NATIONAL IDENTITY ( NATID ) FRAMEWORK IN MULTI-ETHNIC AFFILIATIONS CONSUMER MARKET : THE CASE OF MALAYSIA

This paper examines the transferability of the national identity (NATID) framework in a multiethnic consumer market. Data were collected using heterogeneous purposive sampling from 403 respondents in Malaysia. This study exert also responds to the cross-cultural validation of measures used in international research whereby the authors assessed measurement invariance of the NATID dimensions construct using Malaysian and non-Malaysian samples. The analysis provided strong support for the factor structure, internal consistency, and discriminant validity on the transferability of the NATID framework on the Malaysian data sample. The result also suggested ethnic affiliations are a relevant unit of analysis and indicates that the relationship between the dimensions of NATID and overall national identity, varied across the three major ethnic groups. With reference to originality, this research is pioneering the adaptive etic approach to elicit inputs and to improve the transferability of the NATID framework.


Introduction
The national identity describes a concept of collective behaviour (Balabanis, Diamantopoulus, Mueller, & Melewar, 2001) that endorses economic ethnocentrism, segregation, and isolationism.More significantly, the notion of national identity has been suggested to hamper the international trade economy, and acceptance of foreign-made goods and services (Balabanis et al., 2001;Vida, Dmitrovic, & Obadia, 2008).
However, the very premise of national identity, which is often used to describe similar consumer behaviour at a macro-level within one national market, is gradually being challenged, as many of the current consumer markets are becoming more fragmented and segregated (Craig & Douglas, 2010;Engelen & Brettel, 2010).There has been a growing recognition in the marketing literature that many of the national markets are heterogeneous and culturally diverse (Cui, 1997;Nwankwo, Aiyeku, & Ogbuchi, 2006).The concept of cultural diversity is an expression commonly used to describe salient coexistence of various ethnic affiliations within the one national market (Pires & Stanton, 2000a).
Previous highlighted studies that ethnic affiliation has a significant influence on various consumer attitudinal-behavioural dispositions, and is progressively becoming an important marketing characteristic (Cui, 1997;Hui, Laroche, & Kim, 2006;Singh, Kwon, & Pereira, 2003).Nonetheless, despite the growing importance of both concepts in international marketing, relatively little attention has been paid on the linkage between national identity and ethnic dimension in cross-cultural marketing research.The objective of this study was to examine the marketing concept of national identity in a multi-ethnic consumer market.To achieve this purpose, the study examined the Keillor, Hult, Erffmeyer and Babakus (1996) national identity (NATID) framework across the three major ethnic groups in Malaysia (i.e., Malay, Chinese, and Indian).The coexistence of diverse ethnicity and growing purchasing power may be particularly relevant for the evaluation of national identity within Malaysia (Hirschman, 1987;Teo, 2011).This effort also complements Keillor et al. (1996) and Keillor and Hult (1999) that called for more studies in a wider variety of cultural settings to validate the NATID framework.Correspondingly, this study set forth to expand the understanding of how consumers within the cultural environment that differs significantly from the Western ontology that has profoundly dominated the ethnic marketing research and practices (Engelen & Brettel, 2010).It is believed that this study would offer both scholars and marketers with better insights on the under-presented growing importance of ethnic consumers in Malaysia.
In the following section, the conceptualisation of national identity is reviewed.Building from social identity theory, the next section offers a discussion on ethnic affiliations.Using a survey from Malaysia, the third section explains the methodology used for the present study.The analysis of the results is presented in the next section.Based on these findings, this paper ends with in depth discussion, limitations, and also possibilities for future research.Smith (1991) explicated that the fundamental notion of national identity pertains to (1) common myths and historical memories, (2) common legal rights and duties for all members, (3) mass public culture, and (4) common economy with territorial mobility for members.Similar commonality argument was also suggested by Clark (1990), suggesting that the concept of national identity represents a set of common socialisation, shared understanding, and aspirations, that bind the population together to the homeland.

