Consumer responses to different news items for traditional- versus YouTube endorsers : are consumers more immune to moderately negative publicity about YouTube celebrities than traditional celebrities?

(1)

Consumer responses to different news items for

traditional- versus YouTube endorsers

Are consumers more immune to moderately negative publicity about YouTube celebrities than traditional celebrities?

Master thesis

Student: Janneke Florence de Vries Student number: 11422793

Supervisor: Drs. Roger Pruppers University: University of Amsterdam

Faculty: Economics and Business

Study program: Business Administration – Marketing Track Date: 17th of August 2017

(2)

Statement of Originality

This document is written by Janneke de Vries who declares 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.

(3)

Abstract

Today, celebrity endorsement is one of the most used forms in advertisements. More and more of these brand endorsement deals are closed with ‘new type’ of celebrities, like YouTubers. However, their brand endorsements have barely been studied yet, since most the studies focus on the traditional celebrity. Moreover, the difference in attitude towards a celebrity endorser before and after the exposure to a news item has not been examined in past marketing literature, even though celebrities are written a lot about in the media. More specifically, the impact of negative information about a celebrity has been studied, but the impact of positive information is less thoroughly examined. Besides, the impact of amplitude of information has barely been taken into account in past research within this context.

Based on a theoretical framework, a problem statement was developed to examine the difference in attitude towards the celebrity and brand before and after exposure to a news item. The news item was compared for combinations of valence (i.e. the power of the information: positive/negative), amplitude (i.e. the magnitude of information: moderate/extreme) and type of celebrity (i.e. traditional/YouTube celebrities). A 2x2x2 between-subjects experimental design was used to conduct the research. In a pre-test, the best suitable news items were selected, before the main experiment was filled out by 291 respondents. The results reveal that exposure to the news item, moderated by valence, has a significant impact on difference in attitude towards the celebrity. Contrary to the hypotheses, attitude towards the celebrity does not mediate attitude towards the brand. Besides, amplitude did not prove to be a significant moderator for valence, probably due to the failure of the variable amplitude. Additionally, there was no significant relation found between the moderator type of celebrity and amplitude for

(7)

Chapter 1: Introduction

1.1. YouTuber as brand endorser

One of the most popular forms in advertising is celebrity endorsement. The usage of celebrity endorsers is seen as a very effective way to create a better image or awareness for the brand or product. However, with the evolution of blogs and social media channels like Instagram and YouTube a lot has changed with regard to the traditional celebrity. With the emergence of social media, a new kind of celebrity has arisen: ‘the blogger’, ‘the Instagrammer’ and the ‘Youtuber’. The evolution of these social media channels has given the media landscape a new face and marketers’ new strategies to adopt towards advertising campaigns.

Especially, the ‘YouTuber’ accompanied by his/her self-built community is seen as a new type of celebrity. Video blogging, vlogging, has become a business on itself. Some YouTubers can now even create video content for a living. Hundreds of millions of videos are watched every day on YouTube and the platform has over a billion users (YouTube, 2017). According to Forbes (Ingham, 2015), “product placement is key to the success of a new generation of lifestyle vloggers, who make videos about their daily lives, play pranks on one another, and who, unlike celebrities or film stars, are respected because of their ‘normalness’ rather than their out of this world lifestyles.” An example of the growth in influence of YouTubers is seen in the usage of YouTube vloggers by luxury brands to help build long-term relationships with the consumers. Consumers that consider vloggers resembling themselves will have the same positive brand evaluation as the vlogger (Lee and Watkins, 2016).

1.1.1 Amplitude and valence of information

Celebrities sometimes endure negative publicity and this might impact the brand they endorse. For example, the sexual scandal of Tiger Woods during his collaboration with NIKE led to a

(8)

total shareholder loss of approximately US $5–12 billion (Knittel and Stango, 2010 in Zhou and Whitla, 2013). Negative information about a celebrity could lead towards a lower evaluation of the brand (Till and Shimp, 1998). What happens to the evaluation of the celebrity? What is the case for positive information? Research has proven though, that negative information tends to be evaluated stronger than that of comparably extreme positive information (Ito, Larsen, Smith, and Cacioppo, 1998 in Money, Shimp and Sakano, 2006). Another study shows that moderately negative publicity about a company is moderated by the identification consumers have with the company. Nevertheless, with the exposure of extremely negative information, weakly and strongly identified consumers have the same reactions towards the information (Einwiller, Fedorikhin, Johnson and Kamins, 2006). With the emergence of ‘new-born celebrities’ marketers might need a different approach to target their consumers. Do marketing campaigns work out the same way with YouTubers as with traditional celebrities? Traditional celebrities have a different character compared to YouTube celebrities. For example, traditional celebrities still maintain a distance, whereas YouTube celebrities mainly display accessibility, intimacy and connectedness online (Jerslev, 2016). Most the YouTubers are not controlled by a publicity system and this autonomy creates trust and authenticity (Gamson, 2011). Thus, do YouTubers have a higher level of identification among consumers? In line with the aforementioned theory of Einwiller et al. (2006), a higher level of identification with the YouTube celebrity indicates that consumers are more likely to reject moderately negative information compared to traditional celebrities.

(9)

the celebrity (e.g. Money et al., 2006), but leaves the impact of positive information about the celebrity endorser on attitude towards the celebrity and/or brand largely undiscovered. Additionally, the difference in attitude before versus after exposure to a news item with regard to the celebrity him- or himself, has not been examined yet. Other studies show that moderately negative publicity about a company is moderated by the identification consumers have with the company, but with the exposure of extremely negative information this effect disappears (e.g. Einwiller et al., 2006). However, the impact of amplitude of information has barely been studied within brand endorsements. Moreover, nearly all the brand endorsement literature focusses on the traditional celebrity, while there is still little known about micro-celebrities, like YouTubers. Do these studies apply to YouTubers as well? Is the sense of identification higher among YouTube celebrities than traditional celebrities? In order to fill this gap, this research seeks to test the whether the exposure to a news item about the YouTube- and traditional celebrity, with different levels of valence (i.e. the power of information: positive/negative) and amplitude (the magnitude of information: moderate/extreme), causes a difference in attitude towards the celebrity and brand. The research question for this study is as follows: “How do consumers

respond to the exposure of different levels of information in news items compared for YouTube- and traditional celebrities?”

