 Memorial Way, Stanford University, Stanford, CA 94-305-5015, or electronic mail may be sent to tmKeni-u.i>u>.numuiu.euu or veiutuiuwumiaj.edu. 2 AAKER, BENET-MARTÍNEZ AND GAROLERA is of an abstract, encompassing structure, one that is often indexed by nationality and examined in light of its influence on individuals' behavior. Another perspective is that culture is more fragmented and dynamic, a set of subjective contexts and situations that are constructed and experienced by the individual (Cross & Markus, 1999; Hong et al., 2000). Two key issues within this perspective are that: (a) culture is best conceptualized in terms of the meaning derived from and added to everyday experience, and (b) individuals and culture are inseparable and mutually constitute each other. In light of these views, the study of how cultural meaning and individual psychological tendencies influence each other becomes critical (Shweder & Sullivan, 1990). In the present research, we suggest that one way to study the mutual constitution of the individual and culture is by examining the structural properties of non-human, symbolic objects such as commercial brands. Commercial Brands: Carriers of Cultural Meaning Referred to as "consumption symbols" or cultural icons (McCracken, 1986), commercial brands have significance that goes beyond their physical properties, utilitarian character, and commercial value. This significance rests largely in their ability to carry and communicate cultural meaning (Douglas & Isherwood, 1978; Richins, 1994). Culture-specific meaning typically resides in the more abstract qualities of the commercial brand that provide primarily symbolic or value-expressive functions to the individual (Shavitt, 1990), what is commonly known as 'brand personality' attributes. That is, in contrast to the utilitarian attributes associated with commercial brands (e.g., Levi's jeans are durable) which tend to demonstrate limited variability in meaning or importance across cultures (Aaker & Maheswaran, 1997), the symbolic and expressive functions provided by a brand (e.g., Levi's allows for the expression of independence, strength and masculinity; Solomon, 1986) tend to vary to a larger degree due to the fact that individuals vary in their needs and self-views (Fiske et al., 1998; Han & Shavitt, 1994; Kim & Markus, 1999). The process by which material objects come to possess meaning has been studied in detail by anthropologists (e.g., Douglas & Isherwood, 1978; Levi-Strauss, 1966, p. 116, Solomon, 1986). One institution that has received attention in the context of commercial brands is advertising, which works as a method of meaning transfer by bringing the consumer good and a representation of culture together within the frame of an advertisement (McCracken, 1986). The mechanics of this method begin with the advertising agency charged with the promotion of the commercial brand. Based on marketing research where individuals are asked what characteristics of the commercial brand are important to them and what needs are served by the commercial brand, advertisers determine what characteristics of the brand will be communicated in the advertisement (Lannon, 1993; Plummer, 1985). In this light, individual needs serve to influence the creation of brand meaning. At the same time, however, the communication of these cultural icons in advertisements influences reality and ultimately individuals' attitudes and behavior (Belk & Pollay, 1985; Kim & Markus, 1999; Shore, 1996). Thus, the bi-directional relationship between culture and the individual is captured in both the process of creating the commercial brands and the process by which brands are communicated to and used by individuals. Note that the above processes of cultural meaning creation and redefinition occur over time and involve many different fragments of society (e.g., consumers, companies, technology, political and cultural institutions). Given this complexity, it is difficult to design specific studies to explicitly model these mechanisms and their directionality that are not de-contextualized or over-ambitious. Accordingly, in the present research, we focus instead on providing insight into a slice of this phenomenon by examining some of its perceptual and structural elements: how individuals organize the symbolic and expressive attributes associated with commercial brands and how this organization may vary across cultures. To serve as a basis for the current research, we draw on work that has explored the meaning of commercial brands by examining how brand personality attributes are structured in the minds of individuals in the United States (Aaker, 1997). In this research, the process of meaning identification involved a set of studies whereby individuals were asked to rate a representative set of commercial brands on a battery of personality attributes. Results of exploratory and confirmatory factor analyses showed that American individuals perceive brand perceptual space in terms of five personality dimensions (see Figure 1). These dimensions include: Sincerity, represented by attributes such as down-to-earth, real, sincere and honest; Excitement typified by attributes such as daring, exciting, imaginative and contemporary; Competence, represented by attributes such as intelligent, reliable, secure and confident; Sophistication, represented by attributes such as glamorous, upper class, good looking and charming; and Ruggedness, typified by attributes such as tough, outdoorsy, masculine and western. Note that at least three of the above dimensions (Sincerity, Excitement, and Competence) resemble personality dimensions that are also present in human personality models such as Big Five.1 Specifically, CARRIERS OF CULTURE 3 Sincerity is defined by attributes related to warmth and honesty that also are present in Agreeableness, Excitement captures the energy and activity elements of Extraversion, and Competence denotes dependability and achievement similar to Conscientiousness. The links between Sophistication and Ruggedness and the Big Five are less clear however. Compared to Sincerity, Excitement, and Competence (which seem to capture relatively basic tendencies that may apply to both humans and brands), Sophistication and Ruggedness capture more aspirational images associated with wealth and status (e.g., Lexus automobiles, Monet jewelry) or American individualism (e.g., Levi's jeans, Harley-Davidson motorcycles) that may be more specific to carriers of culture such as commercial brands. In the current research, we examine the extent to which Aaker's (1997) structure of personality attributes associated with commercial brands differs across cultural contexts; that is, how much do Sincerity, Excitement, Competence, Sophistication and Ruggedness connote culture-specific vs. more universal meaning? In addressing this question, we hope to provide insight into the degree to which cultural meaning, as influenced by individuals within a cultural context, is conveyed and consequently communicated to individuals both within and across cultural contexts (Bond, 1994b). Values and Cultural Products One literature that may contribute insight on this question is that on values. Schwartz (1994), for example, proposes a taxonomy of seven distinct types of cultural-level values organized around the two dimensions of: (1) Conservatism vs. Autonomy that relate to social conservatism vs. openness to change and (2) Hierarchy/Mastery vs. Egalitarian Commitment/Harmony that relate to self-enhancement vs. self-transcendence (Schwartz, 1992). The seven value types, Conservatism Intellectual Autonomy, Affective Autonomy, Hierarchy, Mastery, Egalitarian Commitment, and Harmony, were identified through a psychometrically rigorous procedure involving more than 60 cultural groups (Schwartz, 1994; Schwartz & Bilsky, 1990; Schwartz & Sagiv, 1995). Note that although these dimensions capture universal needs, cultures vary considerably in their standing along these dimensions (see Table 7.3 in Schwartz, 1994). These differences in the locations of cultures along the above seven value dimensions reflect differences in the degree to which each value type is embraced by a particular culture. To illustrate, valuing the social aspects of Mastery (self-assertion and getting ahead of other people) seems particularly important in the United States. In contrast, more collectivistic societies such as Asian and Latin cultures stand out as placing particular emphasis on Harmony needs (keeping balance and peace with nature and people). There is some variation within collectivist cultures, however, in their value discrepancies with the United States. Southern Mediterranean cultures such as Spain, Greece, and France, for instance, have particularly high scores compared to both the United States and Asian cultures on Affective Autonomy (valuing novelty, creativity, and having an exciting life) and Egalitarian Commitment (voluntary commitment to promoting the welfare of others). Note that one particularly useful aspect of Schwartz's value taxonomy is that country differences such as those we just described, can be used to interpret cultural differences in norms, attitudes, behavioral patterns, and important macro socio-economic variables (e.g., Gouvenia & Ross, 2000; Schwartz, 1994; 1999; Schwartz & Ross, 1995). The attributes that structure the meaning of commercial brands in the United States (Aaker, 1997) seem to align themselves with several of Schwartz's cultural value types for which the United States has moderate to high scores. For instance, a close inspection of the attributes that define Sincerity (e.g., family-oriented, real, small-town), suggests that this dimension may capture brand perceptions associated to Conservatism needs (emphasis on family security and safety, being stable and polite). Terms defining Excitement (e.g., unique, exciting, young), on the other hand, suggest a link with Affective Autonomy needs (valuing novelty and creativity, having an exciting life). Competence (e.g., reliable, successful, intelligent) appears to be related to Mastery needs (emphasis on being capable and successful, demonstrating competence), and Sophistication (e.g., upper class, glamorous, smooth) to Hierarchy needs (value of social status and prestige, having wealth). Finally, Ruggedness (e.g., masculine, tough, western) appears to be less directly related to a specific value orientation, although some of the attributes may encompass elements from Mastery (being independent, daring) and low Egalitarian Commitment (detachment from others). This dimension is reflected in popular American movies (The Quiet Man, Stagecoach and High Noon; Kim & Markus, 1999) as well as popular American commercial brands (Harley-Davidson, Marlboro, Levi's; Solomon, 1986), and appears to represent institutionalized American values such as strength, masculinity, and ruggedness. One way to assess the particular cultural significance of Ruggedness relative to the other four dimensions and Aaker's (1997) findings in general is to compare the American dimensions against those uncovered in other cultures. By doing so, the possible culture-specific psychological values and 4 AAKER, BENET-MARTÍNEZ AND GAROLERA needs served by commercial brands in the United States and other cultures can be more clearly ascertained. In the present research, we specifically address two potential hypotheses. Both are based on the premise that commercial brands are symbols that can carry cultural meaning (McCracken, 1986; Richins, 1994); however, they differ in their predictions of the degree of cross-cultural similarity in the perceptual representation of the brands. The first possibility is that the perceptual structure may remain largely robust across cultural contexts. That is, since the basic kinds of values held by individuals as well as the organization of these values, i.e., their inter-correlation pattern, tend to be similar across cultural contexts (Schwartz, 1992; 1994), the meaning conveyed in commercial brands may also be largely universal. That