Associational Involvement: an Intervening Concept Between Social Class and Patronage Behavior

Scott Dawson, University of Arizona
Melanie Wallendorf, University of Arizona
ABSTRACT - Social class differences in consumer behavior have been suggested to be a function of differing life styles. Yet, few scholars have devoted attention to what aspects of these life styles are responsible for the observed differences in consumer behavior. This study introduces associational involvement as one aspect of life style. Results indicate that associational involvement is significantly better than social class indicators in predicting patronage of specialty stores for clothing purchases.
[ to cite ]:
Scott Dawson and Melanie Wallendorf (1985) ,"Associational Involvement: an Intervening Concept Between Social Class and Patronage Behavior", in NA - Advances in Consumer Research Volume 12, eds. Elizabeth C. Hirschman and Moris B. Holbrook, Provo, UT : Association for Consumer Research, Pages: 586-591.

Advances in Consumer Research Volume 12, 1985      Pages 586-591


Scott Dawson, University of Arizona

Melanie Wallendorf, University of Arizona


Social class differences in consumer behavior have been suggested to be a function of differing life styles. Yet, few scholars have devoted attention to what aspects of these life styles are responsible for the observed differences in consumer behavior. This study introduces associational involvement as one aspect of life style. Results indicate that associational involvement is significantly better than social class indicators in predicting patronage of specialty stores for clothing purchases.


Seminal work by Martineau (1958a, 1958b) stimulated a stream of research investigating the causal mechanisms linking social class with store choice or patronage behaviors. Facilitated by the availability of an easily applied social class index (e.g., Warner's ISC, 1949), numerous authors further investigated this relationship. Both Martineau and Levy indicated there was a correspondence between patron and store status, Martineau concluding that:

"The woman shopper has a considerable range of ideas about department stores: but these generally become organized on a scale ranking from very high-social status to the lowest-status and prestige. The social status of the department store becomes the primary basis for its definition by the shopper... The shopper is not going to take a chance of going to a store where she might not fit in." (1958a, p. 127)

Levy (1968) further clarified this causal linkage by taking into account the characteristics of store choice decisions. He noted that for merchandise low in risk, upper class individuals would be willing to shop at a lower status store such as Sears, while lower class people were likely to shop at higher status stores such as Marshall Fields only when purchasing gifts. Ryans (1979) found a similar relationship to exist in gift-giving purchase occasions. By utilizing Bauer's (1961) distinction between economic and social risk, Prasad (1975) found social risk to be the most important characteristic of a purchase situation affecting store choice. Regardless of the level of economic risk, when social risk was high, Prasad found high social class respondents unlikely to patronize discount stores.

Other studies have had mixed results concerning the relationship between social class and store choice. For instance, Rich and Jain (1968) obtained a significant and positive association between social class and store preference. In another study, preferences for discount, chain, department, specialty, and boutique stores were compared across women of differing occupational statuses. Housewives, working women, and career women had similar preferences for each store type with the exception of specialty stores (Hirschman 1980). Schaninger (1981) found upper class people patronizing specialty stores more than lower class people when shopping for furniture and clothing. Finally, Bellinger et al. (1976) found income and education to be better predictors than social class of store choice for several types of clothing purchases.

In sum, the literature concerned with social class and patronage behaviors has not produced a conclusive and unified theoretical linkage between these two concepts. To progress further in building such a theory, more explication of why social class should affect store choice is necessary. Yet a concept indicating the potential for further explication emerges from the empirical results of the study by Bellinger et al. (1976). These authors entered membership in several types of organizations into a discriminant analysis of store choice. Whether or not female respondents belonged to a bridge club was the first variable entering the discriminant function for store choice when purchasing suits or dresses. The same type of club membership was also a significant predictor of store choice for women's sportswear.

The irony of the Bellinger et al. study is that the results are consistent with a theoretical explanation of why social class should be indirectly, rather than directly, related to store choice. As the results of this study will show, associational behavior has an intervening effect between social class and store choice A review of these two concePtual linkages follows.


