The risk of transparency in a brand portfolio with contradictory positioned brands

T

HE RISK OF TRANSPARENCY IN A BRAND

PORTFOLIO WITH CONTRADICTORY POSITIONED

BRANDS

By

Maureen Kanter

University of Amsterdam, faculty Economics and Business Business Administration track Marketing Date: August 2015

Student number: 10890211

maureenkanter@gmail.com

+31 6 305 842 97

S

O

This document is written by Maureen Kanter. I declare to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

T

ABLE OF CONTENTS

LIST OF FIGURES AND TABLES ... IV FIGURES ... IV TABLES ... IV ABSTRACT ... VI 1 INTRODUCTION ... 1 1.1BACKGROUND ... 1 1.2 PROBLEM STATEMENT ... 3

1.3 RESEARCH GAP AND RESEARCH QUESTION ... 5

1.4 EXPECTED CONTRIBUTION ... 6

1.5 OUTLINE ... 7

2 THEORETICAL FRAMEWORK ... 9

2.1 BRAND IMAGE ... 9

2.2 NEGATIVE SPILLOVER EFFECTS ... 10

2.3 BRAND ENDORSERS ... 13

2.4INCREASING TRANSPARENCY ... 15

2.5 HYPOTHESES DEVELOPMENT ... 16

2.6 CONCEPTUAL FRAMEWORK ... 19

3 METHODOLOGY ... 20

3.1 PRE-TEST 1: BRAND SELECTION ... 20

3.2 PRE-TEST 2: TESTING SCENARIOS ... 22

3.3 EXPERIMENTAL DESIGN ... 25

4.1 DATA PREPARATION ... 29 4.2 HYPOTHESES TESTING ... 33 4.3 ADDITIONAL ANALYSES ... 41 4.4 OVERVIEW OF HYPOTHESES ... 44 5 DISCUSSION ... 45 5.1 GENERAL DISCUSSION ... 45 5.2 IMPLICATIONS ... 49 6 CONCLUSION ... 51

6.1 LIMITATIONS AND SUGGESTIONS FOR FUTURE RESEARCH ... 52

REFERENCES ... 54

APPENDICES ... 60

L

F

Figure 1. Aaker and Joachimsthaler’s Brand Relationship Spectrum (2000, p. 9) 14

Figure 2. Conceptual framework 19

Figure 3. Logos of step-up brand Express and step-down brand Tenor 24

Figure 4. Core brand image of control group and step-up extension scenarios 35

Figure 5. Core brand image of control group and step-down extension scenarios 36

Figure 6. Core brand image of step-up and down extension scenarios 37

Figure 7. Core brand image of control group and token endorsement scenarios 38

Figure 8. Core brand image of control group and shadow endorsement scenarios 40

Figure 9. Core brand image of token and shadow endorsement scenarios 41

T

Table 1. Means of constructs measured for brands in pre-test 1 21

Table 2. Means of average price/quality for extension scenarios 23

Table 3. Means of neutrality and believability for endorsement scenarios 24

Table 4. Reliability scores of constructs 30

Table 5. Mean scores of quality brands 33

Table 7. Means of indep. constructs trust: Manipulated groups vs. core brand only 41

Table 8. Means of indep. constructs trust: Manipulated groups separate 42

A

Organizations dealing with contradictory positioned brands often follow a brand portfolio strategy that includes individual brands. In this case these brands do not suffer from negative spillover effects. However, consumers are demanding for more transparency at organizations. Therefore, this study tries to find out what the impact is on the brand image of the core brand when transparency increases in this particular brand portfolio. Besides, this study investigates which communication source has the least negative effect on the core brand when the link between the contradictory positioned brands is revealed. An experimental design was setup to test the hypotheses. 159 participants filled out the online distributed questionnaire in order to gather data. Outcomes suggest that the core brand gets negatively impacted when the link with contradictory positioned brands is revealed. However, no significant difference is found between different communication sources. Findings are discussed and implications and suggestions for further research are given.

1

I

For a long time brand extensions, both horizontal and vertical, were recognised as an important and popular strategy for introducing new products to the marketplace (Czellar 2003; Kim & Lavack, 1996). A particular extension type is called vertical extension at maximal distance. Here, a new brand is introduced within the same product category as the core brand, but at a different price or quality point (Keller & Aaker, 1992; Sullivan, 1990). Organizations apply this extension type to eliminate potential negative effects on the core brand (Kim & Lavack, 1996; Kim, Lavack & Smith, 2001). However, customers demand for more openness and transparency at organizations (Dan, 2013). When organisations dealing with vertical brand extensions at maximal distance want to react to this demand, the negative effects they first dodged can become a reality. So question rises: Would the before mentioned extension type become outdated in the near future?

1.1 B

1.1.1BRAND EXTENSIONS

In essence brand extensions involve the application of an established brand name to a new product in a category that is either related or unrelated in order to capitalize on the equity of the core brand name and to capture new market segments (Aaker & Keller, 1990; DeGraba & Sullivan, 1995; Kerin, Kalyanaram, & Howard, 1996; Pitta & Katsanis 1995). Using the established brand name creates consumer acceptance for the new product by associating the new product with the already established core brand (Erdem & Swait, 1998; Moorman 1998). This is called spillover effect and significantly reduces failure rates, marketing research and advertising costs involved in launching the new brand. When the new brand becomes a success reciprocal benefits can arise, which can on their turn lead to an enhanced equity of the core brand (Chen & Liu, 2004). These reciprocal benefits are mainly the reason why

horizontal brand extensions are still popular today. However, the reciprocal benefits do not necessarily occur when the brand extension is vertical. As already said, introducing a new brand in the same product category but at a different price or quality point is called a vertical brand extension (Keller & Aaker, 1992; Sullivan, 1990). Research indicated that the introduction of a vertical brand extension has a negative impact on consumer evaluation of the core brand (Dacin & Smith 1994; Kim et al., 2001; Loken & Roedder John, 1993; Ries & Trout, 2000; Salinas & Pérez &, 2009; Thorbjonson, 2005). The reasons for this negative effect are the inconsistent associations between the brand extension and the core brand hold by the consumer (Loken & Roedder John, 1993; Zimmer & Bhat, 2004).