is, the number and nature of the basic dimensions that organize brand personality perception will be similar across cultures given that the kinds of values people have (and may seek to fulfill through commercial brands) are also universal. Dimensions very similar to those uncovered by Aaker (1997) in the United States should therefore also emerge when the structure of brand personality perception is examined in other cultures. An alternative possibility, however, is that different cultures have somewhat unique organizations of the brand representational space that are reflective of cultural differences in value emphasis. In other words, it is possible that the structure of brand meaning perception is mainly associated with the importance of the value that brands provide for consumers in a given culture. If indeed brand meaning is created to reflect the needs and values held by individuals within a culture (McCracken, 1986), there may be some cross-cultural variance in the meaning connoted in commercial brands and the organization of this meaning (e.g., number and nature of the basic dimensions). For instance, as discussed earlier, Schwartz (1994) shows that Harmony is a value that is endorsed by East Asian cultures to a greater degree than Western cultures such as the United States. Indeed, keeping balance or maintaining harmony is respected as one of the highest virtues by Confucius (Kim & Markus, 1999). Further, the interdependent goal of harmoniously fitting in with others, with its emphasis on fulfilling various social roles and maintaining connections with others, plays a larger role in determining overall life satisfaction in East Asian relative to North American cultures (Kwan, Bond, & Singelis, 1997; Oishi et al., 1999). These findings suggest that, to the degree that a particular value type such as Harmony varies in its importance across cultural contexts, we may observe evidence of culture-specific meaning that relates to this particular value in cultures that embrace allocentric beliefs and harmony-oriented values (Fiske et al., 1998; Schwartz, 1994). In sum, although the research reviewed above does not allow us to predict a specific perceptual structure of commercial brands, it does suggest that there may be some cultural variance in how they are represented in the minds of consumers. Cultural variation in values and needs may influence commercial brand perception in two inter-related ways: by influencing the content of marketing communications that are used to create and develop commercial brands and, at the same time, by influencing the kinds of attributes individuals focus on when perceiving brands (Belk & Pollay, 1985). It is through these processes that cultural differences in the structure of brand personality perception may arise. The current research relies on a combined emic-etic approach to determine the degree to which individuals across cultures share a similar perceptual representation of commercial brands. Methodological Overview Choice of Countries Many cross-cultural researchers have argued that multiple cultural groups are needed in order to disentangle the influences of the various cultural dimensions that may underlie the observed differences (Bond, 1994a). The present research focuses on two countries, an East Asian culture (Japan) and a Latin culture (Spain). These two countries were chosen for several reasons. First, relative to members of Anglo American cultures, individuals from East Asian and Latin cultures tend to be less idiocentric and more allocentric (i.e., higher in desire for interdependence and harmony; Marin & Triandis, 1985; Oishi et al., 1999, Schwartz, 1994; Wierzbicka, 1991, but see Matsumoto, 1999; Takáno & Osaka, 1999). These value differences may relate to variation in brand personality perception. Second, although individuals in Japan and Spain cultures both score relatively high on allocentrism, they differ on other dimensions. Perhaps most notably, individuals in Latin cultures, relative those in East Asian cultures, place special value several socio-emotional behaviors related to Affective Autonomy needs, namely, sensation seeking (McVeagh, 1990), emotional intensity (Benet-Martinez, 1999), and simpatia (Triandis et al., 1984; for more details, see Marin & Marin, 1991; Marin & Triandis, 1985). Therefore, the selection of Japan and Spain provides a context that allows for potential replication since both cultures share an endorsement of allocentric values, but also an extension whereby the values unique to Mediterranean cultures such as Spain may be identified. Finally, from a methodological standpoint, Japan CARRIERS OF CULTURE 5 and Spain are in similar stages of industrial and economic development and spend approximately the same percentage of the GNP on advertising as compared to the United States (1.0% for Japan, 1.5% for Spain, 1.1% for the United States). Thus, several variables that could account for cultural differences in communication styles and possibly bias the results of this research will be kept relatively constant. The Combined Emic-Etic Approach An important issue in cross-cultural research is the distinction between emic (indigenous) and imposed-etic (imported) approaches to data collection (Berry, 1969). Emic approaches explore a particular psychological construct from within the cultural system, whereas imposed-etic approaches study behavior from outside the cultural system. With the emic approach, instruments, and theories indigenous to the target culture are developed by relying on a systematic process that generates a set of culture-specific attributes and stimuli. Imposed-etic approach instruments, in contrast, are either imported in their original form or translated into the local language (Enriquez, 1979). The question of whether imported (i.e., translated) measurement tools overlook important domains of the local culture is the foundation of a classic debate in cross-cultural psychology, the emic-etic issue (Berry et al., 1992). On the one hand, an imposed-etic strategy is useful in that it makes cross-cultural comparisons feasible given that quantitative judgments of similarity require stimuli that are equivalent, but its use may distort the meaning of constructs in some cultures or overlook their culture-specific (emic) aspects. On the other hand, an emic strategy is well suited to identify culture-specific qualities of a construct, i.e., it is ecologically valid. However, its use makes cross-cultural comparisons difficult. Given the opposing advantages and disadvantages of the emic and etic approaches, one solution to the emic-etic debate has been to pool both approaches into what is known as a combined emic-etic approach (Hui & Triandis, 1985). This approach, compared to emic or imposed-etic approaches, provides a more complete and unbiased picture of the degree of cross-cultural overlap and specificity between constructs (for examples, see Benet-Martinez & Waller, 1997; Church & Katigbak, 1988; Yang & Bond, 1990). In our study, the application of a combined emic-etic approach involves the following steps: First, indigenous attributes relevant to the target concept (e.g., commercial brands) are isolated in the new cultures and their underlying dimensional structure identified (Japan in Study 1; Spain in Study 3). Next, using an independent set of participants, this set of emic-based attributes is combined with attributes identified in the United States, and the overlap between the emic and imposed-etic dimensions underlying these two sets of attributes is measured (Aaker, 1997; Studies 2 and 4). This approach does not bias the results in favor of universality, an outcome that is often associated with the imposed-etic approach (McCrae & Costa, 1997). Further, it is more consistent with the perspective of culture adopted in this research, where cultural knowledge is a "lens" that colors people's perception of objects and messages in the environment (McCracken, 1986). By allowing for cultural variations in the form or meaning of personality attributes to be represented (Church & Katigbak, 1988), the emic-derived set of attributes is more likely to reflect the culture-specific lens through which people see. STUDY 1 Identification of Indigenous Japanese Brand Personality Dimensions The objective of Study 1 was to determine how Japanese individuals perceive the perceptual space of commercial brands as defined by personality attributes. We first generate a set of culture-specific attributes and stimuli, and then identify the perceptual representation of brands through a factor-analytic procedure involving attribute ratings on a set of brands by Japanese individuals. Method Stimuli Selection Two criteria guided the selection of commercial brands to serve as stimuli. First, to enhance the representativeness of the sample of stimuli, we selected commercial brands in product categories that serve both symbolic and utilitarian functions. Therefore, we randomly selected 24 product categories that were shown to vary on these two functional dimensions (Ratchford, 1987, Appendix). Six of the categories were highly symbolic or value-expressive (e.g., apparel, alcohol, fragrances), six were utilitarian (e.g., laundry detergent, medication, toothpaste) and twelve scored relatively high on both symbolic and utilitarian dimensions (e.g., automobiles, beverages, toys). Second, to enhance familiarity of the sample of stimuli, well-known commercial brands were selected. Thus, a pretest was conducted where Japanese participants (n = 46, 50% female, mean age = 30.2), who were graduate business students enrolled in a full-time MBA program, were invited to participate in a study on brands. Paid $7 for their participation, the participants were asked, "What is the first brand that comes to mind when you think of this product 6 AAKER, BENET-MARTÍNEZ AND GAROLERA category"? The most frequently listed brands in each of the 24 categories were identified. Although the relatively large number of brands allows for greater variance in brand personality types, it also increases the chance of participant fatigue. Thus, to minimize potential fatigue, the 24 brands were randomly grouped into six sets of four brands. Each group was composed of one symbolic brand, one utilitarian brand and two symbolic/utilitarian brands, such that each brand group contained a similar profile of brands. For example, Group 1 contained Suntory Old whiskey, Denter T toothpaste, Pocari Sweat beverage, and Mercedes Benz automobiles. In this way, the brand groups' profile was similar to that of the total sample of brands. Finally, one well-known brand (Coca-Cola) was used as a control and included in each of the groups in Study 1 to assess the variation of perceptions of personality attributes for a given brand across groups. Thus, the result was a set of 25 brands that were meaningful to the target culture. Personality Attribute Selection The selection of brand-related attributes followed a three-step process similar to the one used in Aaker's (1997) study. First, to ensure familiarity and relevance of the attributes, a free-association task was conducted where Japanese participants (n = 50, 40% female, mean age = 28.2) were asked to write down all the personality attributes that first come to mind when thinking about well-known brands in 10 product categories (three symbolic, three utilitarian, and four symbolic/utilitarian), a process that yielded 138 attributes. Second, to maximize the content representation of personality attributes, 71 additional attributes were compiled from three sources that rely on brand personality research in Japan (Japanese advertising agency, client company, and research supplier) and 44 more that were representative of the Big Five personality dimensions (e.g., John, Donahue, & Kentle, 1991), as in Aaker (1997). Finally, from the total set of 253 personality attributes, three groups of attributes were eliminated because they were redundant (n = 61; e.g., reliable arose from the free association task as well as from John et al., 1991), ambiguous (n = 25; e.g., slight, unfocused, rigid), or relatively irrelevant to the construct of interest (n = 67; e.g., artistically sensitive, fickle, hypochondriacal).11 Thus, the result of this stage was the identification of 100 attributes. Participants To enhance generalizeability, a sample (n = 1495) that represented the Japanese population with respect to five demographic dimensions was used (gender, age, marital status, education level, and occupation; Japan Statistics Bureau and Statistics Center, 1996). To illustrate, 51% of the sample was female, 14% of the sample were 20-30 years of age, 56% of the sample were married, 46% of the sample had a college or graduate school education, and 13% of the sample were professional or technical workers.111 The participants in each of the brand groups were selected to have the same profile as the total sample (n ranged from 243 to 253 in each of the six brand groups), and belonged to a Japanese national mail panel. Procedure Participants, who were paid 500 yen, were asked to participate in a study about people's impressions toward particular brands (names of commodity goods or services). To communicate the brand personality construct and enhance the imaginability of the task (Lannon 1993; Plummer, 1985), participants were asked to think of the brand as a person. Specifically, they were told, "If I asked you to give me your impression of a particular person, you might answer with a set of personality attributes. Now, let's think about brands in the same way. For example, you may be asked to rate the extent to which a set of attributes describes Porsche. Please ask yourself, 'If Porsche was a person, how would you describe him/her?', and then circle a number between to "not at all descriptive" (1) to "extremely descriptive " (5) for the subsequent set of attributes". Then, participants were asked to rate the extent to which the 100 personality attributes describe a specific brand. Participants repeated the rating task for the four additional brands in the particular brand group. Thus, six sub-samples of participants rated five brands (Coca-Cola being common in each group), a task that took approximately 50-60 minutes. To illustrate, Group 1 contained Chanel fragrance, Yomiuri Shimbun newspaper, Nintendo toys, Kuroneko Yamato delivery services, and Coca-Cola soft drinks. To control for primacy and recency effects, the order in which the attributes were presented for each brand was counter-balanced as was the order in which the brands were presented in the questionnaire. Results and Discussion First, to assess the variation of perceptions of personality attributes for a given brand, we examined the mean ratings of Coca-Cola across the groups. No significant differences were found, suggesting high levels of agreement of the human characteristics associated with a particular brand. Second, to examine the systematic individual differences in perceptions of brands in general, the correlation matrix for the brand personality traits (n = 100) across individuals' ratings of each brand was CARRIERS OF CULTURE 7 subjected to a principal component analysis followed by Varimax rotation. The first ten eigenvalues from the 100 x 100 inter-item Pearson correlation matrix were 28.2, 9.5, 5.7, 3.7, 2.9, 1.3, 1.1, .8, .7, and .6. The moderate break after the fifth latent root suggested that a solution with five components was plausible. The adequacy of this solution was supported by the following criteria: (a) shape of the scree plot, (b) stability of the solutions in separate principal components analysis with distinct sub-samples (e.g., males vs. females, older vs. younger individuals), (c) meaningfulness of the dimensions (at least nine traits loaded on each of the first five factors whereas only one trait loaded on the sixth component ("family-oriented"), (d) amount of variance explained by the five components (78%) relative to dimensions six through nine (under 2% each). The five-component solution is reported in Table 1.1V Labels for all the dimensions were selected based on the attributes emphasized within each component. To provide English translations, a six-person back-translation team translated each of the personality attributes. First, a three-person translation team (one native Japanese speaker, one native English speaker and one bilingual speaker) discussed the linguistic meaning of each attribute before final translation into English. Then, the three-person back-translation team followed the same process, translating the attributes back into Japanese to ensure accuracy (Brislin, 1970). Inter-coder agreement was high (94%). That is, for 94 of the 100 traits, the Japanese trait was back-translated as the same word as the original. When discrepancies existed, all six coders discussed them until consensus was formed. A close look at Table 1 reveals that all but one of the Japanese brand personality dimensions are highly isomorphic to the American dimensions reported in Aaker (1997): Dimension I clearly represents Excitement and is primarily defined by attributes such as funny, contemporary, young and energetic. Interestingly, several of these terms are also markers of Excitement in the American brand personality structure. Dimension II (Competence) is defined by attributes such as responsible, reliable, confident and tenacious - consistent with the markers of Competence in the U.S. Dimension III (Peacefulness), on the other hand, is defined by a unique blend of attributes (e.g., shy, peaceful, naive, dependent) reflective of an allocentric and harmony-fostering orientation (Schwartz, 1994). Dimension V (Sincerity) includes warm, thoughtful and kind, markers that are in line with those found in the U.S. for Sincerity. Finally, Dimension IV (Sophistication) is defined by terms such as elegant, smooth, stylish and sophisticated markers that are consistent with those found in the U.S. for Sophistication. Identification of Facets and Markers Because the full set of 100 attributes may be too lengthy to manipulate and measure in subsequent research, a more limited set of attributes that reliably captures each dimension was desired. To achieve this, we first identified the different facets subsumed by each component via separate principal component analyses of the attributes within each brand personality dimension (see also, Costa & McCrae, 1992). Adopting this process, Aaker (1997) found a distinct set of facets that provided a structure to justify which attributes to select to represent each dimension as well as texture to understand the dimensions in greater detail. To illustrate, the American Sincerity dimension consists of four facets: Down-to-earth, Honesty, Wholesomeness and Cheerfulness (see Figure 1). The separate principal component analyses of terms within each dimension yielded a total of 12 facets: four for Dimension I, three for Dimension II, two for Dimension III, one for Dimension IV and two for Dimension V. Within each facet, we then selected the three attributes with the highest item-to-total correlation. Each of the resulting 36 markers (three attributes for each of the 12 facets) had high item-to-total correlations within its corresponding three-item facet and dimension (range from .80 to .94). Cronbach's alphas calculated for each of the five dimensions using the 36-item scale indicated high levels of internal reliability, ranging from .80 (Dimension III) to .90 (Dimension I). The final set of 36 Japanese brand personality markers and their corresponding facets and dimensions are depicted in Figure 2. To further ensure high levels of reliability, a small sample of Japanese participants (n = 60, 50% female, mean age = 31.3) were asked to complete the same questionnaire approximately 8 weeks after completing the original questionnaire. A total of 15 participants each rated four groups of five brands (Groups 1-4) over the two time periods. Test-retest correlations for the five dimensions defined by the 36 final markers were high, ranging from .81 (Dimension III) to .88 (Dimensions IV). In sum, the results of Study 1 suggest that the brand personality space for Japanese individuals is organized in terms of five dimensions representing Excitement, Competence, Peacefulness, Sophistication, and Sincerity. Although four of these dimensions appear to have overlapping meaning with those identified in the United States using North American stimuli (Aaker, 1997), a fifth dimension (Peacefulness) appears to be relatively indigenous to Japan. In Study 2, we empirically test this premise by directly comparing Japanese and North American brand representational structures. 8 AAKER, BENET-MARTÍNEZ AND GAROLERA STUDY 2 Overlap between Japanese and American Brand Personality Dimensions Study 2 was conducted with the primary objective of assessing the conceptual overlap between the Japanese brand personality dimensions identified in Study 1 and American brand dimensions (Aaker, 1997). A secondary objective was to test the robustness of the five Japanese brand dimensions on a different sample of Japanese participants. To accomplish both objectives, an independent sample of Japanese individuals rated a subset of brands using the Japanese attributes (English-translated) identified in Study 1, as well as the attributes that represent the American dimensions. Method Participants To gain confidence that the results found in Study 1 were driven by culture-based perceptions of brands rather than linguistic differences (Enriquez, 1979), the questionnaire was administered entirely in English rather than Japanese. Therefore, a slightly different profile of participants was used, one in which the participants were pre-selected to be bilingual. The sample of 114 Japanese participants was recruited from two sources: (1) Japanese students enrolled at a graduate business program at a large Japanese university (n = 56); (2) Japanese exchange students at a large U.S. western university, affiliated with the Center for East Asian Studies (n = 58). The latter sample had lived in the United States an average of 1.8 years (SD = 2.04). Participants were paid 500 yen (or $5) for their completion of the study. Participants who scored less than 4.0 when rating their written English knowledge ("1" = extremely limited, "5" = extremely good) were eliminated (n = 15), as were those who were not born in Japan (n = 9), leaving a total of 90 Japanese individuals (50% female, mean age = 31.9).v To minimize the problems that often arise from the cultural differences in the meaning and use of personality attributes, all Japanese words were given to participants in personal attribute form (listed in "Kanojo/Kare-wa......da," which corresponds to "It is......")