Brand Engagement in Self-Concept: A Psychometric and Demographic Analysis

Abstract

Brand Engagement in Self-Concept describes the consumer tendency to include important brands as part of one's self-concept. This individual difference variable is operationalized via a previously developed eight-item Likert scale. The purpose of the present study is to assess both the psychometric characteristics and differences in mean scores of the scale across different United States (U.S.) demographic groups. The analyses use data from a national survey of 2,399 adult U.S. consumers. The results show the scale is unidimensional and has high internal consistency across all the different groups: gender, age, ethnic group, income, and education. Mean brand engagement is unrelated to gender and to level of education, but decreases as age increases. Brand engagement increases as household income levels increase. Finally, Black and Asian consumers report higher brand engagement than both Non-Hispanic White consumers and Hispanic consumers who prefer to communicate in Spanish. [PUBLICATION ABSTRACT]

Headnote

Abstract

Brand Engagement in Self-Concept describes the consumer tendency to include important brands as part of one's self-concept. This individual difference variable is operationalized via a previously developed eight-item Likert scale. The purpose of the present study is to assess both the psychometric characteristics and differences in mean scores of the scale across different United States (U.S.) demographic groups. The analyses use data from a national survey of 2,399 adult U.S. consumers. The results show the scale is unidimensional and has high internal consistency across all the different groups: gender, age, ethnic group, income, and education. Mean brand engagement is unrelated to gender and to level of education, but decreases as age increases. Brand engagement increases as household income levels increase. Finally, Black and Asian consumers report higher brand engagement than both Non-Hispanic White consumers and Hispanic consumers who prefer to communicate in Spanish.

Keywords

marketing, branding, demographics, brand engagement.

Introduction

A new concept has appeared in the marketing and branding literature and along with it a new scale to measure it. Brand Engagement in Self-Concept (BESC) is "an individual difference measure representing consumers' propensity to include important brands as a part of how they view themselves" (Sprott, Czellar & Spangenberg, 2009, p. 92). The concept of brand engagement captures the importance of using brands to form and to express consumer selfconcept and identity (Elliott, 2004). Where consumers see brands as relevant to their lives, a relationship or bond grows between the consumer and his or her brand (Fournier, 1998; Uncles, 2008). Sprott, Czellar, and Spangenberg (2009) conceptualize this construct as a fundamental tendency in consumers, an individual difference at the global marketplace, rather than the domain or category specific level. To better understand the phenomenon of branding, researchers need to use such concepts. As Schmitt (2009) argues in the case of brand experience, ". . . we know very little about how consumers experience a brand: how we can measure brand experiences; and whether brand experiences are simply an epiphenomenon or whether they can influence consumer behaviour" (p. 418). While much of our understanding of consumers' relationships with and use of brands comes from a qualitative perspective (cf., Fournier, 1998; Levy, 2005), and even from a post modern point of view (cf., Genosko, 2001a; 2001b) the current study and Sprott, Czellar, and Spangenberg's (2009) original study take a more psychological and measurement focused stance. From the point of view of a business, brand engagement in self-concept is certainly important to makers of branded goods and branding researchers because it describes an important motivator for brand purchase and for brand loyalty.

Brand engagement potentially explains a great deal of consumer decision making, so researchers interested in why consumers select the brands that they do should employ it. Before a self-report scale can be used with confidence, however, its psychometric properties (dimensionality and internal consistency) should be carefully documented; and because demographic characteristics could influence the performance of the scale, psychometric evaluation and means comparison across a variety of demographic groups is advisable:

The interval nature of most marketing and organizational behaviour measures makes the evaluation of mean scores problematic, particularly for new measures. As such, it is important that absolute scale values be interpreted cautiously and that means and standard deviations across studies and samples be recorded in efforts to assist in interpreting subsequent results from scale applications over time and by new researchers. (Netemeyer, Bearden, ¿kSharma, 2003, pp. 164-165)

Thus, the purpose of the present study was to promote the use of this concept and its corresponding scale by evaluating its psychometric properties across many demographic groups of consumers to determine whether the scale was stable enough for researchers to use with confidence. A secondary purpose was to report relationships between demographic variables and BESC scale scores in order to identify potential confounds that can be incorporated in future research. The present paper reports the findings of a survey that begins the process of describing the presence of BESC among culturally diverse groups within the general U.S. population. We feel the scale will be of great value to managers who apply it in real world scenarios. To that end, we develop and test a series of hypotheses about the how we expect the BESC is related to the demographic variables gender, age, level of education and income, and ethnic identification.

