Are bloggers more effective endorsers than celebrities? : finding the right match between endorser and medium

Are Bloggers More Effective Endorsers than Celebrities?

Finding the Right Match between Endorser and Medium

Student: Melissa Lai

Student number: 11375019

University: University of Amsterdam

Faculty: Economics and Business

Supervisor: Roger Pruppers

Study program: MSc in Business Administration

Submission date: June 28

th

, 2017

Specialization: Marketing

Version: Final

Statement of Originality This document is written by Student Melissa Lai 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.

Abstract As the use of celebrity endorsers in marketing communication has become widespread, a new type of endorser has emerged with recent advances in online media: the blogger. Regarded as today’s digital influencers, the growing power of bloggers has caused brands to shift to them as a potential new avenue for endorsements. Due to this being a relatively new phenomenon, though, the use of bloggers for advertising purposes and their effectiveness as endorsers remains largely unexplored. Adding to this, in considering blogger and celebrity endorsers, academic literature has yet to take into account the effect of the advertising medium in which the endorsement is communicated through. While previous studies have automatically considered each endorser in its typical medium, namely celebrities in traditional media and bloggers in social media, brands are increasingly attempting to use their endorsers across different media. The present study, therefore, focuses on the effect of the two different types of endorsers and different advertising media on message credibility and consumer attitudes. A 2 (Endorser type: blogger/celebrity) x 3 (Medium type: personal social media/brand social media/print) between-subjects experiment was conducted. 196 individuals participated in the online questionnaire. The results showed no significant main effect of endorser type or medium type. However, a significant interaction was found with messages endorsed by the blogger through her personal social media account and messages endorsed by the celebrity in a print ad being perceived as the most credible. An important distinction was also found in the sender of the message in terms of the endorser versus the brand. Theoretically, this suggests that a type of endorser and its respective medium are indeed closely linked. From a managerial perspective, the findings highlight the importance for brands to find the right combination of endorser, medium, and sender.

Table of Contents 1. Introduction 1 1.1 Celebrity Endorsers and the Shift Towards Bloggers 1 1.2 Problem Definition 3 1.2.1 Problem Statement 3 1.2.2 Sub-Questions 4 1.2.3 Delimitations 4 1.3 Contributions 4 1.3.1 Theoretical Contributions 4 1.3.2 Managerial Contributions 5 1.4 Outline 6 2. Brands and Secondary Associations 7 2.1 Brand Knowledge 7 2.2 Leveraging Secondary Sources of Brand Knowledge 8 3. Endorsers in Marketing Communication 10 3.1 Endorsement Models 10 3.1.1 Classical Conditioning 10 3.1.2 The Match-up Hypothesis 11 3.1.3 The Source Credibility and Source Attractiveness Models 11 3.2 Celebrity Endorsement 13 3.2.1 The Meaning Transfer Model 13 3.2.2 Effectiveness of Celebrity Endorsers 13 3.2.3 Negative Aspects of Celebrity Endorsement 15 3.3 Bloggers 15 3.3.1 Bloggers as Opinion Leaders 16 3.3.2 Bloggers as Peer Endorsers 17 3.3.3 Sponsored Recommendation Posts 17 4. Advertising Media 19 4.1 Traditional vs. Online Media 19 4.2 The Impact of Different Media 20 4.3 Media Credibility 21 4.4 The Sender 22 5. Hypotheses 23 6. Methodology 27 6.1 Stimuli Development 27 6.2 Qualitative Pretest 28 6.3 Quantitative Pretest 28 6.4 Main Experiment 31 6.4.1 Research Design 31 6.4.2 Procedure 31 6.4.3 Measures 32 7. Results 34 7.1 Respondent Analysis 34 7.2 Reliability Analysis 34