• Procedure The cover story and structure of Study 2 was identical to that of Study 1 with two exceptions. First, participants rated each of the brands on 70 attributes, 42 markers of the five American brand personality dimensions and the 36 markers of the Japanese dimensions identified in Study 1 (minus 8 overlapping attributes: confident, contemporary, friendly, masculine, reliable, smooth, spirited and young). Second, only 10 brands were used; these brands were randomly selected from those used in Study 1 (Levi's jeans, Mercedes automobiles, Chanel fragrance, Coca Cola soft drinks, Mizuno sports apparel, McDonalds restaurants, Sony walkman, Nintendo toys, Seven Eleven stores, and Kleenex tissue). The order in which the attributes were presented was counter-balanced, as was the order in which the brands were presented in the questionnaire. Results and Discussion What is the overall degree of content overlap or specificity between the indigenous Japanese brand personality dimensions and the imported American dimensions? To address this question, we first examined the correlations among the indigenous and imported components. Scale scores representing each participants' rating of each brand on every imported and indigenous dimension were computed. The validity correlations between the conceptually-related dimensions were as follows: Sincerity (Japan) and Sincerity (U.S.) = .63; Excitement (Japan) and Excitement (U.S.) = .75; Competence (Japan) and Competence (U.S.) = .80; Sophistication (Japan) and Sophistication (U.S.) = .81. The size of these convergence correlations (mean = .75) contrasted markedly with the average off-diagonal discriminant correlations (mean = .29), suggesting both convergent and discriminant validity. The correlation patterns for the culture-specific Ruggedness (U.S.) and Peacefulness (Japan) dimensions were as follows: The highest correlation between Ruggedness (U.S.) and any Japanese personality dimension was .39 (with Japanese Competence), and the highest correlation between Peacefulness (Japan) and any American dimension was .41 (with U.S. Sincerity). Using Fisher's Z transformations, a statistical comparison of these two off-diagonal correlations with the validity correlations revealed that the two correlations, although sizable, are significantly smaller, suggesting that Ruggedness and Peacefulness are constructs that mainly capture culture-specific meaning. A question to bear in mind when evaluating the convergent and discriminant correlations reported above is the extent to which they reflect shared measurement error, shared meaningful (i.e., conceptual) variance, or both. One way to address this issue is to explore the latent structure of the variance shared by the indigenous Japanese and imported American scales via confirmatory joint factor analysis (CFA). V1 Relying on this methodology, we investigated the fit for a model with six latent components. Four of these six dimensions CARRIERS OF CULTURE 9 represented brand personality constructs common to Japan and the United States (i.e., Competence, Sophistication, Excitement, and Sincerity) and the other two represented culture-specific brand personality constructs (i.e., Ruggedness and Peacefulness).v" This model yielded adequate fit indices: %2 (20, N= 900) = 163, p < .001; CFI = .91, GFI = .92. Next, we compared the fit of our hypothesized six-component model against a more conservative four-component model that did not include culture-specific dimensions and instead represented American Ruggedness and Japanese Peacefulness as variations of Competence and Sincerity respectively (as suggested by the off-diagonal correlations for Ruggedness and Peacefulness reported above). This four-component model yielded unsatisfactory fit indices, %2 (26, N = 900) = 325, p < .001; CFI = .71, GFI = .79, and a significant decrease in overall fit, A%2 (6) = 626. These results support the idea that two culture-specific and four common latent dimensions may best represent unique and shared variance underlying the Japanese and American scales. In sum, the convergent-discriminant validity patterns derived from the correlational and confirmatory factor analyses suggest that there is considerable overlap between the dimensions organizing the American brand perceptual space and those representing the Japanese brand perceptual space. Specifically, moderate to high convergence was found between the Japanese and American dimensions representing Sincerity, Excitement, Competence and Sophistication. Two other dimensions, however, appeared to be more culture-specific: the Japanese dimensions of Peacefulness and the American dimension of Ruggedness. In other words, although Japanese perceptions of brands include meaning associations related to Peacefulness, Americans perceive brands to carry meaning relating to Ruggedness. These differences are in accordance with research suggesting that attributes and behaviors related to assertiveness are not as likely to be endorsed and nurtured in East Asian cultures (Church & Katigbak, 1988); rather, such associations are often devalued and discouraged (Wierzbicka, 1991). The presence of the Peacefulness dimension, on the other hand, may reflect the visibility that obedience, maintaining harmony, and interdependence has in Asian cultures (Triandis, 1989). Indeed, "wa" (loosely translated into harmony or peace) is "undoubtedly the single most popular component in mottos and names of companies across Japan" (Wierzbicka, 1991; pg. 354), whereas "rugged individualism" is a common theme found among many popular American brands (Solomon, 1986; Vacker, 1992). The patterns of cultural overlap and differences obtained in Study 2 are consistent with theorizing in the consumer behavior literature (McCracken, 1986) suggesting that the creation and nurturance of certain meaning associations in brands (e.g., Excitement) is often culturally-general, although other brand meaning associations may prove highly specific (e.g., Ruggedness). In Study 3, we examine the robustness of this finding by replicating the process adopted in Studies 1 and 2 in a new cultural context: Spain. STUDY 3 Identification of Indigenous Spanish Brand Personality Dimensions The objective of Study 3 was to test the generalizeability of the dimensional structure uncovered in Study 1 in a different cultural context. Specifically, we are interested in the following questions: To what degree will the perceptual space of brand personality in Spain also be organized around five dimensions? More importantly, given the Spanish culture's emphasis on interdependence values and allocentric beliefs, should a dimension similar to the Peacefulness construct uncovered in Japan also be expected? Finally, what is the likelihood that culture-specific Spanish brand personality constructs will emerge given Spain's unique cultural idiosyncrasies (Crow, 1985; McVeagh, 1990)? To address these questions, two studies that relied on emic and combined emic-etic methodology similar to that used in Studies 1 and 2 were conducted. Method Stimuli Selection A set of 25 well-known global brands was selected based on the identical criteria and process used in Study 1. The only difference was the specific brands in the set. For example, one group of brands contained Chanel fragrance (symbolic), Ariel detergent (utilitarian), NH and Melia hotel (symbolic/utilitarian), Volkswagen automobiles (symbolic/utilitarian), and Coca-Cola (constant across all brand groups). Personality Attribute Selection Personality attribute selection was also guided by the criteria used in Study 1. A free association task was conducted where Spanish participants who were economics or business undergraduate and graduate students (n = 36, 55% female, mean age = 25.1) were asked to list the personality attributes that first come to mind when thinking about the most salient brand in 10 randomly selected product categories identified in stimuli selection process (and 10 AAKER, BENET-MARTÍNEZ AND GAROLERA based on the same overall profile as in Study 1), yielding 128 attributes. Next was the addition of 64 attributes compiled from three sources that rely on brand personality research in Spain (Spanish advertising agency, client company, and research supplier), 44 markers that were representative of the Big Five personality markers (John et al., 1991), and 30 personality descriptors representative of Benet-Martinez's (1999) indigenous Spanish personality constructs. Finally, from the total set of 266 personality attributes, three groups of attributes were eliminated because they were redundant (n = 79), ambiguous (n = 16), or relatively irrelevant to the construct of interest (n = 94).vm Thus, the result of this stage was the identification of 77 attributes. Participants and Procedure. To enhance generalizeability, a sample (n = 692) was selected that represented the Spanish population with respect to five demographic dimensions: gender (62% female), age (mean =31.5), marital status (35% married), education level (30% of the sample had a college or graduate school education) and occupation (48% of the sample were professional or technical workers). The participants in each of the six brand groups were selected to have the same profile as the total sample (n ranged from 108 to 131 in each of the six brand groups). Participants belonged to a Spanish national mail panel and were entered into a drawing for a set of electronic products (five televisions and two VCRs). The identical procedure used in Study 1, including counter-balancing, was followed in Study 3. Results and Discussion As in Study 1, no significant differences were found in the mean ratings of Coca-Cola across the groups. To identify the individual differences in perceptions of brand personality dimensions, the correlations among the personality traits (n = 11) across individuals' ratings of each brand were factor analyzed using principal component analysis and varimax rotation. Replicating results from Study 1, a five component solution proved to be the most adequate to organize the covariance among the 77 Spanish brand personality descriptors. The choice of solution, like in Study 1, was based on the following criteria: (a) Scree plot (the first ten components were: 21.2, 6.8, 4.1, 3.2, 2.0, 1.3, 1.0, .9, .8, .6), indicating a moderate break after the fifth latent root, (b) stability of the solution in separate principal components analysis with distinct sub-samples, (c) meaningfulness of the dimensions (at least seven traits loaded on each of the first five factors; only one trait loaded on the sixth component ("ruggedness"), and (d) levels of variance explained (the first five dimensions explained 78%; dimensions six through nine explained under 1.2% each). The five-component solution is reported in Table 2. Labels for the five dimensions were selected based on the content of the dimensions. As can be seen in Table 2, Dimension I (Excitement) include markers such as outgoing, daring, young, unique, several of which are terms that also serve as markers of Excitement in the American and Japanese brand personality sets. Representative markers of Dimension II (Sincerity) include considerate, thoughtful, real and sincere, which are consistent with the markers of Sincerity in the U.S as well as Japan. Dimension III (Sophistication) is depicted by good looking, glamorous, upper class and stylish markers, which are consistent with those found in the U.S. for Sophistication. However, interestingly, another facet of Sophistication included confident, successful and leader (markers of Competence in the American model), which appear to be unique to Spain. Dimension IV (Peacefulness) includes markers such as affectionate, peaceful, naive and dependent, which are consistent with the markers representing Peacefulness in the Japanese cultural context. Finally, Dimension V (Passion) includes fervent, passionate, spiritual and bohemian as representative markers, consistent with the culture-specific findings highlighted in Benet-Martinez and Waller (1997)K Identification of Markers and Facets To identify a smaller set of personality attributes representative of each of the Spanish five dimensions, a facet analysis identical to that adopted in Study 1 was conducted. Second, five facet analyses identical to those in Study 1 were conducted. This analysis yielded a total of 11 facets: three for Dimension I, two for Dimension II, two for Dimension III, two for Dimension IV and two for Dimension V. To maintain high levels of reliability, three attributes with the highest item-to-total correlation were selected from each facet, leaving 33 attributes (three attributes for each of the 11 facets). Each attribute had high item-to-total correlations on the facets and dimensions (range from .70 to .84), thereby ensuring high internal consistency. Further, Cronbach's alphas that were calculated for each of the five dimensions using the 3 3-attribute scale ranged from .80 (Dimension III) to .91 (Dimension I), suggesting high levels of internal reliability. Finally, as in Study 1, an independent set of Spanish participants (n = 58, 60% female, mean age = 21.3) were asked to complete the same questionnaire approximately 7 weeks after completing the original questionnaire. Four versions of the questionnaire were used (n = 14-15 in each cell). The average Pearson correlation of the five CARRIERS OF CULTURE 11 dimensions as measured at