Research has found that the before mentioned negative effects can be reduced by introducing a second brand name alongside the core brand name (Kim et al., 2001). This increases the perceived distance between the core brand and the brand extension, but still demonstrates the link between the brand extension and the core brand name. A common used example in literature is Courtyard Inn by Marriott. When the core brand name is not mentioned in the name of the vertical brand extension, the vertical brand extension is put as far as possible from the core brand. This is called maximal distancing and is for example common practice in the automobile industry (Kim & Lavack, 1996; Kim et al., 2001). Honda, Nissan and Toyota all introduced an entirely new brand, respectively under the names Acura, Infiniti and Lexus. When this strategy is used to introduce a new product, consumers are not aware the relationship between the new brand and another (core) brand exists. One of the advantages of creating a completely new brand name is that the core brand will not be diluted, because consumers will not directly compare the two brands (Kim & Lavack, 1996; Kim et al., 2001). At the other hand, any positive spillover effects that may arise from the potential success of the vertical extended brand do not exist once this strategy is chosen. The

main reason why companies do use this strategy is to serve new target groups without offending the existing target group (Aaker & Joachimsthaler, 2000).

1.1.2 TRANSPARENCY TREND

The concept transparency is often defined in terms of common-sense understandings like openness, insight or clarity, but rarely entails a more precise definition (Christensen & Cornelissen, 2015). For example, when is spoken of organizational transparency one usually refers of the demand for more openness and accessibility in an organization. According to Oliver, who did define transparency precisely, transparency is letting the truth be available for others to see if they choose, or perhaps think to look, or have the time, means, and skills to look (2004, p.3). Fact is that with or without the precise definition of transparency, transparency is revolutionizing every aspect of the economy and its industries and is forcing companies to rethink their fundamental values (Tapscott & Ticoll, 2012). Hence, transparency is a growing concern for al kinds of companies (Tapscott & Ticoll, 2012; Florini, 2007). The reason why transparency is of growing concern for companies, is because it creates trust and respect in the eyes of customers. Besides, transparency is one of the major conditions in establishing positive relationships between customers and companies (Reynolds & Yuthas, 2008; Silver, 2005). Another trend that exacerbates the demand for transparency even more is consumers who are more educated and have higher information needs compared to consumers two decades ago (Labrecque, vor dem Esche, Mathwick & Hofacker, 2013). Hence, it is becoming harder for companies to conceal information from consumers (Peppers & Rogers, 2012).

1.2

P

One can argue that most types of brand extensions are transparent, since the name of the core brand is somehow included in the name of the brand extension. In other words, the customer

gets informed about the connectedness of the brands, which gives the customer clarity. However, this does not apply for vertical brand extensions that placed at maximal distance of the core brand, since the core brand name is not included in the communication of the vertical brand extension. In the eyes of the customer the two brands are independent. In order words, the customer does not see any relationship between the vertical brand at maximal distance and the core brand. If brand transparency increases for this particular introducing strategy, consumers will become aware of the relatedness of these two brands and both brands will be at risk (Kim & Lavack, 1996; Kim et al., 2001). Consumer awareness of the relationship between both brands will have an impact on their brand evaluation, since the consumer will hold inconsistent associations towards the brands (Zimmer & Bhat, 2004). So, it is risky business if companies that deal with this particular introduction strategy want to react on their customers demand for more transparency. However, when these companies decide not to inform customers about the relation of the brands themselves, this does not automatically mean that customers will not become aware of the relation between both brands.

As an integral part of the modern diffusion of information nowadays, consumers do not only have to rely on the information communicated by companies. They can also find out themselves that there is a relationship between the two brands. Consumers only have to click several times on a website of a brand extension at maximal distance, to know the overarching holding of that particular brand. Once they know this information, it is easy for consumers to make the connection between the brands. The foregoing situation is not preferable for both disposed brands, but the information that reaches the customer is still controlled by the company. However, if consumers can find this information easily, the media and competitors can as well. If there is a reason to use this information by an external party in one way or another, the information is not controlled anymore by the company itself. This non-advertising journalism has an enormous effect on the meaning of the brands (Ind, 2003).

Therefore, not communicating the relatedness between brands could also be a risk for companies dealing with vertical brand extensions at maximal distance.

Concluding from the foregoing, when assuming that in the near future consumers will anyhow find out a core brand is related to a vertical brand at maximal distance, these brands are at risk. This leaves these companies with a huge dilemma. Should these companies communicate the relatedness of core brand and the vertical extended brand themselves or not?

1.3

R

To my best knowledge previous research has been focussed on brand evaluations of vertical brand extensions at maximal distance and the core brand, but has not been focussed yet on the impact on the core brand when consumers become aware of the relationship between a vertical brand extension at maximal distance and a core brand. One reason why this impact is not researched yet might be due to the fact that it was not likely this information would be revealed to the mass customers in the past. However, due to the age of transparency consumers become more educated and empowered, because more information is available for consumers (Carter & Curry, 2010). Besides, consumers begin to value transparency more (Buell & Norton, 2011).Therefore, this thesis will set its focus on increasing transparency for a brand architecture with independent brands. This will be done by showing customers a link between the vertical brand at maximal distance and the core brand. This link can be created by the company itself, but can also be created by a third party. Either way, this leads to a change in brand architecture. This study will find out what the impact is of both ways of showing the customer the relationship between the contradictory positioned brands on the brand image of the core brand. Therefore this study will elaborate the following research question:

“To what extent does a shift in brand architecture influence the brand image of the core brand?”

Before the research question can be addressed specifically, it is important to know what already is investigated so far. Therefore, the following sub questions need to be answered:

- What is brand image and why is it important? - What are spillover effects?

- Which types of vertical brand extensions do exist? And what is their impact on the brand image of the core brand?

- Which forms of brand architectures do exist? And what is the difference between a brand architecture with independent brands and brand architecture with endorsed brands?

- What is transparency and how does this influence customer’s opinion about a brand? And does the communication source impact the customer’s opinion?

1.4

E

This study is expected to contribute to the brand extension literature as well as to the brand architecture literature by investigating the effects of revealing the relationship between a core brand and a brand extension at maximal distance. The revelation of the relationship between both brands causes a shift in brand architecture. It is expected to provide a deeper understanding of the impact on the brand image of the core brand when different parties communicate the information.

This research is considered important, because it responds to the trend of the age of transparency. Researchers already revealed that consumers value transparency towards the company’s operations and price (Buell & Norton, 2011; Carter & Curry, 2010). However,

effects of transparency on consumer behaviour are not conclusive yet. Therefore, this research will also contribute to the literature of transparency by offering insights of the consequences of more empowered consumers. It is important to understand what the effect of transparency is on brand image, because consumers demand for more transparency. Besides, it is important to understand whether there is a difference in the core brand image if the link between the core brand and the vertical extended brand is communicated by the company itself or by a third party. When one way of revealing information will impact brand image more negatively, conclusions can be drawn and managerial recommendations can be made.

