J 1\

MARTIN S. ROTH*

Developing and managing brand image is an important part of a firm'smarketing program. However, little research has been done (1) on linking theuse of brand image strategies to product performance or (2) on managingbrand images in giobai markets. The author examines the brand image-performance linkage for consumer goods in two categories marketedinternationally. He also develops a conceptual framework that identifiesvarious cultural and socioeconomic environmental characteristics of foreignmarkets that are hypothesized to affect brand image performance. Resultsfrom a 10 country/60 region study indicate that cultural power distance,cultural individualism, and regional socioeconomics affect the performance offunctional (problem prevention and solving), social {group membership andsymbolic), and sensory (novelty, variety, and sensory gratification) brandimage strategies. The author then discusses the implications for managers

marketing brands internationally and the directions for further research.

The Effects of Culture and Socioeconomicson the Performance of Global BrandImage Strategies

Brand image management is a critical part of a company'smarketing program. Communicating a clearly defined brandimage enables consumers to identify the needs .satisfied bythe brand (Park. Jaworski, and Maclnnis 1986) and differ-entiate the brand from its competUors (DiMingo 1988;Reynolds and Gutman 1984), and has been prescribed byboth marketing practitioners (Ogilvy 1963) and researchers(Gardner and Levy 1955) as a key to product success. Infact, brand image is an integral component ofa brand's eq-uity, that is, the value of a brand in the minds' of consumers(Keller 1993). Developing a needs-based image strategyprovides the foundation for marketing program developmentand enables the brand to create a clear and distinct positionwithin its category. Although much conceptual work hasbeen written on defining and managing brand image (Keller1993; Park, Jaworski, and Maclnnis 1986), very little em-pirical research exists to guide managers' image manage-ment efforts. Two important areas must be addressed bystudies examining and prescribing brand image strategies:(I) tbe effect of brand image .strategies on product perfor-mance and (2) lhe management of brand images acrossglobal markets.

*Manin S, Roih is Associate Professor tif Markeiing. Carroll School ofManagemeni, Boston College. The author thanks Russ Belk. l^rry Feick,['im Heath. Rajiv Grover, Bcmie Jaworski. Sandra Milberg, and GeraldZaltman for their many helpful comments. Insights provided by Bart Weilzand three anonymous JMR reviewers are also gratefully acknowledged.

First, research should link brand image strategies andproduct performance. Although research on brand extensionshas begun to examine the effects of using established versusnew brands on market share (Simon and Sullivan 1993;Smith and Park 1992). the relationship between brand imagestrategies and performance has received little attention. Yetbrand image strategies can affect measures of relative perfor-mance (e.g., market share) as weil. Specifically, a need-basedimage strategy establishes a brand's position (relative tocompetition) within a product category, defining its nicheand establishing its market potential. Research ha.s shownthat serving larger niches typically leads to better long-termproduct performance than serving smaller niches within thesame category (Lambkin and Day 1989; Montgomery 1982;Romanelli 1987). Thu.s brand image strategy should affectproduct pertbmiance, because images appealing to largeniches should capture a larger share of a product categorythan brands whose images are targeted to smaller niches.

Second, research should address international as well asdomestic brand image management. Given that effectivebrand image management centers on satisfying customerneeds (Park, Jaworski, and Maclnnis 1986), and that cultur-al,, social, and economic dimensions of international m;u-ketsshape consumer needs (Dannhaeuser 1987; Daun 1983;Hondrich 1983; Reilly and Wallendorf 1987; Tan and Farley1987; Tse, Wong, and Tan 1988), global aspects of imagemanagement must be underslood. Although there are strate-gic advantages to targeting the same customers and satisfy-

163Journal of Markeiing ResearchVol, XXXII (May 1995), 163-175

164 JOURNAL OF MARKETING RESEARCH, MAY 1995

ing the same types of needs cross-nationally (Levitt 1983),in many cases, the niche or positioning opportunities differacross markets (Douglas and Wind 1987; Jain 1989). For ex-ample, Levi's blue jeans creates different brand images forcustomers in different parts of the world. A very social,group-oriented image is used in the United States, whereasa much more individualist, sexual image is maintained inEurope. Similarly, Reebok varies the image of its athleticshoes on the basis of national and regional differences it per-ceives in consumer tastes and preferences. In the UnitedStates, Reebok tries to balance both life.style and athletic im-ages, whereas in Western Europe, the brand image focus ismore narrowly on athletics and performance. Some brandshave successfully targeted the same types of customersusing the same brand image strategies cross nationally (e.g..Coca Cola, Nike), but many firms have failed international-ly by not adapting their strategies to cultural, economic, andother differences among markets (Hill and Still 1984; Ricks1983). Tlierefore, it is imperative that managers identify, as-sess, and respond to global market conditions that can affectthe size of niches for particular brand image strategies.

My purpose is to provide managers with a framework forselecting brand image strategies for international markets.Specifically, I examine the linkage among brand imagestrategies, cultural and socioeconomic factors, and marketshare in international markets. The international brandimage strategies used by consumer goods firms in two in-dustries were identified, indicating the type of needs thebrands seek to satisfy and their positioning strategy in eachmarket. Because the relative performance of a brand posi-tion depends on the size ofthe niche targeted. I also investi-gated cultural and socioeconomic factors hypothesized toaffect the size of the targeted niche and, therefore, the mar-ket potential of the image strategy. Because marketers havelong noted that countries are heterogeneous and comprisedof unique cultural and socioeconomic regional markets (e.g.,the Flanders and Wallonia regions in Belgium, and the Que-bec province in Canada; cf. Douglas and Wmd 1987; Hilland Still 1984; Jain 1989), the study included both a nation-al and a more micro regional focus.

The article is organized as follows. The next section pre-sents a conceptual framework that links brand image strate-gies, cultural and socioeconomic market characteristics, andproduct perlbrmance. Next, I discuss the research design,which involved secondary data and surveys given to productmanagers. I then present results, followed by discussions ofthe managerial and theoretical implications of the findings.Finally, I present directions for further research.

CONCEPTUAL FRAMEWORK

Much of international marketing research has focused onmarketing mix customization versus standardization (for arecent review, see Aulakh and Kotabe 1993). Yet, before onemakes marketing mix decisions, brand image strategies forinternational markets must be developed. Brand imagestrategies must be designed to appropriately position theproduct for tbe targeted market segments. Aithough man-agers have many image and positioning altematives fromwhich to choose, international marketers must explore howcultural and socioeconomic characteristics affect the perfor-mance of brand image strategies.

Using the normative model proposed by Park, Jaworski,and Maclnnis (1986). I investigate three types of brand im-ages—functional, social, and sensory images. These imagesare based on the fulfillment of basic consumer needs—^prob-lem solving and problem prevention (functional), groupmembership and affiliation (social), and novelty, varietyseeking, and sensory gratification (sensory)—and are con-sistent with other descriptions of relationships involvingpersons, environment, and sociocultural systems (Mali-nowski 1944), Maslow's (1954) need hierarchy, frameworksof motivating forces driving consumer behavior (Rossiterand Percy 1987), and a typology of consumer needs devel-oped from a review of behavioral science research on needsand values (Hanna 1980). It is not uncommon for brandswithin a category to be uniquely positioned along thesetypes of needs. For example in the United States toothpastecategory. Crest's functional image is based on cavity, tartar,and decay prevention. Ultra Brite is positioned toward socialneeds to provide whiter teeth and social acceptability, andAim's image is based on taste and sensory gratification. Al-though the normative model suggests firms should selectonly one type of need when developing brand images (Park,Jaworski, and Maclnnis 1986), research has shown that (1)firms tend not to emphasize only one need, and (2) strategiesemphasizing one need do not necessarily lead to better fi-nancial performance than those based on multiple needs(Roth 1992). The findings indicate that managers feel it iseither necessary and/or advantageous to diversify theirbrand image strategies across two or more needs. The fre-quency with whicb firms try to be "many things to manypeople" indicates a need to better understand the linkagesbetween market environments and consumer needs. Identi-fying environmental factors that favor one type of consumerneed over another will help managers select the brand imagestrategies with the greatest market potential.

