Reconceptualizing Comparative Advertising: a Framework and Theory of Effects

ABSTRACT - Although most researchers have studied comparative ads that explicitly compare two or more brands, we believe that ads need not be explicit or limited to brand-level comparisons before they are considered comparative. In this paper, we identify (1) the level of the comparison object and (2) the explicitness of the comparative claim as dimensions on which comparative ads may vary. We present a conceptual framework and develop propositions, based on the concepts of memory schemata and category knowledge, that explain how differences along these dimensions may affect inferences and claim believability. We conclude by discussing the research and public policy implications of our approach.

Citation:

Beth A. Walker (1991) ,"Reconceptualizing Comparative Advertising: a Framework and Theory of Effects", in NA - Advances in Consumer Research Volume 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT : Association for Consumer Research, Pages: 342-347.

Advances in Consumer Research Volume 18, 1991      Pages 342-347

RECONCEPTUALIZING COMPARATIVE ADVERTISING: A FRAMEWORK AND THEORY OF EFFECTS

Beth A. Walker, Arizona State University

Helen H. Anderson, University of Arizona

ABSTRACT -

Although most researchers have studied comparative ads that explicitly compare two or more brands, we believe that ads need not be explicit or limited to brand-level comparisons before they are considered comparative. In this paper, we identify (1) the level of the comparison object and (2) the explicitness of the comparative claim as dimensions on which comparative ads may vary. We present a conceptual framework and develop propositions, based on the concepts of memory schemata and category knowledge, that explain how differences along these dimensions may affect inferences and claim believability. We conclude by discussing the research and public policy implications of our approach.

INTRODUCTION

Advertisers are making increasing use of comparative advertising both in the print (Jackson, et al., 1979) and television media (Muehling and Kangun 1985). This trend is particularly interesting given the often conflicting and generally negative empirical evidence regarding its effectiveness (Johnson and Horne 1987). We believe that this apparent discrepancy may, in part, be explained by differences in the types of comparative ads that are used by practitioners and those whose effects are typically studied by advertising researchers.

The FTC defined a comparative advertisement as one which "compares alternative brands on objectively measurable attributes or price, and identifies the alternative brand by name, illustration, or other distinctive information (Federal Register 1979, p. 47328). In most studies, researchers have examined comparative ads in which explicit reference is made to one or more brands that are compared on one or more mentioned attributes. Researchers then compare the effects of "comparative advertising" to noncomparative advertising, making conclusions which are presumably generalizable to all comparative ads.

Most empirical evidence suggests that "comparative ads" are no more effective and sometimes less effective than conventional advertising (Turgeon and Barnaby 1989; Walker, et al. 1986). To get more positive results and to avoid a lawsuit that sometimes challenges explicit comparative claims, marketers seem to be resorting to other forms of comparative advertising to persuade consumers. For example, ads may use comparative adjectives without mentioning a competitor ("highest quality"), compare their brand to "other brands," or make a comparison visually without an explicit mention of it in the copy. In addition, ads often compare the sponsored brand to objects other than a competing brand (i.e. the product class in general, noncomparable product classes).

Given these variations in comparative advertising, it seems unlikely that we can make generalizable conclusions about something called "comparative advertising." Thus, we are faced with a more interesting question about how differences in comparative advertising affect information processing and resulting outcomes. In this paper, we propose two dimensions on which comparative ads may vary: (1) the degree of the comparison (implicitness/explicitness of the comparative claim) and the (2) nature of the comparison object (the concreteness/abstractness of the comparison object). We present a conceptual framework, based on the concepts of memory schemata and category knowledge, to explain how differences along these dimensions may affect inferences and claim believability. We develop propositions regarding key relationships. and conclude by discussing relevant research and public policy implications.

A TYPOLOGY OF COMPARATIVE ADVERTISEMENTS

There are probably hundreds of subtle variations in comparative advertisements that have emerged to produce more positive results. To find the dimensions that are most meaningful, we more closely examined the FTC's original definition. This definition rigidly delineates comparative ads-on two basic dimensions: (1) the nature of the comparison object (i.e. traditionally, competing brands), and (2) the explicitness of the comparative claim. We propose the nature of the comparison object and the explicitness of the comparative claim not as defining features, but as dimensions on which comparative ads may vary. Specifically, we conceptualize ads as varying in terms of the level of abstractness of the comparison object and the implicitness/explicitness of the comparative claim.