7.3 Manipulation Checks 35 7.3.1 Perceived Endorser Type 35 7.3.2 Perceived Sender of Message 37 7.4 Control Variables 39 7.5 Hypotheses Testing 41 7.5.1 Hypothesis 1 42 7.5.2 Hypothesis 2 43 7.5.3 Hypothesis 3 44 7.5.4 Hypothesis 4 45 7.6 Additional Analyses 46 7.6.1 Endorser vs. Brand as Sender 46 7.6.2 Hypotheses Testing with Control Variable 49 7.6.3 Perceived Credibility of Endorser 51 7.6.4 Attitude towards the Brand 53 8. Discussion 54 8.1 Interpretation of Results 54 8.1.1 The Effect of Medium Type 54 8.1.2 Interactions between Endorser Type and Medium Type 55 8.1.3 The Effect of Sender Type 56 8.1.4 Attitude towards the Endorser 57 8.1.5 Attitude towards the Message and Brand 58 8.2 Theoretical Implications 58 8.3 Managerial Implications 60 9. Conclusion 63 9.1 Limitations 64 9.2 Recommendations for Further Research 65 10. References 67 11. Appendix 74

Appendix A: Quantitative Pretest Questionnaire 74 Appendix B: Main Experiment Advertisements per Scenario 81 Appendix C: Main Experiment Questionnaire (for Scenario 1) 84 Appendix D: Perceived Sender of Message Manipulation with Brand Social 87 Media and Print Ad as Brand as Sender Appendix E: PROCESS Models for Hypothesis 4 with Effect Sizes and Significance 88

Chapter 1: Introduction 1.1. Celebrity Endorsers and the Shift Towards Bloggers With the immense amount of information and advertising messages directed at consumers daily, companies have increasingly turned to endorsers, specifically celebrity endorsers, to help their products stand out from the surrounding competition (Erdogan, 1999). Famous faces such as David Beckham and Jennifer Aniston are often featured alongside products in order to attract the attention of consumers. A study by Millward Brown found that 15% of advertisements in the U.S. employ a celebrity endorser, with this number reaching above 40% in some markets such as Japan and Taiwan (Crutchfield, 2010). The popularity of this marketing tactic can be seen in the considerable amounts of money companies spend to secure and build advertising campaigns around celebrities today. For instance, Academy Award winner George Clooney is estimated to have earned a total of $50 million from representing Nespresso since 2005 (“10 of the Highest Paid Celebrity Endorsement Deals,” 2015). As the value of these endorsement deals continues to rise, companies are dedicating larger and larger portions of their advertising budgets to aligning their brand with the most recognized celebrities (Agrawal & Kamakura, 1995). According to substantial research, these investments are worth the money. In examining the effectiveness of celebrity endorsements, it has been shown that celebrities can help increase rates of recall and brand name recognition (Petty, Cacioppo, & Schumann, 1983; Friedman & Friedman, 1979), as well as enhance positive evaluations and ultimately purchase intentions (Atkin & Block, 1983; Kamins, 1989; Ohanian, 1991). Thus, due to the fame and success that celebrities command, it can be argued that they are effective promotional tools in influencing a brand’s target audience. Despite the benefits that celebrity endorsements offer, though, O’Mahony and Meenaghan (1997) observe that celebrities may not always be regarded as particularly convincing or believable. In addition, Erdogan (1999) states that there are some potential risks associated with celebrity endorsers such as overshadowing the brand and high costs. Thus, is it possible that some other type of endorser exists nowadays with the potential to be more effective than celebrities? Turning to the Internet and social media reveals a new player in the endorsement space. With the rapid growth of these technologies, ordinary consumers have been given a voice and gained the ability to produce their own content (Labrecque, vor dem Esche, Mathwick, Novak, & Hofacker, 2013). Social media platforms have further allowed members to easily share information with connections, putting even greater power in the hands of individuals. As a result, these developments have given way to the rise and influence of modern day bloggers.

