Social Identity in an Emerging Consumer Market: How You Do the Wash May Say a Lot About Who You Think You Are

Steven M. Burgess, University of the Witwatersrand
Mari Harris, Markinor (Pty) Ltd.
ABSTRACT - This paper calls attention to within-country diversity, an entirely neglected issue in country-of-origin (COO) research, by examining the link between social identity and brand preference in South Africa. Respondents were clustered into 14 social identity clusters on the basis of data collected from a representative national sample of 3493 South Africans as part of a syndicated research study. Previous usage and regular current usage of two types of cigarette brands were examined. The results showed that four social identity groups with European social identity were more likely to have used brands with British or American COO brand images while groups with African social identity were more likely to have used brands with South African or ambiguous international COO brand images. The results suggest that home nation rather than home country may be the basis of the general preference for domestic products oft-noted in the literature. Future research is suggested.
[ to cite ]:
Steven M. Burgess and Mari Harris (1999) ,"Social Identity in an Emerging Consumer Market: How You Do the Wash May Say a Lot About Who You Think You Are", in NA - Advances in Consumer Research Volume 26, eds. Eric J. Arnould and Linda M. Scott, Provo, UT : Association for Consumer Research, Pages: 170-175.

Advances in Consumer Research Volume 26, 1999      Pages 170-175

SOCIAL IDENTITY IN AN EMERGING CONSUMER MARKET: HOW YOU DO THE WASH MAY SAY A LOT ABOUT WHO YOU THINK YOU ARE

Steven M. Burgess, University of the Witwatersrand

Mari Harris, Markinor (Pty) Ltd.

[We thank Markinor (Pty) Ltd. for their support, which made this research possible, and Jan-Benedict E. M. Steenkamp and Glenn Milligan for helpful comments.]

ABSTRACT -

This paper calls attention to within-country diversity, an entirely neglected issue in country-of-origin (COO) research, by examining the link between social identity and brand preference in South Africa. Respondents were clustered into 14 social identity clusters on the basis of data collected from a representative national sample of 3493 South Africans as part of a syndicated research study. Previous usage and regular current usage of two types of cigarette brands were examined. The results showed that four social identity groups with European social identity were more likely to have used brands with British or American COO brand images while groups with African social identity were more likely to have used brands with South African or ambiguous international COO brand images. The results suggest that home nation rather than home country may be the basis of the general preference for domestic products oft-noted in the literature. Future research is suggested.

INTRODUCTION

Although the geneal preference for domestic products is well-established in the literature, the overwhelming bulk of evidence supporting this so-called 'home country bias’ has been collected in North America, Western Europe, and Japan. The current research focuses on social identity and brand preference in South Africa (SA), an emerging consumer market (ECM) experiencing rapid sociopolitical and economic change, and calls attention to an entirely neglected issue in country-of-origin (COO) research: within-country diversity.

Social identity in ECMs

Social identity is "that part of an individual’s self-concept which derives from his knowledge of his membership in a social group (or groups) together with the value or emotional significance attached to that membership" (Tajfel, 1978, p.63). It is a primary influence on the importance of attitudes (Boninger, Krosnick, & Berent, 1995) and a lens through which individuals recode reality (Zavalloni, 1975). Social interaction and differentiation in social position are important characteristics of groups and reasons for affiliating to them (Radcliffe-Brown, 1940). According to social identity theory (SIT, Hogg & Abrams, 1988; Tajfel, 1974), individuals derive a feeling of positive social identity from belonging to desirable social groups; viz., from their affiliation to groups that they perceive to be positively distinctive from other groups. Individuals evaluate in-groups more favorably than out-groups when striving to achieve positive social identity (Tajfel & Turner, 1979). This tendency for more favorable evaluation extends to products, services, and brands that are associated with the in-group (Tajfel, 1981). Product acquisition, consumption, and disposal may feature in the social mobility, social creativity, and social change strategies that negatively-distinctive groups often pursue in order to achieve positive social identity.

SIT is of special interest in ECMs because the more volatile sociopolitical and economic environment can alter the desirability of group membership, create opportunities for social mobility, and relax group boundariesCsometimes almost overnight. In turn, this often leads to new social interaction patterns, lifestyle experiences, and consumption choices that have important symbolic implications for self-perceptions (c.f., Mead, 1934), including the core elements of social identity such as social status, membership groups, social labeling, derived status, personal identity, and social type (Rosenberg, 1979). Values, nation and state equivalence, and historically accentuated group identity are important considerations when conducting social identity research in ECMs.

