An Examination of the Community Identity and Purchase Preferences Using the Social Identity Approach

ABSTRACT - This paper employs principles of the social identity approach to examine the community, or local, identity. The social identity approach, specifically self-categorization theory, proposes that the self-concept is oriented towards both a personal identity and numerous social categories. The community identity is one of these social categories. Individuals feeling a greater degree of this social identity will be more willing to act in support of it. This support is expressed as a consumer ethnocentric tendency which in turn leads to a purchase preference for locally manufactured goods.

Citation:

Garold Lantz and Sandra Loeb (1998) ,"An Examination of the Community Identity and Purchase Preferences Using the Social Identity Approach", in NA - Advances in Consumer Research Volume 25, eds. Joseph W. Alba & J. Wesley Hutchinson, Provo, UT : Association for Consumer Research, Pages: 486-491.

Advances in Consumer Research Volume 25, 1998      Pages 486-491

AN EXAMINATION OF THE COMMUNITY IDENTITY AND PURCHASE PREFERENCES USING THE SOCIAL IDENTITY APPROACH

Garold Lantz, University of Manitoba

Sandra Loeb, Frostburg State University

ABSTRACT -

This paper employs principles of the social identity approach to examine the community, or local, identity. The social identity approach, specifically self-categorization theory, proposes that the self-concept is oriented towards both a personal identity and numerous social categories. The community identity is one of these social categories. Individuals feeling a greater degree of this social identity will be more willing to act in support of it. This support is expressed as a consumer ethnocentric tendency which in turn leads to a purchase preference for locally manufactured goods.

INTRODUCTION

One way of examining the associations people have with others is the social identity approach. This approach considers the proposition that people have a desire and a propensity to hold a more positive self-esteem of themselves through their self-identity and to also hold a positive identity with the various groups with which they associate (Turner 1982; Tajfel 1981). This paper is concerned with examining one such social identity, a community, or local identity, and placing it in the theoretical framework of self-categorization theory. Additionally, the effect of identifying with the community is carried to a purchase preference for locally manufactured products with an illustration of social cooperation.

As the social identity approach has received little exposure in the marketing literature, this paper will begin with a brief review. Following this, the survey applying the social identity approach to the community identity and the community identity to a purchase intent, will be presented.

The Social Identity Approach: Self-Categorization Theory

Social identity theorists posit that the self-concept is made up of two distinct aspects: the personal identity and the social identity (Tajfel 1978; Turner, Hogg, Oakes Reicher and Wetherell 1987; Hogg and Abrams 1990). The personal identity includes specific attributes of the individual such as competence, talent and sociability. The social identity is defined as "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 and emotional significance attached to that membership" (Tajfel 1981, p.225). Personal identity refers to how people see themselves as individuals while social identity refers to how people view themselves with respect to the social groups to which they belong.

The 'social identity approach’ is an umbrella term encompassing social identity theory and self-categorization theory (Hogg and Abrams 1990). Originally, social identity theory was an analysis of intergroup conflict and social change focusing on individuals’ need to have a positive distinctiveness in their ingroup compared to outgroups, achieving a positive social identity (Tajfel 1978). The more recent and closely related self-categorization theory (Turner et al. 1987) represents a theory of group processes based on the idea that a shared social identity depersonalizes the individual identity. The group identity has been internalized within the individual. As Turner points out, "the group process embodies a shift in the level of abstraction at which the individual self operates" (Turner et al. 1987, iv). Self-categorization theory offers to explain the process of acquiring and using social identities.

An individual can be said to have a taxonomy of social identities and the proposition has been accepted that an individual holds several identities contemporaneously and that any one may hold greater importance, depending on characteristics of the individual and the situation. A particular social category will become salient when the combination of personal characteristics and situation creates a condition sufficiently strong to depersonalize the personalidentity. This occurs when the differences of people within a category are minimized and the differences between categories are maximized. The greater the similarities within the group and the greater the differences towards those outside the group, the greater the contrast. This leads to depersonalizing the individual’s identity in favor of the salient social category.