The Level of the Comparison Object

Casual observation of the types of comparative ads featured on television and in magazines reveals that advertisers often compare the sponsored product to a wide variety of other objects, not just to competing brands. For example, a "Saab Turbo" has been compared to "European Sports Cars," "Borden's Cottage Cheese" has been compared to cottage cheeses, "Skippy Peanut Butter" has been compared to other sandwich foods, and "Cadillac" has been compared to a Ming Vase and string of pearls.

A critical issue is how best to conceptualize these differences between comparison objects. Muehling and Kangun (1985) suggested that comparison objects be classified in terms of their similarity to the sponsored product, where comparison objects range from very similar products (i.e. Chevas Regal Scotch comparing itself to Johnny Walker Red) to very dissimilar products (i.e. Chevas Regal Scotch comparing itself to a Rolls Royce). We adopt a similar approach by differentiating comparison objects in terms of their level of abstraction or generality.

It is widely held that knowledge of any object, including products, is organized hierarchically in memory in terms of its level of abstraction (Lingle, Altom and Medin 1984; Mervis and Rosch 1981). Categorization theorists differentiated these levels in terms of categories of knowledge, with superordinate categories being the most abstract, followed by basic categories, and finally, subordinate categories at the most concrete or specific level. Sujan and Dekleva (1987) substantiated that product classes, product types, and brands correspond to each of these three levels, respectively, and suggest that any given product could be grouped according to any one of these levels.

However, comparison objects are often more abstract than even the product class level (Bettman and Sujan 1987). Sponsored brands are often compared to noncomparable alternatives (i.e. a camera to a computer) or even to a noncomparable non-alternative (Muehling and Kangun 1985) such as a wallet to a pair of jeans. Beginning with the most concrete level, we propose the following five levels of comparison objects: brands, product types, product classes, alternative product classes, and nonalternative product classes.

Implicitness/Explicitness of the Comparative Claim

According to the FTC, explicitness is a function of naming your competitors and specifically identifying the attributes on which the brands are compared. However, claims need not be -so explicit. For example, a comparative claim may assert that the sponsored brand is "better than the leading brand" (Jackson, et al. 1979). Jackson, et al. (1979) reported that the use of these implicit comparative claims far exceeds that of the more traditional explicit comparisons which names or showed two or more competing brands.

Other ads may simply describe the sponsored brand as the "freshest" or "highest quality" (Levy 1983; Shimp 1978; Shimp and Preston 1982; Wyckham 1987). Shimp (1978) and Wyckham (1987) studied the effects of these "indirect" comparisons, defined as ads which use a comparative adjectival form (bigger, better, etc.) to compare a brand with an implicit competitive brand on the basis of an implicit attribute (i.e. "Mennen E goes on warmer and dryer"). Both researchers found that, after exposure to an indirect comparison, individuals made comparative inferences about the sponsored brand beyond the manifest content of the incomplete statement, leading them to suggest that indirect comparisons could lead to miscomprehension and/or deception.

Given the wide variety of comparative ads, it may be an oversimplification to categorize ads as either explicit and implicit. We agree with Levy (1983), who suggested that the explicit/implicit dimension be treated as a continuum, so that we may recognize and study the varying degrees of implicit and explicit comparison.

To establish a continuum of explicitness, we have identified specific aspects of advertisements that may contribute to an overall degree of explicitness. After reviewing the ways in which explicitness has been defined in previous research, and analyzing the results of a pilot study where subjects' sorting of comparative ads in terms of their perceived explicitness, we have developed a coding scheme for ad content and structure based on two factors leading to perceived explicitness: (1) the clarity of the claim statement, and (2) the obviousness that the comparison is made. These distinctions are consistent with what Keller (1986) identified as factors of an ad critical to processing (i.e. structure and content).

The clarity of the claim statement refers to the degree of attribute support for the comparative claim. For example, a claim may not mention a specific attribute at all (e.g. "we're the best"), may mention an attribute in a more vague fashion (e.g. "number one in quality, or may feature explicitly stated, measurable attributes (e.g. Hyundai costs $2999 less than Toyota"). The number of attributes that are presented may also increase the perceived clarity of the comparative ad. As the number and specificity of comparison attributes increase, so does the clarity of comparison.

