A New Typology of Brand Image

ABSTRACT - Keller’s 1993) definition of brand image as Aperceptions about a brand as reflected by the brand associations held in consumer memory@ is adopted in this paper. However, both his typology of brand image and Aaker’s typology (1991) are shown to have some weaknesses. In this paper I introduce a new typology of brand image that addresses the weaknesses in these typologies.


Michael Korchia (1999) ,"A New Typology of Brand Image", in E - European Advances in Consumer Research Volume 4, eds. Bernard Dubois, Tina M. Lowrey, and L. J. Shrum, Marc Vanhuele, Provo, UT : Association for Consumer Research, Pages: 147-154.

European Advances in Consumer Research Volume 4, 1999      Pages 147-154


Michael Korchia, ESSEC, France

[I would like to thank my principle advisor, Elyette Roux, and Jerry Olson for valuable discussions and comments, and Pascale Korchia for her patience and her help. An earlier version of this paper was presented at the Corporate Reputation, Image, and Competitiveness Conference, New York University, January 1997. The author can be contacted at watoowatoo@hotmail.com.]


Keller’s 1993) definition of brand image as "perceptions about a brand as reflected by the brand associations held in consumer memory" is adopted in this paper. However, both his typology of brand image and Aaker’s typology (1991) are shown to have some weaknesses. In this paper I introduce a new typology of brand image that addresses the weaknesses in these typologies.

Most researchers and practitioners agree about the importance of stressing brand image. Aaker (1991), for example, says image creates value in a variety of ways, helping consumers to process information, differentiating the brand, generating reasons to buy, giving positive feelings, and providing a basis for extensions. However, there is still a lack of agreement about the definition of brand image (Dobni & Zinkhan, 1990). Keller’s (1993: 3) definition, although it has not, to my knowledge, been tested yet, seems to be a major contribution in this domain. His definition of brand image as "perceptions about a brand as reflected by the brand associations held in consumer memory" is consistent with many other authors’ (Newman, 1957; Dichter, 1985; Aaker, 1991; Engel, Blackwell & Miniard, 1995...); A more simple definition could be "all that a consumer can possibly associate with a particular brand." Mitchell (1982: 46, for a review) considers memory an associative network model: "within this model, the nodes of the network represent concepts while the arcs are linkages (that also define relationships) between concepts." As Keller notes, these associations (Figure 1) can vary in strength (which makes information more accessible), favorability, and uniqueness (the degree to which this association is not shared with other competing brands). Sirgy (1981) thinks that these associations can be characterized by at least eleven dimensions.

Accordingly, this associative network of knowledge for a brand is the image of this brand (Engel, Blackwell & Miniard, 1995) as stored in the long-term memory, which is defined as "a subcomponent of memory which is permanent, virtually unlimited in storage capacity, and well stored" (Dacin & Mitchell, 1986). Brand associations are all the linkages that exist between a brand and the other nodes stored in memory. In measuring the structure (how information related to a brand is organized in memory) and the content (i.e. brand associations) of an associative network of knowledge that a consumer holds for a brand, the perception the consumer has of the brand is also measured.

Note that this view is consistent with the belief that knowledge is organized into packets of information (Mitchell, 1982). For example, schemata are packets of information centered around concepts. As Olson (1979: 5) notes, "the schema notion implies that memory is comprised of many highly organized structures of knowledge or schemata, and is not a vast storage 'bin’ into which all coded information is 'dumped’."


In this paper, the word "brand" is restricted to companies that sell consumer goods. Thus, organizations, charity funds and so forth do not match this definition.

Several ways to classify knowledge have been proposed, including the distinction between declarative versus procedural knowledge (Anderson, 1983). Declarative knowledge involves the facts that are known about a particular domain, whereas procedural knowledge refers to the knowledge of rules to take action. However, to study the perception of brands, the researcher, as well as the practitioner, needs a more relevant typology.

Keller (1993) considers brand image and brand awareness (brand recall and recognition) to be the two components of brand knowledge. He classifies brand associations (and therefore brand image) into three categories that fall along a continuum from concrete to abstract.