1.2.2. Delimitations of the study

This thesis specifically focusses on the impact of a news item, differentiated by valence and amplitude of information, on attitude towards the YouTube-/ traditional celebrity and the brand. For this end, the difference in attitude before and after exposure to the news item is examined. Instead of testing the effect of multiple factors that could influence attitude towards the celebrity via a news item, only valence, amplitude and type of celebrity are investigated. Moreover, only positive and negative information are studied, neutral information is not investigated. The impact of information about the brand on attitude towards the celebrity is not studied in this

(10)

research. Besides, the attitudinal loyalty (e.g. attitude) towards the brand is mainly emphasized in this study instead of the behavioural loyalty (e.g. usage of the brand). This thesis only focusses on YouTube- and traditional celebrities; other type of endorsers are not taken into account due to the limited amount of time given to conduct this research.

1.3. Contribution

1.3.1. Theoretical contribution

The comparison between a YouTube- and traditional celebrity endorser and the impact a news item has on the attitude towards them and the brand they endorse, gives insights into the advantages and disadvantages of these type of endorsements. This research contributes to the suitability of these two endorsers in order to shine a new light on the traditional sorts of endorsements and advertising. It helps to recognize that traditional endorsers might not be as effective as YouTube endorsers, since YouTube celebrities are easier to identify with.

Furthermore, recent studies have not studied the impact of information on the attitude towards the celebrity him- or herself, although this attitude might directly mediate the relationship between the news item and the attitude towards the brand he/she endorses. Besides, the impact of positive information on attitude towards the celebrity or brand has barely been investigated, while positive information might help identifying ways to increase the attitude towards the celebrity and thus towards the brand. Furthermore, amplitude of information remains largely undiscovered in this domain, even though it helps recognizing the way the impact of information is reinforced or weakened. Finally, the difference in attitude towards a celebrity before and after exposure to a news item has not been investigated. Normally, only

(11)

Investigating the brand and celebrity information together helps marketers and publicity managers to understand responses of consumers to different sorts of brand endorsement information in the branding domain.

1.3.2. Managerial contribution

Next to the theoretical contribution of gaining new insights within the field of consumer behaviour, this thesis also contributes to managerial implications on a marketing level. Most research on celebrity endorsement was conducted before the emergence and growth of the vlogging atmosphere. The growth of websites like YouTube and its celebrities, offers new insights and opportunities for researchers and marketers in order to better understand the behaviour of consumers and the new digital media. This thesis provides managers and agencies of traditional- and YouTube celebrities with knowledge, regarding the impact and dynamics of different levels of information on attitude towards the celebrity and the brand they endorse. Thereby, a clear comparison between the effectiveness and impact of the different type of endorsers is made, which helps marketers to choose for the safest and best suitable endorser.

An example that shows the impact of a news item on attitude towards the celebrity, is that of the most popular YouTube celebrity PewDiePie. An investigation by the Walls Street Journal, concluded that he posted several anti-semetic related videos on YouTube. Disney and YouTube immediately closed down their business deals with him (Gilbert, 2017). However, a lot of his followers defend him online for just being funny regardless the accusations: “I'm a Jew myself and I don't see how this harmless YouTuber is anti-Semitic. He was clearly joking #IStandWithPewDiePie” (TheRagingBlazer_) (PewDiePie, 2017). This example shows that the sense of identification his followers have with him, creates opportunities for marketers.

(12)

1.4. Structure outline

In order to answer the central research question, the following sections of this thesis are elaborated. First, a theoretical framework is presented, which is based on previous research and literature. In this framework, sub questions help to explore certain areas. Firstly, the usage, effectiveness and dangers of endorsements in marketing are explored: What is endorsement,

how is endorsement in marketing used to be effective and what are the dangers? Secondly,

valence and amplitude of information are examined: How does valence and amplitude of

information impact brands, celebrities and their endorsements? Lastly, different type of

endorsers and the characteristics of a celebrity are studied in order to be able to compare the traditional celebrity with the YouTube celebrity: What are the different type of endorsers and

what are the differences between a traditional celebrity and a YouTube celebrity?

Accordingly, this literature review is used to design a conceptual model. Thereafter, the method used for this research is presented and elaborated. Eventually, the findings of the analysis of the experiment are presented. A discussion of these results leads to the implications and limitations of this research. Finally, in a concluding chapter, the most important findings are summarized, the research question is answered and directions for future research are given.

(13)

Chapter 2: Celebrity endorsement

In this chapter, an introduction to celebrity endorsement in marketing is given. Accordingly, the effectiveness of endorsement, together with classical condition and secondary knowledge, is explained. Lastly, the dangers of endorsement are given.

2.1 An introduction to endorsement in marketing

According to Erdogan (1999) celebrity endorsement was introduced in the nineteenth century. He illustrates this with Queen Victoria endorsing products from Cadbury's Cocoa (Shennan 1985, in Erdogan, 1999). This advertising technique emerged with the introduction of the cinema, the commercial radio and television (McDonough, 1995, in Erdogan, 1999). However, the endorsement of celebrities was still limited, because people were sceptical about celebrities devoting their status into endorsing products and brands. An increase in celebrities endorsing products took place around the late 1970s when the amount of films and TV programmes increased and the shame among celebrities weakened (Thompson, 1978 in Erdogan, 1999).

2.2 Effectiveness of celebrity endorsement

Approximately 25 per cent of advertisements in the United States use celebrity endorsers (Shimp, 2000, in Amos, Holmes and Strutton, 2008) and they are the most popular form of advertising within the retail industry (Choi and Rifon, 2007 in Thwaites, Lowe, Monkhouse and Barnes, 2012). Celebrity endorsers help to create and maintain the attention of the consumer towards the advertisement and differentiate from the high number of other commercials (Sherman, 1985 in Erdogan, 1999). In addition, celebrities make the advertisement more attractive by adding glamour to the product and make the ad memorable, credible and desirable (Spielman, 1981, in Biswas, Hussain and O'Donnell, 2009). Kaikati (1987, in Biswas et al., 2009) emphasizes the advantage of celebrities in crisis management, brand repositioning, global

(14)

marketing and boosting sales. For example, the sales of Pepsi increased with 8 per cent after the first year Michael Jackson endorsed their product (Gabor, Thorton and Wienner, 1987, in Thwaites et al., 2012). Amos et al. (2008) show that celebrity trustworthiness, expertise, and attractiveness, included in the source credibility model, have the most influential source effects on intention to purchase and attitudes towards the brand and advertisement.