The insights provided by the study will also be useful for marketing practitioners. If results show changes in brand image, companies should review their extension strategy as well as their brand architecture to see whether the advantages of the vertical brand extension at maximal distance outweigh disadvantages. If not, it could be suggested that a different brand strategy should be considered, because it is the brand image that helps to decide about targeting specific market segments and positioning a product (Lee, James & Kim, 2014).

1.5

O

This thesis consists of five sections: Introduction, Theoretical Framework, Methodology, Results and Analysis and Discussion. The first section, the Introduction, outlines the background of the research as well as the research problem and the research question; clarifies the importance and relevance of the topic and delimits the scope of this study. In the Theoretical Framework, based on existing literature the concepts of the research question will be explained and the sub questions will be answered. At the end of the section hypotheses will be drawn based on the reviewed literature. The section Methodology details the research design and the development of measurement for each variable employed for the study. Furthermore it provides a description of the questionnaire and sampling design as well as the

procedure for the data collection. At last the technique used for analysing hypotheses testing is presented. When results are gathered they will be displayed and analysed in the section Results and Analysis. In the last section, Discussion, results derived from the previous sector are discussed and conclusions are drawn. This section provides a detailed answer to the research question outlined in this study. In addition to that, the theoretical and managerial implications, limitations and areas for future research will be presented.

2

T

2.1

B

Drucker and Maciariello (2008) pointed out that branding is the aim of marketing. The stronger the brand, the better it is possible to sell other products for which the consumer does not have to be persuaded with marketing strategies. In this light, brand equity is an important concept (Aaker, 1991; Keller, 1993; Simon & Sullivan, 1993; Shankar, Azar, & Fuller, 2008). Brand equity is defined as customers’ perceived added value, associated with a particular product that is accrued by a brand beyond the functional value of the product (Aaker, 1991; Simon & Sullivan, 1993; Keller, 1993).A company has to have a brand image to establish brand equity, because brand image is a key driver of brand equity (Biel, 1992). Brand image makes it possible to differentiate one brand from another. It is odd that despite all the scientific attention for this concept, there is no clear consensus of the definition of brand image and how to measure it (Lee et al., 2014). Therefore Lee, James and Kim (2014) came up with a consensus definition. According to them brand image is “the sum of a customer’s perceptions about a brand generated by the interaction of the cognitive, affective, and evaluative processes in a customer’s mind” (2014, p.8). In other words brand image is what the consumer thinks of when a brand name is mentioned. However, there are no tools yet to measure this definition effectively. The definition that comes closest to the foregoing definition and of which measurement tools are present is the one of Keller. Keller describes brand image as “perceptions about a brand as reflected by the brand associations held in consumer memory” (1993, p.3). Hence, this is the definition of brand image used for this thesis.

Companies attempt to create positive brand images, because these images generate positive attitudes towards the brand. And at their turn, the favourable attitudes increase the

likelihood of a high level of brand equity. In order words, due to brand image a strong brand equity can be established, which will make selling superfluous and the brand itself desirable (Biel, 1992). Thus, brand image forms the basis for making better strategic marketing decisions about targeting specific market segments and positioning a product (Lee et al., 2014).

2.2

N

Brand leverage is defined as capitalizing on pre-established brand knowledge in the mind of the customer (Keller, 2002). It occurs when associations of one brand transfer to another brand. In order words, brand leverage happens when two brands are related to each other. Leverage through a brand extension can be based on fit or on emerging attributes and cognitive elaboration (Bristol, 2002; Aaker & Keller, 1990; Keller & Aaker, 1992). When a core brand extents a product through its own equity it is called corporate brand leverage. This means the established reputation of the core brand is capitalized through a process of image transfer (James, 2005; Riezebos, 2003). Brand leverage can be used to extent the values of the core brand to reach new market segments (Blackett & Boad, 1999).This is what usually occurs with vertical brand extensions. However, there is no brand leverage at all for vertical brand extensions at maximal distance since consumers do not see a connection with the core brand. In this case, the brand leverage only exists for those consumers who are aware of the connection between the core brand and the brand extension. Once this linkage is created it makes the brands vulnerable to negative spillover (Lei, Dawar, & Lemmin, 2008). For the means of this study it is important to understand this negative spillover effect.

Consumers who are aware of a connection between the core brand and the vertical brand extension at maximal distance will evaluate both brands on their perception of fit (Czellar, 2003). In other words the brand image of both brands depend on the reciprocal

transfer effects between the extended brand and the core brand (Martínez Salinas & Pina Pérez, 2009). Since there is no or little fit between the core brand and the vertical brand extension, the impact on the core brand image is negative (Dacin & Smith, 1994).

This effect is explained by the categorization theory. This theory explains that brands are conceptualized as cognitive categories in the mind of a consumer. According to this theory, new products are mainly evaluated based on the overall attitude of the category to which the new product belongs (Rosch & Mervis, 1975). Consumers associate a brand with a specific set of attributes, which create a particular brand image (Boush & Loken, 1991). Their brand attitude is influenced by the perceived consistency between the core brand and the brand extension (Park, McCarthy & Milberg, 1993). The categorization theory explains that a core brand gets diluted when the set of attributes of the core brand is inconsistent with set of attributes of the extended brand (Meyers-Levy & Tybout, 1989; Sujan & Dekleva, 1987). Brand dilution refers to the negative feedback effects on core brand beliefs and attitudes (Magnoni & Roux, 2012). Brand dilution of the core brand happens especially in combination with a vertical brand extension, since the vertical brand extension differs in price and/or quality. The information that consumers receive about a new brand extension can influence the way consumers perceive the brand extension itself as well as the way consumers perceive the core brand, because consumers’ beliefs about the brand change incrementally when they receive new information about that brand (Loken & Roedder John, 1993). Hence, any inconsistent information about the vertical brand extension will lead to a modification of the current beliefs a consumer hold about the core brand, and therefore the brand image of the core brand becomes diluted.

2.2.1 STEP-UP AND STEP-DOWN BRAND EXTENSIONS

Again, a vertical brand extension is an extension of a new brand in the same product category, but at a different price or quality point (Keller & Aaker, 1992; Sullivan, 1990). This type of brand extension can go in two directions, namely it can go up or go down. If a product at a higher quality level and/or price point than the core brand is introduced it is called a step-up brand extension. The opposite is called a step-down brand extension, so a product at a lower quality level and/or price point is introduced (Kim & Lavack, 1996). It is found that both types of vertical extensions have a negative impact on the consumer’s perception of the core brand. This effect is explained by the categorization theory. Hence, Kim et al. (2001) suggested that a company should put a maximum amount of distance between the core brand and the step-up extension so the extension will not damage the brand image of the core brand.In line with the up extension, greater distance between the step-down brand extension and the core brand is beneficial for the core brand. However, greater distance from the core brand is less effective for the step-down extension (Kim & Lavack, 1996). Closer links are harmful for the core brand, but are beneficial for the step-down brand extension.