Cultural and Socioeconomic Characteristics ofInternational Markets

Social science and management research otfer insightsinto the effects cultural and socioeconomic characteristics offoreign markets may have on consumer needs and the sub-sequent success of brand image strategies. Figure 1 presentsa conceptual model of the potential moderating effects ofculture and socioeconomics on the performance of brandimage strategies. In international markets, environmentalcharacteristics are likely to moderate the image-perfor-mance relationship. As shown, two characteristics, (1) na-tionai culture and (2) regional socioeconomic conditions af-fect the performance of functional, social, and sensory brandimages. In addition, because market and firm conditionsmay also impact performance, the model shows likely co-variates, that is. market experience, extent of competition,and marketing mix implementation problems, that should bemanaged when examining the effects of environmental fac-tors and brand image on performance.

National Culture

A country's culture has long been Identified as an envi-ronmental characteristic that influences consumer behavior,and the many aspects of a culture affect differently the needs

Global Brand Image Strategies 165

Figure 1THE EFFECTS OF CULTURAL AND SOCIOECONOMIC FACTORS ON THE PERFORMANCE OF BRAND IMAGE STRATEGIES

Brand Image Strategy• Fun tional

• Sensory

• 1 _

• 4 -

••

• • L

National Culture• Power distance•Uncertainty

avoidance• Individualism

L

RegionalSocioeconomics• Modernity

•

-

<

ProductPerformance• Market share

n.j.

L

Markel & Firm Conditions• Market experience• Competitive problems• Marketing mix problems

consumers satisfy through the acquisition and use of goodsand services. Hofstede's (1984) influential work on cross-cultural value systems identifies three aspects of culture thatcan be related to consumer needs and brand images—powerdistance, uncertainty avoidance, and individualism. Al-though Hofstede's research was conducted in organizationalsettings, the values he identifies have been associated withconsumer behavior as well as with work-related activities(for a recent example, see Lynn, Zinkhan, and Harris 1993).

Power distance describes the extent to which a culturefosters social inequality. Cultures high in power distancetend to emphasize tbe importance of prestige and wealth inshaping boundaries or vertical relationships between socialand economic classes such as rich and poor, and superiorsand subordinates (Hofstede 1984; Inkeles 1960; Lenski1966). In high power distance cultures, people seek to main-tain and increase their power as a source of .satisfaction. So-cial consciousness is high, and they are motivated by theneed to conform with those in their class or in classes towhich they aspire. In low power distance cultures, peopleare much less focused on class differences and social aspi-rations. If one relates the aspects of cultural power distanceto consumer needs, then it appears that social brand imagesshould be the best fit for high power distance cultures be-cause people are highly motivated by social status and affil-iation norms. On the other hand, when power distance islow, social brand images will not match the cultural normsbecause consumers are not motivated by group-related

needs. More formally, power distance is hypothesized to af-fect brand image performance as follows:

H,: The effects of social brand images on market share will begreater (Iower) when cultural power distance is high (low)than when power distance is low (high).

Uncertainty avoidance captures the cultural pattern ofseeking stability, predictability, and low stress rather thanchange and new experiences (Hofstede 1984). People inhigh uncertainty avoidance cultures are risk averse, resistantto change and variety seeking, and have a low tolerance forambiguity. For brand image management, uncertaintyavoidance can be linked to aversion to a stimulus or event.When consumers experience some type of aversion, theywill be motivated to remove the aversion or at ieast reduce itas far as possible (Rossiter and Percy 1987). For example,when averse to cavities or other dental hygiene problems,consumers will seek a toothpaste capable of reducing therisk of cavities. Functional brand images are positioned toprevent or solve problems and, therefore, reduce perceivedrisk. When cultures are high in uncertainty avoidance, con-sumers will be very focused on risk aversion and problemsolving and prevention and, thus, will be good candidatesfor functional brand images. In contrast, when uncertaintyavoidance is low. risk aversion is diminished, and consumersare more open to variety and novelty behaviors. Thus, in lowuncertainty avoidance cultures, sensory brand images focus-ing on variety, novelty, and sensory gratification are effec-

166 JOURNAL OF MARKETING RESEARCH, MAY 1995

tive. More formally, uncertainty avoidance is hypothesizedto affect brand image performance in the following ways:

Hiai The effects of functional brand images on markel share willbe greater (lower) when cultural uncertainty avoidance ishigh (low) than when uncertainly avoidance is low (high).

H2b: The effects of sensory brand images on market share willbe greater (tower) when cultural uncertainty avoidance islow (high) than when uncertainty avoidance is high (low).

Individualism is an aspect of culture that pertains to peo-ple's tendency to value personal and individuai time, free-dom, and experiences (Hofstede 1984; Parsons and Shils1951; Riesman, Glazer. and Denney 1953). In contrast, cul-tures that emphasize collectivism exhibit patterns of groupor collective thinking and acting. In other words, cultureshigh in individualism tend to seek variety and hedonistic ex-periences, whereas collectivist cultures correlate more withconformity and group behavior. In addition, people in highindividualism cultures tend not to follow social norms, but,rather, form relationships, make decisions, and initiate be-haviors independent of others. In terms of brand imagestrategies, cultures high in hidividualism seem well-suitedfor sensoiy images tbat emphasize variety, novelty, and in-dividual gratification. On the other hand, cultures with lowindividualism will tlnd social brand images tbat reinforcegroup membership and affiliation more attractive. The fol-lowing hypotheses describe how cultural individualism willaffect brand image performance:

H3a: The effects of sensory brand images on market share will begreater (lower) when cultural individualism is high (low)than when uncertainty avoidance is low (high).

H^ : The effects ot" sociai brand images on market share will begreater (lower) when cullural individualism is low (high)than when individualism is high (low).

Regional Socioeconomics

Although country-level factors, such as national culture,affecting brand image strategy are important, micro geo-graphic units, that is, regions within countries, merit investi-gation as well. Although the predominant approach in pastinternational brand management research has been to exam-ine country-level differences (e.g., Boote 1983; Huszagh,Fox, and Day 1986; Martenson 1989), national analyses ofmarketing strategy effectiveness may be too broad in scope,causing important intra- and inter-country similarities anddifferences to be overlooked. Greater homogeneity wil! befound within more localized regions than in countries. (Seeanthropological research on intra- versus inter-cultural vari-ation in Boster 1987: Roth and Moorman 1988.) Althoughproponents of the global standardization approach (e.g.,Levitt 1983) have espoused the growing similarity of con-sumers cross-nationally, domestic managers, in fact, contin-ually look for sophisticated ways to define markets and de-velop micromarketing strategies, such as the integration ofcensus and market research data bases to identify geodemo-graphic, geopsychographic, and other area-based target mar-kets within the United States (McKenna 1992; Mehrotra1990). Furthermore, as trading nations reduce and eliminatestructural, political, and economic barriers, the search for

similar consumers within trading blocs will become moreimportant than national differences.

Although cultures tend to transcend entire nations, socialand economic conditions of en vary substantially withincountries. Within most countries, regions exist that differmarkedly from one another with regards to income, mobility,media access, employment, and other socioeconomic charac-teristics. Sociologists often assess a region's socioeconomiclevel in terms of its "modemity"—a measure of the institu-tions and organizations that provide life experiences for mem-bers of a society (Inkeles 1983; Inkeles and Smith 1974).

One aspect of modemity that relates to consumption andthe viability of brand image strategies is disposable Income.Disposable income indicates the amount of resources con-sumers allocate to goods and services. When resources arelimited, consumers use goods and services to satisfy theirmost basic, functional needs. As resources increase, so doesconsumer willingness to spend money on products that sat-isfy more symbolic and sensory needs. For example, themajor consumption goal of Zinderois Nigerians, a culturewith very limited economic resources, is nutritional self-suf-ficiency, in contrast to the more symbolic and experientialconsumption patterns of more affluent North Americans(Wallendorf and Arnould 1988). Hence, when socioeco-nomic conditions are low, functional brand image strategieswill have the most appeal. When socioeconomic conditionsincrease, consumption of social and sensory goods to satis-fy symbolic and hedonic needs is more likely.