Depersonalization refers to the process of 'self-stereotyping’ whereby people come to perceive themselves more as the interchangeable exemplars of a social category than as unique personalities defined by individual differences (Turner et al. 1987). Given a degree of accessibility to a certain social category, when circumstances or stimuli occur creating a proper fit between the person and the category there will be a merging of the personal identity with the social identity at that category.

According to self-categorization researchers, some outcomes of depersonalization are group cohesion, ethnocentrism and social cooperation (Turner et al. 1987). Group cohesion is the desire for the ingroup to be positively evaluated by group members while ethnocentrism represents this positive evaluation in comparison to outgroups. Social cooperation is an expression of support for the ingroup representing a merging of personal self-interest with group self-interest.

The scope of this paper does not include an examination of the antecedents of depersonalization or of group cohesion. Rather, the empirical portion of this paper is limited to observing the baseline level of individuals’ identity with their community and showing the relationship of this identity to an expression of ethnocentrism and a purchase preference associated with the community-based ethnocentrism. The scope also does not include manipulating the level of a social identity, although this would also be an interesting variation.

Social Identity and the Community

The social identity approach can be usefully employed in the study of outshopping, support for community institutions such as schools, hospitals, etc. and support for locally manufactured products. The community social identity has rarely been the subject of study within the framework of self-categorization theory. While the outshopping phenomenon appears to be closely related to the local identity, a review of outshopping literature shows that outshopping studies generally focus on the shopper’s attitude towards existing shopping facilities, product selection, and demographics (Papadopoulos 1980; Darden and Perreault 1976). So, while the local identity may be useful in explaining outshopping, a literature review of prior outshopping research is not particularly helpful in refining the concept of the local identity.

An analogy to the national identity and the concept of ethnocentrism at the national level may be more straightforward. The local identity could reasonably be taken to be a matter of refining the national identity to accommodate a smaller level of aggregation. However, there are reasons why the national identity does not precisely parallel the community identity. Much research in the theoretical development of the national identity concerns an individual’s ethnocentric tendencies. According to Sumner (1906) ethnocentrism often develops as a result of the perception of a threat from outsiders. Ethnocentrism is also prominent in people who are high in authoritarianism and patriotism (Adorno 1950). The local identity may develop somewhat differently. It is much easier to feel a threat from those of another country than those from another city within the same country.

FIGURE 1

MODEL OF HYPOTHESIZED SOCIAL IDENTITY RELATIONSHIPS

Social dentity theory suggests that feeling a threat from outsiders is not always necessary; a local identity may develop from the positive feelings one receives from one’s association with the community. In other words, collective self-esteem can motivate the elevation of a social identity and also the outcomes of the identity. Nevertheless, feeling a threat can also increase the salience of a particular identity. Citizens of a small town threatened with the loss of stores or jobs may look inward for support. While feeling a threat will increase the sense of identity, it should be sufficient that people simply want to be positively distinguished from others.

While studies may be constructed to manipulate individuals’ level of salience of social categories (Oakes, Turner and Haslam 1991), the present study will simply measure the level of respondents’ community identity without attempting to manipulate it. This could be considered a measurement of the baseline level of the community social category. It is not unlikely that other social categories are typically of greater importance to most individuals. This implies that there may be unseen competition from other social categories within each respondent leading the community identity to be minimized.

The Model, Constructs and Hypotheses

Self-categorization theory is framed in terms of antecedents to depersonalization, of depersonalization, and of outcomes to depersonalization. For a discussion of antecedents to depersonalization in general see Oakes (1987) and specifically in regard to the community identity see Lantz (1997). This study focuses more on depersonalization and its outcomes. Respondents will be presumed to be operating at a baseline level of identity with the community social category, their normal degree of depersonalization when nothing is happening to make it unusually salient. An assessment of the degree of depersonalization will be obtained from each respondent, outcomes of depersonalization will be observed and it will be determined whether or not depersonalization is sufficient to result in a purchase preference. Individuals showing high levels of community identity are showing evidence of being more depersonalized in regard to the community social category.