The obviousness of the claim relates to the location of the claim within the advertisement. Claims made in the headline or in the visual portion of the advertisement are more obvious than those hidden somewhere in the copy. Based on these distinctions, advertisements could be coded according to factors contributing to obviousness and clarity. This score treated as continuous, or used to divide comparative ads into distinct groups according to the explicitness/implicitness of the ad.

In the next section, we consider how the level of the comparison object and the explicitness of the comparative claim may influence consumer's inference processes.

CONCEPTUAL FRAMEWORK

The inferences that consumers make while processing advertisements is increasingly being recognized as fundamental in explaining an advertisement's effects (cf. Ford and Smith 1987; Kardes 1990). Inferences are particularly useful in distinguishing the effects of comparative and noncomparative advertisements (Sujan and Dekleva 1987), and are being used more frequently to identify deceptive advertising claims (Shimp 1978; Shimp and Preston 1982). We identified the concepts of memory schemata and category knowledge to help predict how the level of the comparison object and the explicitness of the comparative claim may affect the source (e.g. memory-based or ad-based) and type of inferences that consumers generate.

Memory Schemata and Inferential Belief Formation

A schema is defined as an associative network of interrelated meanings that represent a person's declarative knowledge about a concept (Alba and Hasher 1983). Schemas about brands and products contain previously learned knowledge about the concept plus the interrelationships between these items of knowledge, stored in an organized, logical framework (Olson 1978). For any product, consumers may have several interrelated schemas which may be organized hierarchically in memory in terms of its level of specificity or abstraction.

There are important differences in the content of the schemas at different levels of abstraction (Lingle, Altom and Medin 1984; Myers-Levy and Tybout 1989; Rosch 1975) which are likely to affect the number and type of inferences that consumers use to frame evaluations (Sujan and Dekleva 1987). Members of superordinate categories, such as product class categories, are distinguished from each other on key attributes, but they share few features. According to Sujan and Dekleva (1987), only a few inferences can be drawn at this level. At the next level are basic, or product type categories. These categories are called basic because they tend to be used most frequently to categorize both natural and social objects. Basic level categories have a greater number of shared attributes than at the superordinate level, as well as attributes that distinguish one basic level category from another. Categorization at this level also allows many more inferences to be drawn which tend to be more evaluative in nature (Sujan and Dekleva 1987). Subordinate categories are at the most concrete level which share only a few attributes in common. Subordinate categories, such as brand-level categories, are have a few more beliefs than basic level categories but the increase in the number of beliefs is small. However, beliefs and attributes at this level tend to be very concrete and specific.

Schema Activation and the Level of Comparison Object

Categorization of an object, such as the sponsored brand, depends on the schema that is activated or evoked from memory (i.e. made available for conscious processing). We propose that the level of the comparison object used in the advertisement will cue or trigger the "matching" schema in memory. For example, an ad that compares a new brand of beer to "Lowenbrau" may activate a person's schema for "Lowenbrau." Once activated, this schema controls what informational cues are selected for processing and how they are encoded in memory (e.g. the ad information is processed relevant to the stored knowledge relevant to "Lowenbrau"). That is, the information in the ad will then be processed "in light of' the activated schema. These activated beliefs will serve as the basis for framing processing and evaluations.

However, this proposition assumes that consumers have acquired schemas at different levels of abstraction for the advertised brand or product and that these relevant schemas are available in memory. This is clearly not the case for all consumers for all brands and products. If a specific schema is not available in memory, the advertisement may activate a more general or abstract schema in memory. For example, if the comparative advertisement compares a new brand of beer to "Lowenbrau," and the consumer is unfamiliar with Lowenbrau beer, the advertisement may then activate a more general schema for imported beers, or beer in general. Alternatively, the consumer could rely primarily on the information available in the advertisement, and form beliefs based on this information as opposed to that which is stored in memory (cf. Bettman and Sujan 1987).

Based on these ideas, we assert that the level of the comparison object will serve as a cue to activate the corresponding or matching schema in memory. If the schema is available in memory, the formation of inferential beliefs will primarily be a function of the contents of that schema. In other words, it will serve as a basis for framing processing and evaluations. When the corresponding schema has not been acquired by the consumer and is not available in memory, processing will be guided by either a schema at a higher level of abstraction and/or primarily by the information available in the advertisement.