Another aspect of modemity that inliuences consumer be-havior is exposure to consumption. The extent to which con-sumers are exposed to Western, material-oriented consump-tion cultures wiil influence their attraction to specific prod-ucts. The demonstration effect, a phenomenon in whichpoorer consumers buy symbolic and sensory products (wilhlittle functional value) to identify with consumption soci-eties to which they have been exposed (Keyfitz 1982;Nurske 1953), explains why consumers in poorer or lesser-developed countries purchase status and hedonic goodswhen functional ones would seem more appropriate (Beik1988; Keegan, StiU, and Hill 1987). In short, what con-sumers see. hear, and access via television, other media, au-tomobiles, and other forms of mobility shapes their needsand attitudes toward the consumption of products and ser-vices (Belk 1988: Jain 1989; Keegan, Still, and Hill 1987;Keyfitz 1982; O'Guinn. Lee, and Faber 1986: Weimann1984). When markets are characterized by low levels of re-gional modemity, consumers will not be familiar with thematerial aspects of consumer culture nor with the ability ofgoods and services to satisfy social and sensory needs. Theirlack of mobility and exposure to media will prevent themfrom leaming about the more symbolic and experiential as-pects of product consumption. Instead, products will be val-ued for their functional capabilities, thus making functionalbrand image strategies more suited to tbe environment thansocial or sensory images. As regions become more modem,however, exposure to material culture increases, motivatingpeople to both desire tbe goods they see being consumed inother cultures and to purchase products that associate themwith other cultures and societies. Hence, social brand im-ages that promote group identification will have greater ap-peal in high modemity regional markets. However, high lev-

Global Brand Image Strategies 167

cis of modemity also foster more individual and self-aware-ness values (Inglehart 1977; Yankelovich 1981), which areassociated witb needs for self-enhancement and personalfuinilment and consistent with arousal and stimulation (sen-sory) need-based brand images.

In most countries, the correlation between per capita in-come and modemity is quite high (e.g., Johansson andMoinpour 1977; Sethi 1971). Measures of mobility andWestem exposure can provide an accurate description of amarket's degree of modemity, capturing both economic anddemonstration effects on consumption. The hypothesized re-lationships between regional socioeconomics (modemity)and brand image performance are summarized as follows:

H4a: The effects of functional brand images on markel share willbe greater (lower) when regional socioeconomics is low(high) than when regional s^xrioecononiics is high (low).

H4h: The eifects of social brand images on market share will begreater (lower) when regional socioeconomics is high (low)than when regional socioeconomics is low (high).

H,4 : The effects of sensory brand images on markel sh;ire will begreater (lower) when regional socioeconomics is high (low)than when regional socioeconomics is low (high).

METHOD

1 identified and collected environmental data from coun-tries and the regions within them that are culturally and so-cioeconomically varied. Then I administered surveys tofirms marketing consumer goods in these countries and re-gions to detemiine botb the brand images used and tbe re-sulting product performance. Collectively, these two datasources provided the means for testing the model and hy-ptnheses outlined previously.

Data Collection

Step I—Socioeconomic data base. I compiled a data baseof socioeconomic characteristics for 60 cities and townswithin 10 countries (Argentina, Belgium, China. France,Germany, Japan, Italy, Netherlands, Peru, and Yugoslavia),which were chosen because of their social, economic, andcultural diversity. Socioeconomic data for the regions withineach country (the number of regions per country ranged fromfour to nine) were drawn from a variety of publicly availablestatistical data sources. Because multiple socioeconomicvariables were collected, factor analysis was u.sed to developa scale indicating each regional market's degree of socioeco-nomics (see the Appendix for a description of the analysis).

Step 2—Survey development. My survey measured allitems at the regional level and included questions about thebrand image strategy used, market experience, problems en-countered, and market share for a brand in a particular re-gional market. The survey was developed, pretested, andmodified for administration to marketing and product man-agers responsible for marketing their products in one ormore of the 10 countries in the data base.

Step 3^Satnple development. United States finns manufac-turing consumer gtxxls in the blue jean and athletic shoe cate-gories were then identified and chosen because of the diversebrand image strategies used to market them. In addition, bothof these categories have many United States-ba-sed multina-tional competitors, thus offering an opportunity to explore

strategic performance across all of the regions in the database.Step 4—Manager identification. I contacted marketing

managers at each company and asked them to identify thecountries in which they marketed their products. The personcontacted often provided the name of another manager re-sponsible for intemational marketing for a particular coun-try. I asked managers of firms marketing their brand in oneor more markets wiihin tbe 10 countries to participate in amarket research study examining the performance of brandimage strategies in international markets.

Step 5—Sun-ey administration. I mailed surveys cus-tomized for each country, a cover letter, and $1 as a token ofgratitude to the manageni. The survey contained items foreach region within the countries served by the fimi. If a re-sponse was not received after two to three weeks, I mailedreminder letters and questionnaires.

Thirty eight managers from 11 firms retumed usable ques-tionnaires, yielding a response rate of 33% (115 managerswere mailed surveys). Many of the participants in the studymanaged a brand in multiple foreign markets. Collectively,they reported on 209 cases of a particular brand's strategy,market, and performance in a particular regional markel. Onaverage, each manager participating in the study reported sixcases. To assess the degree of non-response bias, a sample of10 of the 77 managers who did not retum questionnaireswere contacted by phone and found not to differ from re-spondents with regard to the brand image, covariate, or per-formance survey mea.sures described subsequently. The mostcommon reasons for not responding were proprietary dataconcems, lack of time, and lack of regional-level data.

Measures

Brand images. The survey asked managers to characterizetheir brand's image in each particular regional market by al-locating 100 points across three types of strategies: func-tional, social, and sensory, with more points being allocatedto the more emphasized images (see the Appendix). Man-agers could allocate 100% of the points to one brand image(a depth strategy approach) or allocate the points across twoor three of the strategies (breadth approach).

Culture and socioeconomics. Cultural measures of na-tionai power distance, uncertainty avoidance, and individu-alism were taken from indexes developed by Hofstede(1984), on the basis of his survey of over IO(),OtXJ respon-dents from 66 countries. Hofstede's sample is not represen-tative of Ihc entire population, but of tbe middle classes em-ployed in multinational corporations. Because we are not in-terested in absolute scores, but in cross-national cultural dif-ferences, the indexes are valid because the sampled respon-dents were well-matched demographically, with the onlysystematic difference being their nationality. I measured re-gional socioeconomics in terms of each region's modemity.which was estimated as a factor score determined usingscoring coefficients computed from the factor loadings(refer to the Appendix). The scoring coefficients were esti-mated using the Anderson and Rubin (1956) method, whichproduces accurate estimates for use as independent variablesin regression analyses (Lastovicka and Thamodaran 1991).

Product performance. Managers indicated the brand'sniaikct share during tbe most recent annual period for thebrand in each regional market. Market share was chosen be-

168 JOURNAL OF MARKETING RESEARCH, MAY 1995

cause of its wide use as an indicator of performance amongconsumer product marketers (e.g.. Smith and Park 1992), as-sociation with a business's relative size in its served market(e.g., Buzzell and Gale 1987). and consistent use in other in-ternalional marketing studies of business performance (e.g.Ryans 1988; Szymanski, Bharadwaj, and Varadarajan1993). Although market share differs from typical measuresof communication effectiveness (e.g., brand awareness,brand attitude), the images reported are a function of thefirm's entire marketing program—not jusl tbeir advertising(Park, Jaworski. and Maclnnis 1986) and. tberefore. betteraccounts for the entire marketing program's performance.

Covariates. Tbe survey included tbree covariates: marketexperience, extent of competition, and marketing mix prob-lems. Intemational marketing experience is often a key de-terminant of intemational marketing success (Denis and De-pelteau 1985; Gronbaug and Graham 1987; Johanson andVahlne 1977). In addition, the length of time the brand hast>een available in the market often explains unstable growtbrates and performance measures. In some cases, new marketentrants demonstrate rapid performance growth before pro-gressing into the growth and maturity stages of tbe productlife cycle. On the otber hand, some firms exporting productsto new markets often encounter slow growth rates until theyadjust their marketing programs as more market experienceis gained. Managers reported the number of years the brandhad been available in the regional market.