A blog, which is short for web log, is “a collection of entries by individuals displayed in reverse chronological order” (Agarwal, Liu, Tang, & Yu, 2012). Those who write blogs, bloggers, utilize this space to share experiences, express their opinions, and offer suggestions to readers. Any visit to a social networking site will unveil the vast number of bloggers today, specializing in topics ranging from beauty to food to travel and technology. As of 2009, Technorati had indexed 133 million blogs since 2002, and found an average of 900,000 blog posts being created everyday (“State of the Blogosphere,” 2010). As blogs have gained immense popularity, this has also created new opportunities for marketers. In light of this trend, companies are increasingly shifting their attention to bloggers as an additional marketing communication channel (Uzunoglu & Kip, 2014). Bloggers are being employed as the “new generation of product endorsers” (Mendoza, 2010), who with the size of their online networks are able to quickly reach and directly influence consumers. Take Negin Mirsalehi for example, a fashion blogger. One post about a new product on her Instagram account can reach up to her four million followers. Moreover, according to a study done by The Economist, this one post could earn her up to $75,000 (“Celebrities’ endorsement earnings on social media,” 2016). The unique advantage stemming from this for brands is that they can directly target a specific audience gathered around interests similar to what the blogger represents (Uzunoglu & Kip, 2014). Considering these advantages and the growing power of bloggers, it seems clear that they present a promising new avenue for endorsements. Due to the fact that blogging is a relatively recent phenomenon, though, the use of bloggers for advertising purposes remains largely unexplored. As mentioned, the past decades of literature have focused predominantly on celebrities as endorsers and the variables influencing their effectiveness in advertising. Therefore, in order to begin to fill this gap, the present study seeks to measure the impact that bloggers as endorsers have on consumers in comparison to celebrities. In examining these two different types of endorsers, a significant factor to also be considered is the advertising medium through which the endorsement is communicated. Celebrities, who can be regarded as the traditional endorsers, have long been featured in traditional media such as print and television (Erdogan, 1999), and have become closely tied to this medium as a result. Bloggers, on the other hand, are embedded in the online space and known primarily through their blog and social media posts. They have now even been deemed as the digital influencers of today (Uzunoglu & Kip, 2014). Thus, in comparing the two endorsers, not only is there a difference in the endorser, but also in the medium through which they communicate. This leads to the second gap this study aims to address. Due to the close link of each endorser with its respective medium, academic research has automatically placed celebrities and bloggers in their own respective context.

Without separating the endorser and the medium, it is unknown to what extent each contributes to an endorsement’s effectiveness. The present study, therefore, intends to explore this relationship further and the effect of each on advertisement credibility and consumer attitudes. 1.2. Problem Definition 1.2.1. Problem Statement Traditional marketing communication literature has established that the source, message, and channel are the major factors influencing how an individual processes a message (Belch & Belch, 2004). In looking at bloggers compared to celebrity endorsers, it is clear that they represent two different sources. On top of this, though, there is also a variation in the channel through which each of them delivers endorsement messages. As described, celebrities typically promote a brand through traditional media, while bloggers communicate their endorsements through the Internet or social media. Based on the communication factors, both the source and the channel could potentially have an impact (Belch & Belch, 2004). It is, therefore, necessary to disentangle the two in order to determine whether the endorser or the medium is the main driver of a message’s effectiveness. Or, perhaps, it is a combination of both. In taking the type of medium into account with the type of endorser, it is also interesting to examine if any interactions occur. It is possible that indeed the success of each of the endorsers is closely tied to the medium, meaning celebrities to traditional media and bloggers to online media. Increasingly in advertising, though, brands are also attempting to use their endorsers across different media. For example, celebrity Jessica Alba can be seen promoting the credit card brand American Express on her own social media accounts. As for fashion and beauty blogger, Kristina Bazan, her popularity has landed her as an endorser for L’Oréal, and she is now also featured in the brand’s print advertisements. As these endorsers move to new contexts different than their original medium, it remains to be verified whether this is strategic for brands. Consequently, it is worth investigating if an endorser is effective in different media or only in one particular context. Whether a blogger is seen in a traditional print or social media advertisement can be expected to affect consumers’ perceptions of the endorser differently, and the same can be said for celebrities. Thus, the conditions under which each endorser has the greatest influence need to be further examined together with the medium to determine if there is an ideal combination. Accordingly, the main question to be investigated in this study is as follows: How do bloggers compared to celebrities as endorsers and the type of advertising medium affect consumer evaluations of a message?

