Willa Family Brand Image By Diluted By an Unfavorable Brand Extension? a Brand Trial-Based Approach

Joseph W. Chang, UMIST, UK
ABSTRACT - This research examines the dilution effects of unfavorable brand extensions on family and original brand images in direct experience situations. Research results indicate that a family brand image is diluted by an unfavorable brand extension, regardless of the category similarity of brand extension. However, an original brand image is not diluted by an unfavorable brand extension, regardless of the category similarity of brand extension. The research findings suggest that the favorability, instead of the category similarity, of brand extension determines the dilution effects on the family brand image in direct experience scenario.
[ to cite ]:
Joseph W. Chang (2002) ,"Willa Family Brand Image By Diluted By an Unfavorable Brand Extension? a Brand Trial-Based Approach", in NA - Advances in Consumer Research Volume 29, eds. Susan M. Broniarczyk and Kent Nakamoto, Valdosta, GA : Association for Consumer Research, Pages: 299-304.

Advances in Consumer Research Volume 29, 2002     Pages 299-304

WILLA FAMILY BRAND IMAGE BY DILUTED BY AN UNFAVORABLE BRAND EXTENSION? A BRAND TRIAL-BASED APPROACH

Joseph W. Chang, UMIST, UK

ABSTRACT -

This research examines the dilution effects of unfavorable brand extensions on family and original brand images in direct experience situations. Research results indicate that a family brand image is diluted by an unfavorable brand extension, regardless of the category similarity of brand extension. However, an original brand image is not diluted by an unfavorable brand extension, regardless of the category similarity of brand extension. The research findings suggest that the favorability, instead of the category similarity, of brand extension determines the dilution effects on the family brand image in direct experience scenario.

INTRODUCTION

Research examining how an unsuccessful or unfavorable brand extension dilutes its parent brand has been conducted, but findings in the dilution effects [The dilution effect is defined as belief changes on a parent brand=s attributes (Loken and John, 1993) or image degradations of a parent brand (Keller and Aaker, 1992).] are discrepant. Keller and Aaker (1992) conclude that the core brand image is not affected by unsuccessful brand extensions, regardless of how the brand extensions are perceived as typical of the core brand. Loken and John’s (1993) and John, Loken, and Joiner’s (1998) research reveals that dilution effects on brand beliefs do emerge when brand extension attributes are inconsistent with the family brand, regardless of the category similarity of brand extensions. Romeo’s (1991) research indicates that negative information of brand extensions is most detrimental to evaluations of a family brand image, when extensions are in the same product category as the family brand. A most recent research done by Ahlumalia and Gurhan-Canli (2000) indicates that under higher accessibility, negative information about the extension leads to dilution of brand name. Under lower accessibility, only negative information about a close (vs. far) extension leads to dilution of brand name. The research context of dilution effects is discussed in two levels of brand image (name) and brand attribute and the results of the dilution effects rely on the information relevancy or accessibility of brand extensions, such as the sales result (low relevancy or accessibility) and detailed salient attribute information (high relevancy or accessibility). The similarity of brand extension is important only when the information relevancy or accessibility is low. However, the results of the past dilution effect research are all based on indirect product experience settings, instead of on direct experience settings. [Among consumers= direct experiences are product use from purchase, direct tests, sampling, and other evaluation behaviors. Consumers= indirect experiences include advertising exposure, personal selling presentations, exposure to displays, packages, and point-of-purchase material, word-of-mouth, etc. (Smith and Swinyard, 1983, p. 259).] Accordingly, direct product experience invokes higher order attitudes and the indirect product experience elaborates lower order attitudes (Smith and Swinyard, 1982, 1983). The attitude-behavior (A-B) consistency is high for the direct product experience and low for the indirect product experience (Fazio and Zanna, 1977; Fazio and Zanna, 1981; Fazio, Zanna, and Cooper, 1978; Regan and Fazio, 1977; Smith and Swinyard, 1982, 1983). As different product experiences invoke different order attitudes, research results in dilution effects based on direct product experience should be different from those based on indirect product experience.

Findings from the direct product experience settings are much similar to the real world, where consumers’ buying decisions are dominated by the direct product experience. Therefore, the purpose of this research is to identify the dilution effects based on direct product experience scenario.