Managers also reported the extent to wbicb they encoun-tered competition problems in each regional market. Thisitem was measured using a seven-point Likert scale. Extent ofcompetition is an important covariate because direct competi-tors try to serve the same niche, and fight for miU"ket share. Inaddition, competition affects intemational marketing perfor-mance (Denis and Depelteau 1985; Jain 1989). Hence, brandimage performance could be negatively affected when com-petitive problems are encountered in a regional market.

Effective implementation of the marketing mix brandconveys to consumers tbe image intended by the firm. Firmsexperiencing difficulties managing their marketing mix willhave problems effectively communicating the intendedbrand image to consumers, and. subsequently., tbeir perfor-mance will be adversely affected. The prevalence of market-ing implementation problems can hun a brand's perfor-mance and is. therefore, another important covariate. Thesurvey included four Likert scaled questions on problemsencountered with tbe marketing mix and enabled managersto report the extent of product, distribution, pricing, and ad-vertising and promotion problems experienced in each re-gional market. A principal components analysis showed thefour items to load on a single factor. All items bad factorloadings and communalities greater than .4. Cronbacb'salpha, when combining the four variables into one scale,was .727. Tbe Appendix displays the covariate measures.

Using multiple regression, I examined the hypothesespredicting the effects of the environmental factors on brandimage strategy market sbare. Specifically. I used moderatedregression analysis (Amold 1982; Sbarma, Durand. andGur-Arie 1981) to examine the moderating effects of cultureand socioeconomics on the market share of brand imagestrategies. Prior to hypothesis testing, the data were exam-ined for outlying cases. Sixteen cases bad market shares be-

yond three standard deviations of the mean and, therefore,were not included in subsequent analysis. Tbe sample usedfor hypothesis testing was thus 193 cases. Table I reportsPearson correlations and descriptive statistics for the vari-ables used in this study. Social and sensory images bad asmall, positive correlation, indicating a weak pattem of jointemphasis in the same brand image strategy. In contrast,functional image had a strong, negative correlation withbotb social and sensory images, wbich indicated that theseimage pairs tended not to be emphasized together. Also, nofirms reported using a single-need brand image strategy—the maximum points allocated to any one strategy was 75 of100— thus. all firms incorporated two or more needs intotheir brand image strategy.

Tbe level of intercorrelations among tbe independent,moderator, and covariate variables was, in general, low tomoderate, ranging from .001 to .54. Tbere were two excep-tions—(1) cultural individualism and regional socioeco-nomics (r = .714), and (2) marketing mix problems and cul-tural individualism (r = -.579). Correlations at this level sig-nal the possible presence of multicollinearity. wbicb inflatestbe standard errors of parameter estimates.

To minimize and cbeck for the presence of multi-collinearity. a number of steps were taken. First, I centeredall of the independent, moderator and covariate variables(i.e., put tbem in deviation score form so tbat tbeir means arezero).' As shown by Aiken and West (1991. p. 32). center-ing variables significantly reduces multicollinearity, therebyproviding unbiased parameter estimates. Second. I calculat-ed variance inflation factors (VIF) to examine the extent towbicb nonorthogonality among independent, moderator,and covariate variables inflates standard errors. Tbe averageVIF was 1.94, ranging from 1.32 to 2.95—well below tbecutoff of 10 recommended by Neter, Wassennan, and Kut-ner (1989, p. 409). The centering procedure and the VIFfmdings suggest that multicollinearity is not a threat to the.substantive conclusions to be drawn from the analysis dis-cussed subsequently.

The multiple regression model tested contained the effectsof each independent and moderator viuiable (i.e., three differ-ent brand images; and one regional and four cultural factors,respectively), three covariate variables, and the independent Xmoderator interactions for testing eacb hypothesis. The inter-action temis indicate the moderating effect of the environ-mental factors on the performance of the brand image strate-gies. To control for any correlated residuals due to multipleproduct category and within-category brand observations, 10dummy variables were created (one dummy variable for thetwo product categories, five dummy variables for the six bluejean brands, and four dummy variables for tbe five athleticsboe brands) and included in the regression model.

'This process is recommended when conducting regression analyses(e.g., Dillon and Goldstein 1984; Marquardt 1980), especially when prod-uct terms such as independent X moderator interactions are to be used(Aiken and West 1991; Cronbach 1987).

Global Brand Image Strategies 169

cntum£UJ

IoQ

<a:OCUJoo

tuozUJQ.tuO•COLLcnocnIcnUJ

ccOcnUJ

o

cnzo§tuccccoo

c C 1,

I I

^ ^ C - <u

II gl

•S c Sfj-

f^ O 00

3 r in

g oo — 00 r-o Q ^ >c

rn M p p —•

•g

ff) ^

•3

E "2— «, Q

- S

< e's00 .2 s * — UJ .5

•=13 S

00 fn O

£ ^

: m a

170 JOURNAL OF MARKETING RESEARCH, MAY 1995

Table 2MODERATING EFFECTS ON BRAND IMAGE STRATEGY

PERFORMANCE: STANDARDIZED REGRESSION

COEFFICIENTS

Variables

Main effects nf independent variablesFunctional Brand ImageSocial Brand ImageSensory Brand Image

Main effects of moderatorsNational Cultural Power DistanceNational Cultural Uncertainty AvoidanceNational Cultural IndividualismRegional Socioeconomics

Independent X moderator interactionsFunctional Image X Power DistanceSocial Image X Power DistanceSensory Image X Power Distance

Functional Image X Uncertainty AvoidanceStx-ial Image X Uncertainty AvoidanceSensory Image X Uncertainly Avoidance

Functional Image X IndividualismScKial Image X IndividualismSensory Image X Individualism

Functional Image X Regional SocioeconomicsScxial Image X Regional SocioeconomicsSensory Image X Regional Socioeconomics

Main effects afcovariaiesMarket ExperienceCompetitive ProblemsMarketing Mix Problems

Dummy variablesProduct CategoryBrand 1Brand 2Brand 3Brand 4BrandsBrand 6Brand 7Brand 8Brand 9

Adjusted R-squareF for full model (32 d.f.)

Market Share(n = 193)

.001

.117-.042

.117-.074-.023

.165*

-.074.198**.175**

-.025.094.053

.169*-.339***

.261***

-.203**.305***.286**»

.042-.174**-.365***

.021-.339***

.001

.470***-.104

.012

.333***

.234**

.023-.019

.50224.86***

**p <.05*p<.10

RESULTS

Brand Image and Moderator Main Effects

Table 2 shows the results of the moderator regression anal-yses. Although the hypothesis did not address main effects,all independent, moderator, covariate, and dummy variablemain effects were included in the mode! first to more pre-cisely estimate tbe moderating (interaction) eftects. Tbe maineffects of the brand images on market share were very smalland not statistically significant {p > .10), whicb indicates thatfirms do not necessarily benefit from any one type of brand

image strategy. Market sbare was not directly related to tbethree national culture measures (/j> .10). Marketing productswben regional socioeconomics is bigh, however, does tend toenhance market share (p = .165, p < .10).

Moderating Effects of Culture and Socioeconomics onBrand Image Performance

The sign and signiftcance of tbe independent X modera-tor variable interaction p coefficients are used to accept orreject tbe bypotbeses. I accomplisbed furtber diagnosis ofstatistically significant interactions via post-hoc simpleslope tests (Aiken and West 1991). Following Coben andCohen's (1983) study, I recomputed moderators exhibiting asignificant interaction with brand image into higb, medium,and low values corresponding to (1) one standard deviationabove tbe mean, (2) at the mean, and (3) one standard devi-ation below tbe mean, respectively. The higb, medium, orlow value represents the "conditional value"' of tbe modera-tor (Darlington 1990). 1 tben performed simple regressionsof market share on brand image, the conditional value of tbemoderator, and the brand image X conditional moderator in-teraction for tbe high, medium, and low moderator values. Iused T-tests to assess wbetber tbe simple slopes differ fromzero, wbicb indicated if tbe regression of market sbare onbrand image was positive or negative at different condition-al values of the moderator.