2.2.1 Classical conditioning

Celebrity endorsement can be explained with the classical conditioning paradigm (see figure 1), in which consumers learn associations between the endorser, the unconditional stimulus, and the product, the conditional stimulus, via multiple exposures (Erdogan, 1999). This technique is also known as associative learning (Abimbola, Lim, Halonen-Knight and Hurmerinta, 2010). The association sets of the celebrity and the brand consist of nodes about the brand and the celebrity that are connected to other nodes due to experiences with the two (Till and Shimp, 1998). The better the fit is perceived between the celebrity and brand, the easier and faster an associative link can be developed (Thwaites et al., 2012).

(15)

products across three different stages. This involves the movement of culturally constituted meanings in societies of the consumers. In stage 1, the image of the celebrity is created, in which the endorsement gives the advertisement access to detailed and powerful meanings behind the individual. Celebrities, rather than anonymous models, are considered to own these meanings, because they created them with repetitive performances in public. In the second stage, the endorsement is placed at centre, in which the symbolic features of the product are determined in order to choose a celebrity that will be able to transfer these features. Similarity between the celebrity and the product needs to be considered for the meaning transfer to take place. Accordingly, an advertising campaign will be chosen to transfer these meanings to the product. In stage 3, the symbolic features of the product are recognized by the consumers and transferred into meanings via rituals that partly construct their self-image (McCracken, 1989).

Figure 2: Meaning Transfer Model according to McCracken (1989).

2.2.2 Secondary knowledge

Keller (2003) describes consumer brand knowledge as the personal meaning about a brand stored in the memory of the consumer, like images and experiences. Keller suggests that building customer-based brand equity requires the creation of a familiar brand that has favourable, strong and unique brand associations. One of the options Keller gives to build brand equity is to leverage secondary associations, whereby a link is created with another entity. A way to do so is to link the brand or product to spokespeople and borrow its associations, like celebrity endorsers (Keller, 2003; 2005). If consumers see a high similarity between the entity

(16)

and the brand, they are more likely to transfer their knowledge of the entity to the brand. The match up hypothesis supports this: “endorsers are more effective when there is a fit between the endorser and the endorsed product” (Till and Busler, 2000, p.1). The degree of leverage from secondary knowledge depends on the consumers’ knowledge of the entity, meaningfulness- and transferability of consumers’ knowledge of the entity (Keller, 2005).

2.2.3 Reciprocal meaning transfer

Recently, more studies have focussed on the two-way relationship of endorsement and found that the celebrity is influencing the brand image, while the brand image influences the celebrity as well. Till and Shimp (1998, p.69) argue that “repeated pairings of the celebrity endorser and brand should facilitate the celebrity and brand becoming part of each other’s association set”. Seno and Lukas (2007) support this and argue that within the celebrity endorsement process, brand image mediates brand equity and the celebrity image mediates celebrity equity. Celebrity endorsement should therefore be considered as a brand alliance, wherein a reciprocal meaning and value transfer process takes place between the brand and/or company and the celebrity endorser (Abimbola et al., 2010). This reciprocal process was examined within the collaboration between Sainsbury’s supermarkets and Jamie Oliver, in which negative publicity was involved. Jamie Oliver supported farmed salmon in an advertisement from Sainsbury’s, while it was reported that he refused serving farmed salmon at his own restaurant. A newspaper article analysis showed that this campaign led to negative associations of Sainsbury’s, while negative associations were applied to Jamie Oliver as well: “Oliver, already reeling from complaints about his London restaurant, is now under assault from environmentalists” (The Independent, 2004 in Abimola et al., 2010). Another research found that when negative

(17)

2.3 Dangers of celebrity endorsement

The impact of information about the celebrity is not free from detrimental effects. Despite the benefits of celebrity endorsement, it also has a few hazards. There is a risk that consumers will solely focus on the celebrity and not on the product that is endorsed (Rossiter and Percy 1987, in Erdogan, 1999). The image of the celebrity can change, in which they for instance lose popularity or get involved in an issue of immorality (Cooper, 1984; Kaikati 1987, in Erdogan, 1999). The celebrity can also lose credibility when he/she endorsers too many brands, in which the link with the particular endorsed brand is not unique anymore and the endorsement impact decreases (Mowen and Brown 1981, in Erdogan, 1999).

Lei, Dawar and Lemmink (2008) show this link in relation to brands. They argue that the relatedness between brands is encouraged in the brand portfolios of companies to increase marketing efficiency via positive spillover of brand equity, but these linkages between brands can also cause a negative spillover. They show that the extent of spillover between brands depends on the strength and directionality of brand associations. Brands with fewer associations have stronger linkages to brands with more associations than the other way around. Besides, priming brand associations, which include common and/or distinctive associations, can increase the directional strength of the association. This insight helps to explain why negative information can occasionally impact brands that are considered only weakly related to the affected brand. Thus, the directional strength of association is mainly determined by the number and salience of associations linked to each brand. The authors show that the degree of spillover at the parent brand is positively related to the strength of sub brand-parent brand association. Moreover, they found that the degree of spillover depends on the total strength of association between sub brands, instead of the strength of filial/commonality linkages between sub brands and parent brands.

(18)

Chapter 3: Valence and amplitude of information

The previous chapter elaborated on celebrity endorsement. In this chapter, more attention is given to the impact of valence (positive and negative) and the amplitude (moderate or extreme) of information on the celebrity and brand.

3.1 Valence of information

Negative information is better recalled than neutral information on the long term (Kensinger and Corkin, 2003 in Money et al., 2006). Moreover, negative behaviour from celebrities receives more attention, is better encoded and is remembered easier than positive information (Folkes, 1988; Ybarra and Stephan, 1996 in Money et al., 2006).

3.1.1 On attitude towards the brand

Till and Shimp (1998) found that negative celebrity evaluation led towards a lower evaluation of the brand, which is influenced by the number of associations about the brand and the celebrity, the strength of the link between the two and the timing of the negative information. The impact of the negative information was bigger after the awareness of the endorsement by the consumer. Negative evaluation of the celebrity could even affect competing brands or the product category negatively (Dahlen and Lange, 2006 in Zhou and Whitla, 2013). Moreover, Langmeyer and Shank (1993, in Money et al., 2006) show that a negative view towards the celebrity leads to negative observations about a non-profit agency the celebrity was endorsing.

However, not all the researchers agree on the influence of negative information about the endorser on the brand. Money et al. (2006) argue that negative self-oriented information, that only impacts the celebrity him- or herself, actually leads to a more positive perception of

(19)

negative publicity. They found that high committed consumers to a brand declined any negative information to conserve their established attitudes, while low committed consumers tend to argue less about the information presented and base their attitudes towards the brand on the information valence.