If we examine the conditions under which schemas would most likely be available in memory, we may make more specific predictions about when evaluations are more likely to be framed in terms of inferences generated primarily from ad information versus inferences generated from memory, as well as for the types of memory-based inferences depending on the level of the schema that is activated (e.g. brand, product type, product class, etc.).

Categorization theory suggests that the availability of schema in memory is likely to be a function of the (1) level of the category and (2) the knowledge level of the consumer.

Category Level. Categorization theory suggests that there-is one basic level of inclusiveness at which individuals naturally categorize and spontaneously name objects. Since most individuals have acquired knowledge and communicate at the basic category level, this schema should be fairly large for most consumers. Therefore, basic level categories should be available in memory for processing advertisement information. For products, Sujan and Dekleva (1987) and Myers-Levy and Tybout (1989) found that the product type level constitutes the basic level of categorization.

Because schemas at the basic category level (product class) are generally available in memory, we propose the following:

P1A: When the comparison object is at the basic category level, processing will more likely be framed primarily in terms of inferences generated from the corresponding memory schemata than by ad-based inferences.

Also,

P1B: When the comparison object is at more specific levels (e.g. brand) or more abstract levels (e.g. non-alternative products), processing and evaluations are more likely to be framed in terms of inferences generated from the ad versus memory based inferences.

Knowledge Level of the Consumer. The second factor that should affect the availability of schemas in memory is the knowledge level of the consumer. Although in general, the basic category level is the one which is most likely to be developed and available for processing, categorization theory also suggests that as consumers' knowledge increases in a domain, so does the specificity of their knowledge structure (Mervis and Rosch 1981). Indeed, the research of Bettman and Sujan (1987) and Walker, et al. (1987) supports the notion that more knowledgeable consumers have access to more specific knowledge. Therefore, we would expect that more knowledgeable consumers would have acquired more specific category schemas (at brand and product type levels) with which to process information, and that this information may be more useful to them than more general, basic category knowledge.

However, although more and less knowledgeable consumers may differ in terms of their knowledge at the brand, product type, and product class levels, they are likely to have equal access to schemas for noncomparable alternatives and non-alternative products. Even for experts, these schemas are often unique and newly formed, and may not have ever considered the particular combination of alternatives (Bettman and Sujan 1987). Therefore, schemas for these alternatives are not likely to be cognitively related to the schema for the sponsored brand in memory. Bettman and Sujan (1987) found that experts and novices did not differ in terms of their ability to evaluate noncomparable alternatives, suggesting that they had equal access to these category structures. Based on their research, we propose the following:

P2A: For less knowledgeable consumers, processing and evaluations are more likely to be framed in terms of inferences generated primarily from the advertisement versus from memory for all levels of the comparison object except when the is at the basic category level

P2B: For more knowledgeable consumers, processing and evaluations are more likely to be framed in terms of inferences generated primarily from the advertisement versus from memory only when the level of the comparison object is very abstract (e.g.noncomparable alternatives, non alternative products). Otherwise, memory based inferences will Predominate.

P2C: Processing differences between more and less knowledgeable consumers will increase as the level of the comparison object becomes more specific.

The above propositions suggest to what extent category schemata versus advertisement information direct or control the inferences used to frame processing and evaluations. Basically, the level of the comparison object serves as a trigger to activate the corresponding schema in memory. If this schema has been acquired by the consumer, the activated schema will serve as a basis for generating inferences that frame evaluations. Although we did not offer predictions here, the specific nature of inferences generated will likely depend on the particular level of category schema that is activated (see Sujan and Dekleva-1987). If the schema is unavailable in memory, the inferences used to frame processing will be closely related to the information available in the advertisement. The availability of memory schemata is a function of the level of the category and the knowledge level of the consumer.

INFERENTIAL BELIEF FORMATION AND INFORMATION ACCEPTANCE

In addition to having differential effects on the source and types of inferences that consumers make, different types of comparative ads may also effect the believability of the comparative claim. The degree of brand comprehension depends on the degree of information acceptance by the consumer (Droge and Darmon 1987) which has been of central concern to comparative advertising researchers. Although comparative ads have been reported to be less believable than noncomparative advertising (cf. Wilson 1976; Levine 1976; Shimp and Dyer 1978; Golden 1979; Swinyard 1981; Wilson and Muderrisolglu 1979), an examination of the inferential belief formation process suggests that this may depend on the explicitness of the comparative ad.