Cultural power distance. H] stated that social brand im-ages would perform well wben power distance was higb. Asexbibited in Table 2, tbe power distance X social brandimage interaction term was positive and statistically signifi-cant O = .198, p < .05). Although no formal bypotbeseswere made about the moderating effect of power distance onthe performance of functional or sensory brand images,these effects were also explored and are shown in Table 2.Power distance bad a significant, positive moderating effecton market sbare for sensory brand images O = .175, /? <.05), but did not significantly affect functional images.

Tbe results of simple slope analyses shown in Table 3 pro-vide more insigbt into tbe moderating effects of power dis-tance on brand image. Tbe first line in Table 3 illustrates thebrand image beta coefficient when market sbare is regressedon three variables: (1) brand image, (2) the cultural powerdistance moderator at one standard deviation above its mean(high power distance), and (3) tbe brand image X high cul-tural power distance interaction. If the regression is signifi-cantly different from zero, evidence exists that tbe brandimage's effect on market sbare is indeed moderated at thehigh conditional value of cultural power distance. For socialbrand images, when power distance is higb, tbe effect onmarket sbare is very positive (p = .501, p < .01). Whenpower distance is medium, the effect on market sbare is pos-itive, but much less so tban when power distance is bigb (p= .134, p < .05). When power distance is low, tbe impact onmarket share is negative (p = -.233. p < .05). Thus, serialbrand images enbance performance wben power distance ishigb, and to a lesser extent wben it is medium, whereas so-cial images burt sbare when power distance is low, support-ing hypothesis Hj.

Tbe resuits in Tables 2 and 3 sbow tbat power distance hasa similar moderating effect on tbe performance of sensory

Global Brand Image Strategies 171

Table 3SLOPE ANALYSIS OF SIGNIFICANT INDEPENDENT X MODERATOR INTERACTIONS:

STANDARDIZED REGRESSION COEFFICIENTS AND T-VALUES

Variables (a}

FunctionalBrand tmage

1

.374

.519-.136

-.098.046.215

t-value

I1.S.

4.71*»4.10*"

-1.51

-1.35.25

2.43*

SocialBrand Image

.501

.134-.233

-.402.198.300

.538

.108-.282

t-value

4.64**2.00«

-2.18*

-2.78'*2.26*3.57**

4.50**1.47

-3.07**

SensoryBrand Image

.452-.256-.259

.486-.063-.424

.468-.088-.278

t-value

3.47**-2.92*-3.04**

4.39**-.67

-5.40**

4.27**-.71

^ .02**

Cultural Power Distance X Brand ImageImage at High Power DistanceImage At Medium Power DistanceImage at Low Power Distance

Cultural Individualism X Brand ImageImage at High IndividualismImage at Medium IndividualismImage at Low Individualism

Regional Socioeconomics X Brand ImageImage at High Regional SocioeconomicsImage at Medium Regional .SocioeconomicsImage at Low Regional Socioeconomics

**p < .01*p < .05

brand images. High levels of power distance (p = .452, p <.01) enhance market share, but medium (P = -.256. p < .05)and low (^ = -.259, p < .01) levels reduce share. In summa-ry, the results show that cultural power distance is a rela-tively strong moderator of social and sensory brand imageperformance. When a market's power distance is high, in-creased use of social and sensory brand image strategies en-hances performance, whereas limited use of social and sen-sory images enhances performance when cultural power dis-tance is low.

Cultural uncertainty avoidance. Hj^ and H2b stated thatfunctional brand images would perform well in high uncer-tainty avoidance markets and that sensory brand imageswould perform well when uncertainty avoidance was low.Although not addressed by the hypotheses, the effects of un-certainty avoidance on the performanee of social brand im-ages were examined as well. The uncertainty avoidance Xbrand image interactions found in Table 2 were not statisti-cally significant (p = .025 for functional image, p = .1)94 forsocial image, p = .053 for sensory image; all had p-values >.10). H2a and H2b were not supported.

Cultural individualism. H^ stated that sensory brand im-ages would perform well in high individualism cultures, andH b stated that social brand image strategies would perfonnweli in cultures low in individualism. The results in Table 2show strong, positive individualism X sensory brand imageinteractions (p = .261, /? < .01), as well as a negative indi-vidualism X social brand image interaction (p = -.339, p <.01). Although not included in the hypotheses, cultural indi-vidualism had a positive and marginally significant moder-ating effect on the maiket share of functional brand images(p = . 169, p < . 10). Slope analyses illustrated in Table 3 pro-vide specific insights into the moderating effects of individ-ualism. For sensory images, high individualism has a posi-tive effect on sensory image market share (P = .486, p <.01). Medium individualism has no etfect on sensory imageshare O = -.063, p > .05). and low individualism has a neg-ative effect on sensory image share (p = -.424, p < .01).

Thus, H3a received strong support.For social images, market share was affected negatively

when individualism was high (p = -.402, p< .01), enhancedwhen individualism was medium (p = .198, p < .05), andgreatly enhanced when individualism was low (p = .300, p< .01). Hence, H i, also received strong support.

Although I made no hypothesis about the moderating ef-fects of individualism on functional image performanee.Table 3 shows that functional images are enhanced when in-dividualism is at high (p = .374, p < .01) and medium O =.5\9, p < .01) levels, but not affected when individualism islow (p = -.136, /7 > .05). In summary, when cultural indi-vidualism is high, sensory and functional images should beemphasized, whereas social images should not. When indi-vidualism is low, social images will have the greatest posi-tive impact on market share, whereas sensory images willhurt performance.

Regional socioeconomics. H4a_4,. stated that in !ow so-cioeconomic regions, functional brand images will perfonnwell, and in high socioeconomic regions, social and sensorybrand images will perform well. Table 2 exhibits directionaland statistical support for these hypotheses. The functionalbrand image X regional socioeconomics interaction wasnegative and statistically significant (p = -.203, p < .05), atidthe social and sensory brand image X regional socioeco-nomics interactions were positive and statistically signifi-cant O = .305. p < .01 for social image; p = .286, p < .01for sensory image). The slope analyses in Table 3 providefurther insights into the moderating effects of regional so-cioeconomics. For functional images, low regional socioe-conomics positivly impacts market share (p - .215,/)< .05).At medium and high levels of socioeconomics. functionalimages have little impact on share (p = .046, p > .05 atmedium socioeconomics; p = -.09, p > .05 at high socioe-conomics). Thus. H4a receives partial support because lowlevels of regional socioeconomics enhances functional brandimage performanee, but does not affect performance at highlevels. For social images, market share is enhanced when re-

172 JOURNAL OF MARKETING RESEARCH, MAY 1995

gionaJ socioeconomics is high (P = .538, p < .01), not affect-ed when socioeconomics is mediutn (P = .108, p > .05), andnegatively affected when socioeconomics is low (P - -.282,p < .01). thus providing strong support for H4b. The resultsalso stronly support H4j.: For sensory images, market share isenhanced when regional socioeconomics is high (p = .468. p< .01), not affected when socioeconomics is medium O =-.088,/7 > .05), and adversely affected when socioeconomicsis low (p = -.278. p < .01). In summary, low regionai so-cioeconomics enhances performance when functional brandimages are emphasized as opposed to social and sensory im-ages. High regional socioeconomics achieves better perfor-mance using social and sensory brand image strategies.

Main Effects ofthe Covariates

The main effects of market experience, competitive prob-lems, and marketing mix problems are also found in Table 2.Market experience, operationalized as length of time in themarket, did not affect market share (/ > . 10). however, com-petitive and marketing mix problems did. For compelitionproblems, the effect on market share was negative (P =- . 174) and significant atp< .05, indicating that the presenceof competitive problems hampers market share. For market-ing itiix problems, the beta coefficient was also negative (p= -.365) and significant at p < .01. indicating that whenmanagers experience difficulties implementing their market-ing programs, product performance suffers. In summary, themore (fewer) competitive and implementation problems ex-perienced, the lower (higher) market share tended to be.