1. Attributes: Keller distinguishes between non-product-related attributes (price, packaging, user imagery, usage imagery; the last two can also produce brand personality attributes) and product-related attributes.

2. Benefits: functional (often linked to physiological needs), experiential (what it feels like to use the product), and symbolic (one example: a need for social approval or self-esteem). See Park, Jaworski, & MacInnis (1986) for more details.

3. Brand attitudes, defined as consumers’ overall evaluations of a brand.

Aaker (1991) makes another important contribution with his typology, distinguishing between 11 dimensions: product attributes, intangibles, customer benefits, price, use/ application, user, celebrity, life style, product class, competitors, and country of origin.

Aaker’s and Keller’s typologies share many common categories (see Delamotte, 1996, for a review): price, user imagery, usage imagery, and product attributes. I will show that both have some weaknesses, but first consider how it is possible to tap the content of a consumer’s knowledge...


It is widely accepted that a "spreading activation" process occurs when retrieving information stored in memory (e.g. Collins & Loftus, 1975; Keller, 1991; Mitchell, 1982). When the amount of activation at a node exceeds some threshold level, the information contained in this node is recalled. When a node is activated, the nodes most strongly linked with this node are more likely to be activated.

Olson and Muderrisoglu (1978: 7) define the free-elicitation method as "a procedure in which respondents are free to say anything and everything that comes to mind when presented with a stimulus probe cue." A very general probe attempts to activate a memory schema without inducing any particular answer. When the probe is a brand name, the objective is to activate all the nodes associated with this name in the subject’s memory. [The first statements elicited by a consumer are the most likely to be the "strongest" ones; also, the concepts that are the most repeated are probably connected to many other concepts, indicating that they are central (Michel, 1997, 1998). However, it goes beyond the scope of this paper to take a further look in this direction: our aim is to clarify the content of a brand image, not to show why people have positive/negative or strong/weak brand associations.] However, as subjects are not likely to recall everything they know about a brand in response to a single probe, a multiple elicitation procedure (Brucks, 1986; Finlay, 1996; Kanwar, Olson & Sims, 1979; Olson, 1979) may be used in which each elicited concept is itself subsequently used as a probe. This procedure continues as long as the subject is able to verbalize new concepts.



The elicitation procedure has been shown to be the best way to measure both the content and the structure of knowledge within a domain, because it is quite stable and reliable and it allows researchers to tap most of the content of a subject’s knowledge about a domain (Olson & Muderrisoglu 1978; Finlay, 1996). A study by Steenkamp and Van Trijp (1997) suggests that free elicitation is to be preferred to other attribute elicitation procedures (namely hierarchical dichotomization and repertory grid).

The interviews must be recorded and transcribed before they can be analyzed. It is possible to analyze them in a variety of ways, depending on the goal of the research. Because the objective of this article is to propose a new typology of brand image, and therefore to describe the different kind of brand associations that may exist, I will discuss only the coding of answers.

Coding schemes can be developed to classify as many statements (i.e. associations) as possible. In this case, a coding scheme defines a typology of the domain in question: Bettman and Park (1980) created a coding scheme for classifying the use of prior knowledge during decision making; Brucks (1986) developed a coding schee to classify consumer product knowledge. Brucks defined (1986: 59) three objectives that a coding scheme (and therefore the associated typology) must satisfy:

1. The typology should cover as many of the subjects’ statements as possible while remaining relatively parsimonious.

2. The coding scheme and typology should be easy to use and seem logical to people who code the answers.

3. The categories in the typology should be as distinct from one another as possible.

A new typology of brand image, taking in account Keller’s (1991) and Aaker’s (1991) contributions, and designed to satisfy Brucks’ criteria, has been developed and empirically tested.


Limitations of the existing typologies

If a typology does not satisfy any of Brucks’ three criteria (1986), it must be modified and improved. Both Keller’s and Aaker’s typologies need to be tested conceptually and empirically and subsequently be improved.