THEORETICAL BACKGROUND AND HYPOTHESES: DIRECT PRODUCT EXPERIENCE

Direct Experience vs. Indirect Experience

Theoretically, measures of attitude (A) and behavior (B) variables should be highly correlated because attitude is believed to direct behavior. However, research articles examining the A-B relationship often indicate that the A-B correlation is low. Wicker (1969) concludes that "taken as a whole, these studies suggest that it is considerably more likely that attitudes will be unrelated or only slightly related to overt behaviors than that attitudes will be closely related to actions" (p. 65). Fazio and Zanna challenge Wicker’s conclusions and conduct a series of empirical studies to verify the A-B relationship (Fazio and Zanna, 1977; Fazio and Zanna, 1981; Fazio, Zanna, and Cooper, 1978; Regan and Fazio, 1975). They find that the A-B correlation of the experimental groups with indirect and direct experiences are .20and .54 respectively and conclude that the A-B relationship is significant when the product information is generated from the direct-experience. Therefore, the different results of A-B consistency are mainly induced by the different experimental settings of information sources from the direct- and indirect-experiences. Past brand extension research has provided respondents with semantic brand extension information of product category, brand attributes and sales results in "Consumer Report" formats to evaluate brand extensions and parent brand (e.g. Keller and Aaker, 1992; John, Loken and Joiner, 1998; Loken and John, 1993; Romeo, 1991; Ahlumalia and Gurhan-Canli, 2000). There is no direct product experience involved in the research, which might restrict the applicability of the past research findings to the circumstance of indirect product experience, such as the effects of printed advertisements. The dilution effects research based on direct and indirect experiences might result in different outcomes as that in A-B consistency research.

The unique brand trial of direct experience provides consumers with direct sensory contact with the product and, therefore, predicts attitude-purchase consistency better than indirect experience (Smith and Swinyard, 1982, 1983). The trial processing of direct experience starts with some combination of visual, tactile, olfactory, auditory, and/or taste data that are channeled to the sensory register (Kempf and Smith, 1998, p. 326). The attention to, and memory of, these stimuli should be relatively high, because:

1. Consumers are in evaluative mind-sets and motivated to remember the brand’s performance.

2. Sensory information is self-generated, so the source has maximum trustworthiness.

3. Levels of processing models hold that the degree of personal elaboration used in encoding can affect subsequent recall strongly and that direct experience should create both personal and elaborate memory codes because of its vividness (Fazio and Zanna, 1978).

Wright and Lynch (1995) explain that consumer’s attention to experiential attributes is higher during a product trial than when the same information is presented using indirect experience. Moreover, Fazio and Zanna (1981) depict that an experienced individual might hold attitudes with greater certainty than would an individual whose attitude is formed on the basis of indirect experience. They conclude that direct experience might produce an attitude that is better defined and more confidently held than an attitude formed through indirect means (Fazio and Zanna, 1981, p. 182). Therefore, Smith and Swinyard (1982) propose an intuitively appealing chain of logic: when compared to indirect experience, direct experience (1) leads to more strongly held beliefs, which (2) lead to more strongly held attitudes, which (3) culminate in higher attitude-behavior consistency. Direct experience increases the salience of inconsistent and consistent cues and elaborates more diagnostic brand related thoughts (cognitive responses) that consumers feel more confident (Smith and Swinyard, 1988) and rely more on brand evaluation and purchase decision-making. It is difficult for indirect experience such as advertising and word-of-mouth to create higher order beliefs. Direct experience, in contrast, normally generates high order beliefs since it processes directly through the senses (Smith and Swinyard, 1982, p. 84) and postulates more extreme attitudes (Smith and Swinyard, 1983, p. 258). Piecemeal-based processing and more attribute-oriented thoughts will be evoked by direct product experience than by indirect experience (Sujan, 1985). Moreover, the accessibility of memory of affects and belief through brand trial experience is higher than that through indirect experience, such as reading printed advertisement information. Thereafter, memories of direct product experience dominate the formation of brand attitude and decision-making of brand choice, which results in a significantly high attitude-behavior consistency (Smith and Swinyard, 1983).