Values

Values are enduring, central, conscious, and transituational goals that represent three universal requirements of human existence to which individuals and groups must respond: (1) needs of individuals as biological organisms, (2) requisites of coordinated social interaction, and (3) survival and welfare needs of groups (Schwartz, in press). Values explain in-group perceptions better than the individualism-collectivism dichotomy and also influence the individual’s understanding of national and group differences and readiness for out-group social contact (Rokeach, 1973; Sagiv & Schwartz, 1995; Schwartz, in press). Linked to most aspects of consumer behaviour and capable of explaining the perceived utility products acquire due to an association with social groups (Burgess, 1992; Sheth, Newman, & Gross, 1991), values have played an important role in previous ECM social identity research (e.g., Hofman, 1982). The validation of Schwartz’ (1992; 1994) theory of value content and structure in over 60 countries provides a powerful platform for understanding the cross-cultural motivations underlying social identity in ECMs.

Nation and state equivalence

National identity is a subset of social identity. While nation and state could be considered virtually equivalent terms in many countries where previous COO research was conducted, ECM marketers often must be sensitie to differences in these concepts. A state is a group of people occupying a defined geographic region and recognizing the authority of an organized government. A nation consists of people sharing a common ethnic heritage and a cognate language or languages. In many ECMs, political boundaries were drawn primarily to satisfy colonial exigencies. As a result, nations and states often do not coincide in Africa, Asia, and Eastern Europe. This suggests that social identity could be a core influence on intermarket segment behavior (see Hassan & Samli, 1994).

Historically accentuated group identity

The rapid pace of change has important implications for social identity in many ECMs, especially where group identity has been accentuated, as was the case in many former colonies and in the former Eastern Bloc. For instance, Apartheid accentuated racial and ethnic identity and exerted considerable influence on most aspects of daily life (c.f., Appelgryn & Nieuwoudt, 1987; Christopher, 1991; Dixon & Reicher, 1997; Human & Greenacre, 1987; Louw-Potgieter, 1987; Magwaza & Bhana, 1990). As a legacy, although the elements of social identity that emerge in South African research (e.g., Bennett & Foster, 1996; Duckitt & Mphuthing, 1998) are consistent with other international studies (e.g., Garza & Herringer, 1987; Hofman, 1982; Hooper, 1976), one might expect the influence of ethnicity and nationality on perceived social identity to be more accentuated in South Africa and the influence of social identity on brand choice to be intensified. Indeed, research conducted elsewhere has shown that individuals who place high importance on racial group membership make a greater effort to accurately identify in-group and out-group members, use more caution when making group categorizations, and often use racial prejudice to affirm the self (Blascovich, Wyer, Swart, & Kibler, 1997; Fein & Spencer, 1997; Hopkins, Reicher, & Levine, 1997). Formalized links to a tribe, advantaged political party, liberation movement, culture, or religious group may trigger such rigorous group identification and behavior in other ECMs. Change creates anxiety and threatens the individual’s identity which, in turn, impacts on intergroup bias and aspects of consumer behavior such as the allocation of cognitive resources, information processing, and decision making (Hunter, Platow, Bell, Kypri, & Lewis, 1997; Keinan, 1987; Leary, Barnes, Griebel, Mason, & McCormack, 1987; Martin, Williams, & Clark, 1991; McLeod & Mathew, 1988; Miller, 1987; Mogg, Mathews, Bird, & Macgregor-Morris, 1990; Tesser & Moore, 1989). Thus, there is evidence in the literature to suggest that the rapid change and environmental volatility that characterizes many ECMs may heighten the influence of social identity on consumer behavior (c.f., Mangos & Smith, 1998; Vida & Plassman, 1998; Yergin & Stanislaw, 1998).

All of this suggests that ECM perceptions of social identity sometimes coincide less with current political realities than with an idealized past or future or with a perceived ancestral homeland. This has important implications for understanding the home nation bias in ECMs because it describes an environment in which consumers’ perception of their 'home country’ may not be their geographical home country but rather may be a construct based on national and social identity. The current research addresses the need to understand this phenomenon of within-country diversity in ECMs by examining the influence of social identity on cigarette brand choice in South Africa.