It appears that Keller’s and Aaker’s typologies:

1. Are not exhaustive: Keller’s typology doesn’t include any category for competing brands. It doesn’t seem possible to classify such a statement as "Brand X has a lot of shops everywhere in the country" (this indicates the need for a category called "retailing" or "distribution," see further the discussion about secondary associations), nor to classify any statements about the subject’s own experience (such as "I bought some brand X jeans"). Aaker’s typology doesn’t include attitudes (making a statement such as "I love brand X" difficult to code). If a typology is not sufficiently comprehensive, that is, if too many statements do not fit in any of the categories (say 10 percent or more; Brucks, 1986, reports that 98 percent of the relevant statements could be classified into one of the categories of her typology), it must be improved and/or rewritten.

2. Are too difficult for coders: The fact that some important categories are missing makes the coding task more difficult. Moreover, Keller argues that the typology does not need brand personality attributes (human attributes attributed to the brand) because these attributes arise as a result of inferences about the underlying user or usage situation. Yet, it seems impossible, when a consumer says a brand is "youthful," to determine whether this statement is motivated by the ads, the consumers, or even by his observation of the products. This point is a consequence of the former statement.

3. From a more theoretical point of view, Keller doesn’t emphasize that attitude must be viewed as an unidimensional construct (see Lutz, 1991 for a review)

Keller (1993: 11) distinguishes between brand associations and secondary associations. The latter are associations that are linked to a brand association but not directly related to the product or service. He adds that "Because the brand becomes identified with this other entity, consumers may infer that the brand shares associations with that entity, thus producing indirect or 'secondary’ links for the brand... Secondary associations may arise from primary attributes associations related to the company, the country of origin, the distribution channels, a celebrit... or an event." This means that what Keller calls secondary associations are not directly linked with the brand in memory, but that they can be, if, for example, the brand communicates in this manner. Moreover, the reasons some associations are considered secondary instead of direct are not clearly defined: for instance, why are celebrity spokespersons considered secondary associations while typical users are considered direct associations? In addition, Keller says that the distribution channels are secondary associations because "store images have associations that may be linked with the product they sell," but many associations related to stores are "direct" brand associations, particularly when the store has the same name as the brand. If a customer says that shop assistants working in a shop called "brand X" are good-looking, this is a brand association: "brand X" is directly linked with this node (Dacin & Mitchell, 1986). Customers may make inferences that create a more favorable attitude toward the brand.

Because of the possible limitations of both these typologies [If these typologies can't tap most of a consumer's knowledge about a brand, then they will fail in the study of this brand's image: for example, they neglect the importance of the retailing channels for a brand, leading to an incomplete description.] a new typology of brand image (Figure 2) is proposed and tested. If this typology can satisfy Brucks’ criterion, it can be accepted. This new typology was tested empirically against Keller’s and Aaker’s typologies (see below).

If a large number of statements are classified in categories that don’t exist in Keller’s or in Aaker’s typologies, both typologies will be shown not exhaustive enough, and that the new typology is more comprehensive than the other ones.

The typology is detailed below, and refers in part to previous works by Aaker (1991) and Keller (1993). It was conceptually built taking both Aaker’s and Keller’s typologies into account, and after discussions with experts and personal reflexion. Brands associations can be classified into 6 broad dimensions, or 15 dimensions in total:

1. The company: this category refers to knowledge of facts related to the firm: its country of origin, its strategy, its story, and so forth. Statements relative to the notoriety of the brand are included in this category, because the brand is itself a part of the company (Olins, 1989). This category doesn’t appear in previous typologies (except for the country of origin notion), even though consumers may elicit statements more related with the firm itself than with the brand.

2. Other Organizations: this includes statements referring to the competitors, comparing them with the brand of concern, to government, charity funds, and so forth.

The evoked universe:

3. Brand Personality, lifestyle: Human characteristics associated with the brand (Fournier, 1994).

4. Celebrities/ events: When advertising creates an association between a brand and a celebrity endorser, the celebrity’s associations may then become related with the brand (Rossiter & Percy, 1987). In other words, his/her expertise, attractiveness and so forth, may be shared with the brand. The same thing may happen when dealing with events instead of celebrities.