Background and Hypotheses

The Concept of Brand Engagement in Self-Concept

The idea of using brands to construct the self is grounded in an instrumental view of materialism, whereby people acquire things for a purpose (Csikszentmihalyi & Rochberg-Halton, 1981). However, BESC differs from materialism because its focus is not on amassing quantity of goods, but only on acquiring branded products because they shape and express self-concept. BESC is internally rather than behaviourally driven. BESC is measured with an eight-item scale that was developed by Sprott, Czellar, and Spangenberg (2009) using standard psychometric procedures. Table I contains the items. In the scale development section of their paper, Sprott, Czellar, and Spangenberg (2009) report that BESC is strongly related to material values (Richins, 2004) and marginally related to relational -interdependent Self-Concept (Cross, Bacon, & Morris, 2000), but not to self esteem, life satisfaction, sex role, or self deception, to name a few. Sprott, Czellar, and Spangenberg (2009) test the concept in five studies and find that consumers scoring high on BESC have more accessible memories of favorite brands, feel stronger associations between themselves and those brands, are able to recall more currently owned brands, feel more brand meaning, and are more brand loyal than are low scorers (Sprott, Czellar, & Spangenberg, 2009). The original paper develops the content of the construct and demonstrates its applicability to a variety of marketing situations, but makes no attempt to assess its psychometric properties in the general U.S. population. As a marketplace level individual difference, brand engagement is expected to vary across different demographic segments of the market. Evidence for these variations can improve the use of the concept by managers, and theory-testing researchers can anticipate which demographic variables to include as control variables in their model tests.

From a scale development point of view, it is important to continue the testing of a new scale. Nunnally (1978) calls for norming new scales; it is the last step in Churchill's (1979)scale development process. According to DeVellis (1991), the demonstration of validity is an ongoing process and is "not firmly established during scale development" (p. 113). This is even more important when the findings have some counterintuitive elements, which is the case in studies of materialism to which BESC is closely related. It is also important to see that the scale generalizes across populations (DeVellis, 1991). It is important to know if demographic groups vary in the extent to which they manifest brand engagement. In order to know if an individual or group has a relatively higher or lower score, we need to know if the groups systematically vary from the population score (Spector, 1992). The large, national sample in this study enables us to make such tests.

Theoretically speaking, there is some cleaning up to do as well. Again, when we look the closest relative of BESC, materialism, there are many examples of contradictory findings about how it plays out in the general population. It is not possible to understand the mechanism of a construct when the directions of its correlations with demographic variables switch signs with every study. While the BESC shows stability and predictability in the first study, it has not been thoroughly tested for its relationships with population descriptors.

Managers use population descriptors to construct market segments and to describe customer cohorts. In the case of BESC, we have a scale/concept with great potential for segmentation use in the field. Developing a demographic based profile for the BESC should be of value to managers especially given the versatility the short scale demonstrates in the original paper (Sprott, Czellar, & Spangenberg, 2009) .

BESG and Demographics

Because there is no history of testing or predicting levels of BESC in the population, we have to fall back to using related variables to develop hypotheses. In the original study, Sprott, Czellar, and Spangenberg (2009, p. 94) found that materialism, as measured by Richins's (2004) reduced scale, was the most closely related to the BESC scale (r - .42, p < .01). This makes sense. Constructing identity is an important task for modern consumers (Elliott, 2004)» Logically, a consumer using branded goods to construct a self is somewhat akin to instrumental materialism whereby consumers accumulate goods to achieve goals (Csikszentmihalyi & Rochberg-Halton, 1981). It is probably fair to say that Richins's (2004) scale tends more towards instrumental than terminal materialism. For instance, one item in Richins's (2004) materialism scale reads, "My life would be better if I owned certain things I don't have" (p. 218). The tone of the questions in this scale is more about owning certain things rather than just having a lot of things.

