Schematic Processing of Information: an Exploratory Investigation

Meera Venkatraman, University of Pittsburgh
Angelina Villarreal, San Diego State University
ABSTRACT - In contrast to previous research in consumer behavior, which has studied brand stereotypes from a sociological point of view, this paper adopts a cognitive perspective to the study of brand stereotypes. It proposes the concept of stereotype or schema transfer. This concept is investigated in the context of testimonial advertising, since it may be useful in explaining the "how" and "why" of the effectiveness of testimonial ads.
[ to cite ]:
Meera Venkatraman and Angelina Villarreal (1984) ,"Schematic Processing of Information: an Exploratory Investigation", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 355-360.

Advances in Consumer Research Volume 11, 1984      Pages 355-360

SCHEMATIC PROCESSING OF INFORMATION: AN EXPLORATORY INVESTIGATION

Meera Venkatraman, University of Pittsburgh

Angelina Villarreal, San Diego State University

ABSTRACT -

In contrast to previous research in consumer behavior, which has studied brand stereotypes from a sociological point of view, this paper adopts a cognitive perspective to the study of brand stereotypes. It proposes the concept of stereotype or schema transfer. This concept is investigated in the context of testimonial advertising, since it may be useful in explaining the "how" and "why" of the effectiveness of testimonial ads.

INTRODUCTION

The early research on brand stereotypes in marketing was from a marketing management perspective. The focus was on the marketing and advertising strategy implications of brand stereotyping (Gardner & Levy 1955; Levy 1959). Research in the last decade has focused on the socio-cultural aspects of stereotyping brands. Researchers have studied the phenomenon of acquiring stereotypes as part of the socialization process (Belk et al 1982; Moschis & Churchill 1978; Ward 1974). This paper adopts a cognitive orientation to the study of brand stereotypes.

WHY A COGNITIVE PERSPECTIVE?

At the core of the cognitive argument for stereotyping is the thesis that the environment is complex and a stereotype is a cognitive mechanism to reduce this environment to proportions that can be handled by the individual (Ashmore & Del Boca 1981; Miller 1982). According to this perspective since the demands on the human information processing system are enormous, and the capacity of the system is limited, the individual develops cognitive shortcuts that reduce both the information processing demands and the behavioral demands on the system.

Besides its philosophical foundations, the cognitive orientation differs from the sociological orientation on one major dimension. The latter is content oriented whereas the former is process oriented (Ashmore & Del Boca 1981). The sociological orientation studies what the information content of stereotypes is and how they are influenced by the society of which the individual is a part. The cognitive orientation on the other hand, studies how the stereotype influences the processing of information and what the cognitive consequences of stereotyping are.

STEREOTYPES AS SCHEMAS

The definition of a schema adopted for this paper is "a schema is a cognitive structure that influences all the perceptual cognitive activities that are labeled as information processing, e.g., (perceiving, encoding, storing, retrieving, decision making) with respect to a particular domain" (Ashmore & Del Boca 1981). This definition emphasizes the process dimension of schemas or the influence a schema has on the processing of information (Neisser 1976; Nisbett and Ross 1979).

The process functions of schemas have been classified into two categories (Taylor & Crocker 1981). The first is the function of encoding and retrieval. This includes the influence the schema has on the perception, encoding and retrieval of information from the environment (Bransford & Johnson 1973; Cantor & Mischel 1977). The second is the function of interpretation and inference making. This is the process of filling in data missing from the input stimulus (Taylor & Crocker 1981).

There is substantial research evidence that stereotypes are schemas. It has been established that they perform the two cognitive process functions of schemas. They influence the encoding and retrieval of information. They also exert influence on the process of inference making and interpretation. The following two sections present research evidence that illustrates that stereotypes exhibit the same influences on new input that schemas do.

INFLUENCE OF STEREOTYPES ON THE ENCODING AND RETRIEVAL OF INFORMATION

There is evidence that stereotypes determine what information is encoded and retrieved. Information that confirms the stereotype is more easily attended to, stored in memory and retrieved as compared to information that disconfirms the stereotype. Pichert and Anderson (1977) presented two groups of students with a passage that describes two boys and the house in which they were playing. One group was instructed to read the passage from the perspective of a potential home buyer and the other from the perspective of a burglar. The subjects who read the passage from the perspective of the home owner were more likely to remember that the roof leaked than that the home had a color TV set, while for those who read the passage from the perspective of the burglar, the recall probabilities were reversed. The stereotypes of burglars and prospective home owners influenced the perception and retrieval of information from the passage.