One outcome of depersonalization is ethnocentrism. At the community level ethnocentrism may be manifested by expressions of support for the economic health of the community. This reasoning is precisely parallel to that of national consumer ethnocentrism defined by Shimp and Sharma (1987). Community identity should flow to community ethnocentric tendencies because there would not be an expression of economic support for an identity unless the identity itself is sufficiently strong. In terms of the social identity approach, an individual’s identification with the community should not require an economic incentive to be activated. This is why community identity is characterized as being distinguishable from, and antecedent to community ethnocentrism. Figure 1 illustrates the proposed relationships.

H1: Higher community identity leads to higher community consumer ethnocentric tendencies.

A principle of self-categorization theory is that people wish to see themselves and the groups with which they associate in a positive light. Therefore, it is expected that people with a greater community identity will tend to evaluate a locally manufactured product as being of higher quality. This may occur even under conditions when there is no objective reason t believe the local product is superior.

As will be discussed later, the product selected for examination in this study is of a quite generic quality, so it is not expected that the difference in quality between ingroup and outgroup will be of great magnitude. Nevertheless, a bias in favor of the local product is consistent with self-categorization theory, so a hypothesis relating to quality assessments is included. If it is indeed true that some respondents recognize a difference in product quality, product quality is likely to be related to community identity.

H2: Higher community identity leads to higher perceived product quality assessment for the locally manufactured product.

When an outcome of depersonalization is sufficiently high there may be a marketplace response such as a purchase preference for the locally manufactured product. This is an extension from self-categorization theory per se. A search of published research reveals no instance where social psychologists, who have done most of the work in self-categorization, have carried an outcome of depersonalization to a purchase preference. However, the logical result of consumer ethnocentrism is to show a preference for the ingroup’s products. Therefore, this will be expressed in a hypothesis.

H3: Higher community consumer ethnocentrism leads to a purchase preference for a locally manufactured product.

H4: Assessment of higher product quality leads to preference for a locally manufactured product.

Additionally, a conjoint analysis will be used to demonstrate social cooperation. Social cooperation is defined as an individual making a personal sacrifice in order to support the group. Social cooperation is present when respondents indicate that they will make a personal sacrifice (in this case by paying more) to support the locally manufactured product. Social cooperation will be determined by examining the conjoint analysis scores of two attribute combinations where the local product was higher priced than an otherwise equal, lower priced alternative.

METHOD

Sample

The subjects were comprised of University of Manitoba students enrolled in three introduction to marketing classes who were given course credit for their participation. Use of a student sample is not preferred and limits the generalizability of the results. However, as will be discussed later, the effects of depersonalization are likely to have been minimized rather than enhanced by the use of students so the overall purpose of the study is not compromised.

A question in the survey asked respondents to write in the nam of their community, defined as "the city or town where you consider home to be located." From this question, three foreign respondents and three non-Manitoban respondents were eliminated. Several non-Winnipeg Manitobans were retained as it was determined that due to the rural character of Manitoba, small towns constituted the hinterland of the large city and these residents likely identified with Winnipeg anyway. This reduced the sample size from 114 to 108. Due to some respondents’ purported familiarity with the brand of the product being tested there was a further reduction to 100. The significance of brand will be discussed more later.

Measures

Purchase Preference (Purpref). Purchase preference was determined by the use of conjoint analysis. The product tested was a product which may be unique to northern climates and was familiar to all respondents. It was a block heater cord holder. During winter months automobile engines in northern climates require heat when turned off for several hours so engine block heaters are installed. They are operated electrically with the cord extending from the automobile’s grill. It is common for a holder to be installed on the grill to wrap the cord around. The product is quite generic with essentially nothing to differentiate one from another in terms of quality. They are typically not associated with a brand. The generic quality was considered a desirable property so as to simplify the design. Nevertheless, the survey included questions regarding brand familiarity and product quality.