Conceptualising National Identity and NATID Dimensions
From the cross-cultural marketing perspective, Erdogan and Uzkurt (2010) exerted that consumer sense of national identity relates to social responsibility during the product decision process.This corroborates with Shimp and Sharma (1987) who described consumer sense of national identity as an attitudinal-behavioural disposition that adamantly refuses to buy imported products and reprimands others for doing so, because buying foreign made goods would hurt the domestic economy and it is considered as unpatriotic.
According to Keillor et al. (1996), consumer perception of national identity consists of a multitude of unique dimensions.Building from other research paradigms on national identity, the authors conceptualised and developed the national identity (NATID) framework that is presuppose to measure individual degree of national identity across four dimensions of (1) belief system, (2) national heritage, (3) cultural homogeneity, and (4) ethnocentrism.
For the present study, national identity was conceptualised based on these four dimensions.
The belief system provides an essential utility for decision makers and facilitates decision making by providing ready frameworks for unprecedented situations and provides answers to fundamental questions about the way the system works and the sources of human behaviour.Keillor and Hult (1999) indicated that belief systems pertains to the role of religion as a psychological bridge to the society's aggregate structure.According to Cui and Adams (2002), the belief system cultivates a person's sense of national identity by promoting cultural commonality and also as a mechanism in reducing cultural psychological tension.
Studies suggested that consumer ethnocentrism functions as an antecedent and directly influences national identity (Balabanis et al., 2001;Shimp & Sharma, 1987).Keillor et al. (1996) pointed out that a strong predisposition toward an ethnocentric attitude is also an indication that a particular culture has a high degree of national identity.An ethnocentric disposition pertains to that one group that believes their group to be superior in comparison to other groups.They view other groups from their own perspective and reject those that are different, while at the same time accepting those who are culturally alike (Shimp & Sharma, 1987;Tajfel, 1982).The consumers within this classification consider that it is both inappropriate and immoral for one to buy products from other countries.
The two dimensions of national heritage and cultural homogeneity are theoretically related.National heritage describes the importance of historical figures and events in history, while cultural homogeneity concerns the uniqueness of this national heritage (Cui & Adams, 2002;Thelen & Honeycut, 2004).Different histories and events experienced by an individual produce different perceptions and beliefs about national heritage as a characteristic of national identity.Similarly, a strong sense of cohesion possessed in a group of individuals influences the relationship between cultural homogeneity and national identity.

Social Identity Theory and the Role of Ethnic Affiliations
There has been a significant number of studies in the extant cross-cultural marketing literature that argue on the fundamental assumption of a "country" as the appropriate unit of analysis.As Craig and Douglas (2006) pointed out, although the use of a country as the basis to define geographical boundaries of the research setting may be convenient, the use of a country may also lead to unintended consequences.Engelen and Brettel (2010) also asserted that a country may not be an adequately reliable indicator, as there are often substantial variations within a country that could be even greater than differences between countries.
The explanation on consumers' strong ethnic affiliations can be found in Tajfel and Turner's (1979) theory of social identity.The theory suggested that a person identifies himself as belonging to a social category or group (Stets & Burke, 2000).Through a process of social comparison, persons who are similar to the self are categorised with the self and are labelled the in-group; whereas, persons who differ from the self are categorised as the out-group (Tajfel, 1982).According to Tajfel and Turner (1979), opposing group interests in obtaining scarce resources will promote competition.However, it is from this overtly inter-group competition that helps to promote the intra-group morale, cohesiveness, and cooperation within group.
Building on social identity as a theoretical foundation, the present study proposed that an individual's perception of national identity is characterised by its membership in social and ethnic affiliation, each of which is influenced by a unique and specialised set of values.Previous studies on cross-cultural marketing offer credence on this notion.For instance Burton (2000) indicated that ethnic affiliation has a significant influence on consumer consumption behaviour.Holland and Gentry (1999) further suggested that consumers' perception toward the advertising stimuli differs across different ethnic affiliations.In another related studies, Jamal (2003), and Pires and Stanton (2000b) indicated a person's sense of ethnic affiliation influences his or her service encounter expectations.

Data Collection Procedure
The present study used a sample of respondents from Malaysia.According to the census, the three major ethnic groups in Malaysia are the Malays, Chinese, and Indians (Malaysia Department of Statistics, 2010).The coexistence of multiculturalism contributes to cultural diversity of individuals with different beliefs, values, and customs; hence, offering a suitable setting for the present study.
Online survey was used for data collection.The questionnaire was developed using the web based software-written in the html formatand distributed to respondents via email.The study used heterogeneous purposive sampling for sampling procedure.This method enabled researchers to achieve the study objectives, by selecting cases that were considered as informative (Saunders, Lewis & Manhill, 2007).The survey was distributed to Malaysian respondents working in the public and private institutions in Kuala Lumpur, Malaysia.The respondents were screened to ensure that they were affiliated with the three major ethnic groups.The mailing was followed by a reminder message, and the data collection lasted for three months.Assurances were made that all responses would be confidential.
Out of the 650 online questionnaires that were distributed, 457 of the respondents replied within the specified time frame, indicating a 70% response rate.A total of 54 responses were identified as either not being affiliated with any of the three major ethnic groups or incomplete; therefore, excluded for further thus analysis, thus yielding a final sample of 403 respondents.Table 1 shows a description of the study respondents, where the majority of respondents were male (60%), Malays (45%), and within the range of 40 to 49 years old (47%).