In another experiment, conducted in a "more realistic setting," Cohen (1981) presented to two groups of subjects a videotape of a woman and her husband in conversation. Before viewing the tape, one group was told that the woman was a librarian and the other that the woman was a waitress. The tape was so constructed that it contained half librarian and half waitress features. After viewing the tapes the subjects were given recognition memory tests. The subjects were more accurate in remembering items that were consistent with the stereotypic impression they had of the woman than they were in remembering items that were inconsistent with the stereotypes. Thus, the stereotype of a burglar, home owner, waitress, or librarian facilitates the recall of the information that is consistent with stereotypic cognitions.

INFLUENCE OF STEREOTYPES ON THE PROCESS OF INFERENCE MAKING

Schemas influence the process of inference making in two ways. First, the schema directs a search for information that is missing in the input, but is contained in the existing schema to which the new input is matched. Second, if the new input is incomplete the schema can fill in missing data with best guesses based on prior experience and knowledge (Taylor & Crocker 1981). The function of assigning missing values also Manifests itself as the process of reconstruction. This is the process of systematic adding to information that has been previously acquired so that it is consistent with new information (Cantor & Mischel 1977; Spiro 1977; Loftus & Palmer 1974).

There is substantial evidence that stereotypes influence the process of reconstruction. In the first of a series of interesting experiments, Snyder & Uranowitz (1978) presented different groups of students with a narrative about a woman named Betty K. After reading the narrative, one group of students was told that Betty K is a lesbian and the other that she follows a heterosexual lifestyle. A week later, the students were given a multiple choice recognition test. In responding to the questions about Betty K. they chose stereotype consistent alternatives, although the ability to accurately recognize the stereotype consistent feature actually contained in the passage was clearly not facilitated (Snyder & Uranowitz 1978).

In an extension of this initial experiment, Snyder & Uranowitz (Snyder 1981) studied the role of interpretation in the reconstructive process. For each of their answers the subjects were asked to provide an interpretation of what those events meant for understanding Betty K. They found that "in general, the participants were remarkably adept in interpreting whatever they remembered . . . as manifestations of her emerging sexual orientation." The current stereotype of a lesbian or a heterosexual influenced the reconstruction and interpretation of events from the past.

Ln summary, there is evidence that stereotypes are schemas. They exhibit the two process functions of schemas. They are cognitive structures that exert influence on the encoding and retrieval of information and on the process of reconstruction or inference making.

THE PROPOSED MODEL FOR SCHEMATIC INFORMATION PROCESSING

Current theories of schematic processing propose that the active schema, if it is appropriate, is applied to the processing of a new input. If the active schema is not appropriate, the theories propose that the system searches for an appropriate schema. If an appropriate schema is not found, a new schema is created (Hastle 1981) Thus, according to the current theories. the non-existence of an appropriate schema leads to the creation of a new schema.

In the model presented in this paper, it is proposed that if an appropriate schema is not found, the system could explore the possibility of transferring an existing and related schema to process the new input. This event is termed here "schema transfer." The key premise underlying the process of schema transfer is the same as that for the creation of schemas viz., the need for cognitive economy. It is proposed that schema transfer places less demands on the system than the creation of a new schema. The process of schema transfer is another cognitive mechanism designed to reduce the demands on the human information processing system and reduce the environment to cognitive and behaviorally manageable proportions.

The objective of this paper is to test the hypothesis that schemas of stereotypes can be transferred. To establish that stereotypes can be transferred it must be shown that the process of inference making about the new input is influenced by the existing stereotype (Inference Making Hypothesis). It must also be established that the process of encoding and retrieval of information about the new input is influenced by the existing stereotype (encoding and Retrieval Hypothesis).

It is important to understand the conditions under which schema transfer may occur. It is hypothesized here that the process of schema transfer to a new input is facilitated by the activation of an existing schema for a person/object that is linked to the new input. A crucial question here is: what is the nature of this link? One link that is explored here is the physical proximity of the new and unfamiliar input and the existing stereotype.