Concluded from the foregoing, regardless of the direction of extension the introduction of a vertical extension will have a negative impact on the core brand (Kim et al., 2001). However, this does not imply that the core brand is equally harmed by one of the two directions of the vertical extensions. There are no consistent findings found in literature. Brand image of the core brand partly depends on the brand concept (functional versus luxury concept) (Kim & Lavack, 1996; Kim et al., 2001; Kirmani, Magnoni & Roux, 2012; Sood, & Bridges, 1999). It is found that step-down extensions harm the core brand more than step-up vertical extensions when the core brand is a luxury brand (Kim & Lavack, 1996; Kim et al., 2001; Magnoni & Roux, 2012; Pitta & Katsanis, 1995). The explanation for this effect lies in

the fact that consumers of a luxury brand have a desire to maintain brand exclusivity (Kirmani et al., 1999). On the contrary, functional brands seem to allow step-down but seem not to allow step-up extensions (Pitta & Katsanis, 1995). Overall, literature on vertical extension show that vertical step-down extensions are more likely to damage the original status of the core brand compared to vertical step-up extensions, since the status value is less positive at a step-down extension (Kim & Lavack, 1996; Kirmani et al., 1999; Pitta & Katsanis, 1995).

2.3

B

A company that deals with a core brand and vertical extended brands at maximal distance is considered as not transparent, since the relationship of the two brands is not communicated at all. When the company reveals the link between both brands, they do become more transparent. As a consequence the brand architecture, which defines the relations between the brands, changes. Figure 1 displays the different brand architectures a brand portfolio can have as described by Aaker and Joachimsthaler (2000). When no link is revealed between the core brand and the vertical extension at maximal distance the company applies a house of brands strategy, in which brands are independent of each other. However, when the vertical extended brand at maximal distance communicates in its advertisements that the brand is related to another (core) brand, the vertical extended brand becomes an endorsed brand. Endorsed brands are still independent, but they are also endorsed by the core brand. Aaker and Joachimsthaler (2000) found that an endorsement by an established core brand provides credibility to the offering, but also found that the core brand plays only a minor driver role. Besides, the endorser provides also some useful associations to the endorsed brand. However, this is not the intention of the maximal distance put between the vertical extended brand and

the core brand since associations of the vertical extended brand and the core brand are conflicting.

Figure 1. Aaker and Joachimsthaler’s Brand Relationship Spectrum (2000, p. 9)

Therefore, in order to be transparent but to minimalize the conflicting associations, the core brand should be visible at a minimum. A brand architecture that deals with these concerns is the token endorser, to which the core brand is substantially less prominent than the endorsed brand. The logo of the core brand is visible, but will not have a centre stage. The core brand will provide some reassurance and credibility to the endorsed brand, while the endorsed brand still has its maximum freedom to create own associations (Aaker & Joachimsthaler, 2000). There are several requirements to make a token endorsement work, which among others include that the endorser is already well known and is consistently visual presented.

When an external party reveals the link between the brands, the company becomes more transparent as well. In this case the brand architecture will change to a shadow endorser. A shadow endorser is not visibly connected to the endorsed brand, but many consumers do know about the link between the two brands by other sources than the company itself (Aaker & Joachimsthaler, 2000). Aaker and Joachimsthaler (2000) argue that the shadow endorser has a minimal impact on the image of the brand but still provides credibility and helps in many segments. According to Aaker and Joachimsthaler (2000, p.12) thisarchitecture communicates that the organization realizes that the shadow-endorsed brand represents a totally different market segment.However, it is a enormous risk for the company since they do not control the information that is out there and therefore could be interpret by customers in other ways than the company intents to.On their turn, this could impact the trust in the company and therefore lead to a less positive brand image of the core brand.

2.4

I

According to Silver (2005) in this new age the relationship between the company and its customers depend on transparency and trust, because nothing can be hidden for long. Transparency is one of the major conditions in establishing positive relationships between

consumers and companies (Reynolds & Yuthas, 2008). Transparency and trust are two positive related concepts (Rawlins, 2008). Transparency requires trust. Rawlins (2008, p.2) argues that companies cannot expect trust from stakeholders if they are not willing to trust them first or in return. Hence, organizations must trust their stakeholders first to use their information responsible. The first step of this reciprocal trust is a company’s effort to be transparent in order to establish a positive relationship with their customers. According to Kang and Hustvedt (2014) transparency, as perceived by a consumer, can play a significant role in building trust and a positive attitude toward the company. They furthermore found that trust and positive attitudes built between the company and the consumer could be linked to the consumer’s purchase intent as well as to deliver a positive opinion to others due to transparency. In addition to the foregoing, it is pointed out that consumers gain value from the information communicated by the company via transparent pricing to the extent that they are willing to pay more for the same item (Carter & Curry, 2010). Besides, it is also pointed out that consumers value operational transparency (Buell & Norton, 2011). Both foregoing statements are in line with two-sides advertising. Two-sided advertising includes not only presenting positive information about the product, but also some negative information (Eisend, 2006). This technique is successful when applied in the correct manner.

2.5

H

These foregoing research findings are at odds. At the one hand the theory among vertical brand extensions at maximal distance explains that these brand extensions are put as distant as possible so the core brand does not become diluted. At the other hand the theory among transparency shows that transparency is demanded and valued by customers. Combining these two contradicting insights leads to the following hypotheses.

2.5.1 VERTICAL BRAND EXTENSIONS

When the relationship between the core brand and the vertical step-up extension becomes transparent, the core brand will be diluted (Kim et al., 2001). This is explained by the categorization theory, since consumers hold conflicting associations towards both brands to which they come beware of. Therefore it is expected that the brand image will be more negative for consumers who are aware of the relationship between the core brand and the step-up brand extension at maximal distance, compared to consumers who are not aware of any relationship with another brand.

Hypothesis 1: Consumers who are aware of the relationship between the core brand and the vertical step-up brand extension at maximal distance will have a more negative brand image of the core brand, compared to those consumers who are not aware of the relationship between the core brand and the vertical step-up brand extension at maximal distance.