METHOD

Sample and instrument

A leading professional marketing research company collected data among a representative national sample of 3,493 South Africans. The questionnaire was developed in English and translated into Afrikaans, North Sotho, South Sotho, Tswana, Venda, Xhosa, and Zulu using acceptd back-translation techniques (Brislin, 1986; Brislin, Lonner, & Thorndike, 1973). Respondents were interviewed personally in their homes by interviewers that were from the respondents’ ethnic group. Interviewers were trained to be sensitive to differences in the respondent’s social class where such differences occurred. Data were collected on the respondents’ value priorities, optimum stimulation level, and sociodemographic characteristics. Respondents also indicated their interest in a standard battery of 40 lifestyle activities and product categories.

Operationalizing social identity in the current research

Social identity was operationalized by clustering respondents (see section 2.2.1 below) based on a set of variables culled from previous social identity research (viz., Garza and Herringer 1987, Hofman 1982, and Hooper 1976) and assessed for relevance to South African social identity by the research team. Variables included demographic characteristics, living standard measures, degree of participation in the formal economy, lifestyle interests, and values and the collected data were cluster analyzed. Demographic characteristics included race, home province, degree of urbanization, age, gender, occupation, home language, other languages read and understood, marital status, household income and education. Living standard measures gained wide acceptance for market segmentation as the organized South African marketing community searched for effective alternatives to racial segmentation. Sixteen living standard measures such as the dwelling type, electrification, running water, a water heater, and dishwashing liquid, that have been the subject of much research (see South African Advertising Research Foundation, 1997, pp.88-91), were the basis for operationalizing living standards in the current research. An extensive range of variables measuring financial services products and services usage operationalized degree of participation in the formal economy. Forty lifestyle interests sampled a wide range of interests researched regularly by the research company, such as women’s clothing, interior decorating, sport and outdoor activities, pets, future purchases and investments.

The Portraits Questionnaire (PQ) was employed to measure values (Schwartz, Lehmann, & Roccas, in press; Schwartz, Lehmann, Melech, Burgess, & Harris, 1997). PQ is a new values scale developed by Schwartz for populations that normally may find the Schwartz Value Survey cognitively overly-demanding. Research using PQ reproduces Schwartz’ theoretical value structure and links values to concepts closely related to social identity including gender, religiosity, political orientation, and autocratic behavior (Schwartz et al., 1997). Measuring values only indirectly, PQ is especially well suited to social identity research because respondents indicate "how much like you is this person" in response to 29 statements such as "It is important to her to be rich. She wants to have a lot of money and expensive things". While perceived similarity to others is not social identity, it is no doubt a closely related concept (Jackson, Sullivan, Harnish, & Hodge, 1996). Schwartz (1992; Schwartz et al., 1997) has noted the need to adjust values scale data for response styles. In the current research value importance was standardized within respondents by subtracting each respondent’s mean score (across the 29 PQ items) from each item. In order to transform the values to a dichotomous scale for cluster analysis with the other variables, each value was scored as "1" where these mean-centered scores equaled or exceeded zero (i.e., where the value was of average or above average importance), otherwise the value was scored "0". This is consistent with Rokeach’s (1973) contention that the relative importance of values is most important in understanding differences between individuals and groups.

Creating social identity groups

Social identity groups were created using cluster analysis. he Ward’s method and k-means algorithms were chosen because of superior and robust performance in recovering cluster structure (Milligan, 1981; Milligan & Cooper, 1987). First, a Ward’s minimum variance method cluster analysis was conducted. Although other criteria were inconclusive, the pseudo t2 statistic (Duda & Hart, 1973) suggested 3, 8, and 14 cluster solutions. On examination, it was noted that the 14-cluster solution was far superior to the others. As the objective was to partition respondents into non-overlapping clusters, a K-means cluster analysis was then conducted for a 14-cluster solution. It was noted that the two clustering algorithms produced similar results but the k-means solution was preferred because the clusters were more homogeneous (the detailed characteristics of each cluster’s membership are available from the first author). Black respondents constituted at least 97% of cluster members in nine clusters. Cluster 13 is 97% Coloured. Clusters 3, 8, 11, and 12 include members who speak Afrikaans and English in the home, in an almost equal proportion, and Whites constitute between 60% and 92% of these clustersCthe balance being Asians and Coloureds. A single gender contributes more than 80% of clusters 7, 8, 9, and 11. Over 95% of cluster 1 and 4 members speak Xhosa in the home and 92% of cluster 6 speak Zulu.