Design and Survey Measures
Table 2 shows the four dimensions of NATID (17-item): (1) national heritage, (2) cultural homogeneity, (3) belief system, and (4) consumer ethnocentrism.To improve the transferability of the NATID framework to the Malaysian cultural context, the present study employed Douglas and Craig's (2006) adaptive-etic approach.Specifically, the adaptive-etic approach suggested for both adaptation and incorporation of culturally specific values, which would allow better fitting of the constructs to the Malaysian cultural setting.
The present study followed this suggestion by interviewing a panel of six Malaysian academicians.The panel members were asked to provide comments on the original 17 NATID measurement items.To assure compatibility with the study context, panel members were selected based on their research expertise.Four members were from the business and marketing research domain, and two of the members were familiar with research on general social sciences.The interviews produced four new items that reflect the cultural specific values of national identity in the Malaysian context-and this resulted in a total of 21 measurement items.For example, an item "Malaysian people are proud of historical monuments, as a symbol of their civilization" was included under the national heritage, and three new items "It is a tradition for one to respect others religion and cultural values", "Common language of communication is imperative in cultivating Malaysian solidarity" and "Malaysians are equally the same regardless of their colour and creed" under the cultural homogeneity dimension.
The wording for all the items was changed into "Malaysia" to denote the study context.For instance, "One of the Malaysia's strengths is that it emphasises events of historical importance" and "Malaysian people are proud of their nationality".All the items were measured using a seven-point Likert scale ranging from (1) "disagree strongly" to (7) "agree strongly".All items were measured using seven-point Likert scale, (1) strongly disagree (7) strongly agree.

Measurement Invariance of NATID Dimensions
Measurement invariance represents "whether or not, under different conditions of observing and studying phenomena, measurement operations yield measures of the same attribute" (Horn & McArdle, 1992, p.117).Meredith (1993) has also defined measurement invariance as the parameters of a model depending on group membership.This study has employed the framework by Keillor et al. (1996) where the objective focused on measuring the degree of differences in national identity across nations.
When testing measurement invariance, classically the model fit for the data overall with fixed specific parameters has to be equal across groups, which is compared to the overall data model which is free to vary (Hortensius, 2012).Thereafter, if evidence supporting a measure's invariance is lacking, conclusions based on that scale are at best ambiguous and at worst erroneous (Steenkamp & Baumgartner, 1998).
The measurement of invariance for this study consisted of a comparison between the work of Keillor et al. (1996) and this paper's objective, which understudies the Malaysian case of NATID dimensions.According to Murray, Gao, Kotabe, and Zhou (2007), firstly a two-group confirmatory analysis model was estimated as baseline for the said construct of the study and no cross-group factor constraint was enforced.Table 3 shows the two-group model with free factor loadings that fit the data well.According to Murray et al. (2007) the data proved that there is configural invariance between the two samples in the NATID dimensional construct.
The two group model was then analysed to test metric invariance, and to examine the factor loadings identically for each scale item between the samples.The factor loadings were forced to be equal between the Malaysia and non-Malaysia samples by Keillor et al. (1996).
According to Murray et al. (2007), metric equivalence represents an increase in the degree of measurement invariance compared to the configural equivalence and this model is nested hierarchically on the baseline model.
A chi-square difference test was conducted thereafter to make comparison of the model fit data (∆c² = 104.71,df = 24).The result was not significant.Thus, a conclusion was made where the two-group model of factor loading forced fit the data equally as the baseline model, which was tested earlier.Therefore according to Murray et al., (2007), it supports metric invariance and configural invariance.
The assessment of strength of relationships between the NATID dimensions and multiethnic affiliations of the consumer market in the case of Malaysia can now be conducted.