Based on the above research findings, insignificant dilution effects on parent brand images (Keller and Aaker, 1992) might turn out to be significant if the experimental setting is changed from indirect experience to direct experience. As direct product experience provides vivid and credible product information (Fazio and Zanna, 1977; Fazio and Zanna, 1981; Fazio, Zanna, and Cooper, 1978; Regan and Fazio, 1975; Smith and Swinyard, 1982, 1983), the perceived schematic incongruity between a parent brand and a brand extension, as well as the accessibility of the negative brand extension information, is amplified and induces significant dilution effects on the parent brand. Therefore,

H1: The parent brand image is diluted significantly by an unfavorable brand extension in direct experience settings, regardless of the category similarity.

Moreover, knowledge about brands is organized in an associative network of beliefs and feelings (Anderson, 1981). Beliefs and feelings about brands are represented by nodes in the network and connected by links with various strengths. Brand extensions link directly with their parent brand and link indirectly with each other via the inter-mediator of the parent brand. An original (flagship) brand, such as Classic Coke, of a parent brand, such as Coca Cola, is a brand with well-developed association set that is "encapsulated" (John, Loken and Joiner, 1998), extremely and strongly held (Fazio and Zanna, 1978), and more resistant to change (Petty and Krosnick, 1994). The same as other brand extensions, the original brand links directly to its parent brand and indirectly to the brand extensions. Therefore, the negative effects from unfavorable brand extensions will not affect the original (core) brand directly. Thus, dilution effects on the original brand will not be observed significantly. Specifically,

H2: The original brand image is not diluted significantly by the unfavorable brand extension in direct experience settings, regardless of the category similarity.

METHOD: LABORATORY EXPERIMENT

Experimental Design

The experimental study includes sixty-five students as subjects in a 2x2 factorial design. Each experiment group consists of 15 to 20 subjects. A total of one hundred and forty-nine subjects participate in this research, including fifty-five respondents in pilot and pre-tests and thirty subjects in two control groups.

The first factor is the similarity of the brand extensions toward the parent brand of Sprite products. Two new brand extensions of Sprite orangeades and Sprite washing-up liquids [In the UK, washing-up liquids are detergents for dish-washing. The product category of washing-up liquids is selected as the dissimilar brand extension of Sprite lemonade for three reasons: 1. Sprite products and Sprite lemonades have a distinctive attribute, and are strongly associated with the concept, of lemon/lime, which is firmly linked with the product concept of hygienic cleaning products, such as hand-washing liquid, washing-up liquid, etc. 2. For the convenience of experimental sampling. 3. The products of washing-up liquids and lemonades both belong to the super-ordinate product category of LIQUID, however, they are categorized into different basic categorical levels of DETERGENT and BEVERAGE and very different sub-ordinate categorical levels of washing-up liquids and lemonades (e.g. Meyers-Levy and Tybout, 1989).] in different product categories of orangeades and washing-up liquids represent the similar and dissimilar brand extensions respectively to the parent brand category of lemonades. The second factor is the favorability of brand extensions. There are favorable and unfavorable brand extensions for the similar and dissimilar categories of orangeades and washing-up liquids respectively. Therefore, the experiments include four brand extensions in four experimental groups respectively of favorable Sprite orangeades, unfavorable Sprite orangeades, favorable Sprite washing-up liquids and unfavorable Sprite washing-up liquids. Moreover, two control groups are examined in order to clarify the possible biases of experiment treatment effects. The experiment treatments of the control group I and II are an unfavorable similar and an unfavorable dissimilar brand extensions of 7-Up products, which is an unrelated parent brand in the same product category of lemonades, respectively.

Stimuli

The selection of the parent brand

In line with the brand evaluation research of Aaker and Keller (1990) and Romeo (1991), the strong parent brand of Sprite products is selected from the lemonades product category for reasons that Sprite is a popular brand name with a favorable overall quality image and eliciting relatively specific positive associations, such as lemon/lime flavor, sparkling/fizzy/bubbly, refreshing, etc.