Social Class and Associational Involvement

Recently, Coleman (1983) reviewed two of the most current views of the American status structure, namely the Gilbert-Kahl New Synthesis Class Structure, and the Coleman-Rainwater Social Standing Class Hierarchy. In contrast to the political-social structure basis of the Gilbert-Kahl model (1982), the Coleman-Rainwater (1978) classification is based on patterns of interaction among members of society. The Coleman-Rainwater approach is a broader conception of class structure that not only considers the common indicators of education and occupation, but also the more subjective aspects of individuals' life positions including social skills, status aspirations, recreational habits, and community participation. Coleman (1983) concludes that "ultimately, the proper index of status is a person's social circle of acceptance." (p. 267)

Given that the types of people in social networks effectively define levels of social class, considerable attention has focused on associational behaviors across social classes. Studies have consistently found a strong positive relation between social class and the number of groups to which individuals belong (Kahl 1957, Hodges 1964, Curtis and Jackson 1977). Upper- and Middle-class people are significantly more active in formal and informal organizations than are members of working and lower classes. This difference in extent of social participation by class indicates an overlap in the business and social spheres of upper- and middle-class individuals and a clear separation between the two spheres for lower class individuals. For members of the middle- and upper-classes, business relationships spill over into the social arena, and these two spheres are not recognized as at all distinct. This life style is in sharp contrast to that of the lower and working classes who have fewer intimate friends outside of the extended family. They are less likely to participate socially with their working mates and more likely to interact frequently with siblings (Dotson 1951).

Similar findings have been produced by authors interested in life style, or patterns of cultural choice. This body of literature tends to separate the several components of social class as distinct conceptual entities, rather than combining them into a single index. As such, the results are slightly different for each of the separate components of standard measures of social class. Several studies have demonstrated that education, in comparison to other indicators of status or class taken separately, has the strongest positive relation to number of group memberships, social friends, and neighboring activities (Davis 1982, Greenberg 1983). Conceptually, this relationship exists because greater educational status is associated with a greater breadth of interest. Since these interests are likely to be more abstract, the development of knowledge concerning these interests is more dependent on organizational entities such as journals, conferences, and formal group meetings. Further, education is also positively related to the sophistication of social skills which may make participating in groups less intimidating. Davis (1982) also found occupational prestige significantly and positively related to number of group memberships. Perhaps verifying the difficulties inherent in the use of income as a component of social class (e.g., Coleman 1983), income has not been shown to be significantly related to associational behavior.

In sum, the prestige associated with an individual is primarily a function of his or her circle of friends. The higher an individual is in the social class hierarchy, the greater the extent to which this circle of friends is acquired through multiple associational memberships.

Associational Involvement and Store Choice

Results from the Bellinger et al. (1976) study indicated that membership in a specific group was the most significant predictor of preference for a specific type of store. The prior discussion established that social class is positively related to the extent of associational memberships. The concern here is not with why membership in one specific group should affect choice of a specific store (an appropriate concern of reference group theory), but with why the total number of group memberships, or associational involvement, should affect choice of a category of stores.

Hirschman (1978) described retail structure with a typology that included specialty stores, department stores, mass merchandisers, and discount stores. This typology approximates a price-quality continuum. Further, Jain and Etgar (1976) obtained a multidimensional store map in which the most salient dimension, store status, contained the same continuum. Using the same methodology, Singson (1975) produced a nearly identical store map with a primary dimension which he labelled price/quality. Thus, a general observation is that Hirschman's store typology is one of decreasing status.

As indicated previously, Prasad (1975) found that individuals of higher social class were less likely to patronize discount stores under purchase situations of high social risk. Similarly, when quality of a visible product is the most salient attribute for a consumer, he or she is less positive toward purchasing that product at a discount store than at a department store (Wheatley and Chiu 1974). In sum, this literature suggests that the greater the social risk of a purchasing situation, the greater the patronage of higher status stores. Further, this relationship is more likely for individuals of higher social class than for those of lower social class.

If all members of all social classes faced an equal number of purchasing occasions of high social risk, the proposition just stated concerning store patronage would be observed equally across social classes. Yet, research shows that these differences are more than minimal (e.g., Schaninger 1981). In effect, members of higher social classes may face more purchase situations of high social risk than members of lower social classes. One source for this difference in social risk exposure might be the extent of associational behavior of individuals higher in social class.