3.1.2 On attitude towards the celebrity

An example of negative information about the celebrity on the attitude towards the celebrity is that of supermodel Kate Moss, in which Chanel and Burberry stopped her from endorsing their products after she was alleged for cocaine usage (Wall Street Journal Online, 2005 in Money et al., 2006). Negative information about the celebrity has a significant influence on the credibility and attractiveness of celebrities across different product categories (Thwaites et al., 2012). However, when negative information was exposed to respondents about a celebrity endorsing a product, a negative transference of affect took place towards the product, but the effect of negative information about the brand on the celebrity was mitigated (Runyan et al., 2009).

The authors argue that psychological involvement and attachment theory causes strong relational ties between consumers and celebrities (Atkins and Block, 1983 in Runyan et al., 2009), whereby consumers view a sense of closeness or a type of friendship with a celebrity, which makes it harder to blame them in negative conditions (Ross, 1977, in Runyan et al., 2009). Zhou and Whitla (2013) suggest that when consumers perceive a celebrity responsible for their own actions, consumers directly blame the celebrity for damage to society. Conversely, when the celebrity is seen as a victim of circumstances, consumers consider the damage to society to be out of the celebrity's control. This is partly in line with a theory Money et al. (2006) introduce, in which the behaviour of celebrities is attributed to situational factors. They state that in their research undergraduate students, exposed to self-oriented information, did not commit the fundamental attribution error. The fundamental attribution error states that

(20)

“individuals are overly inclined to attribute another person’s problems or faults to his or her dispositional traits (e.g., he is weak person who lacks the willpower to avoid succumbing to drug usage) and to underestimate the effect of situational factors (e.g., environmental influences) in determining that behaviour” (Ross, 1977, in Money et al., 2006, p.120). The respondents did not ascribe the cocaine usage to the character of the celebrity, but to situational causes like group norms, which justifies the behaviour of the celebrity that would have been considered unacceptable in other environments. This discharge from negative news among celebrities’ damages them less than ordinary people (Money et al., 2006).

Money et al. (2006) suggest another reason why celebrities do not really suffer negative information, namely the sympathy effect. They indicate that respondents felt sympathy for the celebrity endorser with problems concerning cocaine usage and probably identified with the feelings of the celebrity after they became aware of his or her circumstances, which is the core of sympathy (Wispé, 1986, in Money et al., 2006). Sympathy and empathy are related to forgiveness that helps people to deal with un-forgiveness, which is considered as stressful. Forgiveness helps to reduce this stress of un-forgiveness. Hence, participants in the negative self-orientated information, that did not destroy the private and professional lives of the celebrities, group might have become even more positive toward the celebrity because of the sympathy they experienced (Money et al, 2006).

3.2 Amplitude of information

Some studies, besides the valence of information, focus on the amplitude of the information as well. For instance, the evaluation of negative information tends to be stronger than that of

(21)

information (Liu, Wang and Wu, 2010). Bhattacharya and Sen (2003) argue that the identification of an individual with a company can be very powerful, if the consumers do not have a formal membership, like employees have. They give the following definition for consumer-company identification: "the primary psychological substrate for the kind of deep, committed, and meaningful relationships that marketers are increasingly seeking to build with their customers (p. 76)". When consumers experiences identification with a company, they are more likely to have positive feelings and thoughts about it (Einwiller et al., 2006). An increase in the identification the consumer experiences with a company leads to a higher likability of favourably informal word-of-mouth communication about the company (Brown et al. 2005, in Einwiller, 2006).

Even in case of negative information, strongly identified consumers want to defend and preserve their positive beliefs about the company (Bhattacharya and Sen, 2003). The authors state that “just as consumers are highly likely to forgive themselves for minor mistakes, they will forgive the companies they identify with, particularly because identification leads them to trust the company and its intentions (p.84)”. Moderately negative publicity about the company is moderated by the identification consumers have with the company, however, with the exposure of extremely negative information weakly and strongly identified consumers had the same reactions towards the negative information (Einwiller et al., 2006). Bhattacharya and Sen (2003) support that; when the information about a company is extremely negative, strongly identified consumers are less likely to keep a positive view towards the company. Information that is too negative does not serve as a source of positive social identity anymore.

(22)

Chapter 4: Endorser types and characteristics

The previous chapter focused on the valence and amplitude of information, in this chapter attention is given to the different type of endorsers. Moreover, a comparison between the YouTube- and traditional celebrities is made.

4.1 Different types of endorsement

Companies have two options when choosing endorsers: they can opt for a celebrity or a non-celebrity, a created-spokesperson. McCracken (1998, p.310) defines a celebrity endorser as “any individual who enjoys public recognition and who uses this recognition on behalf of a consumer good by appearing with it in an advertisement.” On the other hand, a created spokesperson is created by the organization itself (Tom et al., 1992 in Waldt, Loggerenberg and Wehmeyer, 2009) and they only endorse one particular product, which gives the brand a power of control over this created spokesperson (Tom, et al. 1992, in Erdogan, 1999). Conversely, the controlling power of companies over celebrity endorsers is limited, because they have created their personality themselves (Erdogan, 1999).

Created spokespersons tend to create a better link with the product than celebrity endorsers, because the celebrity is not only associated with the product they endorse, but with many other things (Tom et al., 1992, in Erdogan, 1999). This can be explained with the classical conditioning paradigm as discussed in chapter 2. Thereby, another difference is found in the cognitive response towards the endorsers. People tend to focus more on the brand and its features within non- celebrity endorsements, whereas the celebrity is the main focus within celebrity endorsements (Metha, 1994 in Erdogan, 1999).

(23)

purchase intention and a more positive attitude towards the advertisement than non-celebrity endorsers (Atkin and Block, 1983; Petty, et al., 1983 in Erdogan, 1993). Additionally, it was found that celebrities and created spokesperson were perceived similarly on levels of trustworthiness and expertise, but celebrity endorsers were perceived to be more attractive (Waldt et al., 2009). Lastly, celebrity endorsers that state an argument in advertisements cause a higher attitude towards the endorsed product than non-celebrity endorsers among low involved persons. Low involved persons do not take the effort to think about the product-relevant argumentations in an advertisement they rather consider the credibility of the product endorser for example (Petty, Cacioppo and Schumann, 1983).