The model of stereotype transfer proposes the transfer of both cognitions and affect from one object/person to another. The transfer of affect is based on the schematic processing of affect proposed by Fiske (1982). Fiske proposes that affect is category or schema based. It is stored with the semantic knowledge at the top of the structure. This affect may be generalized from prior instances with the category and transferred to new instances. According to Fiske, "stereotypic or schema based evaluation is a summary of many component evaluations, so it is ar efficient affective processing device."

In this paper, it is proposed that the affect that is linked with an existing stereotype is transferred to the new input along with the transfer of cognitions.

RESEARCH DESIGN

OVERVIEW

The schema transfer hypothesis assumes that the transfer of schematic cognitions and affect is facilitated by a link between the stereotype and the new input. Though the nature of this link is not known, it is hypothesized that the transfer will be facilitated by the physical proximity of the duo. In order to create this condition for the transfer in a context that is relevant for marketers, it was decided to use print testimonial advertising.

The use of testimonial advertising is appropriate for several reasons. First, it has been established that the level of the "role" is the appropriate level at which people stereotype other people (Taylor 1981). Thus the use of a well-known actress o. football player as the endorser is relevant for testing the schema transfer hypothesis. Second, the brand that is endorsed by the stereotyped person here is a new brand for which the respondents have no cognitions. In other words there is no stereotype or schema for the brand. Third, the stereotyped person and the new input are in the same advertisement and thus are physically linked together. They are also linked together by presenting the endorser i.e., the stereotyped individual as a user of the brand.

Construction of the advertisement

A. Creating the profile of the stereotypes

To construct the ad to be used in the experiment the first task was to create the profiles of the stereotyped roles. Thirty-two undergraduate students majoring in Business were asked to write a description of their stereotypes for five different occupations: actress, anchor person, football player, actor, and politician. The respondents were also asked for the name of the person who best fitted the stereotypic description. The descriptions of the stereotypes were analyzed to obtain a list of the attributes and features associated with each stereotype. This analysis was undertaken separately by each author and the lists of attributes were then compared and final profiles of each of the stereotypes were created. These profiles contained the most frequently mentioned attribute.

The respondents showed a high consistency in the descriptions of the actress, politician and football player stereotypes but low consistency in the anchor person and acto stereotypes. Thus, the role stereotypes selected for the proJect were the actress, politician, and football player stereotypes. The actress most frequently mentioned as the person who best fitted the stereotype of an actress was Jessica Lange. The politician and football player cited most often were Richard Nixon and Terry Bradshaw, respectively. These celebrities were selected as endorsers for the brands in the ads. For the stereotypic descriptions that were used for creating the ads refer to Table 1.

TABLE 1

DESCRIPTION OF THE ROLE STEREOTYPES

B. Selection of the new brand

In the selection of the product class from which the new brand was to be included in the advertisement, a major consideration was the value-expressiveness of the product. Products such as automobiles, beer, home furnishings, cigarettes, and magazines have been found to be value-expressive products (Lessig & Park 1978). From these product classes the two selected were magazines and home furnishings.

C. The Research Instruments

Research Instrument for the Inference Making Hypothesis

A set of two ads were created to test the inference making hypothesis. These ads consisted of the same brand (the new magazine) and the same ad copy but different endorsers In the first ad (AD1), the magazine was endorsed by the actress - Jessica Lange and in the second (AD2), the magazine was endorsed by the politician -- Richard Nixon Refer Figure 1.

FIGURE 1

THE DESIGN OF THE STUDY

The copy consisted of just one line: "It is what I have been looking for in a magazine and have not been able to find."

Research Instrument for Encoding and Retrieval Hypothesis

A set of two ads were designed to test the encoding and retrieval hypothesis. These ads consisted of the endorser, the new brand and an advertising copy about the brand. The new brand -- the Drexel Collection, and the ad copy were the same for the two ads, but the endorsers were different. In one ad (AD3), the Drexel Collection was endorsed by the football player and in the other ad (AD4), by the actress --Refer Figure 1. The copy contained equal number of features of the product that were consistent with the two different stereotypes. Thus, for example, the statement that the "collection is strong" is consistent with the football stereotype and the statement that the "collection is glamorous" is consistent with the actress stereotype.