The relative influence of the variables on cluster membership was explored using CHAID (Kass, 1980). The results revealed that the way people heat water, do the wash, get from point A to point B, or sweep the floor may say a lot about who they think they are. The CHAID decision tree split first into two nodes on the presence of hot running water in the home (c2=2265.26, 13, p<.001) and then on the respondents ability to understand Zulu or South Sotho and further on the ability to understand Xhosa or North Sotho or on race. When alternatives were sought for the primary node split, the presence of a vacuum cleaner (c2=2191.74), washing machine (c2=2156.64), or an automobile (c2=2138.22) were the best substitutes and almost identical trees emerged when these variables were selected manually and trees were then grown from the altered primary node. Thus, it appears that living standard measures are most predictive of cluster membership but ethnicity and race also are strong influences.

Product use

Data were collected concerning brand purchase behavior for seven categories as part of the syndicated study (alcoholic beverages, cigarettes, soft drinks, major food chains, major clothing retailers, gas stations, and food products). Three criteria were set for choosing among these categories for the current research: (1) the product category included brands that were positioned as products of foreign countries, (2) a wide spectrum of society could afford to participate in the category, and (3) the respondent would be freely able to make a choice from the entire range of brands in the category. Although alcoholic beverages, automobiles, and gas stations met some of these criteria, only cigarettes met all of the criteria. Cigarette marketing is highly sophisticated in South Africa and brand images are well communicated to all strata of society through electronic, print, outdoor, and point-of-sale media.

Respondents reported the cigarette brands they use regularly now and brands they have ever used from a comprehensive list of ten cigarette brands representing more than 95% of sales in the product category. Other brands not listed could be added but less than 0.5% of respondents chose to do so. Although all respondents were classified into social identity groups, the brand preference analyses were limited to respondents who reported using at least one cigarette brand regularly now (843 respondents) and those who reported ever using at least one brand (1204 respondents) in order to focus on current smokers and those who had ever smoked, respectively. COO was a strong element of the brand position of five brands (hereafter called the "international brands"): Camel and Chesterfield (American) and Benson & Hedges, Dunhill, and Mills (British) The other five brands (hereafter called the "other brands") either were associated with domestic COO images (Lexington), ambiguous international COO images (Consulate, Peter Stuyvesant, and Rothman’s), or did not position on COO (Courtleigh). Lexington, the sole domestic brand tested, has used an American announcer’s voice in its commercials for years and often uses images of American sporting events in its advertising. Thus it may have a bi-national rather than domestic COO image, even though most South Africans realize that it is a domestic brand. The nature of these associations with COO suggests that 'other brands’ are less clearly related to COO. Thus, SIT would predict that the hypothesized influence would be weaker for hypothesis 3 than it is for hypothesis 2.

Hypotheses

The current research tests three hypotheses. The first hypothesis concerns the influence of social identity on brand choice. The previous discussion suggests that consumers should prefer brands that are perceived to be more consistent with their social identity. If this is true, then members of the 14 social identity clusters should differ as regards their likelihood of having used a specific cigarette brand.

Hypothesis 1:   Social identity influences cigarette brand choice.

The second hypothesis concerns the utility of brands with COO images that are perceived to be congruent with those of a perceived ancestral homeland or with another country that has been influenced by that ancestral homeland. More specifically, this hypothesis tests the usage of brands with European or American COO images by the social identity groups for which European identity or a link to European identity is an element of social identity. In the current research, this refers to the five brands with British and American COO images (the 'international brands’) and to clusters 3, 8, 11, and 12.

Hypothesis 2:   When European identity is an element of social identity, South African social identity groups will be more likely to use or to have used the five 'international brands’.