4.2 Traditional celebrities compared to YouTube celebrities

“In traditional formations, followers were subjected to the adulation of culted figures through the anonymous and technocratic efforts of promoters, propagandists and specialists of mass-media technologies, and followers had little involvement in negotiating the imposed personality of these figures. Conversely, personality becomes an item of co-creation in new cults, and adulation becomes elective as contact is made through ostensibly intimate, amateur and do-it-yourself (DIY) technologies. This deepens conversation around and provides an alternative explanation to recent conceptualizations of the influence and power of media personalities in marketing scholarship” (Cocker and Cronin, 2017, p.13).

This quote sums up most the differences between traditional and YouTube celebrities. YouTube is a content community that was founded in 2005, in which users can “post, view, comment and link to videos on the site” (Smith, Fischer and Yongjian, 2012, p.104). Users that register at YouTube display the channels and friends they are subscripted to, recent activities, comments and favourite videos (Smith et al., 2012). The following paragraphs elaborate on the differences between the YouTube and traditional types of celebrities.

(24)

4.2.1 The exceptional versus the ordinary

YouTube celebrities might seem to relate to the traditional celebrity, but they actually differ. YouTube stars compared to traditional celebrities “promote the ordinary over the exceptional” (Gamson, 2011, p.1022). Smith (2014) emphasizes that the YouTube celebrity is seen as an ordinary person that is changed into an extra-ordinary person when he/she become famous online. This fame is mostly based on self-presentation in vlogs: “all vlogger’s have audiences and are stars of their own reality show, their vlogs” (Smith, 2014, p.260).

A celebrity receives his or her status based on an ascription that often is predetermined, achievements via talents and/or accomplishments or the attribution for being exceptional (Rojek, 2001 in Cocker and Cronin, 2017). Conversely, YouTube stars did not achieve or get ascribed their status in the same way, but rather their “day to day actions, thoughts and opinions; purchases and experiences; and interactions with others are packaged and communicated to audiences as their ostensible product” (Cocker and Cronin, 2017, p.3). YouTube’s slogan “Broadcast Yourself” mirrors the self-presentation and self-promotion of the platform’s creators. The slogan also inspires its creators to be the hub of attention in their videos and YouTube creates an environment in which development of micro-celebrities is supported (Smith et al., 2012).

A distinction between celebrity and fame is made: “the celebrity is about personality or persona; fame is circulation of the name of the person who is the concrete embodiment of these traits…and contemporary culture has made a wholesale shift from fame to celebrity, ‘fame emptied of content’” (Elliot, 2011, in Smith et al., 2014, p.259). Keel and Nataraajan (2012)

(25)

does not have talents or qualities that are admired like for the traditional celebrity. The qualities of YouTuber stars are socially activated and deconstructed by their followers and do not really exist on their own (Belk, 2013, in Cocker and Cronin, 2017). YouTube is one of the platforms in which anybody can be a reality star on the Internet (La Ferla, 2010, in Keel and Nataraajan, 2012): “‘anyone could upload anything on YouTube’ versus ‘being a Hollywood movie star is out of the question to most of us’” (Smith, 2014, p.260). Additionally, YouTube celebrities are only famous to a small base of fans (Gamson, 2011). For instance, only viewers that like to watch make-up videos and are interested in makeup perceive the self-made beauty expert on YouTube as a celebrity (Keel and Nataraajan, 2012).

4.2.2 Trust and authenticity

Consumers consider social media to be a more trustworthy source of information about products and services compared to traditional media sources (Foux, 2006, in Mangold and Faulds, 2009). Thereby, YouTuber’s autonomy causes trust and authenticity: “microcelebrity comprises a sense of self-government. Because of the widespread conception of YouTube as a bottom-up social media platform, it is believed that young microcelebrities are not enlisted into the powerful and commercialized system sustaining celebrity culture” (Gamson, 2011; Marwick, 2013 in Jerslev, 2016, p.5240). The traditional celebrities have been “cultivated strategically through traditional mass-aggregate media channels and ‘celebrity industries’ (including publicists, agents and managers)” (Turner, 2004, in Cocker and Cronin, 2017, p.3). On the contrary, YouTube celebrities did bypass this tightly controlled publicity system (Gamson, 2011), although successful YouTubers joined social media talent agencies to help them manage their charismatic personalities, create publicity and negotiating deals with companies (Jerslev, 2016). TV personalities are described as “a coalescence, a composite, a conflation of primal prototypes” (Hackley et al., 2012 in Cocker and Cronin, 2017) and YouTube celebrity as “a

(26)

careful and outwardly transparent or (seemingly) authentic refraction, emulation and bastardization of followers’ own sense of self, normalcy and centeredness” (p.13).

4.3.3 Interaction and co-constitution

Consumers feel more engaged with products and organizations when they can give feedback (Mangolds and Faulds, 2009). YouTube, compared to traditional broadcasting, has an interactive and conservational character, in which YouTubers can respond to their viewers with text comments (Tolson, 2010). YouTuber’s communities connect followers, in which the creators receive feedback via the comment section as well (Thornton, 2014).

The fame of YouTubers depends on this level of interaction with the community (Cocker and Cronin, 2017); when the creators regularly read and react to comments they tend to get more views and retain their subscribers (Chau, 2010). Vlogging features a conversational response in which there is a distinction with traditional media via ‘emphasis on liveness,

immediacy and conversation’ (Burgess and Green, 2009, in Tolson, 2010, p.279). YouTube

celebrities display immediacy with videos that are being uploaded regularly and can be watched via portable devices (Marshall, 2010 in Smith, 2014), which leads to more engagement and sense of involvement. Besides this direct para-social contact, there is a “collectively felt sentiment that it is the followers themselves who are not just the recipients, but the custodians, of their favourite YouTubers’ personalities” (Cocker and Cronin, 2017, p.7).

Shortly, next to consuming content, followers can create content and YouTube stars as well (Thornton, 2014). YouTube celebrities are self-made and created by their audience (Gamson, 2011). The followers have the impression that they know the YouTuber, partly because they

(27)

uploads. They keep their private lives private and rarely have their photographs taken outside of official events” (Jerslev, 2016, p.5238). On the contrary, this scarcity and distance is barely visible among the YouTube stars that mainly display accessibility, intimacy and connectedness online (Jerslev, 2016) and even film in their living room (Gamson, 2011).