Reliability and Validity
With the additional four new items, the analysis shifted into the exploratory mode using exploratory factor analysis (EFA).This procedure had also been suggested as a viable approach in refitting the underlying measurement that could then be evaluated with the confirmatory factor analysis (CFA) (Gerbing & Anderson, 1988;Gerbing & Hamilton, 1997;Hair, Black, Babin, Anderson, & Tatham, 2006).
The EFA was performed as an initial test on the 21 measurement items.As set forth by Hair et al. (2006), the cut-off criteria for the EFA were subjected to: (1) factor loading of larger than 0.60 on each item to indicate a high correlation of items selected, (2) discarding an item having cross-loading lower than 0.30, and (3) Cronbach alpha and item-to-total-correlation of larger than 0.70 and 0.50, respectively.
Table 3 shows the result of the EFA on the pooled sample of 403 Malaysian respondents.Two of the original NATID items (BS4 and CE5) were deleted for failing to meet the specified criteria.
The four NATID dimensions yielded a c² value of 120.9 (df = 89), with c²/df of 1.36.The results of the analysis also indicated a satisfactory goodness-of-fit with the GFI, AGFI, CFI, and RMSEA producing values of .88,.87,.98 and .05,respectively.All the parameter estimates were high with significant t-values, ranging from 0.80 to 0.93 for cultural homogeneity, 0.72 to 0.93 for belief system, 0.70 to 0.95 for consumer ethnocentrism, and 0.71 to 0.76 for national heritage.
An analysis of discriminant validity was performed by comparing the square root of the AVE with the square correlation of the given construct, and with any other construct, in the analysis.Discriminant validity exists if the items share more common variance with their respective construct than any variance that constructs share with other constructs (Fornell & Larcker, 1981).Table 4 provides evidence of discriminant validity for the NATID framework.
The AVE was identified to be higher than the correlation of the given construct or with any other constructs in the analysis.It was concluded that these results provide strong construct unidimensionality, internal consistency, and validity for the NATID framework with the Malaysian data sample.

Developing Norms across Ethnic Groups
The norms were established using the distribution of scores from the final 19 items for the three major ethnic groups in Malaysia (i.e.Malays, Chinese, and Indian) (Churchill, 1979).The distribution of scores was summarised by calculating the means and standard deviation for the group, and subsequently these norms were used for comparison between the ethnic groups.
Based on the results, these mean value for the NATID four dimensions was 17.91.The study further developed five classifications ranging from a low of "extremely weak" to a high of "extremely strong" national identity.As illustrated in Table 5, an overall score above 18.74 (two standard deviations above the mean) is considered to be "extremely strong" national identity.With the same logic, an "extremely weak" national identity is indicated by a score below 16.82.Scores within the range of 18.62 to 18.73 and 16.83 to 17.31 (half to two standard deviations of above and below the mean) are identified as "strong" and "weak", respectively.Finally, a classification of "neither strong nor weak" is for scores ranging from 17.32 to 18.61 (within the half standard deviation).
Table 6 shows the scores for the three ethnic groups used in the present study.The ethnic Malays score higher on the overall NATID four dimensions (score = 22.13), followed by the ethnic Indian (score = 18.34), and Chinese (score = 17.63).

Discussion
For the purpose of assuring the reliability of the NATID framework, an assessment and improvement of the framework in new cultural settings are needed.This endeavour is necessary to validate the representativeness and to ascertain that the theoretical constructs do not vary in their meanings in different contextual settings.
The main objective was to examine the transferability of the NATID framework to the Malaysian multi-ethnic setting.The results indicated that the adaptive-etic approach significantly improves NATID transferability to the Malaysian data sample.The analysis showed that the new items have strong correlation with the study sample.This is supported by the result of psychometric scales (i.e.satisfactory factor structure, internal consistency, and construct validity).
The findings indicated that two of the original NATID items were less relevant to the Malaysian study sample.Specifically, the item that addresses the specific religious philosophy and individual perception toward imported products showed incongruence with the data sample.Perhaps, given the Malaysian multicultural society, enquiries that addresses about a specific religion may be considered contradictory to Malaysian values that endorse cultural diversity.
The study further examined the NATID framework across three major ethnic affiliations in Malaysia.The analysis showed that the notion of national identity significantly varies across ethnic affiliations.Based on the analysis, ethnic Malay consumers exhibited the highest score for the overall national identity within the NATID framework.From the norms' categorisation, the scale score for the ethnic Malays fell within the "extremely strong" category.For foreign business managers, doing business with the Malay consumers may require an adjustment in their strategies by emphasising the knowledge and respect for Malay customs and values.This inward orientation was also found to extend into the Malay consumption behaviour, as reflected by the consumer ethnocentrism and belief system.From the promotional and advertising perspectives, positioning a product with local values that reflect less foreign-ness is imperative for this ethnic segment.
The other two ethnic affiliations fell within the "neither weak nor strong" category.The results showed that both Indian and Chinese respondents have a strong emphasis on national heritage and cultural homogeneity than the other two dimensions of belief system and consumer ethnocentrism.This suggested that both ethnic Chinese and Indian consumers are more receptive to foreign made goods and services.Perhaps, since both of these ethnic groups also belong to a closely-knit community, consuming foreign made goods might also act as an indicator of a person's level of achievement.For the marketers, tailoring for these ethnic segments may require incorporating both ethnic and foreign values.For instance, marketers should consider employing ethnic actors for endorsement, or using the local-ethnic newspapers as a channel for advertisement.