Research Objective

The objective of the research is to show that the schema that is transferred from the endorser to the new brand affects further information processing about the product from the ad. The processes that are researched here are inference making and encoding and retrieval.

Research Objective: Inference Making Hypothesis

The objectives were (i) to test how the transfer of the stereotype from the endorser to the magazine influenced inference making about the magazine, and (ii) to test if the affect for the endorser is transferred to the new brand.

It is expected that in the case of AD1, more inferences will be made about the magazine that are consistent with the actress stereotype. Similarly, in the case of AD2, it is expected that more inferences will be made that are consistent with the politician stereotype. In both cases, it is expected that the affect will be transferred from the role stereotype to the magazine.

Research Objective: Encoding and Retrieval Hypothesis

The objective here was to determine how the transfer of the stereotype from the endorser to the carpet collection influenced the recall of information about the product from the ad. It is expected that in the case of AD3, the recall of information about the product that is consistent with the football player will be higher. While in the case of AD4, it is expected that the recall of information that is consistent with the actress stereotype will be higher.

The Respondents

Eighty-seven undergraduate Business students between the ages of 19 and 22 voluntarily participated in the project. These 87 respondents were randomly divided into four groups two groups (Groups 1&2) to test the inference-making hypothesis and the other two groups (Groups 3&4) to test the encoding and retrieval hypothesis. For the research design, refer to Figure 1.

The questionnaire for the respondents in Groups 1&2 consisted of four sections. The first section was the advertisement. a e second section measured the affect for the endorser. The third and fourth sections, respectively, measured the cognitions and affect of the brand. The questionnaire for Groups 3&4 consisted of three sections. The first was the advertisement and the second section consisted of questions included in the questionnaire to mask the purpose of the study. The third section measured the cognitions for the brand.

Measurement and Statistical Analysis

Inference Making Hypothesis

To elicit the brand cognitions the questionnaire contained a set of 14 statements about the brand. Seven of those statements were consistent with the actress stereotype, and the other seven were consistent with the politician stereotype. The respondents were asked to indicate their agreement with each statement on a five-point scale.

The data was analyzed at two levels, the aggregate scale level and the individual scale item level. For the analysis at the aggregate scale level, the statements based on the actress stereotype were considered as one category of statements. While those based on the politician stereotype were considered as a second category of statements. The vectors of mean responses for each of the categories were computed. Since the objective was to test for significant differences between these vectors of means, the multivariate counterpart of the t-test, the T2 test was used (Tatsuoka 1971). The vectors were compared between and within groups. For the analysis at the level of the individual scale items, the t-test was used.

To measure the affect (attitude) an 18-item affect scale was administered. This scale, based on the one used by Atkin and Block (1983), is a seven-point semantic differential scale. The affect of the respondents for the endorsers and Vanity Fair was measured

Affect transference was measured by comparing the affect for the endorser with the affect for the product. The reliability for each of the affect scales used was measured by the Cronbach's alpha coefficient (Allen & Yen 1979, p. 78). Since the minimum alpha coefficient was .77, the affect scale was considered reliable.

Encoding and Retrieval Hypothesis

To elicit the brand cognitions, the questionnaire contained 2 set of ten statements about the product taken from the copy of the ad. Five of these statements were consistent with the football player stereotype, and the other five with the actress stereotype. The recall of information was measured on a five-Point scale.

Here too, the analysis was undertaken both at the aggregate scale and at the individual scale item level.

RESULTS

Inference Making: Group 1

Analyzing the results for AD1, it was found that the vectors of mean inferences about Vanity Fair that were consistent with the actress stereotype was significantly different from the vector of inferences that were consistent with the politician stereotype (p=0.000) -- Refer Table 2. At the individual scale item level, a comparison of the means shows that the respondents in Group 1 made a higher number of inferences about Vanity Fair that were consistent with the actress stereotype as compared to the inferences they made about Vanity Fair consistent with the politician stereotype -- Refer Table 3. These results support the hypotheses of the transfer of an active schema to a new input.

TABLE 2

COMPARISONS OF MEANS

TABLE 3

COMPARISONS OF MEANS FOR INDIVIDUAL COGNITIVE SCALE ITEMS: INFERENCE MAKING GROUP

For the same groups the affect for the actress stereotype and Vanity Fair were compared. The results showed no significant difference between the vectors of mean responses for the affect for the stereotype and the affect for Vanity Fair (p=O.838) -- Refer Table 2. These results are also generally supported at the individual scale item level -Refer Table 4. These results support the hypothesis of affect transfer.