The package designs of the tested products

The favorable Sprite orangeades and unfavorable Sprite orangeades are products of Fanta orangeades and Asda Farm Stores orangeades respectively. The favorable Sprite washing-up liquids and unfavorable Sprite washing-up liquids are products of Fairy washing-up liquids (85p for a half litre bottle) and Asda Farm Stores washing-up liquids (15p for a one liter bottle) respectively.Four specifically designed hypothetical labels with Sprite brand names replace the package labels of the four products. The label of the favorable Sprite orange (500ml) is similar to that of Sprite lemonades, but the color is changed from green to orange. The label of the unfavorable Sprite orangeades is similar to the label of Asda Farm Stores orangeades, but the "Asda Farm Stores" logo is replaced by the "Sprite" logo. The label of the favorable Sprite washing-up liquids is similar to the Sprite lemonades (500ml), but the label size is enlarged to fit the Fairy washing-up liquids bottle (500ml) and the label color is changed from green to deep blue, which implies a hygienic product. The label of the unfavorable Sprite washing-up liquids is similar to that of Asda Farm Stores, but the "Asda Farm Stores" logo is replaced by the "Sprite" logo. All the four tested products are in 500ml bottles for the purposes of standardization and ease of comparison. The stimuli of unfavorable similar and dissimilar brand extensions of 7-Up are similar to the unfavorable similar and dissimilar brand extensions of Sprite products, despite the "Sprite" logos are replaced by the "7-Up" logos.

Independent Variables

Product favorability of brand extensions is identified by product performances and is measured by six seven-point bi-polar scales of brand attitude measures with endpoints labeled "Bad"/"Good", "Dislike"/"Like", "Unfavorable"/"Favorable", "Low quality"/ "High quality" (Kempf and Smith, 1998; MacKenzie and Lutz, 1989; Smith, 1993), "Not at all likely to try"/"Very likely to try" and "Inferior product"/ "Superior product" (Jacobson and Aaker, 1987; Keller and Aaker, 1992).

Similarity (or typicality) is discussed both on product category (basic category level) and attribute levels (sub-ordinate category level). The attribute-based similarity is partitioned into expected and perceived attribute-based similarities. The category similarity is measured by asking "In my opinion, orangeades (or washing-up liquids) is ____ to lemonades" followed by a seven-point bi-polar scale with endpoints labeled "Dissimilar to" (1) and "Similar" (7) (Chang, 2001; John, Loken and Joiner, 1998; Keller and Aaker, 1992; Loken and John, 1993; Romeo, 1991). The product attribute similarity (or typicality) is measured by asking "In my opinion, the Sprite orangeades (or washing-up liquids) are ______ to the Sprite products" followed by four seven-points bi-polar scales with endpoints labeled "Dissimilar to" (1) / "Similar to" (7), "Inconsistent with" (1) / "Consistent with" (7), "Atypical of" (1) / "Typical of" (7) and "Unrepresentative of" (1) / "Representative of" (7) (Chang, 2001; John, Loken and Joiner, 1998; Keller and Aaker, 1992; Loken and John, 1993).

Dependent Variables

Brand attitudes toward the parent and original brands are measured by the identical measures of product favorability. [While the product favorability and the brand attitudes are measured with the same measures, the product favorabiltiy indicates consumers= attitudes toward brand extensions and the brand attitudes indicate consumers= attitudes toward parent and original brands.]

Subject and Procedure

Participants, with a majority of undergraduate students aged 20 to 24, are selected among universities in Manchester, UK. In the beginning of the experiments, participants are advised that the research purpose is to learn about consumers’ responses toward new products, which will be launched by a famous manufacturer. All respondents are invited volunteers and users of lemonades, orangeades and washing-up liquids. Qualified participants are selected and filtered by screening questionnaires and allocated randomly to one of the experimental groups. Experiments are conducted in small groups with an average size of 3 persons. In order to prevent possible social influences on product evaluations, respondents are advised that there is no discussion allowed during the experiments. Each respondent is asked to taste or use a sample product of orangeades or washing-up liquids, which will be launched by a real manufacturer and, then, fills out Part I questionnaires evaluating the sample product. (For the experiment groups of washing-up liquids, participants are advised to try the sample products of washing-up liquids at home for two days and use them as usual as they used other washing-up liquids for dish-washing) They, then, taste and evaluate the real products of Sprite lemonades. After that, the participants are advised that the Sprite Company will be developing a new brand extension of Sprite orangeades or Sprite washing-up liquids and asked to indicate their expectations on the performance of the brand extension. Later on, subjects are advised that the sample products of orangeades or washing-up liquids that they tried in the beginning of the experiments are exactly the brand extension that the Sprite company will be launching. Simultaneously, the real Sprite orangeades or Sprite washing-up liquids with package designs are shown to them. They, then, re-evaluate the Sprite orangeades or Sprite washing-up liquids.