Social participation is positively related to the degree to which an individual is exposed to the expectations and norms of referents. As exposure to others through participation in groups increases, so does the chance of it being observed that an individual owns a product which is considered inappropriate. Similarly, an informal meeting at a country club or a formal meeting of the Chamber of Commerce are both situations in which ownership of a visible or symbolic product, such as clothing, is a crucial determinant of an individual's acceptance by others within the particular group setting. A gathering in one's home of a business, social, or community group increases risk of others noticing that furniture, a highly visible product, is inappropriate to the expectations of members of the gathering.

In summary, people of higher status interact in groups to a greater extent than do people of lower status. In so doing, they are subject to greater amounts of social risk in their ownership of highly visible and symbolic products. In attempting to meet the expectations of referents of a similar higher status concerning such products, individuals may be more likely to patronize specialty or local department stores.


The research conducted here is an initial attempt to test the previously suggested theoretical linkages. The research question is whether or not introducing amount of associational behavior as an intervening variable will lead to a better explanation of store choices by members of different social classes. As suggested, the greater the number of associations belonged to, the greater the exposure to social risk for visible and symbolic products. When faced with the purchase of such products, individuals are more likely to patronize stores of higher status.

In the current research, these issues are addressed by examining the proportion of clothing purchases made at specialty and at discount stores. Clothing has been widely suggested and used by researchers as a product which is highly symbolic and visible (Belk 1978; Holman 1980, 1981; Belk, Bahn and Mayer 1982; Schaninger 1981; Douglas and Solomon 1983). Specialty and discount stores represent opposite ends of the status and price-quality continuum as indicated previously.

The research question may be addressed by empirically testing the path diagram in Figure l. In that diagram, direct and indirect effects of occupational status, education, and income on choice of specialty and discount stores are hypothesized. If associational involvement does have the suggested effect, the indirect effects of this model should exceed the direct effects.

Occupational status, education, and income were treated as separate constructs, rather than combining them into a single social class index. Research concerned with life styles, of which associational behavior is one facet, has more prevalently treated the components of social class as separate constructs (Sobel 1981, 1983; Gruenburg 1983; Davis 1982). For instance, a recent study found almost no consistency between measures of education, occupational status, and predicted attitudes such as morals. attachments. politics, values, and social issues (Davis 1982). While education may cause an increasing amount of associational behavior through the mechanisms cited earlier, occupational status may lead to such behavior through a separate mechanism. For instance, an elder CEO at a large firm may have begun his or her way up from the bottom with little educational credentials, yet belong to several organizations by the time the CEO status is achieved. Finally, income may very well not have a consistent effect on associational behavior. Two individuals could have identical incomes, yet one could be an overprivileged blue-collar worker and the other a fledgling university professor. The differences in the associational behavior could be pronounced. As Coleman (1982) notes: "Class and income are not really very well correlated. They index two quite different aspects of life circumstances." (p. 272).




A valid argument raised by several marketing scholars has been that studies of class and status in consumer behavior should focus on the household rather than the individual as the unit of analysis (Shrimp and Yokum 1981; Zaltman and Wallendorf 1983). Yet, with the exception of the Coleman-Rainwater (1978) social class index, the rest of the social class indexes utilize only husband's levels of class and status indicators (Warner 1949; Hollingshead 1949).

Although it is clear that husband and wife indicators of occupation, education, and income should both be used in measuring theoretical effects, it is not clear as to how to treat these indicators as a unit. Three intuitively appealing heuristics would be averaging across indicators, taking the highest indicator for each class or status dimension, or adding indicators. However, it is not clear which of these processes would best describe the merging which occurs in a household. This issue was empirically resolved in this research by observing the correlations obtained between each of these three alternatives and the dependent variables. Perusal of these correlations indicated that adding occupational status (seven-point scale), education (interval scale), and income (interval scale) of husband and wife resulted in marginally greater magnitudes. Thus, this heuristic was used for husband-wife households.

Associational involvement was operationalized by asking respondents to name all organizations in which household members participated. Store choice was operationalized by asking respondents to state three stores where the majority of household clothing purchases were made and the percentage of purchases made at each. These stores were classified as specialty, department, and discount. As department stores were not separated into national (mass merchandisers) or local, this category was excluded from the analysis.