Both materialism and BESC have internal, psychological characteristics, and potential external or behavioural characteristics. External characteristics might include status or Visible consumption' (Charles, Hurst, & Roussanov, 2009) . The internal might look more like greed or self-centeredness. It would be interesting to have more correlational information about the BESC, but we can use materialism as our primary surrogate for BESC in forming hypotheses about how the construct is distributed in the population.

Gender and BESC

There is no consistent evidence about how gender and materialism are related. Kamano (1999) found that gender predicted materialistic tendencies in a number of developed countries, with women more materialistic than men. Kamineni (2005) used other countries and found men more materialistic. Studies have found that men are more accepting of materialistic values (Beutel & Marini, 1995; Jiuan, Wirtz, Jung, & Keng, 2001) or score higher on the Richins and Dawson (1992) scale (Pepper, Jackson, &Uzzell, 2009). Still others find no relationship between materialism and gender (Eastman, Goldsmith, Campbell, Calvert, & Fredenberger, 1997; Jusoh Heaney, & Goldsmith, 2001; Richins & Dawson, 1992). BeIk (1984) found women less envious in an early study and no gender differences in a later study (Ger & BeIk, 1996). The Web appendix for Sprott, Czellar, and Spangenberg (2009) reports the BESC is free of gender bias. Still, despite the lack of consistent ties between materialism and gender and realizing materialism is a surrogate for BESC, we find enough studies that show some gender relationship to test that Hl: BESC scores are related to gender.

Age and BESC

Age is another demographic variable that appears in a number of studies of materialism and materialistic values. Sprott, Czellar, and Spangenberg (2009) found no correlation between age and BESC in a sample of 430 undergraduates. This finding is probably not definitive due to the restricted range in the sample. Logically, BESC would decline with age in adults. Studies of how values change with age show that older people tend to become more modest and thrifty as they age (Goldsmith, Flynn, & Kim, 2001). The general consensus is that materialism declines sometime in adulthood (BeIk, 1984; Pepper, Jackson, &Uzzell, 2009; Richins & Dawson, 1992). BeIk (1984) found older consumers less envious but also more non-generous than younger consumers. Non-generosity may be a separate construct and related differently to materialism. Kamano (1999) found older people more materialistic than younger people across seven developed countries, showing disagreement in the literature on the age variable as well. His study, however, used a scale of his own device, and that scale is much different from either Belk's (1984) or Richins and Dawson's (1992). Kilsheimer (1993) found consuming for status declines with age. In a similar measure, Charles, Hurst, and Roussanov (2009) found that spending on Visible goods' declines with age. Logically and anecdotally, we see that as people age they focus less on goods and more on the importance of interpersonal relationships. Given the vast majority of the evidence in materialism and in values, we propose that H2: BESC scores decline with age.

Ethnicity and BESC

It is commonly perceived that there are racial or ethnic differences in consumption of status items (Charles, Hurst, & Roussanov, 2009). Differences in materialism by ethnicity are vague in the literature. Crispell (1993) reports that African Americans are the most materialistic in their consumption behaviour and Asians the least. Korzenny, Korzenny, McGavock, and Inglessis (2006) find African Americans and Asians value wealth more than Hispanics and all three groups more than non-Hispanic Whites do. In a measure closer to the BESC, that same study reports that Asians have the strongest feelings of identification with brands followed by African Americans. Over all, it appears that in the U.S., minority groups express more concern with material goods than the majority does.

The most recent work in the area uses econometric methods along with income and other demographic data to conclude that variance in income explains all differences in materialism and pro-material attitudes and behaviours between ethnic groups. Charles, Hurst, and Roussanov (2009) found that when a measure of income dispersion of the reference group (by this they mean the range of incomes within an ethnic group in a specific geographic area) is added to a regression equation explaining status spending, that all racial or ethnic meaning is lost. Income dispersion, not ethnicity, of the reference group explains consuming for status (Charles, Hurst, & Roussanov, 2009). People who live mostly around others with the same income express less status consumption. This too is logical. Possessions and status items are gathered to express position within a social system. Position means more in diverse groups and less in homogeneous groups. When people are very similar in terms of wealth and status, there is less need for them to display wealth. In the diverse group, there is need to display goods so as not to be mistaken for a poorer person. Thus, we propose that H3: BESC scores do not vary by ethnic group.