After that, subjects re-evaluate the Sprite products and the Sprite lemonades with the identical measures to verify the attitude changes on the Sprite products and the Sprite lemonades. Finally, participants answer selected personal demographics questions, such as age, gender etc., and are dismissed with thanks and rewards. Each participant receives a two-pound worth reward of Coca Cola, Sprite lemonades or Fanta orangeades in successfully completing the experiments.

For the two control groups, respondents experience the same experiment procedures and answer the similar questions with the same measures as those for the four experiment groups, despite the different settings of brand extensions. The experiment treatments of the control group I and II are unfavorable similar and dissimilar brand extensions of an unrelated parent brand of 7-Up respectively, which are expected to rule out the possible biases of the experiment treatments of unfavorable similar and dissimilar brand extensions.

RESULTS

Manipulating Check (independent variables)

Group homogeneity

The Levene’s tests of equality of error variances indicate that both the error variances of the brand attitudes of parent and original brands are equal across the six groups (F=.42, p>.05; F=2.24, p>.05), which indicates that the four experiment groups and the two control groups are homogeneous. Moreover, the brand attitudes of the parent and the original brands before and after the experiment treatments are not significantly different in the two control groups (t=-.36, p>.05; t=.73, p>.05). Also, the brand attitudes of the parent and the original brands are similar (t=1.55, p>.05) and highly correlated (r=.60, p=.00), which suggests that the parent and the original brands are perceived as a same entity before extending a brand extension. Therefore, the original brand will not affect negatively or positively on the parent brand image and the dilution effects on the parent brand image in the four experiment groups are purely induced by the experiment treatments of unfavorable brand extensions, which rules out the possible biases of experiment treatment effect.

Product favorability and similarity/typciality

Results of one-way ANOVA analyses indicate that:

1. The product attitudes of the favorable (Fanta) orangeades are significantly much better than the product attitudes of the unfavorable (Asda Farm Stores) orangeades (M=4.84 vs. 3.77; F=8.59, p<.0) and the product attitudes of the favorable (Fairy) washing-up liquids are significantly much better than the product attitudes of the unfavorable (Asda Farm Stores) washing-up liquids (M=5.20 vs. 4.09; F=14.36, p<.01).

2. The product attitudes of the favorability Sprite (Fanta) orangeades are much better than the product attitudes of the unfavorable Sprite (Asda Farm Stores) orangeades (M=4.98 vs. 3.74; F=7.78, p<.01) and the product attitudes of the favorable Sprite (Fairy) washing-up liquids are significantly much better than the product attitudes of the unfavorable Sprite (Asda Farm Stores) washing-up liquids (M=4.89 vs. 3.50; F=11.76, p<.01).

3. The orangeades category is perceived as extremely dissimilar to the washing-up liquids (M=5.97 vs. 2.14, F=146.96, p=.00).

4. The favorable Sprite (Fanta) orangeades is perceived as significantly more similar (or typical) than the unfavorable Sprite (Asda Farm Stores) orangeades to the Sprite lemonades (M=5.13 vs. 3.20; F=9.92, p<.01) and the favorable Sprite (Fairy) washing-up liquids is perceived as significantly more similar (or typical) to the Sprite products than the unfavorable Sprite (Asda Farm Stores) washing-up liquids is (M=4.86 vs. 2.98; F=12.46, p<.05).

Therefore, the independent variables of product favorability and similarity (or typicality) are properly manipulated both on basic and sub-ordnate categorical levels.

Hypotheses Testing

T-test analyses comparing attitude means before and after the experimental treatments are performed to examine the dilution effects of the favorable and unfavorable brand extensions on the parent and original brands and results are found as the followings.