DISCUSSION AND IMPUCATIONS

My study addressed important limitations of past intema-tional marketing management research by linking brandimage strategies, the management of which should precedemarketing mix customization and standardization decisions,to product performance across a variety of global markets.The findings suggest that environmental characteristics offoreign markets are (1) important indicators of customersegments and market potential and (2) provide insights intothe emphasis managers should place on functional, social,and sensory brand image strategies.

Brand image success relied on several factors. First, re-gional socioeconomics was a strong moderator of brandimage market share. Most intemational marketing researchfocuses on cross-national differences, but the results pre-sented here imply that managers should narrow their geo-graphical focus to cities and towns in addition to examiningentire countries. Heterogeneity within countries makes itdifficult for managers to develop brand image strategieswith strong within-country appeal, yet, the micro-market-ing, regional focus allows managers to develop strategiestargeted at more homogeneous target markets. As the find-ings illustrate, emphasis on functional brand image strate-gies enhances performance when regional socioeconomicsis low. When regional socioeconomics is high, and con-sumers have wide exposure and easy access to Westem con-sumer culture through media and mobility, emphasis shouldbe shifted to sociai and sensory brand images to maximizemarket share.

Second, managers can use knowledge of a market's na-tional cuiture to develop successful brand image strategies.

Two aspects of culture had significant impact on the perfor-mance of brand image strategies—power distance and indi-vidualism. In low power distance cultures (e.g., Gennany,Netherlands, Argentina) in which people are not highly fo-cused on social roles and group affiliation, functional brandimages that de-emphasize the social, symbolic, sensory, andexperiential benefits of products are most appropriate. Whenthe country's degree of power distance is high (e.g., China,France, Belgium), social and/or sensory needs should beemphasized.

In countries with high individualism cultures (e.g., Euro-pean countries), brand images that emphasize functional,variety, novelty, and experiential needs are more effectivethan social image strategies. On the other hand, cultureswith low individualism (e.g., Asian countries) are moreamenable to social brand image strategies that emphasizegroup membership and affiliation benefits than they are tosensory brand images.

Given the strong regional effects on modemity, one couldexpect regional cuiture also to have a moderating effect onbrand image performance. Unfortunately, cultural data forregions within countries are not widely available outside ofNorth America (e.g., Kahle 1986). Nonetheless, further re-search examining subculture effects may enrich our under-standing of brand image performance.

FURTHER RESEARCH

My study provides a starting point for helping managersmarket products intemationaliy so as to best align brand im-ages with customer needs. As with all empirical research, thestudy and results presented have limitations, which provideinteresting avenues for further research in three specificareas: (1) extensions of the present model to other productcategories, (2) the impact of brand image customization onbrand equity, and (3) altemative measures of brand image.

First, the findings can, to some extent, be generalized toconsumer goods positioned on a variety of consumer needs.This study found the moderating effects of culture and re-gionai socioeconomics on market share for the two productcategories examined after accounting for both category andbrand effects. These product categories, however, may besomewhat unique because they are very "American" goodsthat often carry strong cuiturai messages. These goods gen-erate strong country-of-origin effects, because perceptions ofathletic .shoe and blue jean brands encompass the UnitedStates' images of modem, Westem fashion iuid athleticism.Research has shown, in fact, that brand purchase intentionsare elevated when a country's positive image is closely linkedwith benefits desired from a product category (Han and Terp-stra 1988; Roth and Romeo 1992). Here, Because most ofthebrands were marketed abroad using the United States brandname, country image eifects have likely occured. Yet, withintwo categories of country image-sensitive brands, managersimplemented unique brand image strategies, which resultedin varying market shares depending on the environmentalconditions of their foreign markets. Nonetheless, the gener-alizability of these findings might be limited to brands thatcan be tied to a strong, positive country image. Further re-search on categories not so positively related to countryimage will be useful to explore the marketing environment'simpact on brand image strategy effectiveness.

Global Brand Image Strategies 173

Second, further research can examine how changingbrand image and marketing program strategies affects imageperceptions and brand equity. An important feature of glob-al marketing is the mobility of consumers and the frequencyof their exposure (via communication vehicles) to many cul-tures—especially the Westem consumption culture (Beik1988; Quelch and Hoff 1986). Managers must remain awarethat although a product may require brand image cus-tomization (for exampie, shifting an existing brand's relativeemphasis from social to more functional needs when enter-ing a new market), cultural and/or socioeconomic shifts mayrequire further brand image adjustments. For example, as amarket's level of modernity increases, issues of cross-na-tionai brand image consistency must be evaiuated to createand maintain a clear image for consumers exposed to thebrand from various intemational sources. Where previouslythe incidence of cross-border exposure to a brand's multi-re-gional marketing programs may have been low, increasedmodemity wiil make such occurrences more prevalent. Fail-ure to achieve brand image consistency may adversely, inthe long run, affect the brand's equity and the company'sreputation by causing too much confusion about the brand'sposition in the marketplace (Levitt 1983), or by negativelyaffecting perceptions of product quality and value that canaffect market share (Smith and Park 1992). Further studiesare needed that relate environmental changes, consumerperceptions, and the effects of customizing image strategies.

A third area for further research is the measurement ofbrand image. Here, I investigated the emphasis placed onthree image strategies (functional, social, and sensory).Managers have other positioning altematives at their dispos-al (see Crawford i985) that also warrant investigation. Fur-thermore, this research used survey data from managers,thus, the brand image measures represents managers' as-sessments ofthe image they intended consumers to perceive.The meaning consumers attach to the brand may in fact dif-fer from what management intended. To address this possi-ble validity issue, consumer perceptions of brand imagesshould be measured. One procedure used in other intema-tional marketing research (e.g.. Mueller 1987; Tse, Belk,and Zhou 1989) would have consumers evaluate advertisingcontent as a means for assessing perceived brand image. An-other approach that takes into account not just advertisingbut the entire marketing program uses more exploratory in-terviewing techniques. For example, one-on-one interviews,in which consumers elicit image descriptions of the brandand their relationship with it, can be used to develop brandimage measures (Blackston 1992; Levy J985). These ap-proaches would provide richer brand meanings and images,though data collection for subsequent quantitative analysiswould be more expensive and time consuming than survey-based methods. These techniques, when applied cross-cul-turally among diverse regions, may be useful for furtheringour understanding of the marketing environment's moderat-ing effects on image perceptions and brand performance.

APPENDIX

Regional Socioeconomics Measure

At the regional level, socioeconomic "summary" mea-sures, such as disposable income or GNP/capita, are simply

not available for cities and towns except in the few countries(such as the United States) with highly developed census sys-tems. I compiled a data base of 17 variables for each regionto assess each ofthe 60 regions' levei of socioeconomics. Al-though additionai indicators wouid be desirable, most inter-national data bases possess limited numbers of accuratelymeasured variables for the desired cases. The number of andtypes of variables used here appears to capture the same so-cioeconomic depth and breadth as similar classification ef-forts on equal or greater numbers of cases (e.g., Jaffe !974).

I used factor analysis to identify common factors ac-counted for by the socioeconomic variables and eliminatedvariables that did not account for variance across regions.Factor scores for each region were generated for subsequentanalyses of the effect of environmental factors on brandimage performance. I identified factors with the convention-al criteria of eigenvalues greater than one and adherence toa scree test. All variables remaining in the final factor solu-tion met the criteria of loading on a factor at .3 or higher, andhaving communalities of .4 or higher (Acito and Anderson1986). I eliminated variables with low factor ioadings, lowcommunalities, and/or high cross-ioadings from the analysisand found a two factor solution. The ratio of number of re-gions analyzed to number of factors extracted was greaterthan 20 (60 regions ^ 2 factors = 30), thus indicating ade-quate sample size and a stable factor solution (Arrindeil andvan der Ende 1985). The first factor accounted for 42% ofthe data base variance, capturing the market's modemity—its mobility, communication exposure, sector employment,growth rate, and household size. The more autos and televi-sions per capita, the lower the birthrate, population growth,and percent employed in agriculture, and the fewer peopleper dwelling, the higher the region's modemity or socioeco-notnic level.