Income and BESC

Income and materialism are hard to disentangle. If we measure materialistic tendencies as money spent on categories of objects, we will always see spending increasing absolutely as income increases (Charles, Hurst, & Roussanov, 2009). We are also likely to find that spending in discretionary categories as a percentage of income varies predictably with income. When materialism is measured as an individual difference variable, we should find it unrelated to income as psychological characteristics are independent of demographics. Kilsheimer (1993) found income uncorrelated with status consumption measured as an individual difference variable. Pepper, Jackson, and Uzzell (2009) found that materialism is not related to income. Nguyen, Vu, Moschis, and Shannon (2009) tested the relationship between adolescent socioeconomic status and adult materialism and found no relationship for Thai consumers. The only evidence to the contrary we found was in a study that used materialism as a component of a measure of utility; Corfman, Lehman, and Narayanan (1991) found income having only a small direct effect on the utility measure. We will test, however, for BESC correlations with income as it is important to at least demonstrate that BESC is free from confounds of income level. Again, a good psychometric variable is not confounded with demographics. H4: BESC does not vary with income.

Education and Materialism

The general consensus is that more education is associated with less materialism. Singaporeans with less education were more materialistic than their more educated counterparts (Jiuan, Wirtz, Jung, & Keng, 2001). Moors (2003) found the same relationship among Europeans from nine countries. Others have found no significant relationship between education and materialist tendencies (Kilsheimer, 1993; Pepper, Jackson, & Uzzell, 2009). We were unable to find any research finding education to be positively related to materialistic tendencies. Education levels are known to be related positively to income, implying no correlation with materialism, but logically, or at least thinking along the lines of a hierarchy of motivation, à la Maslow, it might be expected that, as people achieve more monetary and educational success, they might become less driven by material desires. As success is achieved, there is less felt need to make material displays to others in the social system. That is essentially the position taken by Moors (2003). It is reasonable to think that might translate over to BESC. So we propose H5: BESC varies negatively with level of education.

Method

Survey Method

Data were collected via online surveys during March of 2009. A Center for Hispanic Marketing Communication at a large southeastern U.S. university conducted the survey. It was a part of an annual survey examining the behaviour of Hispanic consumers in the United Stated. Respondents were intercepted online, asked to complete a survey, and offered a small incentive. DMS Research Opinion Place was used for English speaking respondents. DMS Research Tu Opinion Latina, a Hispanic online panel, was used to recruit Spanish speakers. English speaking respondents were originally sampled via the Opinion Place online 'river' methodology. This method also has been referred to as 'RDD for the Web' as it uses broadcast promotional intercepts to generate a flow of respondents to the Opinion Place site. Respondents are screened and assigned to surveys in real-time, and are not considered registered panellists since most do not return to the site for ongoing survey participation.

In addition, given the quota requirements for this study, a random sample of past respondents to the annual survey was selected based on their demographic characteristics and invited to participate in this survey via a custom e-mail invitation. They were chosen based on their willingness to respond in the past and to boost the number of Hispanic participants. Respondents completed the survey by clicking on a link in the e-mail invitation, which connected them with the online questionnaire. Respondents were required to be 18 years of age or older, and the final sample included 505 non-Hispanic whites, 541 English speaking Hispanics, 351 Spanish speaking Hispanics, 500 African Americans, and 502 Asians. The surveys took an average of 20 minutes for English speakers and 29 minutes for Spanish speakers to complete. The number of Spanish speakers is smaller due to the difficulty recruiting these subjects and is possibly compounded by the longer time it took Spanish speakers to complete the survey. The completion rate was 74*5% for English speakers and 64*4% for Spanish speakers.