The same effect occurs for step-down extensions. Therefore it is also expected that the brand image will be more negative for consumers who are aware of the relationship between the core brand and the step-down brand extension at maximal distance, compared to consumers who are not aware of the relationship.

Hypothesis 2: Consumers who are aware of the relationship between the core brand

and the vertical step-down brand extension at maximal distance will have a significant more negative brand image of the core brand, compared to those consumers who are not aware of the relationship between the core brand and the step-down vertical brand extension at maximal distance.

As seen in the vertical extension literature, the core brand is not equally harmed by one of the two types of the vertical extensions. Overall, vertical step-down extensions are likely to damage the status of the core brand more compared to vertical step-up extensions, since

the status value of a step-down extension is less positive (Kim & Lavack, 1996; Kirmani et al., 1999; Pitta & Katsanis, 1995). Only in case the core brand is a luxury brand, the core brand is more damaged when it concerns a downward extension (Magnoni & Roux, 2012). It is not the intention of this study to investigate at a luxury brand level, since both ways of extensions (step-up and step-down) need to be manipulated in the experimental design. It is supposed that a step-up extension is harder to make for a luxury brand and may be difficult to imagine for a respondent. Therefore, it is assumed that the brand image of the core brand is more negative after revealing the relationship with a up extension compared to a step-down extension.

Hypothesis 3: The brand image of the core brand will be more negative when the core brand is related to a step-down extension in comparison with a step-up extension.

2.5.2 SHIFT IN BRAND ARCHITECTURE

Research insights about transparency stated that consumers value various forms of transparency. However, when transparency increases for brand extensions at maximal distance a shift in brand architecture will unconsciously occur. This shift automatically leads towards exposing the conflicting associations of the core brand and the vertical brand extension. Therefore the following is supposed:

Hypothesis 4: Awareness of the relationship between the vertical extended brand and

the core brand will have a negative effect on the brand image of the core brand when revealed by a token endorsement strategy.

Hypothesis 5: Awareness of the relationship between the extended brand and the core

brand will have a negative effect on the brand image of the core brand when revealed by a shadow endorsement strategy.

In line with transparency lies trust. As already explained, reciprocal trust can be created by being transparent. This can create a more positive relationship with their customers (Rawlins, 2008). Therefore it could be argued that if a company reveals the relationship between the core brand and the vertical brand extension itself, the customers will trust the company more compared to when a third party reveals the relationship. Therefore it is assumed that the core brand is less harmed by a token endorsement strategy than by a shadow endorsement strategy.

Hypothesis 6: Consumers who became aware of the relationship between the

extended brand and the core brand through the token endorsement strategy will be less negative than consumers who became aware of the relationship between the extended brand and the core brand through the shadow endorsement strategy.

2.6

C

The conceptual framework will look like the figure below.

Figure 2. Conceptual framework

Awareness of relationship between

vertical extension and core brand

Brand image core brand

3

M

In order to test the hypotheses an experimental field study was conducted. Prior to the main study several preliminary studies were conducted. The brand that is selected for the study is Esprit. The main survey consisted of three parts. First a short introduction was given to ensure confidentiality of the respondent. Then one of the scenarios was outlined, which was followed by a three-paged questionnaire.

3.1

P

-

1:

B

Two preliminary tests were undertaken prior to the main study. The first pre-test was conducted to determine the core brand that would be used in the main study. The three core brands used in this pre-test were: Nike, Esprit and Mango. These brands were chosen for this pre-test because of their average price and quality perception, so it was possible to set up fictive scenarios including a step-up extension and a step-down extension for the main study. Besides, these brands have not introduced a vertical brand extension yet, so respondents cannot easily refer to other brands when reading the manipulated scenario in the main study.

The first pre-test had to ensure that there is brand familiarity of the core brand among the participants. An average to high familiarity of the core brand is needed for the main study since the initial brand image will be compared with the post brand image after the relatedness between the core brand and the vertical step-down or step-up brand extension at maximal distance is explored. If there is no or little initial brand image of the core brand, a comparison is hard to make. To calculate the brand familiarity of these three brands the participants were given two questions about the brands, namely: Whether they know the brand; and whether they could indicated on a 7-point Likert scale (1 = very unfamiliar and 7 = very familiar) their level of familiarity with the brand. Besides, it is important to eliminate the brands that

are viewed by the participants as either extremely favourably or extremely unfavourably. Extreme consumer attitudes toward the core brands could result in ceiling or floor effects on consumer attitudes toward the newly introduced step-up or step-down extension which can lead to biases (Kim et. al., 2001). Hence, the participants were asked on 7-point Likert scale how favourable they view each core brand (1 = not very favourable and 7 = very favourable). Last, in the first preliminary test the participants were asked to indicate on 7-point Likert scale the level of perceived price and quality of the three brands. This should be as neutral as possible, so the vertical step-up and step-down brands can be equally distance placed. The perceived quality of the brand is measured via two questions. One focused on opinions about the products (1 = bad products and 7 = good products) and the other directly evaluated brand quality (1 = poor quality and 7 = good quality). To measure the perceived price, the participants were asked on 7-point Likert scale how they perceived the given price range of the brand (1 = extremely low and 7 = extremely high).

The first preliminary test involved 18 Dutch participants (8 men and 10 women) who were between 21 and 51 years old (M = 24.6, SD = 6.56). All participants knew about the existence of the brands Nike and Esprit, but two male participants didn’t know the brand Mango. This may be explained by the fact that Mango mostly offers women’s clothing. Hence, Mango wasn’t a good fit for the main study. Other results of the first preliminary test can be found in table 1.

Table 1. Means of constructs measured for brands in pre-test 1

Nike Esprit Mango

Constructs M SD M SD M SD

Brand familiarity 6.22 .73 5.39 1.04 4.50 1.98

Favourability 5.89 .90 4.89 1.60 4.72 .89

Perceived quality 6.09 .92 5.42 .86 4.86 .81

As can be seen from table 1 Nike scored highest on every construct and Mango lowest. Mango was already eliminated for usage in the main study because not each respondent did know the brand. Nike is eliminated because results showed that respondents were highly familiar with Nike and had a strong favourability towards the brand. If a respondent has already too many associations towards the brand it can influence the results of the main study, since it can cause a minimalized change in attitude towards the brand. The results of Esprit are satisfying for the use of this particular brand for the main study, since the brand is highly familiar and has no extreme favourability. Besides the perceived price and quality were satisfying for Esprit. Hence, as can be concluded from the first preliminary test Esprit will be the brand used in the main study.