Implications
Several implications may be drawn from this study.First, the study offered important theoretical implications on the replication and transferability of NATID framework.As noted by Douglas and Craig (2006), many constructs and measurement instruments are replication of studies, which are often conducted in one country.Therefore, they may not be adequate when tested in different cultural settings.The results further substantiated this argument.
Since it was originally developed and tested on individualistic and relatively homogeneous social settings, the framework was initially less fitting with Malaysian culture values, which are predominantly collective and heterogeneous.The result, however, significantly improved after the inclusion of four cultural specific values.The ramifications of these findings are of great importance for researchers who intend to test the NATID framework in other cultural settings.
As asserted by Herche, Swerson, and Verbeke (1996), the use of measurements assumed to possess etic, but are actually "pseudoetic", is a violation to cross-cultural research.By simply replicating a model to another cultural setting, a researcher may be drawing erroneous answers, which might be of no relevant value to the context of interest.
The second implication of this study highlighted issues of examining consumer variations from a micro-level.Correspondingly, this study provided support for ethnicity as a reliable unit of analysis.The results suggested that there are distinguishable perceptions of national identity across the major ethnic affiliations in Malaysia.
For international business managers, the findings suggested that ethnic affiliation plays a significant role in the Malaysian consumer market.For foreign firms that plan to develop a marketing strategy, consideration on operational adjustments is necessary to account for the varying strengths that have been identified to be present across these ethnic groups.Specifically, developing a marketing programme based on "one-fits-all" for Malaysian consumers may not be feasible, since different ethnic consumers are suggested to maintain a significant degree of ethnic values.

Limitations and Future Research
Some limitations should be noted in the present study.Firstly, since the self-report scales were used to measure the study variables, it is possible that a common method bias will have affected some of the results.Secondly, though the model demonstrated satisfactory results with the data sample, it is posited that many of these measurement items were not entirely sufficient to reflect the Malaysian national identity.For that reason, identifying other cultural specific values that may offer a better fit to the Malaysian national identity are warranted in future investigations.
Thirdly, there is a possibility for the "universal cultural values" of the NATID framework to be attenuated as a result of inclusion of the new items.Therefore, future research should seek to validate this possibility using a new Malaysian data sample, and identify whether similar results can be achieved.
Fourthly, the majority of respondents were from high-income and highly educated categories.Therefore, the interpretation from this study should be used with caution since generalisations can only partially explicate the actual behaviour for each of the ethnic segments.Future research may want to extend this by using a larger sample to further examine each of the ethnic groups used in the present study (e.g., from low level income, and/or different geographical areas).Moreover, since this is a one-country study, it is also likely that the results may not be easily generalised with other countries.
Finally, the present study used the mean and standard deviation to establish norms across the three major ethnic groups.Future research should consider using more advance statistical methods, such as using structural equation modelling for second-order analysis.

Table 2
Dimensions of the NATID framework CE5Only those products that are unavailable in the Malaysia should be imported CE Note.NH = national heritage, CH = cultural homogeneity, BS = belief system, CE = consumer ethnocentrism.

Table 3
Model Comparison for Measurement of NATID Dimensions.

Table 5
Discriminant validity of the NATID dimensions

Table 6
Classification of Scores

Table 7
Scores Across Three Ethnic Groups Mean scores are presented first followed by the standard deviation in parenthesis 2. The overall NATID score is the summated score for the four dimensions.