5. User imagery: brand associations about the typical user or other users. Several distinctions can be made: age, physical appearance, job...

6. Usage imagery: associations about the typical usage situation: the location, personal experience or information search.

"Attributes are those descriptive features that characterize a product -what a consumer thinks the product is or has and what is involved with its purchase or consumption... Product-related attributes are defined as the ingredients necessary for performing the product function sought by consumers. Non product-related attributes are external aspects o the product that relate to its purchase or consumption" (Keller, 1993: 4).

Non product-related attributes:

7. Product category: associations about the product category to which some of the products of the brand belong. Positioning of the brand as perceived by consumers. The level of abstraction (from "Kenzo makes some jackets," the most concrete, to "Kenzo makes ready-to-wear," the most abstract) may vary widely (Kanwar et al., 1981; Conover, 1982).

8. Price: consumers often strongly associate the price, for example, with the quality of the brand.

9. Communication: all associations, mainly about the ads and the catalogs.

10. Distribution: associations about the distribution networks, the decoration of the stores, the shop assistants.

11. Product-related attributes: they are "the ingredients necessary for performing the product function sought by the consumer. Hence, they relate to a product’s physical composition" (Keller, 1993: 4). Note that, contrarily to Keller, packaging is considered here a product-related attribute, because for many products (for example perfumes, ties, cultural goods), it is one of the necessary ingredients sought by consumers. [In this article, packaging can be seen a product-related attribute because the product category is clothes. For some other product categories, packaging may be a non product-related attribute.]

"Benefits are the personal value consumers attach to the product attributes -that is, what the consumers think the product can do for them" (Keller, 1993: 4; see also Park et al., 1986).

12. Functional benefits: Refer mainly to physiological and safety needs, as well as to desires for problem removal or problem avoidance.

13. Experiential benefits: refer to what it feels like to use the product. They are related with sensory pleasure, variety and cognitive stimulation.

14. Symbolic benefits: relate to underlying needs for social approval or personal expression and outer-directed self-esteem.

15. Attitude: a quite narrow definition of attitude will be used: "an attitude is an index of the degree to which a person likes or dislikes an object, where 'object’ is used in the generic sense to refer to any aspect of the individual’s world" (Ajzen & Fischbein, 1980: 64). Thus, the attitude is considered here unidimensional (Lutz, 1991), consistent with some recent writings (Engel, Blackwell & Miniard 1995; Machleit, Allen & Madden, 1993). Note that our interest is not the predictive power of attitude, but rather an affective feeling toward the brand.

The relationships between attributes, benefits and attitude need to be detailed: attributes are objective items that do not depend of the consumer’s point of view (a clothe in wool is in wool for everyone); however, the perceptions of these attributes may lead to different perception of benefits and attitudes. The main difference between attitude and benefits is that attitude is a global level of liking of disliking about a brand or a product while benefits refer to what the product can do for the consumer. If a person says "I like the taste of this soft drink," this statement [This statement contains actually two units to code: the first is "I like the taste of...", and the second is "(This is a) soft drink," which can be classified in the product category class.] refers to an experiential benefit (whose valence is positive), because the product can bring her/him a pleasant taste; an attitudinal statement would be "I like this soft drink" (attitude is considered unidimensional, hence related statements must be straight and simple).



The relationships between these conceptscan be viewed as causal too: some benefits arise because of some attributes (a clothe is hot because it is in wool), and the level of attitude depends on the valences associated to the attributes and benefits. A consumer will like a product because it is in wool, which is comfortable and hot, if, according to the Ajzen & Fischbein model (1980), she/he likes wool and hot and comfortable clothes, and if she/he thinks this is important for an article of clothing.

Each of these 15 categories can be subdivided into several sub-categories, to increase the depth of analysis. The "short version" (with 15 categories) of the coding scheme must be used as is for any brand and any product category. However, the detailed coding scheme [The detailed coding scheme can be found in the lengthier version of this paper.] which features 65 categories, must be adapted to the specificity of each brand and/or product category.