3.2

P

-

2:

T

The second pre-test had to determine whether participants see the both brand extensions as equally distant to the core brand, as well as it had to ensure that nothing else would manipulate the brand image than the message in the main study. Hence, non-existing brand names were chosen for the extensions so the respondents have not formed an attitude yet. Since Esprit has not introduced other vertical brand extensions yet and the extensions are at maximal distance, the brand name for the step-up and step-down extension did not have to correspond to Esprit. Hence the brand name ‘Express’ is chosen for the step-up extension and ‘Tenor’is chosen as a brand name for the step-down extension in the main study.

Both brand extensions should be equally distant so they can be compared with each other in the analysis of the main study. To equally manipulate the degree of the distance of the extensions, this second pre-test is conducted to investigate appropriate price point of a step-up and a step-down extension. From the first pre-test was determined that the given price point of Esprit (€50) is considered a moderate price point which is used as a starting point

for the second pre-test as well as for the main study. In pre-test two, the respondents were again asked to indicate on 7-point Likert scale the level of perceived price and quality of both vertical extended brands with the same questions as in the first preliminary test. For each extension three different price points were given. At pre-test two the same 18 Dutch participants (8 men and 10 women) were involved. Table 2 displays the results of the first test of the second pre-test. Price point 2 of the up extension and price point 4 of the step-down extension are rated as equally distant from the core brand, so these were best suited to use in the main study.

Table 2. Means of average price/quality for extension scenarios Average price/quality

Price point M SD

Price point 1 for step-up extension 4.85 1.24

Price point 2 for step-up extension 5.47 .93

Price point 3 for step-up extension 6.39 1.16

Price point 4 for step-down extension 2.53 .74

Price point 5 for step-down extension 3.89 .99

Price point 6 for step-down extension 4.70 1.03

Then six different communication scenarios were tested that will manipulate the attitude towards the brands in the main study. The first three scenarios included token endorsement strategies and second three scenarios involved shadow endorsement strategies. These scenarios had to be tested because it had to be sure that nothing else could manipulate the brand image of the respondent in the main study. Hence, the participants were asked on 7-point Likert scale how neutral they view each core brand (1 = not neutral and 7 = neutral). Besides, the believability of the scenarios was also important for the reliability of the measurements in the main study. Therefore, the participants were asked on a 7-point Likert scale whether they thought the six different scenarios were believable (1 = unbelievable and 7 = believable). Results of this test can be found in table 3. Results showed that all scenarios

were perceived as neutral since every scenario scored above 4. However, not all scenarios were considered believable since scenario 3 and 5 did not had scored above 4.

Table 3. Means of neutrality and believability for endorsement scenarios

Neutrality Believability

Scenarios M SD M SD

Scenario 1 for token endorsement 4.88 .97 5.17 1.20

Scenario 2 for token endorsement 4.36 .91 6.29 .91

Scenario 3 for token endorsement 5.15 1.06 3.46 1.43

Scenario 4 for shadow endorsement 4.29 .89 5.79 .88

Scenario 5 for shadow endorsement 5.34 1.30 3.91 1.12

Scenario 6 for shadow endorsement 4.74 .86 4.89 .95

Based on the results of the second preliminary test, the price point for a step-up and step-down extension used for the main study respectively were €125 and €10. Besides communication scenarios 2 and 4 were chosen as endorsement strategies for the main study. At scenario 2 the respondents were told they saw a random advertisement of the step-up (Express) or step-down brand (Tenor) and a picture was viewed with the logo’s of both the step-up or step-down brand and a small logo of Esprit (see figure 3). Scenario 4 was a newspaper article in a fictive newspaper were respondents only red the heading and the first line of the article that said that the step-up or step-down brand extension is related to Esprit.

3.3

E

To test the proposed hypotheses the main study involved a 2 (Awareness of brand extensions: step-up vs. step-down) x 2 (Communication style: token endorsement vs. shadow endorsement) research design. Four surveys include manipulated information. In addition to these surveys, there is one survey that only consists questions about the core brand. This means that five different surveys were distributed to collect all the data needed to test the hypotheses.

At the general introduction of the survey the participant were welcomed and were made sure that their anonymity was secured. Besides, they were told that this research was part of the researcher’s Master thesis and that the questionnaire would not take longer than five minutes. After this introduction, the respondent was randomly assigned to one of the following five scenarios:

1. The relationship between core brand and step-up brand extension is created through token endorsement strategy

Esprit is een kledingmerk dat al tientallen jaren bestaat. De prijs voor een kledingstuk is gemiddeld €50. Het merk Express is ook een kledingmerk dat al tientallen jaren bestaat. De prijs voor een kledingstuk is bij Express echter gemiddeld €125.

Beeld je in dat je een advertentie ziet in een willekeurig tijdschrift, waarop het onderstaande logo wordt afgebeeld. Via dit logo in de advertentie ben je erachter gekomen dat Express al die tijd al was gelinkt aan Esprit.

2. Thee relationship between core brand and step-down brand extension is created through token endorsement strategy

Esprit is een kledingmerk dat al tientallen jaren bestaat. De prijs voor een kledingstuk is gemiddeld €50. Het merk Tenor is ook een kledingmerk dat al tientallen jaren bestaat. De prijs voor een kledingstuk is bij Tenor echter gemiddeld €10.

Beeld je in dat je een advertentie ziet in een willekeurig tijdschrift, waarop het onderstaande logo wordt afgebeeld. Via dit logo in de advertentie ben je erachter gekomen dat Tenor al die tijd al was gelinkt aan Esprit.

3. The relationship between core brand and step-up brand extension is created through shadow endorsement strategy

Esprit is een kledingmerk dat al tientallen jaren bestaat. De prijs voor een kledingstuk is gemiddeld €50. Het merk Express is ook een kledingmerk dat al tientallen jaren bestaat. De prijs voor een kledingstuk is bij Express echter gemiddeld €125.

Via het onderstaande (deel van het) nieuwsbericht in krant X ben je erachter gekomen dat Express al die tijd al was gelinkt aan Esprit:

4. The relationship between core brand and step-down brand extension is created through shadow endorsement strategy

Esprit is een kledingmerk dat al tientallen jaren bestaat. De prijs voor een kledingstuk is gemiddeld €50. Het merk Tenor is ook een kledingmerk dat al tientallen jaren bestaat. De prijs voor een kledingstuk is bij Express echter gemiddeld €10.

Via het onderstaande (deel van het) nieuwsbericht in krant X ben je erachter gekomen dat Tenor al die tijd al was gelinkt aan Esprit:

5. Core brand is not related to any other brand

Voor dit onderzoek betreffende merk-imago’s wil ik u graag wat vragen stellen over het merk Esprit. Esprit is een kledingmerk dat al tientallen jaren bestaat. De prijs voor een kledingstuk is gemiddeld €50.