According to Gamson (2011, p.1063) “the most common narrative strategy— ‘come see what they’re really like’—invites identification with celebrities” and gives the imagination of intimacy with the famous (Dyer, 1979; Schickel, 1985, in Gamson, 2011). This sense of identification can be illustrated with YouTube vloggers that help build long-term relationships with the consumers for luxury brands. Consumers that consider vloggers to be like themselves, are proven to have the same positive attitude towards the brand as the vlogger (Lee and Watkins, 2016). YouTube facilitates virtually tourism “which allows us to extend our senses beyond the range of our body’s geographic environment, introducing us to people and places, sights and sounds that we would not otherwise have the opportunity to perceive” (Grusin 2009, in Thornton, 2014, p.2). The costliest asset of YouTube stars is the creation of videos that draw attention, in which they show their private authentic self (Jerslev, 2016).

(28)

Chapter 5: Conceptual Model and Hypotheses

In this chapter hypotheses are developed, based on the theoretical framework that helps answering the research question. A conceptual model is presented to illustrate the concepts and the relationships between these concepts. Accordingly, concepts of the model are elaborated and hypotheses are formulated, based on assumptions derived from the theoretical framework.

5.1 Conceptual Model

As seen in the literature review, previous research has mainly studied the impact of, mainly negative, information of the celebrity on the brand or information about the brand on the celebrity. However, the impact of information of the celebrity on attitude towards the celebrity and the impact of valence and amplitude of information about a celebrity taken together within the field of brand endorsement remains largely undiscovered. Thereby, the difference between YouTube and traditional celebrity endorsers has not been studied yet within this context. This thesis studies the difference in attitude towards the celebrity and the brand before and after

exposure to a news item, differentiated by Valence and Amplitude of the information and type of celebrity (see figure 3).

Figure 3: Conceptual model including the relationship between valence, amplitude, type of celebrity, attitude towards the celebrity and attitude towards the brand.

(29)

5.1.1 Difference in attitude towards the celebrity and the brand

The dependent variables attitude towards the celebrity and attitude towards the brand are influenced by the independent variable of news item, which is differentiated by Valence, Amplitude and Type of Celebrity. The variable Valance of information is moderated by Amplitude of information, which in turn is moderated by Type of Celebrity. Besides serving as a dependent variable, attitude towards the celebrity mediates the attitude towards the brand. The dependent variables depend on the total of eight possible combinations of Type of Celebrity, Valance and Amplitude of information.

5.1.2 Valence and amplitude

The exposure to a news item impacts the difference in attitude towards the celebrity and the brand. This news item is moderated by Valence (positive or negative), which is moderated by Amplitude (moderate or extreme). Amplitude enhances or decreases the impact of Valence.

5.1.3 Type of celebrity

Successful YouTube creators are transformed into celebrities via social media. The comparison between the types of celebrity (YouTube and traditional) enduring different levels of information is the core of this thesis, since YouTube celebrities differ from traditional celebrities on a few fronts. A higher level of identification with a YouTube celebrity causes this difference. Customers might therefor perceive different levels of information differently for YouTube- and traditional celebrities and accordingly create a different attitude towards the celebrity and the brand.

5.2 Hypotheses

The theoretical framework together with the derived conceptual model helps to develop the hypotheses. The relationships between the variables in the conceptual model is tested via the hypotheses.

(30)

5.2.1 Valence on attitude towards the celebrity and attitude towards the brand

Most of the literature in which the association transfer from brands to brands (e.g. Lei et al., 2008) or the celebrity towards the brand has been studied, investigated this with negative information (e.g. Runyan et al., 2009). For instance, Till and Shimp (1998) found that negative celebrity evaluation led towards a lower evaluation of the brand. However, the transfer of positive information about celebrities on attitude towards the celebrity has barely been studied yet. The following hypothesis is developed to test if the exposure to a news item, moderated by positive and negative valence, impacts the difference in attitude towards the celebrity:

• H1: Positive valence leads to a more positive difference in attitude towards the celebrity

before versus after the exposure to the news item than negative valence.

The theoretical framework provided that celebrities transfer their meaning and values towards the brand they endorse (e.g. Till and Shimp, 1998; Keller, 2003; 2005). This happens via the meaning transfer model (McCracken, 1989), also known as associative learning (Abimbola et al., 2010) and classical conditioning, wherein consumers learn associations between the endorser and the product via multiple exposures (Tom et al., 1992, in Erdogan, 1999). The association sets of the celebrity and the brand consist of nodes about the brand and the celebrity that are connected to other nodes due to experiences with the two (Till and Shimp, 1998). The better the fit between the celebrity and the brand is considered, the faster an associative link can be developed (Thwaites et al., 2012). Therefore, the following hypothesis is created:

• H2: A favourable attitude towards the celebrity leads to a more positive attitude towards

(31)

information in consideration is that of Einwiller et al. (2006). They show that moderately negative publicity about a company is moderated by the identification consumers have with the company. In short, strong identification immunizes consumers against moderate negative publicity. Strongly identified consumers want to defend and preserve their positive beliefs about the company. Besides, they are more likely to forgive the company, because identification causes trust towards the company and its intentions (Bhattacharya and Sen, 2003). However, with the exposure of extremely negative information weakly and strongly identified consumers have the same reactions towards the information (Einwiller et al., 2006; Bhattacharya and Sen, 2003). The impact of amplitude is tested with the following hypothesis:

• H3: The interaction pattern between exposure to the news item and valence (H1) is more

pronounced when amplitude is extreme than when amplitude is moderate.

5.2.3 YouTube- compared to traditional celebrities

YouTube stars are seen as ordinary persons that change into extra-ordinary persons when they become famous online (Smith, 2014). Social media, like YouTube, is considered to be a more trustworthy source of information about products and services by consumers compared to traditional media sources (Foux, 2006, in Mangold and Faulds, 2006). YouTuber’s autonomy causes trust and authenticity (Jerslev, 2016). It is believed that YouTube celebrities are not enrolled in the powerful celebrity culture, which is perceived as commercialized and tightly controlled by publicity systems (Jerslev, 2016; Gamson, 2011). Whereas traditional celebrities keep their private life private and cultivate a distance, YouTube stars mainly display accessibility, intimacy and connectedness online (Jerslev, 2016). This “come see what they’re really like” appeals to identification with celebrities (Gamson, 2011, p.1063). Moreover, YouTubers interact with their viewers via text comments underneath their videos (Tolson, 2010). It has been found that consumers feel more engaged with products and organizations when they can give feedback (Mangolds and Faulds, 2009). YouTube’s interactive and

(32)

conservational character gives followers the impression that they know the YouTuber, partly because they have helped to co-create and co-constitute the YouTuber (Cocker and Cronin, 2017).