Community Identity (Comid). The measure for community identity was derived from several sources and their use is partly exploratory. Items were adapted from the identity subscale of Luhtanen and Crocker’s collective self-esteem scale (1992) and pretested for this study. An additional item was taken from Hawes and Lumpkin (1984), an outshopping paper, because it appeared to focus on community identity. All items are included in the appendix.

Community Consumer Ethnocentric Tendencies (Comcet). The community consumer ethnocentrism items were adapted from Shimp and Sharma’s (1987) Cetscale, customized to be appropriate for the local identity. Only four items are readily adaptable. Nevertheless, as reported below, they show good reliability and appear to measure the desired construct.

Product Quality (PQdif). Product quality was assessed using a five item battery of questions with a seven point agree-disagree format for the product from each city. A difference variable was created by subtracting the sum of the product quality assessments of the Belleville, Ontario block heater cord holder from the sum of the Winnipeg, Manitoba block heater cord holder. A difference variable was deemed appropriate because the point of interest is in observing the preference for one product over another based on the city of manufacture.

The Conjoint Design

Conjoint analysis is an appropriate means of determining purchase preference because it can elicit the desired information about a specific product attribute without broadcasting the purpose of the study to repondents. The product attributes selected for inclusion in conjoint analysis were brand and place of manufacture, price, and retail stores. Three prices were included for the purpose of quantifying the preference. Price will be used later in this paper to illustrate social cooperation. The retail stores were Zellers and Wal-Mart, two competing discounters. Stores were included to disguise the true purpose of the survey. Further, the place of manufacture was presented to respondents with the brand so as to disguise the purpose of the survey. Due to the possibility of brand names influencing respondents’ choice, questions were included in the survey to confirm that they were unfamiliar with them. Indeed, one was fictitious while one was real. A real brand was used so that the label information (showing it was made locally) could be authenticated to the respondents.

Since the city where the product was manufactured was presented to respondents together with a brand name, there was a possibility that product preferences were expressed for the brand rather than the city. To rule this out, questions inquiring about brand familiarity were included. Respondents claiming to have substantially greater knowledge of either brand were eliminated from the sample. On a seven point scale with one indicating no familiarity with the brand the mean scores were 1.6699 for the local brand and 1.6565 for the fictitious other brand. A t-test shows an insignificant difference (t=.200, sig.=.839). It can be reasonably concluded that the product importance of the attributes described as "made in Winnipeg by Temro" and "made in Belleville by KarGard" referred to the city and not the brand.

The conjoint exercise was done first in the survey so that respondents did not know the purpose of the survey. A full profile plan was used. Respondents were asked to consider themselves to be in need of a block heater cord holder and were asked to express their preference by ranking each combination of product attributes. In order to avoid unduly burdening respondents, only eight of the possible twelve attribute combinations were used. The combinations were orthogonal to each other, having been selected by the ORTHOPLAN procedure of SPSS. Conjoint analysis yields utilities for each level of each product attribute. The utilities for each product attribute sum to zero. In other words, when there are two levels of an attribute (such as two cities for the "made in" attribute) one utility will be a positive number and the other utility will be the same number but negative. The only exception to this is price. Prices were specified to have a linear and negative relationship. In other words, a lower price is viewed as more preferred as compared to a higher price. The relative importance of each attribute is simply the combined utilities expressed as a percentage (see Table 1). This allows an easy comparison of the contribution each attribute makes to the overall preference. Reliability of the estimates is tested by Kendall’s tau. It should be close to one.

TABLE 1

CONJOINT ANALYSIS RESULTS

Reliability and Unidimensionality of Measures

Community identity was measured with an eight item scale. Reliability analysis showed that improvements could be made with the deletion of two items. Also, an exploratory factor analysis showed that when these two items were deleted, the remaining six items became unidimensional. Reliability of the six items was .84 and an exploratory factor analysis showed the scale to be unidimensional explaining 56.2% of the variance.