The status or prestige of the stores where respondents purchased clothing was not measured. A key assumption made in this study is that specialty and discount stores represent opposite ends of a status continuum. Although the store maps produced by Singson (1975) and Jain and Etgar (1976) indicate this assumption to hold in general, there are store types which may violate the assumption. For instance, it is not clear where off-price merchandisers or catalog showrooms fit on the status continuum. Yet, it is unlikely that these anomalies contaminate the results of this investigation. Catalog showrooms were coded in a separate category and off-price merchandisers were infrequently if at all encountered in the responses. Indeed, this latter type of retail institution is in its infancY in the market where the data was collected.


In-home personal interviews were conducted in a Southwestern city using a multi-stage sampling plan. Verification of the interviews was indirectly assessed by having interviewers photograph each respondent. As the household was the sampling unit, respondents less than 18 years of age were eliminated from the sample. A total of 346 interviews were completed. A preliminary examination of age and income distributions indicated the sample to be representative of the local population.


The path model was submitted to a submodel of Joreskog and Sorbom's (1981) LISREL V, producing the coefficients shown in Figure 2. The majority of the coefficients are not significant at the p < .05 level. Partial support is received for the positive relation between social class indicators and amount of associational behavior. While occupational status and income are significant causal variables of amount of associational behavior, education is surprisingly insignificant.



Encouragingly, the data strongly supports the new causal linkage introduced in this study. That is, amount of associational behavior is a significant and positive predictor of proportion of clothing purchases made at specialty stores. Yet, while associational behavior does appear to increase specialty store patronage, it does not affect discount store patronage in an opposite manner as might be expected. Amount of associational behavior does not appear to be related to discount store patronage in the context of clothing purchases. Nevertheless, the predicted relationship with specialty store purchases is the strongest coefficient in the entire model.

Two other general observations are noteworthy. First, of all variables examined here, only education is a moderately significant determinant of clothing purchases at discount stores. Neither social participation, income, nor occupational status is a significant predictor of discount patronage. Apparently those variables that cause individuals to patronize this type of store have not been tapped. In fact, the equations in the model explain a trivial 3% of the variance of clothing purchases made at discount stores.

The second observation is that, with the single exception of education as mentioned above, the social class variables considered here have no significant direct effects on patronage of either specialty or discount stores. For that matter, Table l indicates that neither direct or indirect effects are significant in predicting specialty store patronage for any of the three variables. In all cases, however, the indirect effects are greater than the direct effects. This result leads to the tempered conclusion that class or status variables do not directly affect patronage behaviors. Rather, it appears that these variables cause other life style behaviors which in turn affect patronage. It is in this sense that amount of associational behavior acts as an intervening variable between social class and store choice for products of a visible and symbolic nature.



On the basis of the significant coefficients reported in Table 1, the revised model illustrated in Figure 3 was submitted to the structural equations routine. This model is readily accepted according to the current criteria for accepting covariance models (X2 with 7 degrees of freedom - 6.69, p - .461). The results reaffirm the interpretations just made. The class variables of income and occupation are positively associated with group activity, which in turn is positively related to specialty store patronage. Education is a significant predictor of not patronizing discount stores.



An attempt to lend better clarity to the somewhat mixed results was made by separating total associational memberships into a typology that included business, recreational, social, and religious memberships. This typology was chosen as it had been used successfully by Hirschman and Wallendorf (1979) in the context of stimulus variation.

Several conclusions may be drawn from the correlations between the social class variables and memberships in the different types of associations displayed in Table 2. First, business and community associations are by far the most prevalent types of memberships. Few individuals participate in social or religious groups and this participation is poorly predicted by any of the social class variables. Secondly, as earlier studies indicate (e.g., Davis 1982; Greenberg 1983) business and community group involvement is strongly predicted by occupational status, education and income.

The correlations of Table 3 illustrate the contribution offered by this study. With the exception of social group memberships, all of the types of associational involvement are significantly and positively correlated with specialty store patronage. In fact, purchases of clothing at specialty stores is more highly correlated with each associational indicator (except for social) than it is with occupational status, education, or income. Conversely, none of the associational memberships is more significantly correlated with clothing purchases at discount stores than any of the social class variables.