As stated before, strong identification immunizes consumers against moderate negative publicity, but weakly and strongly identified consumers have the same reactions towards extremely negative information (e.g. Einwiller et al., 2006). Taking the aforementioned characteristics of the YouTube celebrity into consideration, YouTube stars are easier to identify with in comparison to traditional celebrities. Therefore, people are more immunized towards a moderately negative news item about the YouTube celebrity, compared to the traditional celebrity (e.g. Einwiller et al., 2006). However, this effect disappears in case of exposure towards an extremely negative news item. Thus, the difference in attitude towards the celebrity before versus after the exposure towards a moderately negative news item is probably smaller for YouTube celebrities than traditional celebrities, but this difference in attitude is similar for both celebrities in case of extremely negative information. Thus, the overall interaction pattern between Amplitude and Type of Celebrity is less pronounced for the YouTube celebrity compared to the traditional celebrity. The following hypothesis is developed:

• H4: In case of exposure to a negative news item, the interaction pattern between amplitude

and type of celebrity is less pronounced for the YouTube celebrity than for the traditional celebrity.

(33)

Chapter 6: Methodology

In this chapter, the methodology is presented. Firstly, requirements of the stimuli for the experiment are explained based on the theoretical framework. A focus group and pre-test helped in finalizing the stimuli and in creating the questionnaires. Subsequently, the research design is elaborated and shows how the data is collected.

6.1 Stimuli development

The stimuli were developed from the theoretical background. The independent variables of this study are type of celebrity, valence and amplitude of information. These variables are manipulated in order to create eight different stimuli. These stimuli are tested within the category of personal care products. Product care products have been chosen, because a lot of brand endorsements take place within the market of personal care products. This makes the questionnaire more realistic for the participants and eases answering the questions.

The stimuli require certain elements in order to strengthen the reliability. Before the main experiment questionnaire could be set up, the stimuli needed to be discussed in a focus group and tested in a pre-test. Within a focus group (n=5), different combinations of celebrities, brands and type of news stories, based on positive/negative valence and moderate/extreme amplitude, were discussed. The focus group of 5 people aged between 20 and 30 was gathered via convenience sampling.

6.1.1 Stimuli requirements

In the focus group, together with the theoretical background, the following requirements for the stimuli have been determined. Firstly, the brand and the celebrity in the experiment should be well known. The brand and the celebrity need to hold a relatively high awareness amongst the respondents, in order to compare the celebrity and brand accurately. Secondly, the brand and the celebrity should be perceived neutral in order to prevent extreme positive or negative

(34)

associations among the respondents, because these associations influence the way people look at the brand and the celebrity. In this way, it can be tested how a positive or negative story about the celebrity changes the attitude of the participants and the image of the brand (Runyan et al., 2009). Preferably, the celebrity needs to be perceived positive and meaningful (McGuire, 1969; Till and Shimp, 1998). Lastly, the information about the celebrity should be realistic; the respondents need to have the impression that the fictional news story could be true.

6.1.2 News items

With the help of the focus group and the aforementioned requirements, the brands Elvive and Lancôme had been chosen to study in the pre-test. Brand endorsement for Lancôme was dedicated to their facial day care cream line and for Elvive to their shampoos. The traditional celebrities that were chosen with the help of the focus group were Angelina Jolie and Beyoncé. Angelina Jolie is an American actress and Beyoncé is mainly known as an American singer and songwriter. The YouTube celebrities that took part in the pre-test were Mascha Feoktistova, better known as Beautygloss, and Anna Nooshin. They are two of the biggest Dutch YouTube stars in the category of fashion, lifestyle and beauty. Both the YouTubers post videos on their channels on a weekly basis and have won vlogging awards. British YouTuber Zoella was taken into consideration for the experiment as well. Due to the higher number of Dutch students that would probably take part in the questionnaire, Dutch YouTubers were considered to better fit the experiment, since the celebrity should be well known. The following scenarios (see table 1) were developed for the pre-test:

(35)

Table 1: Scenarios used in the pre-test. Elvive shampoo of L’Oréal

VAL E NCE

AMPLITUDE TYPE OF CELBRITY

YouTube Traditional

Positive

Moderate 1. Anna donates €3.000 to _charity 2. Beyoncé donates €3.000 to _charity

Extreme 3. Anna organizes a benefit _{gala to raise money for charity} 4. Beyoncé organizes a benefit _{gala to raise money for charity}

Negative

Moderate 5. Anna cheats on her _boyfriend 6. Beyoncé cheats on her husband

Extreme 7. Anna is accused of domestic _violence 8. Beyoncé is accused of domestic _violence

Lancôme facial day crème

VAL

E

NCE

Positive

Moderate 9. Mascha becomes an _{ambassador for Pink Ribbon} 11. Angelina Jolie becomes an _{ambassador for Pink Ribbon}

Extreme 11. Mascha makes a movie _{about climate change} 12. Angelina Jolie makes a movie _{about climate change}

Negative

Moderate 13. Mascha is sniffs cocaine _{in a club} 14. Angelina Jolie sniffs cocaine _{in a club}

Extreme

15. Mascha has been involved in a car accident, while being drunk

16. Angelina Jolie has been involved in a car accident, while being drunk

6.2 Research design: Pre-test

In order to test the stimuli, the aforementioned news scenarios were presented in a pre-test to check which stimuli would be best suited for the main experiment. This was needed to ensure the reliability of the main experiment. The pre-test was distributed online via Qualtrics, an online platform to fill out surveys, and was filled out by 64 people. The pre-test experiment had a within-subject design, whereby the participants were randomly assigned to four different scenarios out of the total of sixteen scenarios. In these scenarios, a news messages about Anna Nooshin and/or Beyoncé related to Elvive or Mascha Feoktistova and/or Angelina Jolie related to Lancôme was presented. After general information about the brand and celebrity, this news scenario was presented to the respondents. After each news item, they were asked a few

(36)

questions on a 7-point-Likert scale, ranging from (1) strongly disagree to (7) strongly agree, about the celebrity, brand and the news item (for the questionnaire, see Appendix A). These questions (see table 5) were related to the perceived realism and understanding of the news item and the degree of perceived valence and amplitude of the information. First, two questions were posed with regard to the familiarity with the brand and the celebrity. Secondly, two questions measuring the perception of the celebrity as a traditional celebrity and as blogger/vlogger were asked. Accordingly, a question about perceived realism and one about the understanding of the message were included in the questionnaire. Lastly, three questions were asked in order to check for perceived valence of the news item and four questions for perceived amplitude.