Dilution effects on parent brand, Sprite products

Similar brand extension: Sprite orangeades

Analysis results indicate that the brand attitudes of Sprite products before and after the experimental treatment of favorable Sprite orangeades is not significantly different (Mean=5.30 vs. 5.48, t(1,13)=-.521, p>.05), which suggests that the brand image of Sprite products is not diluted, or enhanced, by the favorable similar brand extension of Sprite orangeades. However, the brand attitudes of Sprite products before the experimental treatment of unfavorable Sprite orangeades is significantly higher than that after the experimental treatment (Mean=5.29 vs. 4.66; t(1,14)=3.68, p<.05), which suggests that the parent brand image of Sprite products is diluted by the unfavorable similar brand extension of Sprite orangeades.

Dissimilar brand extension: Sprite washing-up liquids

Analysis results indicate that the brand attitude of Sprite products before and after the experimental treatment of the favorable Sprite washing-up liquids is not significantly different (Mean=4.99 vs. 5.19, t(1,14)=-1.25, p>.05), which suggests that the brand image of Sprite products is not diluted, or enhanced, by the favorable dissimilar brand extension of Sprite washing-up liquids. However, the brand attitudes of Sprite products before the experimental treatment of unfavorable Sprite orangeades is significantly higher than that after the experimental treatment (Mean=5.33 vs. 4.78; t(1,19)=2.94, p<.05), which suggests that the parent brand image of Sprite products is diluted by the unfavorable dissimilar brand extension of Sprite washing-up liquids.

Dilution effects on original brand: Sprite lemonades

Similar brand extension: Sprite orangeades

Analysis results indicate that the brand attitude of Sprite lemonades before the experimental treatment of the favorable similar brand extension of Sprite orangeades is significantly lower than that after the experimental treatment (Mean=5.32 vs. 5.71, t(1,14)=-2.65, p<.05), which suggests that the brand image of Sprite lemonades is not diluted, but enhanced, by the favorable Sprite orangeades. However, the brand attitudes of Sprite lemonades before and after the experimental treatment of unfavorable Sprite orangeades is not significantly different (Mean=5.29 vs. 5.43; t(1,14)=-1.00, p>.05), which suggests that the original brand image of Sprite lemonades is not diluted, or enhanced, by the unfavorable similar brand extension of Sprite orangeades.

Dissimilar brand extension: Sprite washing-up liquids

Analysis results indicate that both the brand attitude of Sprite lemonades before and after the experimental treatment of favorable and unfavorable dissimilar brand extensions of Sprite washing-up liquids are not significantly different (Mean=4.97 vs. 5.23, t(1,12)=-1.82, p>.05; Mean=5.58 vs. 5.54 t(1,19)=.40, p>.05), which suggests that the original brand image of Sprite lemonades is not diluted, or enhanced, by the favorable and unfavorable dissimilar brand extension of Sprite orangeades.

As expected the brand image of the parent brand of Sprite products is diluted both by the unfavorable Sprite orangeades and the unfavorable Sprite washing-up liquids (H1). However, the unfavorable Sprite orangeades or the unfavorable Sprite washing-up liquids (H2) does not dilute the brand image of the original brand of Sprite lemonades. Therefore, both hypotheses of H1 and H2 are supported.

IMPLICATIONS AND LIMITATIONS

Theoretical Implications

Research results indicate that the strong parent brand image of Sprite products is diluted by the unfavorable brand extensions of Sprite orangeades and Sprite washing-up liquids, regardless of the category similarities of brand extensions. However, the core (or original) brand of Sprte lemonades is not diluted by unfavorable brand extensions, regardless of the similarities of the brand extensions. The research findings partially match the research findings of Keller and Aaker (1992) and Romeo (1991) and are of good match with the research finding of Loken and John (1993) and Ahluwalia and Gurhan-Canli (2000). Moreover, the brand image of the original brand, such as the Sprite lemonades, is not diluted by unfavorable brand extensions, regardless of the category similarities of brand extensions. The result parallels the research findings of Keller and Aaker (1992) and is similar to those of John, Loken and Joiner (1998). In conclusion, regardless of the category similarity, a family brand image will be diluted whenever there is an unfavorable brand extension being launched into the markets, where consumers’ buying decisions are dominated by direct product experiences. However, the original (or core) brand image will not be diluted by an unfavorable brand extension, regardless of the attribute or category similarity of the brand extension.