Measures

BESC items were interspersed with other measures for a study unrelated to the present one. The response scale was six points from ? = completely disagree' to '6 = completely agree.' The screening portion of the questionnaire asked participants to indicate their gender and ethnic identification (Caucasian/White, African American/Black, Asian or Pacific Islander, or Hispanic/ Latino/Spanish, or other). The Hispanic participants were further subdivided into primarily English-Speaking and Spanish-Speaking groups respectively. The final section of the interview asked participants to report their highest level of education completed (1 = elementary school to 6 = Graduate Degree) and their household income (1 = $19,999 or less to 12 = $150Kor more) (see Table 2). Education and income were treated as interval level variables in the subsequent analyses.

Analyses

Psychometric Stability

In order to deepen the understanding of the functioning of the scale and to further test the relationship between BESC scores and demographics, we test its factor structure (dimensionality) and reliability (internal consistency) across the different population groups. Next, we assess the relationships between the demographics and BESC via regression.

An important aspect of scale construction is its factor structure (Nunnally, 1978; Spector, 1992). In the original article, Sprott, Czellar, and Spangenberg (2009) find a unidimensional structure in multiple administrations. However, it is useful to see if this structure holds across demographic groups (Packman, Brown, Englert, Sisarich, & Bauer, 2005; Rahim & Magner, 1996). We performed a series of Exploratory Factor Analyses (EFAs) to assess the expected, unidimensional factor stability across demographic groups. The results showed that the eight items formed a single factor in each analysis. The factors explained from 59.0 to 71.4 percent total variance in the items. This is impressive factor stability. The same approach was taken for internal consistency. Cronbach's alpha was computed for the same demographic subgroups. Alpha varied from .92 to .95. Again, this is solidly consistent and remarkably high.

Hypothesis Tests of Group Differences

The bivariate differences in mean BESC scores for the demographic groups appear in Table 2. These results support Hl, H2, and H4* Men (M = 27.9) scored higher than women (M = 26.9); mean BESC scores declined with age, and they did not vary by income. Contrary to our hypotheses, the results showed significant differences in mean scores for ethnic group and for level of education. However, none of the differences were very large, and in the case of education, there was no directional difference, and the observed difference may not be reliable because of very small cell sizes for the two lowest education categories.

To assess the multivariate relationship between brand engagement and the demographic variables, we first mean-centered participant age and income so that interaction terms could be computed that would mitigate multicollinearity in the analysis. Next, we regressed the brand engagement scores across the variables gender, age, education, and income plus dummy variables representing the five ethnic categories for the 1951 participants for which we had complete data. In this first stage of the analysis, the results showed significant relationships (adj R2 = .07, F ^1942) = 20.5, p < .0001). Standardized regression coefficients were significant (p < .05) only for age (ß = -.176), income (ß = .054), and the dummy variables standing for White (ß = -.114), Hispanics who prefer Spanish (ß = -.076), and Asian (ß = .054)« In the next stage of the analysis, we added interaction terms for age and gender, gender and income, and age and income, but the change in R2 was not significant and none of the regression coefficients for these variables were significant. None of the Variance Inflation Factors (VIFs) for the independent variables were larger than 2.0, indicating freedom from multicollinearity. The residual statistics did not reveal any outlying observations. The residual plots confirmed that the regression analysis met the assumptions of linearity, homoscedasticity, independence of the residuals, and normality.

The multivariate analysis gives us additional insight into the hypotheses. With the effects of the other variables in the analysis, men and women showed no differences in BESC scores in the regression. The significant, negative coefficient for age supports H2; BESC does seem to decline with age. Ethnicity was the subject of Hypothesis 3, which proposed no ethnic differences. However, there were significant differences for Whites, Blacks, English, and Spanish speaking Hispanics, and Asians. Hispanics who prefer Spanish had the lowest scores (negative) on the BESC and Asians the highest. H4 predicted no relationship between income and BESC. This hypothesis also was rejected as income showed a positive coefficient; BESC increased with income. H5 stated that BESC should vary negatively with education levels. We did not find education related to BESC, and thus that hypothesis also was rejected.