Time 1 and Time 2 was .80 (ranging from .77 to .83). In sum, the results of Study 3 suggest that five dimensions representing Excitement, Sincerity, Peacefulness, Sophistication, and Passion organize brand personality attributes in Spain. Three findings appear particularly noteworthy. The first was the emergence of several components that convey meaning similar in nature to those previously found in the United States (i.e., Excitement, Sincerity) and Japan (i.e., Excitement, Sincerity, Peacefulness). Second, one dimension that appears to carry culture specific meaning emerged, Passion. Third, there was a blending of Competence associations into the Sophistication dimension in Spain. Study 4 was conducted to determine the degree to which these findings are robust across stimuli and participants, and to explicitly examine the degree of overlap between these indigenous dimensions and those found in the United States. STUDY 4 Overlap between Spanish and American Brand Personality Dimensions In Study 4, we compare the Spanish and North American brand representational structures by assessing their conceptual overlap at the dimension level via correlational and confirmatory joint factor analysis. Method Participants As in Study 2, a sample of 101 Spanish individuals was recruited from two sources: (1) Spanish students enrolled at a graduate program in Spain (n = 42) and (2) Spanish individuals living in the United States, affiliated with the Association Espanola de Silicon Valley (n = 59). The average time that the latter sample lived in the U.S. was 2.8 years (SD = 2.66). Participants were paid approximately $5 for their participation. Like in Study 2, participants who scored less than 4.0 when rating their written English knowledge were eliminated (n = 12), as were any participants not born in Spain (n = 3), thereby leaving 87 Spanish individuals (39% female, mean age = 25.3). Procedure A total of 10 brands from the overall set used in Study 3 were randomly selected (Rolex watches, Chanel fragrance, Marlboro cigarettes, Armani suits, Coca Cola soft drinks, Nintendo toys, Unicef, Sony CD player, Kodak film). Participants rated each of these 10 brands on 65 attributes (33 markers of the Spanish dimensions and 42 markers of the American dimensions), minus the 10 overlapping attributes (daring, young, spirited, unique, real, sincere, down to earth, good looking, upper class, tough, leader). The final set of Spanish markers was back translated through the process outlined in Study 1. Inter-rater agreement was 89%; discrepancies were resolved through discussion. For the purposes of assessing the convergent validity, we also included the three markers of Peacefulness (Japan) that did not appear in the Peacefulness (Spain) dimension (childlike, shy, dependent). Finally, attribute and brand order were counterbalanced. Results and Discussion First, we assessed the conceptual overlap between the indigenous Spanish and imported American brand personality dimensions (plus the Japanese Peacefulness dimension) by examining the patterns of inter-correlations among all the scales representing these constructs. Correlations between corresponding dimensions were as follows: Sincerity (Spain) and Sincerity (U.S.) = .85; Excitement (Spain) and Excitement (U.S.) = .87, Sophistication (Spain) and Sophistication (U.S.) = .83. The correlation between the Spanish and Japanese Peacefulness dimensions was .78. These validity correlations (mean = .83) contrasted with the off-diagonal correlations (mean = .32), suggesting moderate-to-high levels of convergent and discriminant validity. A close examination of the off-diagonal correlations revealed that their relatively large absolute mean value was mostly driven by the presence of a large (.79) correlation between Sophistication (Spain) and Competence (U.S.). This result supports our previous comment that Sophistication in Spain appears to comprise a unique mixture of Sophistication and Competence attributes (i.e., Competence in Spain appears to be a facet of Sophistication instead of defining a separate dimension). The correlation patterns for the culture-specific Passion (Spain) and Ruggedness (U.S.) dimensions were as follows: The highest correlation between Ruggedness (U.S.) and any Spanish personality dimension was only .42 (with Spanish Sophistication), and the highest correlation between Passion (Spain) and any American dimension was .51 (with American Sophistication). Comparisons using Fisher's Z transformations revealed that these two off-diagonal correlations are significantly smaller than the four validity pair-wise correlations. Like in Study 2, we also examined the latent structure the Spanish and American scales (plus the Japanese Peacefulness scale) via confirmatory joint factor analyses. We first examined a model that 12 AAKER, BENET-MARTÍNEZ AND GAROLERA specified seven latent dimensions: Three dimensions representing brand personality constructs common to the United States and Spain (i.e., Sophistication, Excitement, and Sincerity), one dimension representing the one brand personality construct common to Spain and Japan (Peacefulness), and two dimensions representing Spain- and US-specific brand personality constructs (Passion and Ruggedness, respectively). This seven-component model fitted the data adequately, %2 (23, N = 870) = III, p < .001; CFI = .92, GFI = .91. We also tested a more conservative four-component model where dimensions not shared by the United States and Spain would load as follows: Ruggedness and Competence on Sophistication, Passion on Sophistication, and Peacefulness on Sincerity (reflecting the patterns of off-diagonal correlations discussed above). This four-component model yielded unsatisfactory fit indices, %2 (43, N = 870) = 392, p < .001; CFI = .55, GFI = .74, and a significant decrease in overall fit, A%2 (20) = 281. These results corroborate that four culture-specific and three common dimensions are needed to capture the major sources of variance underlying the Spanish and American data. GENERAL DISCUSSION The overarching goal of our studies is to gain insight into how cultural meaning is represented in individuals' perceptions of symbolic objects such as commercial icons. Findings from Studies 1 and 2 identified a set of brand personality dimensions that share similar meaning in Japan and the United States (Sincerity, Excitement, Competence, and Sophistication), as well as relatively culture-specific Japanese (Peacefulness) and American (Ruggedness) dimensions. Studies 3 and 4 extended this set of findings to Spain. Results from these studies also revealed brand personality dimensions that shared similar meaning in both Spain and the United States (Sincerity. Excitement, and Sophistication), plus nonshared Spanish (Passion) and American (Competence and Ruggedness) dimensions. Consistent with the premise that individuals in Japanese and Spanish cultures are more likely to embrace harmony-oriented value types than individuals in the United States (Schwartz, 1994), Peacefulness emerged in Spain as it did in Japan. These results are consistent with the proposition that consumption symbols such as commercial brands may carry both relatively culturally-common and -specific meaning. Consider, for example, the meaning of the Japanese and Spanish Peacefulness dimensions. Considerable research has demonstrated that members of East Asian and Latin cultures tend to place greater weight on cooperation and harmony relative to members of North American cultures, who give more value to mastering the social environment through self-assertion and independence (Hsu, 1983; Marin & Marin, 1991; Triandis et al. 1984). The emergence of Peacefulness in Japan and Spain is consistent with these countries' significantly higher scores relative to the U.S. on Harmony values (see Table 7.3, Schwartz, 1994). The consequences of this cultural variance in value endorsement range from preferences in persuasion appeals that convey harmony (e.g., Kim & Markus, 1999), to subjective assessments of one's happiness that covary with perceptions of harmony in one's relationships (e.g., Kwan, Bond, & Singelis, 1997), to preference for conflict resolution strategies that involve mutual coordination of feelings (e.g., Gabrielidis et al., 1997; Markus & Lin, 1999). In contrast, individuals in the U.S. value self-assertion and personal achievement, as demonstrated in both preferences toward persuasive appeal (Han & Shavitt, 1994) and correlates of life satisfaction (Oishi et al. 1999). Our results indicate that another potential consequence of cultural variations in the emphasis placed on cooperation and harmony relative to individualism and self-assertion involves the emergence of unique configurations in the meaning embedded in commercial brands. For example, the culture-specific status of Ruggedness, with its associations with institutionalized American values such as strength, masculinity and toughness (Solomon, 1986), seems to align well with the findings on value endorsement, whereby the United States has relatively higher scores on Mastery and lower scores on Egalitarian Commitment as compared to Japan and Spain (Schwartz, 1994). Let's now consider the meaning of the Passion dimension. The emergence of this dimension in Spain is supported by recent findings from cultural studies suggesting links between Latin cultures' characteristic higher levels of felt and communicated emotions (Basabe et al., 2000; Zummuner and Fisher, 1995) and several socio-cultural and psychological factors such as honor- and Catholic-related values (Rodriguez, Manstead, & Fischer, 2000; Zubieta et al., 1998), differences in temperament development (Axia, Prior, & Carelli, 1992) and personality (Benet-Martinez, 1999). Portrayals of Spaniards and Latin individuals as 'intense and passionate' abound not only in the social sciences, but also in the popular media. Spain is frequently advertised to the visitor as a land of intense and pleasurable experiences; a country that not only celebrates gastronomy, art, socializing, and risk-taking, but also performs them intensely (e.g., www.cyberspain.com/passion/; see also McVeagh, 1990). Novelists (Hemingway, 1926), travel journalists (Gibson, 2000), film experts (Pally, 1991), and sociologists (Crow, 1985; Hooper, 1987; Shubert, 1990) call attention to the centrality of CARRIERS OF CULTURE 13 passion in Spanish culture. Further, supporting our premise that cultural values penetrate the creation and perception of commercial symbols, we found several Spanish companies that engage in branding efforts and marketing campaigns where the construct of 'passion' is central (e.g., Osborne Group, 2000). A particularly interesting feature of Passion is how affective experience (e.g., fervent, passionate and intense) and spirituality (e.g., spiritual, mystical and bohemian) are combined in a single construct, an association that, according to anthropologists of religion (Mitchell, 1990), is commonly found in Southern European, Catholic cultures. In Spain, for instance, the association between affective-sensual experiences and religion is captured in the following quote: "Religious events in Spain are celebrated with wine and dance and every excess that goes hand in hand with merriment" (McVeagh, 1990; pg. 73). The two components of Passion (emotional intensity and spirituality) can also be interpreted in light of Schwartz's values, particularly in relation to Affective and Intellectual Autonomy. The unique emphasis on unrestrained affect captured by Passion denotes an orientation towards the enjoyment and exploration of one's emotional life that, according to Schwartz's value theory, is at the core of Affective Autonomy. Not surprisingly, Spain has a remarkably high score on this value dimension, one that is significantly above Japan and United States' scores and close to other Mediterranean societies such as France and Greece (Schwartz, 1994). The other emphasis of Passion, spirituality and worldly life style, on the other hand, seems particularly related to Intellectual