TABLE 1

SYNTHETIC T-TEST RESULTS OF THE RECIPROCAL EFFECTS OF EXPERIMENTAL TREATMENTS ON THE BRAND IMAGES OF SPRITE PRODUCTS AND SPRITE LEMONADES

Managerial Implications

The cannibalizing effect from the dissimilar brand extension actually exists as the unsuccessful similar brand extension does and should be identified by the attribute favorability of brand extension, instead of simply by the category similarity. Therefore, whenever deciding to develop new brand extensions, marketers should control product quality by all means and make every effort in promotions to secure the successes of the new extensions, regardless of how similar the brand extensions are to the parent brand.

Moreover, as the original (or core) brand image is not diluted by an unfavorable brand extension, brand leveraging for established core brand should be encouraged, because brand extending helps market penetration for profit-maximizing and the threats to the original brand image and attribute beliefs, which are induced by unsuccessful brand extensions, are extremely low. But, as the threat of dilution effect to the parent brand is concerned, brand extending should be well controlled and prepared to secure the success of brand extensions.

Limitations and Future Research

This research manipulates the two independent variables of favorability and similarity and the selected parent brand is a strong brand with strong brand image. However, the brand extending activities are not restricted to strong brands. Some brands with fair or weak brand images are still engaging in their brand leveraging, such as Asda Farm Stores brands, Cresta lemonades, etc. The parent brand image of lower quality brands might be much more difficult to be diluted, because of consumers’ low performance expectations on the brand extensions of the lower quality brand. Moreover, the lemonades are low-involvement frequent-purchased goods, therefore the findings of this research are applicable for the products with similar features as the tested products in this research. As consumers spend much time and do more effort in their purchase decision-making, they will be more loyal to the high-involvement products they selected and were satisfied with. The brand attitudes of the parent brand shoud be more difficult to be diluted by unfavorable brand extensions. Therefore, the applicability of the research findings for high-involvement products still needs to be explored. Also, the negative reciprocal effects of the unfavorable brand extensions on the parent brand in this research are based on the parent brand (Sprite) that has not ever been extended. In this research, the dilution effect on the parent brand can be more specifically explained as the dilution effect of the first unfavorable brand extension on the brand image of a frequently purchased strong parent brand. Accordingly, as the sequential dilution effects of brand extensions on a parent brand image are observed to be diminishing in past research with indirect product experience settings (Keller and Aaker, 1992), the negative reciprocal effects of the second and so forth brand extensions on the parent brand in direct product experience settings can also be diminishing. Moreover, the respondents of this research are those who have positive attitudes toward the Sprite products. The consumer loyalty is not manipulated as an independent variable. As consumer loyalty is partitioned into cognitive, affective, conative and action loyalties with different corresponding vulnerabilities (Oliver, 1999), dilution effects of unfavorable brand extensions on family brand images may be different among different consumers with different brand loyalty. Therefore, further research is needed to verify the dilution effects among consumers with different loyalty phases.

REFERENCES

Aaker, David A. and Kevin Lane Keller (1990), "Consumer Evaluations of Brand Extensions," Journal of Marketing, 54 (January), 27-41.

Ahlumalia, Rohini and Zeynep Gurhan-Canli (2000), "The Effects of Extensions on the Family Brand Name: An Accessibility-Diagnosticity Perspective," Journal of Consumer Research, 27(December), 371-381.

Anderson, Norman H. (1981), Foundations of Information Integration Theory. New York: Academic Press.

Chang, Joseph W. (2001), "The Determination Roles of Category and Attribute Factors on the Reciprocal Effects of Brand Extensions," European Advances in Consumer Research, forthcoming.

Fazio, Russell H. and Mark P. Zanna(1977), "On the Predictive Validity of Attitudes: The Roles of Direct Experience and Confidence," Journal of Personality, 46, 228-243.