Community consumer ethnocentrism was measured using a modified version of Shimp and Sharma’s Cetscale (1987). Relibility analysis showed that the four item scale had a Cronbach’s Alpha of .8199. This is somewhat lower than in the pretest (.865) but still acceptable. Exploratory factor analysis shows the scale to be unidimensional, as well. The five item product quality scales duplicated for the products from Winnipeg, Manitoba and Belleville, Ontario had reliabilities of .944 and .939 respectively and they were unidimensional.

RESULTS

Conjoint Results

The conjoint analysis results are presented in Table 1. In terms of the overall results of the conjoint exercise, price was by far the most important attribute. This was certainly due to the near generic nature of the product. The favored store was Wal-Mart. Retail quality measures included in the survey support the conjoint analysis results. The utilities associated with city of manufacture also show a significant preference for the locally manufactured product. The purchase preference variable was defined as the difference between the utility of the local product and the outgroup product. This represents the preference respondents expressed for the local product in relation to all attributes in the conjoint. Purchase preference will be used as the de

Regression Results

The means, standard deviations and correlations of the variables of interest are summarized in Table 2. H1 predicts a relationship between community identity and community consumer ethnocentrism, and H2 predicts a relationship between community identity and product quality assessment. Regression analysis supports H1, a significant relationship was obtained (F=16.93, sig.=.0001) with a beta of .4174 and an R-square of .1739. As seen in Table 2 the correlation between community identity (Comid) and community ethnocentric tendencies (Comcet) is .445. This supports the proposition offered by self-categorization theory that a high level of identity with a social category leads to an outcome of higher ethnocentric tendencies within that social category.

H2, however, is not supported. This relationship has insignificant correlation and regression results. This was not the expected result. Several reasons may account for this lack of association. While regression is robust in regard to violations of normality, PQdif is not normal and this may have influenced the result. Additionally, the proposition that community identity represents depersonalization is still a concept under development. It is possible that the depersonalization of the community social category is not fully captured by an idenity variable. And finally, it is possible that the community identity simply ranks too low in order of importance to most people for it to be properly manifested.

A regression was run to examine H3 and H4 with purchase preference as the dependent variable and community ethnocentric tendencies and product quality as independent variables. The results show a significant relationship (F=20.48, sig.=.000), R square=.3036. Both Comcet (t=4.418, sig.=.000; beta=.3807) and PQdif (t=4.430, sig.=.000; beta=.3816) make significant contributions to purchase preference, confirming both hypotheses. This result supports the role of self-categorization theory in explaining purchase preference. While social psychologists have not extended studies to an expression of purchase preference, the support for H3, following H1, shows that a purchase preference is a reasonable final outcome of an elevated community identity.

TABLE 2

MEANS, STANDARD DEVIATIONS AND CORRELATIONS

FIGURE 2

SAMPLES OF CONJOINT ANALYSIS CARDS

Regarding product quality assessments, 62% said that product quality was precisely equal between the two cities. But those who perceived a difference generally favored the local product with means of 26.55 and 25.67 for Winnipeg and Belleville, respectively. The difference is statistically significant (t=3.07, sig.=.003). This result is somewhat surprising in view of the standard quality of the product. The preference for the local product is, however, quite consistent with the expectation of self-categorization theory. Even with products in which the quality is observably equal to the great majority of observers, some people may express their group cohesion by giving the local product higher quality ratings.

Social Cooperation

Social cooperation was not included as a formal hypothesis because it is a construct still being developed. Nevertheless, the results will be presented simply to make a clear demonstration of the significance of local preferences. Social cooperation was determined by examining the conjoint analysis scores of two cards where the locally manufactured product was priced higher. Store, the remaining factor, was the same on both cards. The two cards are illustrated in Figure 2. The results showed that 27 of 100 respondents indicated that they would buy the local product when the price of the local product was higher.

CONCLUSION

The findings provide overall support for a model that uses the social identity approach, specifically self-categorization theory, to illustrate the relationship between the community social category and the outcomes of this community association. Greater levels of community identification lead to greater levels of community consumer ethnocentrism. Greater levels of community-based ethnocentrism, in turn, result in a tendency to express a preference for locally manufactured products. Even with an extremely mundane product a substantial portion of the respondents expressed a willingness to spend more in support of the locally manufactured product. In a practical sense, this paper suggests tht strategies are available which may be effective in building local identity which can be used to encourage local economic health.