The third hypothesis concerns the utility of domestic brands and brands with ambiguous international COO images. Whereas those with European identity may view the new South Africa with some suspicion or fear, Black South Africans are the greatest beneficiaries of the birth of a new South Africa. Because their national identities are more likely to coincide with the new political reality, SIT suggests that they would be more likely to prefer domestic brands. Blacks also perceive international political and economic pressure to have contributed to the fall of the previous government. Neither the British nor the Americans stand out as exceptional supporters of the liberation movement; in fact, the opposite may be perceived by many, and there is a wide perception that all countries participated in ending Apartheid. Thus, this hypothesis tests the usage of the "other brands’ by the social identity groups for which African identity is an element of social identity (i.e., all groups except 3, 8, 11, and 12).

Hypothesis 3:   When African identity is an element of social identity, South African social identity groups will be more likely to use or to hav used the five 'other brands’.

RESULTS

The hypotheses were tested using a series of logistic regression models in which the log-odds of using a specific brand were postulated to be a linear function of social identity and an intercept term, so that:

Log-odds (cigarette brand X usage | social identity cluster)

        =b0 +b1"1+b2"2++b13"13,

where A1"13 are dummy coded respondent groups clustered on social identity.

Cluster 14 was chosen as the reference group for dummy coding because it is a relatively homogeneous group of Zulu women living in poverty in rural Kwa-Zulu Natal. The large size and representativeness of the sample allowed the cut-off for classification of brand users/non-users to be set to the proportional usage reported by the sub-samples (of respondents reporting to have ever used a cigarette brand or to use one or more brands regularly now, respectively).

TABLE 1

LOGISTIC REGRESSION OF BRANDS ON SOCIAL IDENTITY CLUSTERS

Table 1 reports the results for each brand including the Nagelkerke R2 (a measure of relationship strength analogous to the R2 in linear regression, see Nagelkerke, 1991), the B2 log likelihood for the fitted model, the c2 statistic for test of the difference between the -2 log likelihood of the null model and the fitted model, the proportion of users classified correctly by the model, the combined proportion of users and non-users classified correctly by the model, the clusters more likely to use the brand, and the clusters less likely to use the brand.

Social identity and brand choice

The results support the first hypothesis and confirm the influence of social identity on brand choice. The c2 test for goodness of fit indicates that the fitted model is significantly better than the null model (p<.001) for every cigarette brand. The Nagelkerke R2 suggests a moderate/moderately weak relationship that is somewhat stronger, on average, for brands used regularly now (.198) than for brands ever used (.146) and strongest for Benson & Hedges, Camel, and Chesterfield. Consistent with SIT, the international brands (.231, .184) are more strongly related to brand choice than the local brands; (.164, .108), for brands used regularly now and brands ever used respectively. Perhaps the most compelling evidence for the goodness of fit of the model is the high level of accuracy with which the model predicts respondents who use the international brands. The accuracy exceeds 77% when predicting regular brand usage now and 64% when predicting brand usage ever. Results are less impressive when considering the accuracy with which users and non-users are predicted. However, results remain impressive when Mills and Dunhill, the only two brands regularly used now by less than 1% of the total sample, are ignored.

Brand choice and identity

The second hypothesis postulates that clusters 3, 8, 11, and 12 will be more likely to have used the five international brands. The final two columns of Table 1 report the clusters for which the Wald statistic was significant (p<.05, 1 d.f.) and for which betas were positive or negative, respectively, for each brand. The general pattern of results suggests support for the second hypothesis, with the four clusters being more likely to use or to have used the international brands and less likely to use or to have used other brands. The general pattern of results also supports hypothesis 3; which proposes that all clusters, except 3, 8, 11, 12, and 13, will be more likely to use other brands. The reference group, Cluster 14, is less likely to have used most brands andmore likely to have used the two largest brands, Peter Stuyvesant and Rothmans, as expected. However, the results are not entirely supportive of either hypothesis. A significant cluster-brand relationship fails to emerge in some cases, the two clusters which include a majority of Xhosas living in the Eastern Cape region (viz., 1 and 4) fail to emerge as significantly related to any brand, and the positive relationship between cluster 8 and regular use now of Courtleigh is in opposition to the hypothesized direction. Nevertheless, the results show compelling support for both hypotheses.