Fazio, Russell H. and Mark P. Zanna(1978), "Attitudinal Qualities Relating to the Strength of the Strength of the Attitude-Behavior Relationship," Journal of Experimental Social Psychology, 14, 398-408.

Fazio, Russell H. and Mark P. Zanna (1981), "Direct Experience and Attitude-Behavior Consistency," in Advances in Experimental Social Psychology, Vol. 14, L. Berkowitz, ed. New York: Academic Press, 14, 161-202.

Fazio, Russell H., Mark P. Zanna, and J. Cooper (1978), "Direct Experience and Attitude-Behavior Consistency: An Information Processing Analysis," Personality and Social Psychology Bulletin, 4, 48-52.

Jacobson, Robert and David A. Aaker (1987), "The Strategic Role of Product Quality," Journal of Marketing, 51 (October), 31-44.

John, Deborah Roedder, Barbara Loken, and Christopher Joiner (1998), "The Negative Impact of Brand Extensions: Can You Dilute Flagship Products?" Journal of Marketing, 62 (January), 19-32.

Keller, Kevin L. and David A. Aaker (199), "The Effects of Sequential Introduction of Brand Extensions," Journal of Marketing Research, 29 (February), 35-50.

Kempf, Deanna S. and Robert E. Smith (1998), "Consumer Processing of Product Trial and the Influence of Prior Advertising: A Structural Modeling Approach," Journal of Marketing Research, 35 (August), 325-338.

Loken, Barbara and Deborah Roedder John (1993), "Diluting Brand Beliefs: When Do Brand Extensions Have a Negative Impact?" Journal of Marketing, 57 (July), 71-84.

MacKenzie, Scott B. and Richard J. Lutz (1989), "An Empirical Examination of the Structural Antecedents of Attitude Toward the Ad in an Advertising Pre-testing Context," Journal of Marketing, 53 (April), 48-65.

Meyers-Levy, Joan and Alice M. Tybout (1989), "Schema Congruity as a Basis for Product Evaluation," Journal of Consumer Research, 16 (June), 39-54.

Oliver, Richard L. (1999), "Whence Consumer Loyalty?" Journal of Marketing, 63 (special issue), 33-44.

Petty, Richard C. and Jon A. Krosnick (1994), Attitude Strength: Antecedents and Consequences, Hillsdale, NJ: Lawrence Erlbaum & Associates.

Regan, D. T. and R. H. Fazio (1975), "On the Consistency Between Attitudes and Behavior Look to the Method of Attitude Formation," Journal of Experimental Social Psychology, 32, 850-856.

Romeo, Jean B. (1991), "The Effect of Negative Information on the Evaluation of Brand Extensions and the Family Brand," in Advances in Consumer Research, Vol. 18, Rebecca H. Holman and Michael R. Solomon, eds. Provo, UT: Association for Consumer Research, 399-406.

Smith, Robert E.(1993), "Integrating Information from Advertising and Trial: Processes and Effects on Consumer Response to Product Information," Journal of Marketing Research, 30 (May), 204-219.

Smith, Robert E. and William R. Swinyard(1982), "Information Response Models: An Integrated Approach," Journal of Marketing, 46 (Winter), 81-93.

Smith, Robert E. and William R. Swinyard (1983), "Attitude-Behavior Consistency: The Impact of Product Trial versus Advertising," Journal of Marketing Research, 20 (August), 257-267.

Smith, Robert E. and William R. Swinyard (1988), "Cognitive Response to Advertising and Trial: Belief Strength, Belief Confidence and Product Curiosity," Journal of Advertising, 17 (3), 3-14.

Sujan, Mita (1985), "Consumer Knowledge: Effects on Evaluation Strategies Mediating Consumer Judgements," Journal of Consumer Research, 12 (June), 31-46.

Wicker, A. W. (1969), "Attitudes vs. Action: The Relationship of Verbal and Overt Behavioral Responses to Attitude Objects," Journal of Social Issues, 25 (4), 41-78.

Wright, Alice Ann and John G. Lynch Jr. (1995), "Communication Effects of Advertising Versus Direct Experience when Both Search and Experience Attributes Are Present," Journal of Consumer Research, 21 (March), 708-718.

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