Autonomy, another value orientation for which Spain scores quite highly, and above Japan and the United States. Note that the emergence of Passion can not be interpreted as an indication that Affective and Intellectual Autonomy are emphasized in Spain only, given that Excitement, which also seems to relate to Affective Autonomy needs (see p. 8) emerged in all three cultures. Most likely, the emergence of Passion reflects culture-specific meanings and needs related to Excitement (i.e., in Spain, fulfilling one's needs for novelty and excitement may be largely achieved by having intense emotional and spiritual experiences) that are powerful enough to define their own dimension and that may help to explain Spain's remarkably high scores on Autonomy. In contrast to Peacefulness, Ruggedness and Passion, the Sincerity, Excitement and Sophistication dimensions appear to be more similarly construed across cultures. This suggests that, in addition to potential cultural variance in consumer needs, commercial brands may reflect more universally held individual needs. However, despite the cross-cultural stability of the above dimensions, the results of the cross-cultural correlations in Studies 2 and 4 indicate that the correspondence is not unitary. To illustrate, Excitement is associated with being young, contemporary, spirited, and daring across the cultural contexts. However, it also conveys imaginativeness, uniqueness and independence in North America and Spain. In contrast, in Japan, it contains a "talkativeness" facet (e.g., talkative, funny and optimistic). This idiosyncratic meaning is consistent with the relativist argument that constructs shift in meaning when examined in different cultural contexts (Shweder, 1990). This implication is particularly important in the context of Sophistication, where there is considerable overlap in certain attributes across the cultures (e.g., glamorous, good looking, stylish, smooth). However, unique to Spain was a secondary facet that contains attributes more closely associated with Competence in Japan and the United States. This finding indicates that Sophistication takes on a different meaning in Spain than it does in North America or Japan. In other words, the interpretation of the meaning of a commercial brand must take into consideration the particular cultural lens through which the brand is being seen. This result highlights the notion that absolute equivalents and universals may not be as useful as understanding and investigating the idea of partial equivalents and partial universals (Wierzbicka, 1991). In other words, the dimensions that emerged in these four studies appear to simply vary in the degree to which they contain universal meaning relative to culture-specific meaning. Our findings also have implications for the understanding of human personality. As discussed earlier, our culture-specific brand personality dimensions (Ruggedness, Passion, and Peacefulness) can be related to particular patterns of human personality traits, emotions, and value orientations characteristic of American, Spanish, and Japanese cultures respectively. These links between brand and human personality differences are to be expected if one acknowledges the inseparability of culture and psyche (Markus & Lin, 1999) and the largely socially-constructed nature of personality (Hampson, 1988). Namely, culture can be see as a network of shared meaning that influences how social perception is organized, from the way commercial symbols are seen to how human personality is described and even experienced. Furthermore, it is interesting to note that our work shows that, as with human personality, brand personality appears to be consistently organized around five dimensions. The robustness of a five-dimensional structure across these two kinds of personality perception suggests that, functionally, social perception may be influenced by cognitive-economy processes similar to those affecting memory (Miller, 1956), where information is best organized 14 AAKER, BENET-MARTÍNEZ AND GAROLERA and retrieved around seven 'chunks' of information (plus or minus two). Caveats and Future Research The contributions of this research involve a substantive focus on identifying and examining the culturally-similar and culture-specific meaning carried and conveyed in consumption symbols, as well as the methodological emphasis on a combined emic-etic to compare this approach. However, despite these contributions, there are limitations that reveal areas for future research. First, from a methodological perspective, this research relied on a limited number of attributes and commercial brands to create the perceptual space of brands. Future research is needed to determine the degree to which the results found in the current research are generalizeable across contexts, persons and brands. Second, the current research took a single picture of individual's perceptions of the meaning of commercial brands at a static point in time. Therefore, although the conceptualization of culture put forth in this research is dynamic, the nature of this dynamism was not explored. Future research is needed to determine the degree to which exposure to the constructs represented by the indigenous dimensions, and market globalization efforts more broadly, makes all of us psychologically more similar (Hermans & Kempen, 1998). Shore (1996), for example, comments that as Coke and Pepsi quickly make their way to the recently liberated South Africa, "a global mass culture with Western commodities at its heart was created" (pg. 9). To what degree do these Western commodities subsequently shape the new culture in which they are distributed? The answer may depend not only on the meaning of those Western brands (e.g., Coca-Cola), but also on the nature of the interaction between the brand and individuals in the culture (e.g., South Africa). If meaning construction is an ongoing process, one that involves the active interaction with people (Kim & Markus, 1999; Shweder, 1990; Shore 1996), the distribution of these Western commodities may not in fact lead to psychologically more similar individuals. Rather, it may lead to individuals who are exposed to multiple cultural models, and a commercial brand whose meaning is jointly created by advertisers and individuals in the culture. Finally, the current work is a first step towards understanding the link between the culture and psyche in the context of commercial symbols; however, it remains exploratory in nature. Future research is needed to elucidate the specific mechanisms by which commercial symbols are imbued with meaning, as well as how that meaning characterizes perceptions of human attributes and values (Roccas et al. 2000). In this way, the more dynamic nature of culture may be taken into consideration, and the movement of meaning from culture to the individual may be explored more directly. Conducting longitudinal research is one way to address this question; another is to examine the process of frame-shifting. Recent research on biculturalism has shown that individuals have the ability to "frame shift", that is to view things from different cultural vantage points (Hong et al., 2000). In this light, the culture-as-a-lens metaphor (McCracken, 1986) is extended to one where multiple glasses with different color lenses can be put on and taken off. Given this perspective, the question arises, to what degree does a particular perceptual representation of brands in consumers' minds lead to different evaluations of brands? That is, what happens when a consumer holds a Japanese perception of the structural space of brands? What are the consequences of holding such a mental representation, particularly as it compares to when one holds the American (or Spanish) mental representation of brands? To address these questions, a set of priming experiments that manipulate the salience of one cultural frame over another may be conducted. For example, in Japan, kanji is perceived as a relatively traditional Japanese writing system, whereas katakana is perceived as more modern or westernized. Therefore, one might examine the degree to which brand names or personality attributes written in kanji (katakana) may evoke a Japanese (American) perceptual structure, thereby leading to potentially different sets of consequences. In this way, the more dynamic nature of culture may be taken into consideration, and the movement of meaning from culture to the individual may be explored more directly. In conclusion, the work presented here shows that the study of consumption symbols, such as commercial brands, is a useful approach to the understanding of how cultural beliefs and values are represented and institutionalized. In accordance with an ethno-psychological perspective (Wierzbicka, 1991), our results indicate that the meaning embedded in commercial brands has both culturally-specific and -common elements. Above all, our studies underscore the mobile quality of culture and the bi-directional relationship between the individual and culture. CARRIERS OF CULTURE 15 REFERENCES Aaker, J. (1997). Dimensions of brand personality. Journal of Marketing Research, 34, 342-352. Aaker, J., & Maheswaran, D. (1997). The effect of cultural orientation on persuasion. Journal of Consumer Research, 24, 315-328. Axia, G., Prior, M., & Carelli, M.G. (1992). Cultural influences on temperament: A comparison of Italian, Italio-Australian, and Anglo-Australian toddlers. Australian Psychologist, 27, 52-56. Basabe, N., Paez, D., Valencia, J., Bernard, R., Pennebaker, J., Diener, E., & Gonzalez, J.L. (2000). Sociocultural factors predicting subjective experience of emotion: A collective level analyses. Psicothema, 12, 55-69. Belk, R. W., & Pollay, R. W. (1985). Images of ourselves: The good life in twentieth century advertising. Journal of Consumer Research, 11, 887-898. Benet-Martinez, V. (1999). Exploring indigenous Spanish personality constructs with a combined emic-etic approach. In J. C. Lasry, J. G. Adair, & K. L. Dion, (Eds.), Latest contributions to cross-cultural psychology (pp. 151-175). Lisse, The Netherlands: S wets and Zeitlinge. Benet-Martinez, V., & Waller, N. G. (1997). Further evidence for the cross-cultural generality of the Big Seven Factor model: Indigenous and imported Spanish constructs. Journal of Personality, 65, 567-598. Bentler, P.M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238-246. Berry, J. W. (1969). On cross-cultural comparability. International Journal of Psychology, 4, 119-128. Berry, J. W., Poortinga, Y. H., Segall, M. H., Dasen, P. R. (1992). Cross-cultural psychology: Research and applications. New York, NY: Cambridge University Press. Bond, M. H. (1994a). Into the heart of collectivism: A personal and scientific journey. In U. Kim, H. C. Triandis, C. Kagitcizbasi, S. Choi, & G. Yoon (Eds.), Individualism and collectivism: Theory, method, and applications (Vol. 18, pp. 66-84), Cross Cultural Research and Methodology Series. Thousand Oaks, CA: Sage Publications, Inc. Bond, M. H. (1994b). Trait theory and cross-cultural studies of person perception. Psychological Inquiry, 5, 114-117. Briley, D., Morris, M. W., & Simonson, I. (in press). Culture, reasons, and compromise in a choice dilemma: Chinese and American cultures bring different reasons to mind. Journal of Consumer Research. Brislin, R.W. (1970). Back-translation for cross-cultural research. Journal of Cross-Cultural Psychology, 3, 185-216. Church, T., & Katigbak, M. (1988). The emic strategy in the identification and assessment of personality dimensions in a non-Western culture. Journal of Cross-Cultural Psychology, 19, 140-163. Cooper, C. R., & Denner, J. (1998). Theories linking culture and psychology: Universal and community-specific processes. Annual Review of Psychology, 49, 559-584. Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory and NEO five-factor inventory professional manual. Odessa, FL: Psychological Assessment Resources, Inc. Cross, S., & Markus, H.R. (1999). The cultural constitution of personality. In L. Pervin and O. John (Eds.), Handbook of personality: Theory and research (2nd ed.) (pp. 378-396). New York, NY: The Guilford Press. Crow, J. A. (1985). Spain: The root and the flower. Berkeley, CA: University of California Press. DiMaggio, P. (1997). Culture and cognition. Annual Review of Sociology, 23, 263-287. Douglas, M., & Isherwood, B. (1978). The world of goods: Towards an anthropology of consumption. New York NY: W. W. Norton. Enriquez, V. (1979). Towards cross-cultural knowledge through cross-indigenous methods and perspectives. Hong Kong Psychological Society Bulletin, 3, 7-21. Fiske, A., Kitayama, S. Markus, H. R., & Nisbett, R. (1998). The cultural matrix of social psychology. In D. T Gilbert & S. Fiske (Eds.), The handbook of social psychology (pp. 915-981). Boston: McGraw-Hill. Gabrielidis, C, Stephan, W.G., Ybarra, O., Pearson, V.M., & Villareal, L. (1997). Preferred styles of conflict resolution: Mexico and the United States. Journal of Cross-Cultural Psychology, 28, 661-677. Gibson, V. (2000). Zest for life: From flamenco to flirting, Spain sizzles with the passion of its. Downloaded on September 10, 2000. http://www.canoe.ca/TravelEurope/spain.html. Gouvenia, V.V., & Ross, M. (2000). Hofstede and Schwartz's models for classifying individualism at the cultural level: Their relation to macro-social and macro-economic variables. Psicothema, 12,25-33. 16 AAKER, BENET-MARTÍNEZ AND GAROLERA Hampson, S.E. (1988). The construction of personality: An introduction. New York, NY: Routledge and Kegan. Han, S., & Shavitt, S. (1994). Persuasion and culture: Advertising appeals in individualistic and collectivistic societies. Journal of Experimental Social Psychology, 30, 326-350. Hemingway, E. (1926). The sun also rises. New York: Grosset & Dunlap. Hermans, J. M, & Kempen, H. J. G. (1998). Moving cultures: The perilous problems of cultural dichotomies in a globalizing society. American Psychologist, 53, 1111-1120. Hofstede, G. (1980). Culture's consequences. Beverly Hills, CA: Sage Publications, Inc. Hong, Y., Morris, M. W., Chiu, C, & Benet-Martinez, V. (2000). Multicultural minds: A dynamic constructivist approach to culture and cognition. American Psychologist, 55, 709-720. Hooper, J. (1987). The Spaniards: A portrait of the new Spain. London: Penguin Books. Hsu, J. (1983). Asian family interaction patterns and their therapeutic implications. International Journal of Family Psychiatry, 4, 307-320. Hui, C.H., & Triandis, H.C. (1985). Measurement in cross-cultural psychology: A review and comparison of strategies. Journal of Cross-Cultural Psychology, 16, 131-152. Japan Statistics Bureau and Statistics Center (1996), Statistical Abstracts on the web site: http://www.stat.go.jp/Lhtm. John, O.P., Donahue, E.M., & Kentle, R.L. (1991). The "Big Five" Inventory—Versions 4a and 54. Technical Report. IP AR. University of California at Berkeley. Kim, H., & Markus, H. R. (1999). Deviance or uniqueness, harmony or conformity? A cultural analysis. Journal of Personality and Social Psychology, 77, 785-796. Kitayama, S., Markus, H. R., Matsumoto, H., & Norasakkunkit, V. (1997). Individual and collective processes in the construction of the self: Self-enhancement in the United States and self-criticism in Japan. Journal of Personality and Social Psychology, 72, 1245-1267. Kwan, V. S. Y., Bond, M. H., & Singelis, T. M. (1997). Pancultural explanations for life satisfaction: Adding relationship harmony to self-esteem. Journal of Personality and Social Psychology, 73, 1038-1051. Lannon, J. (1993). Asking the right questions: What do people do with advertising? In D. Aaker and A. Biel (Eds.), Brand Equity and Advertising, (pp. 163-176). New Jersey: Lawrence Erlbaum Associates. Leung, K., & Bond, M.H. (1989). On the empirical identification of dimensions for cross-cultural comparisons. Journal of Cross-Cultural Psychology, 20, 133-151. Levi-Strauss, C. (1966). The Savage Mind. Chicago, IL: University of Chicago Press. Marín, G., & Marín, B. V. (1991). Research with Hispanic populations. In Applied Social Research Methods Series (Vol. 23). Newbury Park, CA: Sage Publications, Inc. Marin, G., & Triandis, H. C. (1985). Allocentrism as an important characteristic of the behavior of Latin Americans and Hispanics. In R. D. Guerrero (Ed.), Cross-cultural and national studies in social psychology (pp. 85-104). Oxford, UK: Elsevier Science Publishers. Markus, H.R., & Lin, L.R.. (1999). Conflictways: Cultural diversity in the meanings and practices of conflict. In D.A. Prentice and D.T. Miller (Eds.), Cultural divides: Understanding and overcoming group conflict (pp. 302-333). New York, NY: Russell Sage Foundation. Matsumoto, D. (1999). Culture and self: An empirical assessment of Markus and Kitayama's theory of independent and interdependent self-construal, Asian Journal of Social Psychology, 2, 289-310. McCracken, G. (1986). Culture and consumption: A theoretical account of the structure and movement of the cultural meaning of consumer goods. Journal of Consumer Research, 13, 71-84. McCrae, R. R., & Costa, P. T., Jr. (1997). Personality attribute structure as a human universal. American Psychologist, 52, 509-516. McCrae, R.R., Costa, P.T., Ostedorf, F., Angleitner, A., Hrebickova, M., Avia, M., Sanz, J., Sanchez-Bernardos, M.L., Kusdil, E., Woodfield, R., Saunders, P.R., & Smith, P.B. (2000). Nature Over Nurture: Temperament, Personality and Life Span Development. Journal of Personality and Social Psychology, 78, 173-196. McVeagh, V. (1990). People. In K. Wheaton (Ed.), Inside Guides: Spain. Singapore: APA Publications. Miller, G.A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63, 81-97. Mitchell, T. (1990). Passional culture: Emotion, Religion, and Society in Southern Spain. Philadelphia, PE: University of Pennsylvania Press. Nunnally, J. C. (1978). Psychometric theory. New York, NY: McGraw-Hill, Inc. Oishi, S., Diener, E. F., Lucas, R. E., & Suh, E. M. (1999). Cross-cultural variations in predictors of life satisfaction: Perspectives from needs and CARRIERS OF CULTURE 17 values. Personality and Social Psychology Bulletin, 25, 980-990. The Osborne Group (2000). http ://www. osborne. es/english/presen/campa. ht m. Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning. Chicago, IL: University of Illinois Press. Pally, M. (1991). The Politics of Passion: Pedro Almodovar and the Camp Esthetic. Cineaste Magazine, Vol. 12, Issue 3, 32-39. Park, B. (1986). A method for studying the development of impressions in real people. Journal of Personality and Social Psychology, 51, 907-17. Plummer, J. T. (1985). How personality makes a difference. Journal of Advertising Research, 24, 27-31. Ratchford, B. (1987). New insights about the FCB grid. Journal of Advertising Research, 27, 24-26. Richins, M. L. (1994). Valuing things: The public and private meaning of possessions. Journal of Consumer Research, 21, 504-521. Rodriguez, P.M., Manstead, A.S., & Fischer, A.H. (2000). The role of honor-related values in the elicitation, experience, and communication of pride, shame, and anger: Spain and the Netherlands compares. Personality and Social Psychology Bulletin, 26, 833-844. Roccas, S., Sagiv, L., Schwartz, S., & Knafo, A. (2000). The big five personality factors and personal values. Unpublished manuscript. Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In M. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 25). Orlando, FL: Academic Press. Schwartz, S. H. (1994). Beyond individualism/collectivism: New cultural dimensions of values. InU. Kim, H.C. Triandis, C. Kagitcibasi, S. Choi, and G. Yoon (Eds.), Individualism and collectivism: Theory, method, and applications (Vol. 18) (pp. 85-119), Cross Cultural Research and Methodology Series. Thousand Oaks, CA: Sage Publications, Inc. Schwartz, S.H. (1999). A theory of cultural values and some implications for work. Applied psychology: An international review: Special issue: Values and work, 48, 23-47. Schwartz, S. H., & Bilsky, W. (1990). Toward a theory of the universal content and structure of values: Extensions and cross-cultural replications. Journal of Personality and Social Psychology, 58, 878-891. Schwartz, S.H., & Ross, M. (1995). Value priorities in West European nations: A cross-cultural perspective. Studies in psychology in honor of Solomon Kugelmass, 36, 322-347. Schwartz, S.H., & Sagiv, L. (1995). Identifying culture-specifics in the content and structure of values. Journal of Cross-Cultural Psychology, 26,92-116. Shavitt, S. (1990). The role of attitude objects in attitude functions. Journal of Experimental Social Psychology, 26,_124-148. Shore, B. (1996). Culture in mind: Cognition, culture and the problem of meaning. New York, NY: Oxford University Press. Shubert, A. (1990). A social history of modern Spain. Boston, MA : Unwin Hyman. Shweder, R. A. (1990). Cultural psychology: What is it? In J. Stigler, R. Shweder, & G. Herdt (Eds.), Cultural psychology (pp. 1-43). Cambridge, UK: Cambridge University Press. Shweder, R. A., & Sullivan, M. A. (1990). The semiotic subject of cultural psychology. In L. A. Pervin et al. (Ed.), Handbook of personality: Theory and research (pp. 399-416). New York, NY: Guilford Press. Solomon, M. (1986). Deep-seated materialism: The case of Levi's 501 jeans. Advances in Consumer Research, 13, 619-622. Takáno, Y. & Osaka E. (1999). An unsupported common view: Comparing Japan and the U.S. on individualism/collectivism. Asian Journal of Social Psychology, 2, 311-411. Triandis, H. C. (1989). The self and behavior in differing cultural contexts. Psychological Review, 96, 506-52. Triandis, H. C, Marín, G., Lisansky, J., & Betancourt, H. H. (1984). Simpatia as a cultural script of Hispanics. Journal of Personality and Social Psychology, 47, 1363-1375. Vacker, B (1992). The Marlboro man as a twentieth century David: A philosophical inquiry into the Aristotelian aesthetic of advertising. Advances in Consumer Research, 19, 746-755. Wierzbicka, A. (1991). Japanese key words and core cultural values. Language in Society, 20, 333- 385. Yang, K., & Bond, M. H. (1990). Exploring implicit personality theories with indigenous or imported constructs: The Chinese case. Journal of Personality and Social Psychology, 58, 1087- 1095. Zubieta, E., Fernandez, I., Vergara, A.I., Martinez, M.D., & Candia, L. (1998). Cultura y emocion en America. Boletin de Psicologia, 61, 65-89. Zummuner, V.L., & Fischer, A.H. (1995). The social regulation of emotions in jealousy situations: A comparison between Italy and the Netherlands. Journal of Cross-Cultural Psychology, 26, 189- 208. 18 AAKER, BENET-MARTINEZ AND GAROLERA TABLE 1: Japanese Brand Personality Dimensions Varimax-rotated principal factors Abbreviated Original English Translation Ex Co Pe Si So Japanese Terms Fun 76 13 -06 -03 14 (DD^l^ Humorous 72 09 14 06 09 I>± 0%kt£ Talkative 72 09 15 06 08 mitte Optimistic 72 09 14 02 02 s&ftjftft Free 71 12 -08 05 15 ěsft Funny 70 -04 28 -07 01 ^—^Tfrfo& Chatty 69 -00 18 00 07 *5L^*?