DISCUSSION AND SUGGESTIONS FOR FUTURE RESEARCH

The current research has investigated the influence of social identity on brand usage in an ECM environment. ECMs emerged as important trading regions during the past decade and promise to receive much attention from marketers in the new millennium. Very little is known about the effect of rapid change on social identity in ECMs or about the influence of social identity on ECM consumer behavior. The findings of the current research are important and exciting. Membership in various groups based on demographic characteristics, living standard measures, degree of participation in the formal economy, lifestyle interests, and values were the basis for operationalizing social identity. Although the "value or emotional significance attached to that membership (Tajfel, 1978, p.63)" was not measured, there was substantial support for the hypotheses and social identity clusters predicted brand choice with impressive accuracy.

Although we recognize that South African within-country diversity was accentuated by some five decades of Apartheid rule, SIT suggests that a previous colonial heritage or rapid sociopolitical or economic change potentially creates social forces in any ECM that may intensify awareness and processing of social identity perceptions and the striving for positive social identity. It seems very likely that the influence of the national identity on consumer behavior is moderated by the rapid change that characterizes life in ECMs and that national identity may hold special importance where nation and state are less equivalent.

There is much to do. The current research suggests further research in three main areas: (1) what we call the 'home nation bias’, (2) the structure and content of social identity in ECMs, and (3) the influence of social identity on consumer behavior. Home nation bias is evident in the reported brand choice behavior of groups with European and African identity in the current research. This finding, that national identity influences brand preference, is important and suggests the need to consider what we would call 'the home nation bias’ when conducting COO research in ECMs. Presently, we do not know if home nation is a universal concept consisting of structural elements that assume varying influence on consumption in different geographic regions and cultures or if its structure and content vary. The current research should encourage researchers to define the home nation bias better by exploring the cognitive elements that shape the meaning of 'home’ and 'nation’ in ECMs and the environmental influences that moderate the relative influence of those elements. Future research also should concentrate on defining reliable instruments for measuring these elements and their relative influence on behavior within a product category or categories of interest.

The structure and content of social identity in ECMs requires particular attention. We need to understand the relative influence of national identity and other elements of social identity in ECMs and how environmental influences and individual differences shape the relative influence of these elements on the gestalt of social identity and on subsets of social identity, such as national identity. Measuring group membership and the value or significance of that membership will lead to an impovement in our understanding of social identity in ECMs. Finally, there is a significant opportunity to understand the influence of social identity on consumer behavior, despite the impressive stream of SIT research. This is especially true in ECM environments. Personal influence and reference groups have benefited from important streams of consumer research but there has been little recognition of the importance of social identity. We need to understand the nature of social identity influence during the decision process and especially its influence on consumer attention, information processing, and the diffusion of innovation. Research that focuses on industrialized and less industrialized regions will be of the greatest value.

BIBLIOGRAPHY

Appelgryn, A. E. M., & Nieuwoudt, J. M. (1987). Relative Deprivation and the Ethnic Attitudes of Blacks and Afrikaans-Speaking Whites in South Africa. The Journal of Social Psychology, 311-323.

Bennett, J., & Foster, D. (1996). Social Identity: Comparing White English-Speaking South African Students in 1975 and 1994. South African Journal of Psychology, 26(4), 203-211.

Blascovich, J., Wyer, N. A., Swart, L. A., & Kibler, J. L. (1997). Racism and Racial Categorization. Journal of Personality and Social Psychology, 72(6), 1364-1372.

Boninger, D. S., Krosnick, J. A., & Berent, M. K. (1995). Origins of Attitude Importance: Self-Interest, Social Identification, and Value Relevance. Journal of Personality and Social Psychology, 68(1), 61-80.

Brislin, R. W. (1986). The Wording and Translation of Research Instruments. In W. J. Lonner & J. W. Berry (Eds.), Field Methods in Cross-Cultural Research (Vol. Eight, pp. 137-164). Beverly Hills, California: Sage.

Brislin, R. W., Lonner, W. J., & Thorndike, R. M. (1973). Cross-Cultural Research Methods. New York: Wiley.

Burgess, S. M. (1992). Personal Values and Consumer Research: An Historical Perspective. In J. N. Sheth (Ed.), Research in Marketing (Vol. 11, pp. 35-80). Greenwich, Connecticut: JAI Press.

Christopher, A. J. (1991). Urban Segregation Levels in South Africa Under Apartheid. SSR, 75(2), 89-94.

Dixon, J. A., & Reicher, S. (1997). Intergroup Contact and Desegregation in the New South Africa. British Journal of Social Psychology, 36, 361-381.