Inferential Belief Formation

According to Olson (1978), the inferential process itself can be likened to an attributional process in which incoming encoded information from the advertisement environment is matched or fitted into the established beliefs that are stored in memory schemata. Inferences may then be drawn about other concepts not present in the immediate environment. For example, if an advertisement describes a winter coat as "down-filled," this descriptive belief will be compared to the schema for winter coats, and will enable the consumer to then infer certain other qualities of the coat (i.e. light weight, very warm).

However, what happens if the information in the ad does not match or "fit" the belief that is already stored in the existing schemata? For example, a comparative advertisement that states that "Skippy Peanut Butter has more protein than any other sandwich food," may generate the inference that "Skippy" is the best sandwich food. Represented in the consumer's activated schema for peanut butter, however, are the strongly held beliefs that peanut butter is high in fat, high in calories, and is a food to be avoided. Since the information in the advertisement may not "fit" with the schema that is activated in memory, we may expect the information in the ad to be rejected, affecting the acceptance and believability of the comparative claim.

Given these ideas, the explicitness of the comparative claim may determine when information in a comparative advertisement is more or less likely to be accepted. While processing comparative advertisements that are implicit and ambiguous, (i.e. "Mennen goes on warmer and dryer" or "Ragu beats all other brands"), it is likely that the basic level category is automatically activated. Since the implicit claim does not contain specific points with which to conflict with the general or basic level category schema, the ad will be processed naturally and automatically, with minimal conflict. Researchers have found that consumers believe their own inferences after exposure to an implicit comparative advertisement as if they were stated in the advertisement itself (Shimp 1978; Wyckham 1987).

On the other hand, more explicit comparative advertisements that name the competitor and compare the objects based on explicitly named attributes, more opportunity exists for the presence of assertions in the ad that do not fit with the beliefs already stored in the activated memory structure. This situation would lead to the generation of more negative inferences, and stimulate processing as mismatches of ad and memory information have been shown to do (cf. Sujan 1985). Thus, we assert that:

P3: As comparative advertisements become more explicit, the likelihood that the advertisement information will not fit with beliefs that are stored in the activated schemata increases.

Consequently, we are likely see a greater number of total inferences, and more negative inferences, than if the advertisement is more implicit.

Finally, the explicitness of the comparative advertisement will interact with the level of knowledge of the consumer. Because more knowledgeable consumers have more complex and well-defined schemata, they are more likely to conflict with the advertisement than less knowledgeable consumers who do not have schemata available in memory with which to compare the incoming information, whether the comparison is explicit or implicit. Consequently, we propose the following:

P4: Category knowledge will moderate the effects of the explicitness of the comparative ad. Specifically, the explicitness of the comparative claim will have a greater effect on the inferences of more versus less knowledgeable consumers.

RESEARCH IMPLICATIONS

We hope that our "reconceptualization" of comparative advertising will give a new direction to comparative research. Although understanding the effects of explicit brand to brand comparisons that has predominated research is important, there are likely to be dozens of different types of comparative ads, the effects of which are virtually unexplored. Indeed, expanding our view of comparative advertising offers an almost endless prospect of research exploring comparative ads in their numerous degrees, conditions, and outcomes.

We hope that our typology and framework will provide a useful theoretical foundation for future research on comparative advertising. The concepts of memory schemata and category knowledge provide a very useful theoretical foundation, particularly for predicting the inferences that consumers use to frame processing. Inferential beliefs are a powerful indicator of how ads are processed and may provide great insight into understanding the differences between types of comparative ads.

Finally, we believe that our conceptualization may have important implications for public policy. First, our conceptualization suggests that the FTC broaden its definition of comparative advertising, so that all forms of comparative ads be subject to their regulations. Second, it suggests that inferences generated-after processing a comparative ad may be an excellent indicator of the deceptive nature of an ad. Recent research suggests that many ads subtly deceive consumers by encouraging them to make inferences about the sponsored brand which may be misleading or completely untrue (Shimp and Preston 1982). We believe that deception through inference or implication may be especially likely to occur with implicit advertisements and ads that compare the sponsored brand to objects with which the consumer is relatively unfamiliar. In these cases, inferences are likely to be generated automatically, without much conscious awareness because they do not conflict with information stored in memory. Understanding and measuring this subtle form of deception and identifying what ads are most likely to mislead is a challenge for future research.