With the results of the pre-test, the scenarios for the main experiment were selected. In order to do this, the familiarity with the celebrity and the brand preferably had to score above the midpoint of the 7-point Likert scale (>4). Thereby, the traditional celebrity should score above the midpoint of the scale for the perception of traditional celebrity and below the midpoint of the scale for the perception of blogger/ vlogger and vice versa for the YouTube celebrity.

Table 2: The familiarity of the celebrities and brand and the perception as traditional celebrity and blogger/vlogger for the celebrities in percentages.

Familiar with Perceived as traditional celebrity Perceived as blogger/ vlogger Beyoncé (n=64) 98.40% 96.90% 3.1% Angelina Jolie (n=57) 83.90% 94.60% 3.6% Mascha Feoktistova (n=57) 35.10% 5.30% 64.90% Anna Nooshin (n=64) 57.80% 23.40% 81.20% Elvive (n=128) 90.60% - - Lancôme (n=114) 89.40% - -

(37)

blogger/ vlogger than Mascha Feoktistova. Since the questions regarding the news items only focused on realism, understanding, valence and amplitude of information, the celebrity and the brand were compatible for the news items used in the main experiment.

In order to select the most representative scenarios to measure the independent variables, a comparison between the 16 scenarios had to be made on the level of realism, amplitude and valence. The higher the score of perceived realism on the 7-point Likert scale, the better. Scenarios with positive valence and/or high amplitude were aimed to score above the midpoint of the scale (>4), whereas negative valence and/or low amplitude were aimed to score below the midpoint of the scale (<4). Amplitude turned the final scale and preponderated when choosing between two scenarios. A One Sample T-Test was conducted in order to check for significance difference of the midpoint of the scale, namely 4 (see table 3). Regardless the fact that the moderate scenarios scored above the midpoint of the scale of 4, instead of below, they were still considered as accurate news items, because they did not score significantly above 4.

Table 3: One sample T-test for the winning scenarios for main experiment: mean scores and standard deviations for Valence, Amplitude and Realism.

VAL

E

NCE

YouTube Traditional Positive Moderate Scenario 1 (n=13) Valence: 5.82 (0.70)* Amplitude: 4.54 (0.83) Realistic: 5.62 (0.77)* Scenario 2 (n=15) Valence: 5.93 (1.16)* Amplitude: 4.52 (1.22) Realistic: 5.53 (1.64)* Extreme Scenario 11 (n=14) Valence: 5.88 (0.43)* Amplitude: 4.71 (0.93)* Realistic: 4.57 (1.56) Scenario 12 (n=14) Valence: 6.02 (0.92)* Amplitude: 5.39 (1.00)* Realistic: 6.07 (1.00)* Negative Moderate Scenario 13 (n=14) Valence: 1.71 (0.69)* Amplitude: 4.25 (0.84) Realistic: 4.36 (1.00) Scenario 14 (n=17) Valence: 2.00 (1.47)* Amplitude: 4.46 (1.15) Realistic: 4.12 (2.00) Extreme Scenario 15 (n= 14) Valence: 1.55 (0.59)* Amplitude: 4.73 (0.83)* Realistic: 3.64 (1.60) Scenario 16 (n=14) Valence: 1.50 (0.43)* Amplitude: 5.14 (0.64)* Realistic: 3.71 (1.54) *Significant at the proposed direction, p = <.05 =4

(38)

Based on analysis of the pre-test, some elements of the questionnaire and the stimuli required change before the main experiment could be conducted. First of all, the text regarding the introduction about the celebrity and brand was shortened, because a few respondents considered it too long. Secondly, the amount of €3000 was doubled to €6000, since some respondents commented that they considered the amount a little too low. Lastly, when looking at the ‘Chronbach’s α if item deleted’ (see appendix A) it is recommended to delete item 2 of the scale to increase the overall Chronbach’s alpha from .57 to .60. Therefore, item two “I perceive the

information as important” was deleted from the scale of Amplitude for the main experiment.

6.3 Research design: Main Experiment

A 2 (Valence: positive/negative) x 2 (Amplitude: moderate/extreme) x 2 (Type of celebrity: YouTube/traditional) design was used for the main experiment to test the hypotheses. A between-subject design was used, which is known as an independent design; “an experimental design in which different treatment conditions utilize different organisms and so the resulting data are independent” (Saunders et al., 2009, p.787). Since the respondents only go through one treatment, the carry over effect and the contamination effect of irrelevant factors is limited. The following news items have been chosen for the main experiment.

Table 4: News items for the main experiment.

VAL

E

NCE

Positive

Moderate 1. Anna Nooshin donates €6.000 _{to charity} 2. Beyoncé donates €6.000 to _charity

Extreme 3. Anna Nooshin makes a movie _{about climate change} 4. Beyoncé makes a movie _{about climate change}

(39)

6.3.1 Procedure

First of all, the respondents were welcomed on the online platform Qualtrics with a short introduction about the questionnaire. Accordingly, the participants were presented a photo and general information about the celebrity, the brand and the brand endorsement. After this short introduction, they were asked questions concerning their familiarity with the celebrity and the brand, whether they perceived the celebrity as a traditional celebrity or blogger/vlogger and attitude towards the celebrity and the brand.

Subsequently, they were randomly divided into one of the ten groups. These ten groups were based on the eight scenarios (see table 4) and two control conditions, one for a YouTube celebrity and one for the traditional celebrity. Each celebrity, YouTube- or traditional, represented 4 fictional news items. For the control condition, the participants were not presented a scenario, but were only asked general questions regarding: familiarity, perception of the celebrity as a traditional celebrity and blogger/vlogger, attitude towards the brand and celebrity and questions about the brand endorsement. The respondents assigned to a fictional news scenario, had to read the concerned news item and had to fill out a questionnaire concerning questions about the news item (i.e. amplitude, valence and realism of the information) and attitude towards the celebrity, brand and brand endorsement. Lastly, the respondents were thanked for filling out the survey and a disclaimer was shown at the end (for the questionnaire and news items, see appendix B).

6.3.2 Measurements

A 7-point Likert scale is used for measuring all the items in this experiment, except for the ones regarding cognition and demographics. A 7-point Likert scale “allows the respondent to indicate how strongly she or he agrees or disagrees with a statement” (Saunders, Lewis and Thornhil, 2009, p.594). In the following table (see table 5), an overview of the role of the variables, the source the items were adopted from or based on and the items used, are presented.

Consumer responses to different news items for traditional- versus YouTube endorsers : are consumers more immune to moderately negative publicity about YouTube celebrities than traditional celebrities?