TABLE 4

COMPARISON OF MEANS BY INDIVIDUAL AFFECT SCALE ITEMS

Inference Making: Group 2

Analyzing the results for AD2, it was found that the vectors of mean inferences about Vanity Fair that were consistent with the politician stereotype were significantly different from the vector of inferences that were consistent with the actress stereotype (p=O 0398) -- Refer Table 2. A comparison of the means at the individual scale item level shows that the respondents in f,roup 2 made a higher number of inferences about Vanity Fair that were consistent with the politician stereotype as compared to inferences they made about Vanity Fair based on the actress stereotype -- Refer Table 3. These results support the hypothesis of transfer of cognition from a known object (or person) to a new input.

For the same group, the affect for the politician stereotype and Vanity Fair were compared. The results showed t at there is no significant difference between the affect for the stereotype and the affect for Vanity Fair (p=0.3081- Refer Table 2. These results are also generally supported at the individual item level -- Refer Table 4. Again, these results support the affect transfer hypothesis.

Inference Making: Group 1 vs. Group 2

A comparison was made between the vector of mean cognitions consistent with the actress stereotype in Group 1 and those consistent with the actress stereotype in Group 2. As was expected, a significant difference between the two groups was found (p=O. 049) -- Refer Table 2. The respondents in Group 1 made a higher number of inferences about the product that were consistent with the actress stereotype Refer Table 3. Similarly, the set of cognitions consistent with the politician stereotype were compared. As was expected, a significant difference between the two groups was found (p=0.015). The respondents in Group 2 made a higher number of inferences about the brand that were consistent with the politician stereotype -- Refer Table 3.

Summarizing, the results for the inference making process support the schema transfer hypothesis.

Encoding and Retrieval: Group 3

Analyzing the results for AD3, it was found that information encoded and retrieved based on the football player stereotype was significantly different from the information encoded and retrieved about the Drexel Collection on the actress stereotype (p=0.0130) - Refer Table 2. A comparison of the means at the individual scale item level showed that the respondents in Group 3 retrieved significantly more information about the Drexel Collection that was based on the football player stereotype as compared to information retrieved that was consistent with the actress stereotype -- Refer Table 5. The results support the hypothesis of schema transfer.

TABLE 5

COMPARISON OF MEANS FOR INDIVIDUAL COGNITIVE SCALE ITEMS: ERODING AND RETRIEVAL GROUP

Encoding and Retrieval: Group 4

Analyzing the results for AD4, it was found that information retrieved about the Drexel Collection based on the actress stereotype was not significantly different from the information retrieved based on the football player stereotype (p=O. 1342) -- Refer Table 2. Thus, in this case, the schema transfer hypothesis is not supported. Hence the results of the more detailed analysis between Groups 3&4 are not presented.

DISCUSSION

The results generally support the hypothesis of cognition transfer. They show that the schema that is transferred to the new brand influences both the processes of inference making, and encoding and retrieval. The support for the process of inference making was stronger than that for encoding and retrieval. But, overall there is empirical support for a thesis that a schema can be transferred from one object (of person) to another. The results also support the thesis that affect is transferred along with the transfer of cognition.

The concept of schema transfer has interesting implications for testimonial advertising. The main focus of the research in testimonial advertising has been on the effectiveness of different types of endorsers (experts, celebrities, etc.) in generating cognitive and affective responses towards the product. Although the results are promising, very little work has beer done in explaining "how" and "why" the endorser-product relationship impacts consumer beliefs and attitudes about the sponsored brand. Here the schema transfer hypothesis may be useful in providing a theoretical basis for this phenomenon.

REFERENCES

Allen, Mary J. & Yen, Wendy (1979), Introduction to Measurement Theory, Monterey, CA: Brooks/Cale Publishing Co.

Ashmore, Richard D. & Del Boca, Frances K. (1981), Conceptual Approaches to Stereotypes and Stereotyping. In Hamilton, David L. (Ed.), Cognitive Process in Stereotyping and Intergroup Behavior, Hillsdale. N.J.: Lawrence Erlbaum Associates

Atkin, Charles & Block, Martin (1983), Effectiveness of Celebrity Endorsers. Journal of Advertising Research, 23(1), 57-61.