Survey Measures

Brand image. How would you characterize your brand'simage in this market? Please allocate 100 points to each ofthe following types of images. Allocate the most points to themost emphasized image (up to 100 points), and the fewestpoints to the least emphasized image (as little as zero points).

,. Mean (S.D.)

Functional image fproblem solving,problem prevention)

Social image (conveys status, socialapproval, accreditation)

Sensory image (provides variety,stimulation, sensory gratification)

100%

48.52 (24.24)

18.16(13.98)

33.32(16.67)

Market experience. For how many years has your brandbeen available in this market?

Mean= 14.63; S.D. = 5.56

Extent of competition. To what extent are you experienc-ing competitive problems in this market?

(7-point Likert scale, where 1 = no problem.s, 7 = manyproblems.)

Mean = 3.60; S.D, = .97

174 JOURNAL OF MARKETING RESEARCH, MAY 1995

Marketing implementation problems. To what extent areyou experiencing marketing problems in each ofthe follow-iflg areas?

(7-point Likert scale, where 1 = no problems, 7 = manyproblems.)

Mean fS.D.)

a) Distribution 4.07 (.99)b) Pricing 3.39 (.92)c) Advertising and promotion 3.70 (1.11)

d) Product characteristics 3.23 (.79)

Cronbach's alpha = .727.

Scale mean = 3.65; S.D. - .71

REFERENCES

Acito. Frank and Ronald D- Anderson (1986). "A Simulation Studyof Factor Score Indetenninancy," Journal of Marketing Re-.search, 23 (May), 111-18.

Aiken. LeonaS. and Stephen G. West {\99\\ Mitliiple Regression:Testing and Interpreting Interactions. Newbury Park, CA: SagePublications.

Anderson, T. W. and H. Rubin (1956), "Statistical Inference in Fac-tor Analysis," in Proceedings ofthe Third Berkley Symposium onMathematical Statistics and Probability. Berkeley: University ofCaiifomia Press, 111-50.

Amold. Hugh J. (1982), "Moderator Viiriabies: A Clarification ofConceptual, Analytic, and P.sychometric Issues," OrganizationalBehavior and Human Performance. 29 (April), 143-74.

Arrindeil, Willem A. and Jan van der Ende (1985), "An EmpiricalTest of the Utility of the Observation-To-Variables Ratio in Fac-tor and Components Analysis," Applied Psychological Measure-ment, 9 Ount), 165-78.

Aulakh, Preel S. and Masaaki Kotabe (1993), "An Assessment ofTheoretical and Methodological Development in InlemationalMarketing: \9H(>-199O." Journal of Intemational Marketing, I(2), 5-28.

Belk, Russell W. (1988), "Third World Consumer Culture." in Mar-keting's Contribution to Development. Research in Marketing.Suppl. 4, E. Kumcu and A. F. Firat, eds. Greenwich. CT: JAIPress, 103-27.

Blackston. Max (1992), "Building Brand Equity by Managing theBrand's Relationships," Journal of Advertising Research, 32(May/June), 79-83.

Boote, Alfred S. (1983), "Psychographic Segmentation in Europe,"Journal of Advertising Research, 22 (December/January),19-25.

Boster. James S., ed. (1987), Intracultural Variation, American Be-havioral Scientist, 31 (November/Etecember). Newbury Park,CA: Sage Publications.

Buzzell, Robert D. and Bradley T Gale (1987), The PIMS Princi-ples: Unking Strategy to Performance. New York: The FreePress.

Cohen. Jacob and Patricia Cohen (1983), Applied Multiple Regres-sion/Correlation Analysis for the Behavioral Sciences, 2nd ed.Hillsdale, NJ: Lawrence Eribaum.

Crawford, Merle (1985), "A New Positioning Typology," Journalof Product Innovation Management. (December), 243-53.

Cronbach, Lee J. (1987). "Statistical Tests for Moderator Variables:Flaws in Analysis Recently Proposed," Psychological Bulletin,102(3), 414-17.

Dannhaeuser. Norbert (1987), "Marketing Systems and Rural De-velopment: A Review of Consumer Goods Distribution," HumanOrganization. 46 (2). 177-85.

Darlington. R. B. (1990), Regression and Linear Models. NewYork: McGraw-Hill.

Daun, Ake (1983), "The Materialistic Life-Style: Some Socio-Psy-chological Aspects," in Consumer Behavior and EnviwnmentalQuality, L. Uusitalo, ed. New York: St. Martin's Press, 6-16.

Denis, Jean-Emile and Daniel Depelteau (1985), "Market Knowl-edge, Diversification and Export Expansion," Journal of Inter-national Business Studies. 16 (Fall), 77-89.

Dillon, William R. and Matthew Goldstein (1984), MultivariateAnalysis: Methods and Applications. New York: John Wiley &Sons, Inc.

DiMingo, Edward (1988), "The Fine Art of Positioning," Journalof Business Strategy. 9 (March/April). 34-38.

Douglas, Susan P and Yoram Wind (1987). "The Myth of Global-ization," Columbia Journal of Wortd Business. 22 (4), 19-29.

Gardner. Burleigh B. and Sidney J. Levy (1955), 'The Product andthe Brand." Harvard Business Review, 33 (March/April), 33-39,

Gronhaug, Kjell and John L. Graham (1987). "Intemational Mar-keting Research Revisited." in Advances in International Mar-keting, Vol. 2, S. Tamer Cavusgil, ed. Greenwhich, CT: JAIPress. Inc., 121-37.

Han, C. Min and Vem Terpstra (1988), "Country-of-Origin Effectsfor Uni-National and Bi-National Products," Journal af Interna-tional Business Studies, 19 (Summer). 235-55.

Hanna, Janice G. (1980), "ATypoIogy of Consumer Needs," in Re-search in Marketing, Vol. 3, Jagdish Sheth. ed. Greenwich, CT:JAI Press, 83-104.

Hill, John S. and Richard R. Still (1984), "Effects of Urbanizationon Muhinational Product Planning: Markets in LDCs,"Columbia Journal of World Business. 19 (Summer), 62-67.

Hofstede, Geert (1984), Culture's Consequences. Beverly Hills.CA: Sage Publications.

Hondrich, Kar! Otto (1983), "How Do Needs Change?" in Con-sumer Behavior and Environmental Quality, L. Uusitalo, ed.New York: St. Martin's Press, 56-73.

Huszagh, Sandra M., Richard J. Fox, and Ellen Day (1986), "Glob-al Marketing: An Empirical Investigation," Columbia Journal ofWorld Business, 21 (4), 3 1 ^ 3 .

Inglehart, Ronald (1977), The Silent Revolution: Changing Valuesand Political Styles Among Westem Publics. Princeton, NJ:Princeton University Press.

Inkeles. Alex (I960), "Industrial Man: The Relation of Status toExperience, Perception, and Value," American Journal of Soci-ology, 75. 208-25.

(1983), Exploring Individual Modernity. New York:Columbia University Press.

and David Horton Smith (1974), Becoming Modem. Cam-bridge, MA: Harvard University Press.

Jaffe, Eugene D. (1974). Grouping: A Strategy for InternationalMarketing. New York: American Management Association.

Jain, Subhash C. (1989), "Standardization of International Market-ing Strategy: Some Research Hypotheses," Journal of Market-ing, 53 (January), 70-79.

Johanson, Jan and Jan-Erik Vahlne (1977), "The Internationaliza-tion Process of the Firm—A Model of Knowledge Developmentand lncrea.sing Foreign Market Commitments," Journal of Inter-national Business Studies, 8 (Spring/Summer), 23-32.

Johansson, Johny K. and Reza Moinpour (1977), "Objective andPerceived Similarity of Pacific Rim Countries," Columbia Jour-nal of World Business, 11 (Winter), 65-76.