& Energetic 69 20 -18 21 -01 TĚA,^ Youthful 68 04 -16 07 27 %* LV^ Laid-back 67 06 11 09 -04 Ztzfrb^^ Spirited 67 19 -15 17 05 tfeféft Cheerful 67 06 -20 28 16 K 31^ Friendly 66 06 09 36 -02 A^ooi^ Active 66 29 -27 08 07 mmmte Easy-going 63 05 17 -15 07 (Dk^fš. Positive 61 43 -16 06 14 mmmi Happy 61 08 15 33 10 Uičbfrft Curious 61 30 -02 -05 17 0^'b■%>&%: Nice 40 10 -11 31 31 X^ Cooperative 40 31 14 38 11 mn\í(Dhz> Note. N — 1495 Japanese. All loadings multiplied by 100; loadings | .40 | or larger are set in bold. Ex Co = Competence, Pe = Peacefulness, Si = Sincerity, So = Sophistication. Excitement, TABLE 1: Japanese Brand Personality Dimensions 19 Varimax-rotated principal factors Abbreviated Original English Translation Ex Co Pe Si So Japanese Terms Easygoing 40 35 19 09 -03 -?4^<~ xft Ordinary 38 00 08 32 -31 f&mw Reliable 13 71 04 26 15 Lo^t) L/c Determined 22 71 09 15 08 IčiĚW^ Dignified 29 68 -06 07 19 ^ t L/c Patient 11 66 23 17 01 UWfä^ Tenacious 18 65 18 17 -02 *Ě!9&^ Responsible 02 64 15 38 12 Äffi&Ofcä Respectable 09 64 11 18 34 ÄMft Confident 30 63 -10 07 22 ěífi-tt^/c Strong 36 63 00 00 04 usi^ Sharp 28 63 02 -08 23 f^ Consistent 02 59 20 41 09 —Iifc Courageous 39 58 04 -05 09 Mm Tough 07 56 25 -13 -01 fcft Assertive 19 55 00 09 25 á^Á/Cft Masculine 27 54 -01 -20 -07 %mw Clear 43 52 -12 06 10 í±o# D L/c Precise -02 51 30 36 24 JUSffift Stable 11 50 12 41 05 žčaĚVŤz Self-composed -17 49 28 37 26 íi^^Wc Dependable -04 46 38 34 -09 $Btb£ Rational 25 43 06 11 -05 G-mm* Tolerant 38 42 22 26 13 Ä^ft Realistic 35 37 -04 17 -02 «wft Mild-mannered 00 04 74 20 15 fco t D L/c Timid -03 09 73 10 12 íŕ/f^L^^O Shy 09 07 67 12 13 ft^lft Reserved -05 12 66 21 10 t>^*L*6ft Peaceful -13 18 64 31 19 ¥fPft Modest -18 32 55 20 -09 iÉÄft Clumsy 17 14 55 -16 -14 ^fsfflft Dependent 08 17 51 -07 12 ÄLá^DMft Chüdlike 30 04 50 07 -03 ŤÄooV 20 AAKER, BENET-MARTINEZ AND GAROLERA Note. N = 1495 Japanese. All loadings multiplied by 100; loadings | .40 | or larger are set in bold. Ex = Excitement, Co = Competence, Pe = Peacefulness, Si = Sincerity, So = Sophistication. TABLE 1: Japanese Brand Personality Dimensions Varimax-rotated principal factors Abbreviated Original English Translation Ex Co Pe Si So Japanese Terms Calm 12 18 49 43 21 ^rf^ft Naive 20 16 42 33 -12 -r4--ýt£ Cute 32 -12 41 31 37 frf}\^\^ Feminine 11 -12 39 34 38 -kmfc Kind 19 20 32 55 30 ÄL^ Family oriented 18 10 30 54 -02 flW4 Thoughtful 21 31 32 53 -02 mm < Sincere 09 49 21 53 18 M^ft Clean 17 29 05 51 35 WMte True 47 22 12 49 09 Wft Warm 39 16 26 49 12 flgäM^ Honest 39 39 10 47 04 lEtÜft Healthy 37 19 -07 46 08 ftÄfôfr Considerate 35 33 17 40 21 EĚM^ Stylish 29 11 -01 10 68 ŽoL^tiŕj: Elegant -10 31 19 26 65 -Cpn'cČ Romantic 18 05 31 16 63 w?y=?--y?řj: Smooth -12 38 16 08 60 lÄft Extravagant 39 15 02 -01 58 «ft Sexy 18 05 36 -02 55 -feoíäV Delicate -05 29 25 27 51 ftíflft Stunning, cool 44 29 -12 -04 50 ^oi^t^ Sophisticated 39 19 -20 00 47 VMŽtiŤc Poised 12 39 03 -02 41 ¥#ft Note. N — 1495 Japanese. All loadings multiplied by 100; loadings | .40 | or larger are set in bold. Ex — Excitement, Co = Competence, Pe = Peacefulness, Si = Sincerity, So = Sophistication. TABLE 2: Spanish Brand Personality Dimensions 21 Varimax-rotated principal factors Abbreviated Original Spanish English Translation Ex Si So Pe Pa (Castillian) Terms Happy 80 06 -03 10 16 alegre Fun 78 02 -05 12 20 divertida Spirited 75 05 12 13 07 animosa Outgoing 75 13 -02 16 19 extrovertida Sharp, shrewd 73 10 02 01 19 avispada Young 72 -06 14 05 05 j oven Energetic 68 07 26 11 04 llena de vida y energia Daring 67 -09 26 04 24 atrevida Cool 67 00 18 12 10 fresca Active 67 21 07 -06 16 activa Spirited 65 24 14 04 12 viva Imaginative 60 17 19 05 22 creativa Popular 58 10 11 05 -15 popular Original 57 15 24 05 24 originál Contemporary 55 08 41 -04 03 contemporánea Unique 53 07 39 10 13 única Playful 53 02 04 11 44 picara Familiar 52 35 -08 19 -11 familiar Independent 47 17 38 -03 17 independiente Free 46 -01 15 24 12 libre Likable 41 18 41 34 11 simpatica Fiesty 37 06 24 14 13 peleona Considerate -01 76 19 09 14 considerada Thoughtful -06 72 17 14 19 atenta Well-mannered 00 72 20 15 10 correcta Orderly -01 71 24 11 11 ordenada Moderate -06 70 19 14 13 moderada Balanced 01 67 27 06 11 equilibrada Down-to-earth 16 65 16 14 00 realista Trustworthy 10 64 17 27 01 honrada Sincere 31 57 11 28 03 sincera Real 33 55 12 11 00 real Logical 11 51 32 25 -01 lógica Rational 13 49 41 17 00 racional Hardworking 29 46 39 11 -01 trabajadora Practical 28 42 27 15 -10 practica Flexible 28 42 13 29 06 flexible Good looking -06 16 67 22 21 elegante Glamorous 05 10 60 24 29 glamorosa Note. N = 692 Spania» ds. All loading rs multipliec lby 100; loadings .40 or larger are set in bold. Ex = Excitement, Si — Sincerity, So — Sophistication, Pe — Peacefulness, Pa — Passion. 22 AAKER, BENET-MARTINEZ AND GAROLERA TABLE 2: Spanish Brand Personality Dimensions Varimax-rotated principal factors Abbreviated Original Spanish English Translation Ex Si So Pe Pa (Castillian) Terms Upper-class 02 23 59 07 15 de clase superior Strong 19 19 59 -02 12 fuerte Secure 34 37 55 -05 00 segura Tough 09 18 54 08 13 dura Leader 35 22 53 -10 05 dirigente Confident 39 32 51 -04 07 segura de si misma Persistent 38 22 49 02 00 persistente Successful 39 24 47 06 08 exitosa Stylish 27 36 46 17 00 moderna Reliable 26 42 46 07 -08 fiable Reflective 34 33 42 18 11 pensativa Feminine 01 07 41 38 21 femenina Western 13 18 40 -01 03 occidental Masculine -01 08 36 19 19 masculina Naive 12 05 03 64 07 ingenua Mild mannered 07 27 19 61 -01 apacible Good natured 22 39 -06 60 14 buenaza Shy -05 07 01 59 12 timida Peaceful 08 34 19 59 -06 paciftca Affectionate 30 36 -02 58 19 carinosa Sweet 29 26 04 56 18 dulce Docile 05 36 -00 54 18 dó cil Calm 10 16 27 48 06 tranquila Chüdlike 33 -06 -14 47 00 infantil Gentle 27 39 01 46 11 amable Dependent 01 09 08 41 12 dependiente Rugged -03 -03 10 31 28 ásp era Normal 07 29 06 30 -06 normal Fervent 34 06 12 11 68 fervorosa Passionate 37 07 16 11 65 apasionada Impulsive 47 03 09 03 62 impulsiva Temp eramental 36 13 18 00 61 temperamental Emotional 41 14 07 13 56 emocional Intense 39 15 19 01 55 intensa Mystical -06 13 10 43 52 mistica Spiritual -05 16 09 44 51 espiritual Bohemian 10 04 12 32 51 bohemia Extravagant 25 -07 18 17 42 extravagante Note. N = 692 Spaniards. All loadings multiplied by 100; loadings | .40 | Si = Sincerity, So = Sophistication, Pe = Peacefulness, Pa = Passion. or larger are set in bold. Ex — Excitement, 23 FIGURE 1: American Brand Personality Dimensions American Brand Personality Dimensions Sincerity Excitement Competence Sophisticatiot Ruggedness Down-to-Earth: down-to-earth, family-oriented, small town Honesty: honest, sincere, real Wholesomeness: wholesome, original Cheerfulness: cheerful, friendly, sentimental Daring: daring, trendy, exciting Spiritedness: spirited, cool, young Imagination: imaginative, unique Contemporary: up-to-date, independent, contemporary Reliability: reliable, hard working, secure Intelligence: intelligent, technical, corporate Success: successful, leader, confident Class: upper class, good-looking, glamorous Charm: charming, feminine, smooth Masculinity: outdoorsy, masculine, western Toughness: tough, rugged 24 AAKER, BENET-MARTINEZ AND GAROLERA FIGURE 2: Japanese Brand Personality Dimensions (Study 2) Japanese Brand Personality Dimensions Excitement Competence Peacefulness Sincerity Sophistication Talkativeness: talkative, funny, optimistic Freedom: positive, contemporary, free Happiness: friendly, happy, likable Energy: youthful, energetic, spirited Responsibility: consistent, responsible, reliable Determination: dignified, determined, confident Patience: patient, tenacious, masculine Mildness: shy, mild-mannered, peaceful Naivity: naive, dependent, childlike Warmth: warm, thoughtful, kind Elegance: elegant, smooth, romantic Style: stylish, sophisticated, extravagant 25 FIGURE 3: Spanish Brand Personality Dimensions (Study 3) Spanish Brand Personality Dimensions Excitement Sincerity Sophistication Peacefulness Passion Happiness: happy, outgoing, fun Youth: daring, young, spirited Independence: unique, imaginative, independent Thoughtfulness: considerate, thoughtful, well-mannered Realness: real, sincere, down-to-earth Style: good looking, glamorous, stylish Confidence: confident, persistent, leader Affection: affectionate, sweet, gentle Naivity: naive, mild-mannered, peaceful Intensity: fervent, passionate, intense Spirituality: spiritual, mystical, bohemian 26 AAKER, BENET-MARTÍNEZ AND GAROLERA FOOTNOTES 1 Although the conceptualization of brand and human personality may be similar, the two constructs vary in their antecedents as well as the distinct roles that they serve. In the case of individuals, personality traits are inferred from observable and stated attitudes and behavior as well as physical characteristics (Park, 1986). In this light, people develop their own personalities, thereby reflecting a relatively basic process of personality development (McCrae et al., 2000). In contrast, brands are inanimate objects imbued with personality trait associations through marketing communications, thereby reflecting a more impressionable process of personality development. For example, marketers rely on user imagery (defined as the set of human characteristics associated with the user of the brand), celebrity endorsers (e.g., Michael Jordan), symbols, logos and slogans (e.g., AT&T's "Reach Out and Touch Someone") and personification (e.g., Pillsbury Doughboy) to develop the personality associations of a brand (Plummer, 1985). a To identify the relatively irrelevant attributes, Japanese participants (« = 140, 55% female, mean age = 35.3) rated how descriptive the 167 attributes (253 minus the redundant and ambiguous attributes) were of the most salient brand in 10 product categories that spanned the symbolic-utilitarian framework. To isolate the most relevant attributes for this set of stimuli, the cutoff for the final list of attributes was a scale rating of 4 (very descriptive), thereby leaving 100 attributes for Study 1. Of those 100 attributes, 68% were indigenous (plus 15% from the Big Five and 17% from Aaker 1997). m The participants were representative of the geographic regions in Japan (e.g., 30% of the participants were from the Kanto region), although no one from the islands outside of Honshu participated. lv One limitation of a disaggregated analysis (i.e., making each individual's ratings of each brand the unit of analysis) relative to an aggregated analysis (where brands are the unit of analysis after averaging across individuals' ratings of each particular brand) is that the correlations among attributes are likely to also reflect individual differences in scale use. To assess the impact of this methodological issue, we also examined factor solutions obtained from aggregated data (« = 25 brands). Interestingly, these factor structures were similar to those obtained with the disaggregated data (see Leung & Bond, 1989 and Schwartz, 1994 for a discussion of why structures obtained from aggregated and disaggregated data tend to be closely related, and the rationale for using aggregated vs. disaggregated data in factor analyses). We also examined structure obtained using an oblique rotation (Promax), which proved nearly identical to the orthogonal solution (Varimax). v By relying on a different sample than in Study 1, Study 2 provides more support for the robustness of the findings. However, it also suffers from the limitation of small sample size. ^ In study 2, facets were used as indicators of the latent factors (which were allowed to correlate). The same was the case in Study 4. ™Many different indices are available to assess the degree to which a hypothesized model is consistent with observed data. The chi-square statistic is the most widely used but is highly dependent on sample size so it can be significant even for models that fit the data well (Bender, 1990). Another index is the comparative fit index (CFI; Bender, 1990), which ranges from 0 to 1 and is relatively independent from sample size. The rule of thumb is that a CFI of .90 or greater indicates that the specified model fits the data well. vm "j<0 identify the relatively irrelevant attributes, Spanish participants (« = 75, 46% female, mean age = 34) rated how descriptive the 171 attributes (266 minus the redundant and ambiguous attributes) were of the most salient brand in 10 product categories that spanned the symbolic-utilitarian framework. To isolate the most relevant attributes for this set of stimuli, the cutoff for the final list of attributes was a scale rating of 4 (very descriptive), thereby leaving 77 attributes for Study 3. Of those 77 attributes, 67% were indigenous (plus 7% from the Big Five and 26% from Aaker 1997). Lx As in Study 1, two additional analyses were run to gain insight into the robustness of the results. We examined the structure obtained using the disaggregated data and an oblique rather than an orthogonal rotation, and as well as the factor solutions obtained from aggregated data (« = 25 brands). The results provided structures that were similar to that reported above.