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

Bransford, J.D. & Johnson, M.K. (1973), Considerations of Some Problems of Comprehension. In Chase, W.G. (Ed.) Visual Information Processing, New York: Academic Press.

Cantor, N. & Mischel, W. (1977), Traits of Prototypes: Effects on Recognition Memory. Journal of Personality and Social Psychology, 35, 38-48.

Cohen, C.E. (1981), Person Categories and Social Perceptions: Testing Some Boundaries of the Processing Effects of Prior Knowledge, Journal of Personality and Social Psychology, 40, 441-452.

Fiske, Susan T. (1982), Schema Triggered Affect: Applications to Social Perception. In Affect and Cognition, (Ed.) Clark, M.S. & Fiske, S.T., Hillsdale, NJ: Lawrence Erlbaum Associates

Gardner, Burleigh B. & Levy, Sidney J. (1955), The Product and the Brand. Harvard Business Review, March-April, 33-39.

Hastie, Reid (1981), Schematic Principles of Human Memory. In Higgins, E.T., Herman, C.P. & Zanna, Mark P. (Eds.) Social Cognition: The Ontario Symposium, Vol. 1, Hillsdale, NJ: Lawrence Erlbaum Associates.

Lessig, V.P. & Park, C.W. (1978), Promotional Perspectives of Reference Group Influence: Advertising Implications. Journal of Advertising, XX, 41-47.

Levy, Sidney (1959), Symbols for Sale. Harvard Business Review, 37, 117-124.

Loftus, E.F. & Palmer, J.C. (1974), Reconstruction of Automobile Destruction: An Example of the Interaction Between Language and Memory. Journal of Verbal Learning and Verbal Behavior, 13, 585-589.

Miller, Arthur G. (1982), Historical and Contemporary Perspectives on Stereotyping. In Miller, A.G. (Ed.) In the Eve of the Beholder: Contemporary Issues in Stereotyping, Praeger Publishing.

Moschis, George P. & Churchill, Gilbert A., Jr., (1978), Consumer Socialization: Theoretical and Empirical Analysis. Journal of Marketing Research, 15, November, 599-609.

Neisser, V. (1976), Cognition and Reality: Principles and Implications of Cognitive Psychology, San Francisco: Freeman Press

Nisbett, R.E. & Ross, L. (1979), Human Inference: Strategies and Shortcomings in Social Judgment. Englewood Cliffs, NJ: Prentice-Hall.

Pichert, J.W. & Anderson, R.C. (1977), Taking Different Perspectives on a Story, Journal of Educational Psychology, 69, 309-315.

Spiro, R.J. (1977), Remembering Information from Text: The "State of Schema" Approach. In Anderson, R.C., Spiro, R.J. & Montague, W.E. (Eds.), Schooling and the Acquisition of Knowledge, Hillsdale, NJ: Lawrence Erlbaum Associates.

Snyder, Mark (1981), On the Self-Perpetuating Nature of Social Stereotypes. In Cognitive Processes in Stereotyping and Intergroup Behavior (Ed.), Hamilton, D.L., Hillsdale, NJ: Lawrence Erlbaum Associates.

Snyder, M. & Uranowitz, S.W. (1978), Reconstructing the Past: Some Cognitive Consequences of Person Perception. Journal of Personality and Social Psychology, 36, 941-950.

Tatsuoka, M.M. (1971), Multivariate Analysis: Techniques for Educational and Psychological Research, New York: John Wiley & Sons

Taylor, Shelly E. & Crocker, Jennifer (1981), Schematic Bases of Social Information Processing. In Social Cognition: The Ontario Symposium, Vol. 1, (Eds.) Higgins, E.T., et al., Hillsdale, NJ: Lawrence Erlbaum Associates.

Taylor, Shelley E. (1981), A Categorization Approach to Stereotyping in Cognitive Processing in Stereotyping and Intergroup Behavior, (Ed.), Hamilton, D.L., Hillsdale, NJ: Lawrence Erlbaum Associates.

Ward, Scott (1974), Consumer Socialization. Journal of Consumer Research, 1(2), 1-15.

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