Marketing scholars have claimed that "The concept of life style is the essence of social class..." (Myers and Gutman 1974, p. 236). One aspect of life style that differs across social classes is the degree of involvement in different types of associations. This study provides initial evidence that this component of life style intervenes in the relationship between social class and patronage of specialty stores for a visible and symbolic product. As such, the theoretical argument and results concerning amount of associational behavior provide one answer to the question of why social classes should differ in patronage behaviors. Individuals higher in social class, as part of their life style, tend to belong to more associations than individuals of lower social class. For products which are symbolic and visible, such as clothing or furniture, individuals with greater associational involvement are subject to greater social risk. When purchasing occasions arise for such products, people of high associational involvement are likely to alleviate social risk by making purchases at higher status, specialty stores.

Although the results of this study are encouraging, they are subject to limitations and, thus, point to several avenues for future research. Perhaps most importantly, the theoretical argument has centered around the concept of social risk, while not directly measuring this concept. The greatest contribution toward establishing the validity of this argument would be direct measurement of social risk for clothing or furniture purchases, in addition to the variables considered here. As the validity of measuring all forms of risk is in question (e.g., Ross 1975), using product involvement or importance may produce better results. These concepts are closely related to social risk and have enjoyed the development of valid and reliable scales (Bloch 1981; Bloch and Richins 1983). If associational involvement is a significant predictor of high social risk and/or product involvement purchases, the relationship between this concept and patronage of specialty stores could be made with a higher degree of confidence.

A second limitation of this study concerns the procedures used for operationalizing patronage. The results could be generalized further by measuring dollar purchases made at each of the four store categories put forth by Hirschman (1978). A crucial point in the theory posited here was that high social risk purchases would be made at high status stores. While the approximate price-quality continuum in Hirschman's typology does provide a useful surrogate of store status, a more valid procedure would be to measure the status of stores in which respondents made purchases. This measurement could be accomplished by the ranking or rating of stores in terms of status by a separate sample.





Finally, practitioners can benefit from the results of this study in developing segmentation and promotional strategies. For retailers who seek a merchandise mix consisting of highly visible and symbolic products, further development of that mix and promotion ought to be directed at those who belong to many business, community, religious, and recreational groups. The most salient outcome of this study is that concentrating on these entities, rather than directly on individuals defined according to indicators of social class, is a more efficient and perhaps effective use of managerial resources. The suggestion is the opposite for discount retailers. Given the results of this study; these retailers would be better off concentrating marketing efforts on those individuals lower in occupational status. education. and income.


Bauer, Raymond (1960), "Consumer Behavior as Risk Taking," in Proceedings, ed. M. L. Bell, Chicago: American Marketing Association.

Belk, Russell W. (1978), "Assessing the Effects of Visible Consumption on Impression Formation," in H. K. Hunt (ed.), Advances in Consumer Research, Vol. 5: Association for Consumer Research.

Belk, Russell W., Kenneth D. Bahn, and Robert N. Mayer (1982), "Developmental Recognition of Consumption Symbolism," Journal of Consumer Research 9 (June): 4-17

Bellinger, Danny N., Dan H. Robertson, and Elizabeth Hirschman (1976), "Age and Education as Key Correlates of Store Selection for Female Shoppers," Journal of Retailing 52 (Winter): 71-78.

Bloch, Peter H. (1981), "An Exploration Into the Scaling of Consumers' Involvement with a Product Class," in Advances in Consumer Research, ed., K. Monroe, Ann Arbor: Association for Consumer Research, 61-65.

Bloch, Peter H., and Marsha L. Richins (1983), "A Theoretical Model for the Study of Product Importance Perceptions," Journal of Marketing 47 (Summer): 69-81.

Coleman, Richard P. and Lee Rainwater (1978), Social Standing in America: New Dimensions of Class, New York: Basic Books, Inc.

Coleman, Richard P. (1983), "The Continuing Significance of Social Class to Marketing," Journal of Consumer Research, 10 (December): 265-280.

Curtis, Richard F. and Elton F. Jackson (1977), In- equality in American Communities, New York: Academic Press.

Davis, James (1982), "Achievement Variables and Class Cultures," American Sociological Review 47 (October): 569-586.

Dotson, Floyd (1951), "Patterns of Voluntary Association Among Urban Working Class Families," American Sociological Review 16: 687-93.