REFERENCES

Bettman, J. R. and M. Sujan (1987), "Effects of Framing of Evaluation of Comparable and Noncomparable Alternatives by Expert and Novice Consumers," Journal of Consumer Research, 14, 141-154.

Droge, C. and R.Y. Darmon (1987), "Associative Positioning Strategies Through Comparative Advertising: Attribute vs. Overall Similarity Approaches," Journal of Marketing Research, 24, 377-389.

Federal Register (1979), "Comparative Advertising: Issuance of a Policy Statement," 4, no. 157 (August 13), 47328-29.

Jackson, D., S. Brown and R. Harmon (1979), "Comparative Magazine Advertisements," Journal of Advertising Research, 19(Dec), 21-26.

Johnson, Michael and David Home (1987), "Subject/Referent Positioning in Comparative Advertising: A Pilot Study," in Advances in Consumer Research, vol. 14, eds. Paul Anderson and Melanie Wallendorf, Ann Arbor, MI: Association for Consumer Research, 164-167.

Keller, K. L. (1986), "Memory in Advertising: The Effect of Advertising Memory Cues and Brand Evaluations, " unpublished doctoral dissertation, Duke University, Durham, NC.

Levy, Sidney (1983), "How Comparative is Comparative Advertising?," in Richard P. Bagozzi and Alice M. Tybout, eds., Advances in Consumer Research, 381-382.

Lingle, John, M. Altom, and D. Medin (1984), "Of Cabbages and Kings: Assessing the Extendibility of Natural Object Concept Models to Social Things," in Handbook of Social Cognition, eds. Robert S. Wyer and Thomas K. Srull, Hillsdale, NJ: Lawrence Erlbaum Associates, 71 -118.

Mervis, C. B. and E. Rosch (1981), "Categorization of Natural Objects," Annual Review of Psychology, 12, 89-115.

Muehling, D. and N. Kangun (1985), 'The Multidimensionality of Comparative Advertising: Implications for the FTC," Journal of Public Policy and Marketing, 112-128.

Myers-Levy, J. and A. M. Tybout (1989), "Schema Congruity as a Basis for Product Evaluation," Journal of Consumer Research, 16, 39-54.

Olson, Jerry (1978), "Inferential Belief Formation in the Cue Utilization Process," in Advances in Consumer Research, vol. 5, H. Keith Hunt, ed., Ann Arbor, Michigan: Association for Consumer Research, pp. 706-713.

Rosch, Eleanor (1978), "Principles of Categorization," in Cognition and Categorization, eds. Eleanor Rosch and Barbara B. Lloyd, Hillsdale, NJ: Lawrence Erlbaum Associates, 2748.

Shimp, T. A. (1978), "Do Incomplete Comparisons Mislead?," Journal of Advertising Research, 18(December), 21-27.

Shimp, T. A. and Ivan Preston (1982), "Deceptive and Nondeceptive Consequences of Evaluative Advertising," Journal of Marketing, 45, 22-32.

Sujan, Mita and C. Dekleva (1987), "Product Categorization and Inference Making: Some Implications for Comparative Advertising," Journal of Consumer Research, 14, 372-378.

Turgeon, N. and D. J. Barnaby (1989), "Comparative Advertising: Two Decades of Practice and Research," in J.H. Leigh and C. R. Martin (eds), Current Issues and Research in Advertising, vol 11., Ann Arbor, MI: University of Michigan, Graduate School of Business, 4145.

Walker, Beth, John Swasy, and Arno Rethans (1986), 'The Impact of Comparative Advertising on Perception Formation in New Product Introductions," in Advances in Consumer Research, vol. 13, ed. Richard Lutz, Ann Arbor, MI: Association for Consumer Research, 121125.

Walker, Beth, R. Celsi and J. Olson (1987), "Exploring the Structural Characteristics of Consumers' Knowledge," in Advances in Consumer Research, Vol. 14, eds. M. Wallendorf and P. Anderson, Provo, UT: Association for Consumer Research, 17-21.

Wyckham, R. G. (1987), "Implied Superiority Claims," Journal of Advertising Research, February/March, 54-63.

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