This study is in part a replication of Brucks’ works (1986: 60) about product class knowledge, adapted to the case of brands. Hence, the objectives of the study are:

1. To provide a test of the comprehensiveness of the typology;

2. To assess the inter-judge reliability of the coding-scheme; and,

3. To provide an estimate of the inter-correlations between the amount of knowledge in each of the categories.

As explained earlier, the multiple elicitation procedure was used. Using a probe such as "Tell me all that comes to mind when I say brand X" has been shown to be a good way to tap the content of knowledge structures. "Constructive recall" was allowed: subjects were allowed to make inferences during the interview and to verbalize them, as they were asked to verbalize the "true" knowledge stored in their memory. This runs contrary to some previous works (Dacin & Mitchell, 1986; Olson & Muderrisoglu 1978). Brucks (1986: 60), however, argues that constructive recall and stored knowledge seem to "better represent the knowledge that people actually use during decision making than stored knowledge alone."


Before selecting some brands to study, a product class had to be chosen. Ready-to-wear clothing is a product class for which there exists a wide variance of knowledge and involvement, and in which some strong brands exist. Kookan and Kenzo were chosen: both sell clothes for women (Kenzo makes clothes for men too, but the focus of this study is on women). Kookan’s target market is primarily young women between the ages of 15 and 25. Their prices are quite low, and their adverts often show attractive and impertinent young women who are supposed to be typical consumers. Kenzo sells high-priced clothes, but their target market is older (mostly between the ages of 30 and 40). Kenzo styles itself according to the world of French and Japanese "haute couture " (Kenzo is a Japanese who settled in France 30 years ago). Some pre-test interviews and interviews with experts showed that women had very clear (and sometimes very contrasted) ideas and images of these brands, which spend a lot on advertising. Working with these brands (one mass-market brand and another brand considered "haut-de-gamme" -a step below luxury), that communicate very differently, and whose clothes are very different, may be a good way to elicit some various statements while remaining in the same product category (and therefore reduce product bias).


Twenty women were interviewed, but, because of tape machine failure, two interviews were lost. Two subjects refused to be interviewed about Kenzo (because they thought they didn’t know enough about this brand). Subjects were between the ages of 19 and 39. About half of them were students, while the remaining half were employed. Most had at least a master’s degree or were enrolled in graduate programs. Two of them were married, and one had children.

Before being questioned about the two brands of interest, people were asked to practice the elicitation procedure by saying all that came to mind about a brand of their choice, that is to verbalize their knowledge about a brand. Then half the subjects were first questioned about Kenzo, then Kookan (and vice versa). The elicitation procedure lasted between 10 and 40 minutes for each brand. After the elicitation task, subjects had to answer a questionnaire about their knowledge of the brands, their attitude toward these brands, and some other questions unrelated to this paper.

Interjudge Reliability

The transcription of seven interviews about Kookan were coded by the author and a person not involved in the study. The transcription represented 40 percent of the elicited statements. Perreault and Leigh (1989, p. 146) recommend to * start the coding process on a random sample of responses using multiple judges (coders) and to evaluate the reliability of the coding process + when it is not practical that multiple judges code every statement.

Kassarjian (1977: 14) defines interjudge reliability as "the percentage of agreement between several judges processing the same communications material." The judges agreed on 81 percent of the classifications for the coding scheme involving 15 categories, which is quite satisfying; the agreement in the Brucks’ study (1986) was 72 percent, with only 8 categories. Moreover, the Perreault and Leigh’s Ir, which indicates reliability, is high at .89 (Perreault and Leigh 1989). Due to these results, the coding of the author was used for all the interviews.


Respectively 1730 statements (96.1 per subject) and 1512 (94.5 per subject) were obtained for Kookan and Kenzo. A statement repeated twice or more by a subject was only counted once. Only 25 statements for Kookan and 8 for Kenzo were judged to be irrelevant to the brand or could not be classified, demonstrating that the coding scheme and the related typology are exhaustive. The number of associations an individual has about an object ranged between 17 and 176 for Kookan and between 28 and 177 for Kenzo.