After reading one of the scenarios carefully, which had to be confirmed the respondent did, the respondent had to fill in two or three pages with questions about the brand(s) and about his/her characteristics. The core brand image was measured through eight questions about the respondents’ associations towards the brand. This attitude and association measurement was already been tested and used by Keller & Aaker (1992) and Keller (1993). In these nine questions the respondent had to indicate their attitude towards the core brand on a 7-point Likert scale (1 = strongly disagree and 7 = strongly agree). The separate questions indicated how trustworthy the brand was perceived; how appealing the brand was perceived; how well the brand was known; how positive the respondent felt towards the brand; how reliable the brand was perceived; how the image of the brand was perceived; how trustworthy the claims of the brand were perceived; how the quality of the brand’s products were perceived and how likely it was to purchase a product of the brand. Besides, each respondent had to indicate on a 7-point Likert scale how familiar he/she was with the brand Esprit. This is asked for the same reason it was asked in the pre-test.Extreme consumer attitudes toward the core brands could result in ceiling or floor effects on consumer attitudes toward the newly introduced step-up or step-down extension which can lead to biases (Kim et. al., 2001).

The previous measurements were also used to measure the brand image of the extended brands. However, the statement about how well the brand is known is left out in the measurement of the extended brands, since the brand is a hypothetical brand. The brand familiarity of the extended brand is obviously not measured either. All the questionnaires with corresponding logos, the newspaper articles and the questions asked can be found in appendix I.

The link to the web-based survey was distributed via e-mail and posted on the social network site Facebook. The snowball sampling technique was used to recruit even a wider

range of potential participants. Those who were initially contacted by e-mail were asked to forward the link to their friends and relatives who met the eligibility requirements. The only requirement to fill out the questionnaire was that the respondent was aware of the brand Esprit. The survey was available to fill in from June 7 till June 15. To run statistical analysis a sample size of 150 participants was required, based on a minimum of 30 respondents per condition (Saunders, Lewis & Thornhill, 2009).

4

R

This chapter will describe all results.

4.1

D

All data had to be prepared before correctly analyse it.  There were no negatively keyed items in the variables, so there was no need to recode any variable. Besides, all respondents who succeed to complete the survey showed no missing data, since an error would appear if the respondent left a question unanswered.  

4.1.1 SAMPLE SIZE

A total of 172 respondents agreed to take part in the survey. Twelve respondents failed to complete the survey and are therefore are excluded from further analysis. This leaves the completion rate of this survey to 93,0%. Besides, one respondent had an age of thirteen. It is assumed thatthis respondent does not buy clothes for himself yet and therefore is unable to form proper brand images for the brands in this study. Hence, this respondent is left out the analysis as well. This leaves 159 randomly assigned respondents for the upcoming data analysis. The response rate cannot be calculated, since the survey is shared on social media. Therefore it is unknown how many people have seen the invitation to fill out the survey.

4.1.2 RESPONDENTS

The respondents were distributed among the five different surveys as followed: 33 respondents for awareness of relationship between step-up extension and core brand through an advertisement, 34 respondents for awareness of relationship between step-down extension and core brand through an advertisement, 29 respondents for awareness of relationship between step-up extension and core brand through an news paper article, 31 respondents for awareness of relationship between step-down extension and core brand through an news

paper article and 35 respondents for the group that only had to answer questions about the core brand. The sample included more women (59,7%) than men (40,3%). The age range of the respondents varied between seventeen and 62 years, with an average of 26 years (M = 25.6 and SD = 8.2). Most respondents had a university degree (60,4%), followed by 28,9% having a degree from the university of applied science (in Dutch: HBO). 10,7% of the respondents had a degree that is lower than a degree from the university of applied science. As expected more women filled out the questionnaire, because generally speaking they fancy fashion more than man. Besides, since respondents were mostly approached via Facebook it was expected that higher educated people of an age between 20 and 30 filled out the survey most. Therefore, this sample group is considered representative for the population that has been approached for this study. The results derived from this study can be generalized back to this population.

4.1.3 RELIABILITY CHECK

Different constructs (the brand image of different brands) in the model are measured in various questions. Therefore, it is important to see which questions correlate to each other to make a total variable that shows the total score of the particular construct. A reliability test (Cronbach’s Alpha) was conducted to see if each item measured the constructs. Internal consistency on a scale is considered to be appropriate as a Cronbach’s Alpha coefficient is at least .70 (Pallant, 2010). Table 4 shows the Cronbach’s Alpha including the number of items the scale constructs of.

Table 4. Reliability scores of constructs

Construct Number of items Cronbach’s Alpha coefficient

Brand image core brand 9 α .884

Both coefficients are above the limit of α = .70 and therefore are considered as reliable constructs. Since the Cronbach’s Alpha coefficient did not improve when an item was deleted, no adjustments were made. Hence, the items of both constructs are combined into two constructs that measures brand image of either the core brand or the extension. These two constructs will be used for further analyses.

4.1.4 CONDITION SIMILARITY

Next step in analysing the gathered data is comparing the five different scenarios to each other. In order to do so, it is important that there are no differences in the characteristics of the respondents between the groups, since dissimilarities could hinder the comparison of the different groups. In this study there were three characteristics known from the respondents, which are ‘gender’, ‘age’ and ‘education’. First, a Pearson Chi-square test for independence was conducted to see whether there was a significant difference in gender distribution among the various scenarios. Two conditions have to be met in order to interpret the outcomes of this test, namely: (1) the total number of frequencies need to be equal or higher than fifty and (2) the expected frequencies in the cells need to be equal or higher than five (Hair, Black, Babin, Anderson & Tatham, 2006). Since 159 respondents did completely finish the survey, the first assumption is met. However, the Pearson Chi-square test showed that 95% of the cells have expected to count less than five. Therefore the second assumption is not met, so another test has to be conducted to determine whether there are differences in gender in the different scenarios. Hence, a Fisher’s exact test is performed instead. Results indicate that there are no differences between the groups in terms of gender (p=0,635>0.05). In addition to that, a Pearson Chi-square was conducted to see whether there was a significant difference in age distribution among the various scenarios. For this test both assumptions are met, since age is categorized into five different age groups. Results indicate that there are no significant differences between the groups in terms of age (χ!(5) = 2.74, p = .65). Last, it is been

checked whether there was a significant difference in education distribution among the various scenarios. Here both assumptions to perform a Pearson Chi-square are also met, since education is categorized into five different groups. Results indicate that there are no significant differences between the groups in terms of age (χ! (5) = 3.46, p = .71). From the foregoing analyses it can be concluded that there are no differences in the characteristics of the respondents between the groups.