Douglas, Susan P. and Michael R. Solomon (1983), "Clothing the Female Executive: Fashion or Fortune," in AMA Educator's Proceedings, Murphy et al. (eds.), Chicago: American Marketing Association, 127-132.

Gilbert, Dennis and Joseph A. Kahl (1982), The American Class Structure: A New Synthesis, Il.: Dorsey Press.

Hirschman, Elizabeth C. (1978), "A Descriptive Theory of Retail Market Structure," Journal of Retailing 54 (Winter): 29-48.

Hirschman, Elizabeth C. (1980), "Women's Self-Ascribed Occupational Status and Retail Patronage," in Advances in Consumer Research, ed., Kent Monroe, Ann Arbor: Association for Consumer Research. 648-654.

Hodges, Harold M. (1964), Social Stratification: Class in America, Cambridge, Mass.: Schenkman.

Hollingshead, August B. (1949), Elmtown's Youth: The Impact of Social Class on Adolescents, New York: John Wiley.

Holman, Rebecca (1980), "Clothing as Communication: An Empirical Investigation," in Advances in Consumer Re- search, ed. Jerry C. Olsen, Vol. 7: Association for Consumer Research, 372-77.

Holman, Rebecca (1981), "Apparel as Communication," in Symbolic Consumer Behavior, eds., Elizabeth Hirschman and Morris B. Holbrook, Ann Arbor: Association for Consumer Research, 7-15.

Jain, Arun K. and Michael Etgar (1976), "Measuring Store Image Through Multidimensional Scaling of Free Response Data," Journal of Retailing 52 (Winter): 61-70.

Joreskog, K.G. and D. Sorbom (1981), LISREL V: Analysis of Linear Structural Relationships by Maximum Likelihood and Least Squares Methods, Chicago: National Educational Resources.

Kahl, J.A. (1957), The American Class Structure, New York: Hole, Rinehart, and Winston.

Levy, Sidney J. (1968), "Social Class and Consumer behavior," in On Knowing the Consumer, ed ., Joseph Newman, New York: John Wiley.

Martineau, Pierre (1958a), "Social Classes and Spending Behavior," Journal of 'Marketing 23 (October): 121-130.

Martineau, Pierre (1958b), "The Personality of the Retail Store," Harvard Business Review 36 (January): 47-55.

Prasad, V. Kanti (1975), "Socioeconomic Product Risk and Patronage Preferences of Retail Shoppers," Journal of .Marketing 39 (July): 42-47.

Rich, Stuart A. and Subbash Jain ( 968), "Social Class and Life Cycle as Predictors of Shopping Behavior," Journal of Marketing Research 5 (February): 41-49.

Ross, Ivan (1975), "Perceived Risk and Consumer Behavior: A Critical Review," in Advances in Consumer Re- search, ed., M.J. Schlinger, Chicago: Association for Consumer Research. 1-19.

Ryans, Adrian (1977), "Consumer Gift Buying Behavior: An Exploratory Analysis," in Contemporary Marketing Thought, eds., D. Bellinger and B. Greenberg, Chicago: American Marketing Association. 99-104.

Schaninger, Charles M. (1981), "Social Class Versus Income Revisited," Journal of Marketing Research 18 (May): 192-208.

Shimp, Terence A. and J. Thomas Yokum (1981), "Extensions of the Basic Social Class Model Employed in Consumer Research," in Advances in Consumer Research Vol. 8, ed., Kent Monroe, Ann Arbor, MI: Association for Consumer Research.

Sobel, Michael (1981), Lifestyle and Social Structure, New York, Academic Press.

Tauber, Edward M. (1972), "Why Do People Shop?," Journal of Marketing 36 (October): 46-59.

Warner, Lloyd, Marcia Meeker, and Kenneth Eels (1949), Social Class in America: A Manual of Procedure for the Measurement of Social Status, Chicago: Social Science Research Associates.

Wheatley, John J. and John Chiu (1977), "The Effects of Price, Store Image, and Product and Respondent Characteristics on Perceptions of Quality," Journal of Marketing Research 14: 57-69.

Zaltman, Gerald and Melanie Wallendorf (1983), Consumer Behavior: Basic Findings and Management Implications, New York: John Wiley and Sons.