The number and the percentage of elicited statements per category (Table 1) show the composition of knowledge in memory (but not its salience). Approximately half of the statements fall into three categories: Users, product-related attributes, and communication. Most of the subjects had a very precise (and often convergent) image of the typical Kookan user, called a "kookanette." Because Kookan is one of the leaders of ready-to-wear (Le Figaro Economie, 1994) in France, and has a very large advertising budget, and because the subjects were exposed intensively to the brand (many of them belonged to the target market), subjects had a great deal of knowledge about these two latter categories in memory. Because all subjects thought Kenzo is a high quality brand, most of the statements elicited concerned product related attributes.

Note that these results are very similar for both brands. Kendall and Spearman correlation coefficients were computed. They are both very high (0.82 and 0.90), which indicates that the ranks (i.e. the number of associations elicite for each brand) between the 15 categories (i.e. the number of associations elicited) are the same for both brands.

For both brands, approximately 40 percent of elicited statements fall into some categories not mentioned as brand associations by Keller (1993), and 19 percent fall into categories he neither mentioned as brand associations nor as secondary associations (competitors, brand personality, product category, and communication).

Approximately 16 percent of elicited statements fall into some categories not mentioned by Aaker (1991): the company, communication and attitude. As some of the categories presented in this article have the same name as some categories of Aaker’s categories, but are much wider (i.e., Aaker’s definition of the category "Usage, experiences" is narrower than that used in this study), this percentage may be raised.

These empirical results show the lack of comprehensiveness of both Aaker’s (1991) and Keller’s (1993) typologies, at least for these brands (the question of validity of this study will be discussed later). Moreover, they demonstrate that the new typology is more exhaustive.



As Brucks (1986: 61) noted in her article about product class knowledge: "it may be erroneous to conclude, however, that these frequencies represent the actual amount of knowledge that people have stored in memory in each of these categories." The reasons are that:

1. Information is often "chunked" in memory: to reduce the number of salient concepts in a structure, and therefore to require less of a person’s limited cognitive capacity, "a new code is assigned to represent several other usually less abstract codes" (Kanwar et al., 1981: 123). The multiple elicitation procedures may have reduced this bias, because when an abstract code (or concept) was mentioned by a subject, this code was subsequently used as a probe, often resulting in an activation of the related (more concrete) associations.

2. Because attitude is considered an unidimensional concept, and because the number of attitudinaly different statements is limited, the number of elicited statements in this category may seem low. On another hand, a subject who has only been in a Kookan shop once may have a sharp impression of the experience and therefore elicit a lot of statements about the shop and the vendors, even though these associations are weak or confused. The number of associations elicited does not reflect the strength or the favorability of these associations.

Relationships Between the Categories.

If the levels of the correlations between the categories vary widely, subjects’ perceptions of brand image aren’t made on the same basis. Moderate correlations were expected for most of the categories, because the categories are conceptually dependent. Some high correlations are however expected, because consumers may interrelate beliefs about some of the different concepts. For example, we could expect relationships between the strategy of the company (category 1) and descriptions of its ads (category 9).

The correlations between the number of statements elicited in each category range from -0.26 (non significant) to 0.77, the mean is a moderate correlation of 0.31 (and non significant because of the small sample). For Kenzo, they range from -0.37 (non significant) to 0.74, the mean is a moderate correlation of 0.24. Additional factor analysis is needed to better interpret this set of correlations. Factor analysis is the best method to use in this situation, because the data to analyze are in a contingency table (Lebart & Salem, 1994).

It has been shown (Korchia, 1994; Lebart & Salem, 1994) that it is often easier, and more precise to make a classification based on the results of the factor analysis when the matrix to analyze is a contingency table crossing texts with semanticunits, and to interpret the classes obtained with this classification. The result of the dendrogram (not shown because of space limitations) indicates that the best partition is to use three classes for Kookan, and four classes for Kenzo. It is not surprising if we get different results for these brands because:

* No association related to celebrities or spokespersons was elicited for Kookan, while 51 were elicited for Kenzo (because some subjects knew that Kenzo is the name of the brand and the name of the creator).