Duckitt, J., & Mphuthing, T. (1998). Group Identification and Intergroup Attitudes: A Longitudinal Analysis in South Africa. Journal of Personality and Social Psychology, 74(1), 80-85.

Duda, R. O., & Hart, P. E. (1973). Pattern Classification and Scene Analysis. New York: John Wiley.

Fein, S., & Spencer, S. J. (1997). Prejudice as Self-Image Maintenance: Affirming the Self Through Derogating Others. Social Psychology, 73(1), 31-44.

Garza, R. T., & Herringer, L. G. (1987). Social Identity: A Multidimensional Approach. Journal of Social Psychology, 127(3), 299-308.

Hassan, S. S., & Samli, A. C. (1994). The New Frontiers of Intermarket Segmentation. In S. S. Hassan & R. D. Blackwell (Eds.), Global Marketing: Perspectives and Cases (pp. 76-100). New York: Dryden.

Hofman, J. E. (1982). Social Identity and the Readiness for Social Relations Between Jews and Arabs in Israel. Human Relations, 35(9), 727-741.

Hogg, M., & Abrams, D (1988). Social Identifications. London: Routledge.

Hooper, M. (1976). The Structure and Measurement of Social Identity. Public Opinion Quarterly, 40(2), 154-164.

Hopkins, N., Reicher, S., & Levine, M. (1997). On the Parallels Between Social Cognition and the 'New Racism’. British Journal of Social Psychology, 36, 305-29.

Human, L., & Greenacre, M. J. (1987). Labour Market Discrimination in the Manufacturing Sector: The Impact of Race, Gender, Education and Age on Income. South African Journal of Economics, 55(2), 150-64.

Hunter, J. A., Platow, M. J., Bell, L. M., Kypri, K., & Lewis, C. A. (1997). Intergroup Bias and Self-Evaluation: Domain-Specific Self-Esteem, Threats to Identity and Dimensional Importance. British Journal of Social Psychology, 36, 405-26.

Jackson, L. A., Sullivan, L. A., Harnish, R., & Hodge, C. N. (1996). Achieving Positive Social Identity: Social Mobility, Social Creativity and Permeability of Group Boundaries. Journal of Personality and Social Psychology, 70(2), 241-254.

Kass, G. V. (1980). An Exploratory Technique for Investigating Large Quantities of Categorical Data. Applied Statistics, 29, 119-127.

Keinan, G. (1987). Decision Making Under Stress: Scanning of Alternatives Under Controllable and Uncontrollable Threats. Journal of Personality and Social Psychology, 52(3), 639-44.

Leary, M. R., Barnes, B. D., Griebel, C., Mason, E., & McCormack, D., jr. (1987). The Impact of Conjoint Threats to Social- and Self-Esteem on Evaluation Apprehension. Social Psychology Quarterly, 50(4), 304-11.

Louw-Potgieter, J. (1987). The Authoritarian Personality: An Inadequate Explanation for Intergroup Conflict in South Africa. The Journal of Social Psychology, 128(1), 75-87.

Magwaza, A. S., & Bhana, K. (1990). Stress, Locus of Control, and Psychological Status in Black South African Migrants. The Journal of Social Psychology, 13(2), 157-64.

Mangos, N. C., & Smith, M. (1998, September 17-20, 1998). Application of the Ethnocentrism Paradigm to the International Advertising Debate for Russia and the Newly Independent States: A Matter of Convergence. Paper presented at the 1998 Multicultural Marketing Conference, Montreal.

Martin, M., Williams, R. M., & Clark, D. M. (1991). Does Anxiety Lead to Selective Processing of Threat-Related Information? Behavoir Research and Theory, 29(2), 147-60.

McLeod, C., & Mathew, A. (1988). Anxiety and the Allocation of Attention to Threat. The Quarterly Journal of Experimental Psychology, 40A(4), 653-70.

Mead, G. H. (1934). Mind, Self and Society. Chicago: University of Chicago Press.

Miller, S. (1987). Monitoring and Blunting: Validation of a Questionnaire to Assess Styles of Information Seeking Under Threat. Journal of Personality and Social Psychology, 52(2), 345-53.

Milligan, G. W. (1981). A Review of Monte Carlo Tests of Cluster Analysis. Multivariate Behavioral Research, 16(July), 379-407.