Kahle, Lynn R. (1986), 'The Nine Nations of North America andthe Value Basis of Geographic Segmentation," Joumal of Mar-keting, 50 (April), 37-47.

Keegan, Warren J., Richard R. Slill, and John S. Hill (1987).'Transferability and Adaptability of Products and PromotionThemes in Multinational Marketing—MNCs in LDCs," Journalof Global Marketing. I (FallAVinter), 85-103.

Keller, Kevin Lane (1993), "Conceptualizing, Measuring, andManaging Customer-Based Brand Equity," Joumal of Market-

Global Brand Image Strategies 175

ing, 57 (January), 1-22.Keyfitz. Natfian (1982). "Development and lhe Elimination of

Poverty," Economic Development and Cultural Change. 30 (3).649-70.

Lambkin. Mary and George S. Day (1989), "Evolutionary Pro-cesses in Competitive Markets: Beyond the Product Life Cycle,"Joumat of Marketing, 53 (July). 4-20.

Lastovicka, John L. and Kanchana Thamodaran (1991), "CommonFactor ,Score Estimates in Multiple Regression Problems." Jour-nal of Marketing Research, 28 (February). 105-12.

Lenski. G.E. (1966), Power and Privilege: A Theory of SocialStratification. NewYork: McGraw-Hill.

Levitt, Thecxlcre (1983), 'The Globalization of Markets," HarvardBusiness Review. 61 (3), 92-102.

Levy, Sidney (1985), "Dreams, Fairy Tales, Animals, and Cars,"Psychology and Marketing, 2 (Summer), 67-81.

Lynn. Michael, George M. Zinkhan, and Judy Harris (1993), "Con-sumer Tipping: A Cross-Country Study," Joumal of ConsumerResearch, 20 (December). 478-88.

Malinowski, Bronislaw (1944), A Scientific Theory of Culture.Chapel Hill: University of North Caroiina Press.

Marquardt, D. W. (1980). "You Should Standardize Predictor Vari-ables in Your Regression Models." Journal ofthe American Sta-ti.stical Association, 75, 87-91.

Martenson, Rita (1989), "Intemational Advertising in Cross-Cul-tural Environments." Joumal of Intemational Consumer Mar-keting. 2 (]).1-\S.

Maslow, Abraham H. (1954), Motivation and Personality. NewYork: Harper-Collins.

McKenna, Shawn (1992), The Complete Guide to Regional Mar-keting. Homewood, IL: Business One Irwin.

Mehrotra, Sunil (1990), "Strategic Regional Marketing: Two PlusTwo Equals Six," Joumal of Advertising Research, 30 (Decem-ber/January), 9-17.

Montgomery, Cynthia A. (1982), 'The Measurement of Firm Di-versification: Some New Empirical Evidence," Academy ofManagement Joumal, 25, 299-308.

Mueller. Barbara (1987). "Reflections of Culture: An Analysis ofJapanese and American Advertising Appeals," Journal of Adver-tising Research. 27 (June/July), 51-59.

Nctcr. John, William Wassennan, and Michael H. Kutner (1989),Applied Linear Regression Models, 2nd ed. Homewood, IL:Richard D. Irwin, Inc.

Nurske, Ragnar (1953), Problems of Capital Formation in Unde-veloped Countries. New York: Oxford University Press.

Oglivy, David (1963), Confessions of an Advertising Man. NewTork: Ballantine.

O'Guinn, Thomas, Wei-Na Lee, and Ronald J. Faber (1986), "Ac-culturation: The Impact of Divergent Paths on Buyer Behavior,"in Advances in Consumer Re.search, Vol. 13, R. J, Lutz., ed.Provo UT: Association for Consumer Research, 579-83.

Park, C. Whan, Bernard J. Jaworski. and Deborah J. Maclnnis(1986). "Strategic Brand Concept- Image Management," Jour-nal of Marketing. 50 (October). 135-45.

Parsons, Talcott and E. A. Shils (1951), Toward a General Tlieoryof Action. Cambridge, MA: Harvard University Press.

Quelch, John A. and Edward J. Hoff (1986), "Customizing GlobalMarketing," Harvard Bu.sine.ss Review. 64 (May/June), 59-68.

Reilly, Michael D. and Melanie Wallendorf (1987), "A Comparisonof Group Differences in Food Consumption Using HouseholdRefuse," Journal of Consumer Research, 14 (September),289-94.

Reynolds, Thomas J. and Jonathan Gutman (1984), "Advertising asImage Management." Joumal of Advertising Research, 24

(February/March), 27-38.Ricks, David A. (1983), Big Business Blunders. Homewood, IL:

Dow Jones-lrwin.Riesman, Daniel, N. Glazer, and R. Denney (1953), The Lonely

Crowd: A Study of the Changing American Character. NewYork: Doubleday.

Romanelli, Elaine (1987), "New Venture Strategies in the Micro-computer Industry," California Management Review, 30 (Fall),160-75.

Rossiter. John R. and Larry Percy (1987), Advertising and Promo-tion Management. New York: McGraw-Hill.

Roth. Martin S. (1992), "Depth Versus Breadth Strategies for Glob-al Brand Image Management." Joumal of Advertising, 21 (June)25-36.

and Christine Moorman (1988), "The Cuiturai Content ofCognition and the Cognitive Content of Culture: Implicationsfor Consumer Research," in Advances in Consumer Research,Vol. \5, M. J. Houston, ed. Provo, UT: Association for Con-sumer Research, 403-10.

and Jean B. Romeo (1992), "Matching Product Categoryand Country Image Perceptions: A Framework for ManagingCountry-of-Origin Effects," Joumal of International BusinessStudies. 23 (Fall), 411-97.

Ryans, Adrian B. (1988), "Strategic Market Entry Factors andMarket Share Achievement in Japan," Journal of IntemationalBu.viness Studies. 19 (Fall), 389^U)9.

Sethi, S. Prdkash (1971). "Comparative Cluster Analysis for WorldMarkets," Journal of Marketing Re.seatch, 8 (August). 348-54.

Shanna, Subhash, Richard M. Durand, and Oded Gur-Arie (1981),"Identification and Analysis of Moderator Variables," Joumal ofMarketing Research. 18 (August), 291-300.

Simon, Carol J. and Mary W. Sullivan (1993), 'The Measurementand Determinants of Brand B<:]uity: A Financial Approach,"Marketing Science, 12 (Winter). 28-52.

Smith, Daniel C. and C. Whan Park (1992), 'The Effects of BrandExtensions on Market Share and Advertising Efficiency," Jour-nal of Marketing Re.search, 29 (August), 296-313.

Szymanski, David M., Sundar G. Bharadwaj, and P RajanVaradarajan (1993), "Standardization Versus Adaptation of In-temational Marketing Strategy: An Empirical Investigation,"Joumal of Marketing, 57 (October), 1-17.

Twi, Chin Tiong and John U. Farley (1987), "The Impact of Cul-tural Pattems on Cognition and Intention in Singapore," Joumalof Consumer Research. 13 (4), 540-44.

Tse, David K., John K. Wong, and Chin Tiong Tan (1988), 'TowardSome Standardized Cross Cultural Consumption Values," in Ad-vances in Consumer Reseanh,Vol.i5, M. J. Houston, ed. Provo,UT: Association for Consumer Research. 387-95.

, Russell W. Belk, and Nan Zhou (1989), "Becoming a Con-sumer Society: A Longitudinal and Cross-Cultural ContentAnalysis of Print Ads from Hong Kong, the People's Republic ofChina, and Taiwan," Joumal of Consumer Research, 15(March), 457-72.

Wallendorf, Melanie and Eric J. Amould (1988), '"My FavoriteThing.s": A Cross-Cultural Inquiry into Object Attachment, Pos-sessiveness. and Social Linkage" Joumal of Con.mmer Re-seanh, 14 (March), 531-47.

Weimann, Gabriel (1984), "Images of Life in America: The Impactof American T.V in Israel," Intemational Joumal of Intercultur-al Relations. 8. 185-97.

Yankelovich, Daniel (1981), New Rules: Searching for Self-Fulfill-ment in a World Tumed Upside Down. New York: RandomHouse.