* There is no reason why the 15 categories should interrelate the same way for both brands, because these brands are different.

This partition in three classes indicates, for the case of Kookan, that brand image perception is multidimensional and based on three dimensions:

* The company, brand personality, users, communication, distribution. The first dimension represents a functional and "clinical" description of Kookan.

* Product-related attributes. This dimension is strictly related to product-related attributes.

* Competitors, usage, product category, price, benefits (functional, experential, symbolic), attitude. The last dimension is related to knowledge that arises out of personal experience. It represents the non-descriptive elements that may, or may not, make a consumer buy certain Kookan products.

The results are different, yet interesting, for Kenzo (dimensions are the brand and its products, brand personality through communication, the designer Kenzo, and other categories).


The results of this research are quite similar to previous works by Brucks (1986) who made a typology of consumer knowledge content.

The new typology of brand image I proposed in this article has been shown more comprehensive and easier to code than previous ones proposed by Aaker (1991) and Keller (1993). It indicates that brand image can be classified into 15 distinct categories. In the case of Kookan, a French ready-to-wear brand, these categories can be themselves classified along three broader dimensions. For Kenzo, a * haut-de-gamme + ready-to-wear brand, four dimensions have been found. However, the external validity of this typology cannot be assessed: the case of Kookan and Kenzo, and of ready-to-wear in general, is limited in generalizability. Other limitations arise from the small sample and from the elicitation procedure (although it seems to be the best way to tap consumers’ knowledge).

Because the basis on which consumers perceive brand image have been identified and classified, it is easier to build exhaustive questionnaires designed to measure brand images. These questionnaires must contain some questions designed to tap the fifteen dimensions of brand image. Some of these questions must be open-ended, because of the qualitative nature of brand image. Knowing how they are perceived, brands can adjust their communication to change consumers’ beliefs (Park et al., 1986).

Effects of specific dimensions of brand image (for example, beliefs about product related attributes) on consumer decision-making or brand equity need also to be explored.



Aaker, D. A. (1991), Managing Brand Equity, New York: The Free Press.

Ajzen, I., Fischbein, M. (1980), Understanding Attitudes and Predicting Social Behavior, Englewood Cliffs, NJ: Prentice-Hall

Anderson, J. R.. (1987), "A Spreading Activation Theory of Memory," Journal of Verbal Learning and Verbal Behaviour, 22: 261-295.

Bettman, J. R., & Park C. W. (1980), "Implications of a Constructive View of Choice for Analysis of Protocol Data: A Coding Scheme for Elements of Choice Processes," Advances In Consumer Research, 7: 148-153.

Brucks, M. (1985), "The Effects of Product Class Knowledge on Information Search Behavior," Journal of Consumer Research, 12 (June): 1-15.

Collin, A. M., & Loftus, E. F. (1975), "A Spreading Activation Theory of Semantic Processing," Psychological Review, 82 (6): 407-428.

Conover, J. N. (1982), "Familiarity and the Structure of Consumer Knowledge," Advances In Consumer Research, 9: 494-498

Dacin, P. A, & Mitchell A. A.(1986), "The Measurement of Declarative Knowledge," Advances In Consumer Research, 13: 454-459.

Delamotte, M. (1996), "Le Capital de Marque du Point de Vue du Consommateur: Le Point Sur les Mesures," Actes du CongrFs de l’AFM, Poitiers, 681-695.

Dichter, E. (1985), "What’s in an Image," Journal of Consumer Marketing, 2 (1, Winter): 75-81.

Dobni, D., & Zinkhan, G. M. (1990), "In Search of Brand Image: a Foundation Analysis," Advances In Consumer Research, 17: 110-119.

Engel, Blackwell, & Miniard, (1995), Consumer Behavior. The Dryden Press, Harcourt Brace College Publishers.

Finlay, Karen, (1996), "Reliable and Valid Measurement of Memory Content and Structure as a Function of Brand Usage Pattern," Advances In Consumer Research, 23: 282-288.