The regression results suggest that when several demographic variables were considered simultaneously, brand engagement is unrelated to gender and to level of education. There is a relationship, however, with age, such that as age increases, brand engagement decreases, suggesting that older consumers use brands less frequently to construct self-concept. Brand engagement also increases as household income levels increase. The results also showed that Black and Asian consumers express higher brand engagement than Non-Hispanic White consumers and Hispanics consumers who prefer Spanish. We will mention English speaking Hispanics in the next section.

While testing the psychometric properties of the BESC, we saw that mean scores varied with gender. However, this relationship failed to materialize in the regression. A follow-up, post hoc 2X4 between-groups analysis of variance comparing average brand engagement scores for men and women and for the five ethnic groups (Whites, Blacks, Asians, and English speaking and Spanish speaking Hispanics) showed a more complicated pattern of differences (see Figure 1).

The main effect for gender was not significant, consistent with the regression results. The main effect for ethnic group also was consistent with the regression results; the mean brand engagement scores for Blacks, Asians, and English speaking Hispanics were significantly higher than the mean score for Whites and Hispanics who speak Spanish. However, a significant interaction term and plot of the mean scores showed that for Whites, Blacks, and Asians, the mean brand engagement scores for women were higher than for men (statistically significant for Whites and Blacks, but not for Asians) , while the mean brand engagement scores for Hispanic men, both English and Spanish speaking (M = 29.7 and 25.3, SD = 10.1 for both groups ) were higher than for both groups of Hispanic women respectively (M(^sub English^) = 26.7, SD = 11.1 and M(^sub Spanish^) = 24.0, SD = 11.1). Because age was a significant predictor in the regression equation, we tested it as a covariate in the ANOVA. We found, however, no effect for age as a covariate.

Discussion

The notion that consumers use brands to express who they are is an important concept dating from the early days of consumer theory (Levy, 1959) that has been incorporated into a variety of accounts of consumer brand behaviour ever since. McCracken (1988) promoted the notion that consumers use material goods to "construct notions of the self (p. xi). Kapferer (2008) introduced the concept of the 'brand prism' in 1992 to describe the components that make up a brand's image. Among the elements is a facet identified as what the brand says about the user. Recent empirical research shows how consumers in the less industrialised countries eagerly adopt brands to express their status and personal identity (Ustuner & Holt, 2009). To more completely understand this phenomenon, however, researchers need reliable and valid tools. We believe that Sprott, Czellar, and Spangenberg's BESC scale provides one of these.

We conducted a large scale test of the new BESC scale. We found remarkably stable psychometric properties. The eight items of the scale showed stable, unidimensional, structure when the sample was sliced and diced. Different age, gender, racial/ethnic, income, and educational groups all exhibited the single factor structure. The same groups had high internal consistency scores. Never did a single item fail to contribute to alpha. This shows that across the population, this scale holds up psychometrically. It is unidimensional and internally consistent.

In order to understand how groups vary in their possession of the characteristic that is brand engagement with self-concept, we regressed BESC scores on demographic variables. The results give us valuable information about BESC. First, it is not the same as materialism, at least in as much as we can say anything consistent about who is or is not likely to be materialistic. The variation in BESC was not what might be expected from looking at the materialism literature. Only age was related as proposed. Second, the regression accomplishes more than a just a test of the hypotheses. We see that demographics do not define BESC. The R2 was only .08. Demographic characteristics account for a very small part of the variation in BESC. This is good. If demographics were enough to define personality or individual difference variables, we would not need individual difference variables. Psychological measures vastly expand what we can say about consumers. They vastly expand how we can relate to consumers as well. Finally, the post hoc test showed us something unexpected. We found an interaction of gender and ethnicity. For White, Black, and Asian participants, women scored higher, either significantly or nominally, on BESC. It is possible that women in these groups are less cynical about brand communications in specific product categories. We investigated this further.