4.1.5 MANIPULATION CHECK

If the brand extensions have been successfully manipulated, then the quality of the step-up extension should be evaluated higher than the core brand, and the quality of the step-down extension should be evaluated lower than the core brand. Hence, a comparison is made between the evaluation of the perceived quality of the brand extensions and the quality of the core brand. The means and standard deviations are shown in table 5. These outcomes suggest that the step-down brand is seen as a downward extension compared to the core brand (Mtenor = 3.54< Mesprit = 5.23; t(63) = 1.376, p<0.001). However, the respondents did not see the step-up brand extension as an upward extension compared to the core brand (Mexpress = 4.47 < Mesprit = 5.23). The quality of the vertical extension is measured after the participant knew the vertical extension is related to the core brand, which could explain the lower rated step-up extension. The relationship between the core brand and the extended brand could be considered as a direct link. This direct link could have had a negatively influence on the evaluation of the step-up extension. In order to conclude whether the manipulation was successful or not, the outcome has to be compared to an evaluation of the vertical brands rated by a control group. Unfortunately, in this study the control group did not evaluate the extended brands and therefore the conclusion cannot be drawn. What is clear is that the respondents did not evaluate the step-up extension as higher than the core brand. Therefore it is doubtful whether this manipulation did succeed or not.

Table 5. Mean scores of quality brands

Brand Number of respondents Mean SD

Esprit 35 5.23 1.11

Express 60 4.47 .93

Tenor 63 3.54 1.10

4.2

H

This paragraph will report the outcomes of the analysis. The experiment consisted of four different questionnaires that represented two different extension relationships, namely a step-up extension and a step-down extension. The four experiment grostep-ups were exposed to the relationship between the core brand and the extension through an advertisement of the extension (token endorsement strategy) or through an article in a newspaper (shadow endorsement strategy). In each of the four questionnaires the respondents had to evaluate the core brand and the extended brand. In addition to the four groups, there was a group that answered only questions about the core brand Esprit. To test the hypotheses the different outcomes of the five scenarios are compared and analysed. To test the hypotheses an univariate ANOVA test was conducted. The results of this test are displayed in table 6.

Table 6. Means and standard deviations of the brand image at different scenarios Step-up extension Step-down extension Core brand only

M SD M SD M SD

Token endorsement 4.40 .94 4.60 .91

Shadow endorsement 4.43 .76 4.53 .81

Average 4.41 .86 4.57 .86 4.97 .81

From table 6 can be concluded that the group who only evaluated the core brand gave the core brand the highest brand image score (M = 4,97 and SD = .81). To analyse whether the brand image of the core brand is different among the different scenarios a one-way ANOVA

is conducted. This test resulted in a significant outcome (F(4, 279) = 2,64, p = 0,03). Hence, it can be concluded that the brand image of the core brand is significant different in the various scenarios. Now, the hypotheses will be tested independent of each other. After each hypothesis is tested, several constructs are further explored in detail.

4.2.1 HYPOTHESIS ONE

Hypothesis one stated that consumers who are aware of the relationship between the core brand and the vertical step-up brand extension at maximal distance will have a more negative brand image of the core brand, compared to those consumers who are not aware of the relationship between the core brand and the vertical step-up brand extension at maximal distance. Hence, to test this first hypothesis the scenarios that included the step-up extension are compared to the scenario that only included the evaluation of the core brand. The communication source is not of importance for this hypothesis and is therefore ignored.

It can be assumed that there is a normal distribution, since both groups have more than 30 subjects (control group 35 and step-up extension 62). As can be seen from table 6, the mean of the core brand is evaluated higher by the group that only reviewed the core brand, compared to the group was exposed to the step-up extension scenarios (Mcontrol = 4.97 > Mstep-up extension scenarios = 4.41). This is graphically displayed in figure 4. Results of an independent-samples t-test showed that there was a significant difference between the perception of brand image of the group that only had to evaluate the core brand and the group that was aware of a relationship with a step-up brand (t(96) = 3.048, p = .003). The mean difference is -.54 with a 95% confidence interval ranging from .19 to .90. These outcomes suggest that the brand image of Esprit is significant lower evaluated when a respondent was aware of the relationship between the step-up extension and the core brand.

Figure 4. Brand image core brand of the control group compared to the step-up extension scenarios

4.2.2 HYPOTHESIS TWO

Hypothesis two stated that consumers who are aware of the relationship between the core brand and the vertical step-down extension at maximal distance will have a significant more negative brand image of the core brand, compared to those consumers who are not aware of the relationship between the core brand and the step-down vertical extension at maximal distance. Hence, to test the second hypothesis the scenarios that included the step-down extension are compared to the scenario that only included the evaluation of the core brand. The communication source is again not of importance for this hypothesis and is therefore ignored.

It can be assumed that there is a normal distribution, since both groups have more than 30 subjects (control group 35 and step-down extension 65). As can be seen from table 6, the mean of the core brand is evaluated higher by the group that only reviewed the core brand,

4,1 4,2 4,3 4,4 4,5 4,6 4,7 4,8 4,9 5 5,1

control group step-up extension

Mean

 of  perc

eption

b

ra

n

d

 im

age

compared to the group that was exposed to the step-down extension scenarios (Mcontrol = 4.97 > Mstep-down extension scenarios = 4.57). This is graphically displayed in figure 5. Results of an independent-samples t-test showed that there was a significant difference between the perception of brand image of the group that only had to evaluate the core brand and the group that was aware of a relationship with a step-down brand (t(96) = 2.293, p = .024). The mean difference is .41 with a 95% confidence interval ranging from .05 to .76. These outcomes suggest that the brand image of Esprit is significant lower evaluated when a respondent was aware of the relationship between the step-down extension and the core brand.

Figure 5. Brand image core brand of the control group compared to the step-down extension scenarios

4.2.3HYPOTHESIS THREE

In the third hypothesis is assumed that brand image of the core brand is more negative when the core brand is related to a step-down extension compared to a step-up extension. However, table 6 shows that the brand image of the core brand is better evaluated when the relationship includes a step-down extension (Mstep-down = 4.57 > Mstep-up = 4.41). This is graphically

4,1 4,2 4,3 4,4 4,5 4,6 4,7 4,8 4,9 5 5,1

control group step-down extension

M

ean

of p

er

ce

p

ti

on

b

ran

d

i

mage