Fournier, S. (1995), "A Brand as a Character, a Partner, and a Person: Three Perspectives on the Question of Brand Personality," Advances In Consumer Research, 22: 393-395

Kanwar, R., Olson, J. C., & Sims, L. S. (1981), "Toward Conceptualizing and Measuring Cognitive Structures," Advances In Consumer Research, 8: 122-127.

Kassarjian, H. H. (1977), "Content Analysis in Consumer Research," Journal of Consumer Research, 4 (June): 8-18.

Keller, K. L. (1993), "Conceptualizing, Measuring, and Managing Customer-Based Brand Equity," Journal of Marketing Research, 29: 1-22.

Korchia, M. (1994), Recherche d’Indicateurs Statistiques en Veille Technologique, Unpublished DEA Dissertation, Laboratoire Conception Produits Nouveaux, Ecole Nationale SupTrieure des Arts et MTtiers, Paris, France.

Le Figaro Economie. (1994), Kookan, la mode sous haute tension. May 9: 16.

Lebart L., & Salem A. (1994), Statistique Textuelle, Paris, France: Dunod.

Lutz, R. J. (1991), "The Role of Attitude Theory in Marketing," In Kassarjian & Robertson (Eds.), Perspectives in Consumer Behavior: 317-339, Glenview, IL: Scott, Foresman & Company.

Machleit, K. A., Allen, C. T., & Madden, T. J. (1993), "The Mature Brand and Brand Interest: An Alternative Consequence of Ad-Evoked Affect," Journal of Marketing, 57: 72-82.

Michel, GTraldine (1997), "L’Impact de l’Extension de Marque sur la Marque: Contribution de la ThTorie du Noyau Central a l’Analyse de la Structure Interne de la Marque," Doctoral Dissertation, IAE de Paris, UniversitT Paris I.

Michel, G`raldine (1998), "L’Extension de Marque et son Impact sur la Marque MFre," DTcisions Marketing, 13:25-35.

Mitchell, A. A. (1982), "Models of Memory: Implications for Measuring Knowledge Structures," Advances In Consumer Research, 9: 945-951.

Newman, J.W. (1957), "New Insight, New Progress, for Marketing," Harvard Business Review, November-December: 95-102.

Olins, W. (1989), Corporate Identity BMaking Business Strategy Visible Through Design, London, U.K.: Thames and Hudson.

Olson, J. C. (1979), The Importance of Cognitive Processes and Existing Structures for Understanding Food Acceptance, Working Paper 89, College of Business Administration, The Pennsylvania State University.

Olson, J. C., & Muderrisoglu, A. (1979), "The Stability of Free responses Obtained by Free Elicitation: Implications for Measuring Attribute Salience and Memory Structure," Advances In Consumer Research, 6: 269-283.

Park, C. W., & Jaworski B. J., MacInnis D. (1986), "Strategic Brand Concept-Image Management," Journal of Marketing, 50 (October): 621-635.

Sirgy, J. K. (1981), "Product Familiarity: Critical Comments on Selected Studies and Theoretical Extensions," Advances In Consumer Research, 8: 156-160.

Steenkamp, J-B. and Van Tripp, H.C.M. (1997), "Attribute Elicitation in Marketing Research: A Comparison of Three Procedures," Marketing Letters, Vol. 8, No. 2: 153-165.



Michael Korchia, ESSEC, France


E - European Advances in Consumer Research Volume 4 | 1999

Share Proceeding

Featured papers

See More


Using a Meta-Analysis to Unravel Relative Importance of Postulated Explanations for the Endowment Effect

Peter Nguyen, Ivey Business School
Xin (Shane) Wang, Western University, Canada
David J. Curry, University of Cincinnati, USA

Read More


N3. Emotion Regulation and Memory for Negative Emotion Ads

Sandra Segal, Ben Gurion University, Israel
Hila Riemer, Ben Gurion University, Israel
Shai Danziger, Tel Aviv University, Israel
Gal Sheppes, Tel Aviv University, Israel

Read More


Human or Robot? The Uncanny Valley in Consumer Robots

Noah Castelo, Columbia University, USA
Bernd Schmitt, Columbia University, USA
Miklos Sarvary, Columbia University, USA

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.