The results are limited in that university students were used as respondents. While the results cannot be generalized to the entire population it is more likely than not that community identity will increase with age as older people have been found to value stability and security (Kahle 1983). Additionally, a student sample largely precludes the use of income as a meaningful characteristic, although income and perception of economic conditions have been found to be useful in explaining consumer ethnocentric tendencies (Lantz 1997). In other words, use of a student sample is likely to have minimized the effect rather than enhanced it.

Future studies may attempt to manipulate the level of community identity in order to learn more about how it develops and how it may be strengthened. Social marketing causes may be able to benefit, as well. If people will pay more to support local manufacturers they may also be willing to show support for parks, public health issues, schools, etc.. It is not expected that the community identity will ever be particularly strong in terms of magnitude compared to other social categories. Nevertheless, a measure of the local identity is important in defining the range of social identities to which a person may subscribe.

APPENDIX

SCALE ITEMS FOR COMMUNITY IDENTITY AND COMMUNITY ETHNOCENTRIC TENDENCIES

REFERENCES

Darden, William R. and William D. Perreault (1976), "Identifying Interurban Shoppers: Multiproduct Purchase Patterns and Segmentation Profiles," Journal of Marketing Research, 13 (February), 51-60.

Hawes, Jon M., and James R. Lumpkin (1984), "Understanding the Outshopper," Journal of the Academy of Marketing Science, 12 (4), 200-218.

Hogg, Michael A. and Dominic Abrams (1990), Social Identifications: A Social Psychology of Intergroup Relations and Group Processes. London: Routledge.

Hozier, George C. and Donald E. Stem (1985), "General Retail Patronage Loyalty as a Determinant of Consumer Outshopping Behavior," Journal of the Academy of Marketing Science, 13 (1), 32-46.

Lumpkin, James R., Jon M. Hawes and William R. Darden (1986), "Shopping Patterns of the Rural Consumer: Exploring the Relationship between Shopping Orientations and Outshopping," Journal of Business Research, 14, 63-81.

Lantz, Garold (1997), "An Assessment and Scale Development of the Community Identity Using the Social Identity Approach." American Marketing Association Winter Educator’s Conference Proceedings.

Luhtanen, Riia and Jennifer Crocker (1991), "A Collective Self-Esteem Scale: Self-Evaluation of One’s Social Identity," Personality and Social Psychology Bulletin, 18 (3), 302-318.

Oakes, Penelope (1987), "The Salience of Social Categories," in Rediscovering the Social Group: A Self-Categorization Theory, ed. John C. Turner et al., Oxford, England: Basil Blackwood Limited.

Oakes, Penelope, John C Turner and S. Alexander Haslam (1991), "Perceiving People as Group Members: The Role of Fit in the Salience of Social Categorizations," British Journal of Social Psychology, 30, 125-144.

Papadopoulos, Nicholas G. (1980), "Consumer Outshopping Research: Review and Extension," Journal of Retailing, 56 (4), 41-58.

Shimp, Terece A. and Subhash Sharma (1987), "Consumer Ethnocentrism: Construction and Validation of the CETSCALE," Journal of Marketing Research, 24 (August), 280-289.

Sumner, G. A. (1906), Folkways. New York: Ginn Publishing Company.

Tajfel, Henri (1978), Differentiation Between Social Groups. London: Academic Press.

Tajfel, Henri (1981), Human Groups and Social Categories. Cambridge, England: Cambridge University Press.

Turner, John C. (1982), "Towards a Cognitive Redefinition of the Social Group," in Social Identity and Intergroup Relations, ed. Henri Tajfel, Cambridge, England: University Press.

Turner, John C., Michael A. Hogg, Penelope J. Oakes, Stephen D. Reicher and Margaret S. Wetherell (1987), Rediscovering the Social Group: A Self-Categorization Theory, Oxford, England: Basil Blackwood Ltd.

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