Milligan, G. W., & Cooper, M. C. (1987). Methodology Review: Clustering Methods. Applied Psychological Measurement, 11(4), 329-354.

Mogg, K., Mathews, A., Bird, C., & Macgregor-Morris, R. (1990). Effects of Stress and Anxiety on the Processing of Threat Stimuli. Journal of Personality and Social Psychology, 59(6), 1230-7.

Nagelkerke, N. J. D. (1991). A Note on General Definition of the Coefficient of Determination. Biometrika, 78, 691-692.

Radcliffe-Brown, A. R. (1940). On Social Structure. Journal of the Royal Anthropological Institute, 70, Reprinted in Radcliffe-Brown, A. R. (1952) Structure and Function in Primitive Society. London: Cohen & West.

Rokeach, M. (1973). The Nature of Human Values. New York: Free Press.

Rosenberg, M. (1979). Conceiving the Self. New York: Basic Boks.

Sagiv, L., & Schwartz, S. H. (1995). Value Priorities and Readiness for Out-Group Social Contact. Journal of Personality and Social Psychology, 69(3), 437-448.

Schwartz, S. H. (1994). Are There Universal Aspects in the Content and Structure of Human Values? Journal of Social Issues, 50, 19-45.

Schwartz, S. H. (in press). Cultural Dimensions of Values: Toward an Understanding of National Differences. Applied Psychology: An International Review.

Schwartz, S. H., & Bardi, A. (1997). Cross-Cultural Similarities in Value Hierarchies: Evidence and Implications (Working Paper ). Jerusalem: Hebrew University of Jerusalem.

Schwartz, S. H., Lehmann, A., Melech, G., Burgess, S. M., & Harris, M. (1997). Validation of a Theory of Basic Human Values with a New Instrument in New Populations (Working Paper ). Jerusalem: Hebrew University of Jerusalem.

Schwartz, S. H., Lehmann, A., & Roccas, S. (in press). Multimethod Probes of Basic Human Values. In J. Adamopoulos & Y. Kashima (Eds.), Social Psychology and Cultural Context . Newbury Park, Ca.: Sage.

Schwartz, S.H., Roccas, S., & Sagiv, L. (1992). Universals in the Content and Structure of Values: Theoretical Advances and Empirical Tests in 20 Countries. Advances in Social Psychology, 25.

Schwartz, S. H., & Ros, M. (1995). Values in the West: A Theoretical and Empirical Challenge to the Individualism-Collectivism Cultural Dimension. World Psychology, 1(2), 91-122.

Sheth, J. N., Newman, B. I., & Gross, B. L. (1991). Why We Buy What We Buy: A Theory of Consumption Values. Journal of Business Research, 22, 159-170.

South African Advertising Research Foundation. (1997). All Media and Products Survey 1997 (Technical Report, Volume 1). Johannesburg: Author.

Tajfel, H. (1974). Social Identity and Intergroup Behavior. Social Science Information, 13, 65-93.

Tajfel, H. (1978). Differentiation between social groups: Studies in the Social Psychology of Intergroup Relations. London: Academic Press.

Tajfel, H. (1981). Human Groups and Social Categories. Cambridge, U.K.: Cambridge University Press.

Tajfel, H., & Turner, J. C. (1979). An Integrative Theory of Intergroup Conflict. In W. G. Austin & S. Worchel (Eds.), The Social Psychology of Intergroup Relations (pp. 33-47). Monterey, California: Brooks/Cole.

Tesser, A., & Moore, J. (1989). Independent Threats and Self-Evaluation Maintenance Processes. Journal of Social Psychology, 677-89.

Vida, I., & Plassman, V. S. (1998, September 17-20, 1998). Consumer Attitudes Toward Importation and Purchasing of Foreign Made Products: A Study of Young Estonian Consumers. Presented at the 1998 Multicultural Marketing Conference, Montreal.

Yergin, D., & Stanislaw, J. (1998). The Commanding Heights: The Battle Between Government and the Marketplace that is Remaking the Modern World. New York: Simon & Shuster.

Zavalloni, M. (1975). Social Identity and the Recoding of Reality: Its Relevance for Cross-Cultural Psychology. International Journal of Psychology, 10(3), 197-217.

----------------------------------------