For both groups of Hispanic subjects, men scored higher, and in the case of English speaking Hispanic men, higher by a large margin, roughly 10 percent. It is a dramatic reversal. This is very interesting. One possible explanation is that Hispanic men, both English and Spanish speaking, are more acculturated than Hispanic women are and, therefore, more brand aware as well as brand engaged (Korzenny & Korzenny, 2005). This is a result of immigration patterns where the men arrive first to find work and the women follow later. The men are in the workforce, and many immigrant women stay home where they acculturate more slowly (Korzenny & Korzenny, 2005). Differences in brand engagement by ethnic groups and by gender in these groups may be explained by breaking down brand engagement by category. It is likely that different product categories behave differently in consumers' minds. When researchers have a reliable scale, they can look for just this type of consumer difference. The mean differences for groups really are different.

The findings lead us to conclude that the BESC measure is psychometrically sound and can be used with confidence by researchers interested in studying its effects. We also can conclude that researchers should have little concern for the confounding influences of demographic variables on its relationships. Consistent with Sprott, Czellar, and Spangenberg (2009), we found few consistent demographic relationships with BESC, and those we did detect were very small. For researchers interested in how demographic groups such as ethnic groups might differ, however, we did find interesting patterns that deserve further study.

As with most studies, ours is limited in generalizability to the sample and measures used. Other samples and additional measures might reveal different findings. However, the large national sample of survey participants does give us confidence that at least for the variables we measured, the findings are robust. Only replication with additional data can assess how robust they are. Future research into BESC could follow this path. In particular, ethnic differences in BESC might be of considerable interest to those studying non-White market segments. In addition, other demographic variables could be studied, such as marital status and country region. Cross-cultural research with BESC will have to be done to assess demographic influences in other countries. Provisionally, we feel that researchers in the U.S. can use the scale to reliably and validly study consumer brand relationships with confidence that demographic influences will not confound their findings.

References

References

Belk, R. (1984) Three scales to measure constructs related to materialism: Reliability, validity, and relationships to measures of happiness. Advances in Consumer Research, 1 1, 291-297.

Beutel, A. M., & Marini, M. M. (1995). Gender and values. American Sociological Review, 60(3), 436-448.

Charles, K. K., Hurst, E., & Roussanov, N. (2009). Conspicuous consumption and race. The Quarterly Journal of Economics, 124(2), 425-467.

Churchill, G. A. (1979, February). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16(1), 64-73.

Corfman, K. P., Lehman, D. R., & Narayanan, S. (1991). Values, utility, and ownership: Modeling the relationships for consumer durables. Journal of Retailing, 67(2), 184-204 ?

Crispell, D. (1993). Materialism among minorities. American Demographics, 15(8), 14.

Cross, S. E., Bacon, P. L., & Morris, M. L. (2000). The relational-interdependent self construal and relationships. Journal of Personality and Social Psychology, 78(4), 791-808.

Csikszentmihalyi, M., & Rochberg-Halton, E. (1981). The meaning of things. New York: Cambridge University Press.

DeVellis, R. F. (1991). Scale development: Theory and applications. Newbury Park, CA: Sage Publications.

Eastman, J. K., Goldsmith, R. E., Campbell, E., Calvert, S., & Fredenberger, B. (1997). Gender differences in status consumption and materialism: A study of international consumer behaviour. In E. W. Stuart & E. M. Moore (Eds.). Enriching marketing practice and education (pp. 128-132). Proceedings of the Southern Marketing Association, Atlanta, Georgia.

Elliott, R. (2004). Making up people: Consumption as a symbolic vocabulary for the construction of identity. In K. M. Ekstrom & H. Brembeck (Eds.). Elusive consumption (pp. 129-143). Oxford: Berg.

Fournier, S. (1998). Consumers and their brands: Developing relationship theory in consumer research. Journal of Consumer Research, 24(4) , 343-373.

Genosko, G. (2001a). McLuhan and Baudrillard: The masters of implosion. London: Routledge.

Genosko, G. (2001b). Uncollected Baudrillard. London: Sage Publications.

Ger, G., & BeIk, R. W. (1996). Cross-cultural differences in materialism. Journal of Economic Psychology, 17,55-77.

Goldsmith, R. E., Flynn, L. R., & Kim, W. -M. (2001). Age and gender differences in self-image: A cross cultural study. Journal of Social Behavior and Personality, 1 6(1), 35-62.

Jiuan, T., Wirtz, J., Jung, K., & Keng, K. (2001). Singaporeans' attitudes toward work, pecuniary adherence, materialism, feminism, environmental consciousness, and media credibility. Singapore Management Review, 23(1), 1st Half, 89-86.

Jusoh, W. J. W., Heaney, J.-G., & Goldsmith, R. E. (2001). Self-ratings of materialism and status consumption in a Malaysian sample: Effects of answering during an assumed recession versus economic growth. Psychological Reports, 88, 1142-1144.

Kamano, S. (1999). Comparing individual attitudes in seven countries. Social Science Research, 28, 1-35.

Kamineni, R. (2005). Influence of materialism, gender and nationality on consumer brand perceptions. Journal of Targeting, Measurement and Analysis for Marketing, 14(1), 25-32.

Kapferer, J. -N. (2008). The new strategic brand management: Creating and sustaining brand equity long term (4th ed.). London: Kogan Page.

Kilsheimer, J. C. (1993). Status consumption: The development and implications of a scale measuring the motivation to consume for status, Ph.D. dissertation, Florida State University, Tallahassee, FL.

Korzenny, F., & Korzenny, B. A. (2005). Hispanic marketing: A cultural perspective, Burlington, MA: Elsevier.

Korzenny, F., Korzenny, B. A., McGavock, H., & Inglessis, M. G. (2006). The multicultural marketing equation: Media, attitudes, brands, and spending (Research Report). Tallahassee, Florida: Center for Hispanic Marketing Communication.

Levy, S. J. (1959). Symbols for sale. Harvard Business Review, 37(4)> 117-124.

Levy, S. J. (2005). The evolution of qualitative research in consumer behavior. Journal of Business Research, 58(3), 34 1-347.

McCracken, G. D. (1988). Culture and consumption. Bloomington, IN: Indiana University Press.

Moors, G. (2003). The two faces of (post) materialism. International Journal of Public Opinion Research, 15(4), 396-412.

Netemeyer, R. G., Bearden, W. O., & Sharma, S. (2003). Scaling procedures: Issues and applications. Thousand Oaks, CA: Sage Publications.

Nguyen, H., Vu V., Moschis, G. P., & Shannon, R. (2009). Effects of family structure and socialization on materialism: A life course study in Thailand. International Journal of Consumer Studies, 33(4), 486-495.

Nunnally, J. C. (1978). Psychometric theory. New York: McGraw-Hill.

Packman, T., Brown, G. S., Englert, P., Sisarich, H., & Bauer, F. (2005). Differences in personality traits across ethnic groups within New Zealand and across an international sample. New Zealand Journal of Psychology, 34(2), 77-85.

Pepper, M., Jackson, T., & Uzzell D. (2009). An examination of the values that motivate socially conscious and frugal consumer behaviours. International Journal of Consumer Studies, 33(2), 126-136.

Rahim, M. A., & Magner, N. R. (1996). Confirmatory factor analysis of the bases of leader power: First-order factor model and its invariance across groups. Multivariate Behavioral Research, 31 (4), 495-516.

Richins, M. L. (2004). The material values scale: Measurement properties and development of a short form. Journal of Consumer Research, 3 1 (1) , 209-2 19.

Richins, M. L., & Dawson, S. (1992). A consumer values orientation for materialism and its measurement: Scale development and validation. Journal of Consumer Research, 19(3), 303-316.

Schmitt, B. (2009). The concept of brand experience. Journal of Brand Management, 16(1), 417419.

Spector, P. E. (1992). Summated rating scale construction: An introduction. Newbury Park, CA: Sage.

Sprott, D., Czellar, S., & Spangenberg, E. (2009). The importance of a general measure of brand engagement on market behavior: Development and validation of a scale. Journal of Marketing Research, 46(1), 92-104.

Uncles, M. (2008, March). Know thy changing consumer. Journal of Brand Management, 15(4), 227-231.

Ustuner, T., & Holt, D. (2009). Toward a theory of status consumption in less industrialized countries. Journal of Consumer Research, 37(1), 37-56.