1.2.2. Sub-Questions In examining the main research question, the effect of the endorser and the medium first have to be separated. Comparisons between the two different endorsers and between the different media can then be made to see if a particular type of endorser is inherently more credible, or if a particular medium is viewed as more credible than others. When looking at the medium, a distinction to also consider is in the sender of the message as this could be the endorser him/herself or the brand. The key sub-question thereafter is: what sort of interactions take place between the type of endorser and the type of medium? This will help provide a more thorough understanding on what contexts a certain endorser is most credible and effective in. By evaluating bloggers and celebrities in their own medium and in new channels different than normal, it can be determined how consumer perceptions change. A certain combination of endorser and medium may prove to be the most impactful on consumers, or conversely, a not so effective combination may be found. This can then help explain why consumer evaluations of the same advertised message may differ depending on the endorser or medium used. 1.2.3. Delimitations To draw a clear distinction between bloggers and celebrities, the present study will focus on bloggers specifically who have built up their presence through writing blog posts and through social media channels. In this way, those who may have attained recognition first through other avenues will be excluded, and those who may now be considered micro-celebrities will still fall under the category of bloggers. By this definition, vloggers who blog through videos can be considered a special kind of blogger, but will be excluded from this study since their posts are not written, but consist of a different form of media content. In terms of the types of advertising media, traditional media and the Internet will be compared as this tends to be the main division and corresponds to the typical media of celebrities and bloggers, respectively (Yoon & Kim, 2001). Specifically, within the Internet, social media will be focused on as these platforms have become the primary space for delivering endorsements online (Li, Lee, & Lien, 2012). 1.3. Contributions 1.3.1. Theoretical Contributions From a theoretical perspective, this study extends the current literature beyond celebrity endorsers to a new type, namely bloggers. With the considerable research on aligning celebrities with brands,

effect on consumers. In particular, the meaning transfer theory of McCracken (1989) will be drawn upon and the source credibility and attractiveness models used to explain differences in how the endorsers are perceived. Furthermore, by taking the medium separately into account, it can be determined if differences in an endorsement’s influence are due to the endorser or the medium. Since this is an important distinction that has yet to be made in literature, this research will add theoretically by showing the significance of each. Thus, by having the endorser type and medium type as different variables, their impact on the effectiveness of an endorsed message can be identified. This then also allows the interaction between them to be studied. As virtually no research up to this point has integrated the advertising medium in with the endorsement, this interaction between different endorsers and different media remains largely unknown. It is possible that due to the characteristics and credibility of a medium, a blogger endorsing a product in one medium is seen as more credible than in another. Similarly, this can be applied to celebrities. By discerning this interaction, it can be observed how the medium contributes to how an endorser is perceived in different contexts. 1.3.2. Managerial Contributions Managerially, the present study aims to provide a clearer understanding of bloggers and their emerging role as endorsers in marketing today. Due to the long-standing practice of employing celebrities in advertising, companies are still searching for the best ways to utilize the influence bloggers wield to generate interest and drive customer action (Uzunoglu & Kip, 2014). The results from this research can, therefore, first provide proof as to whether bloggers are effective as endorsers of products and brands. More importantly, by investigating the endorsers in the context of different media, this study can provide firms with a better indication of which situations a blogger versus a celebrity will be seen as most credible and, thus, persuasive. As celebrity endorsers are becoming more active on social media and bloggers are crossing over into traditional media, the insights can help managers in determining if this is in fact an effective use of the endorsers. It could be that with social media, endorsements from bloggers still generate more positive attitudes, while celebrities have a greater influence on consumers in their original context of print ads. Thus, depending on the medium being leveraged, managers can decide when to use celebrities and when to use bloggers, or to exploit a combination of the two endorsers to achieve maximum effectiveness. In many cases, bloggers may even prove to be a more cost-effective means as collaborating with celebrity endorsers typically entails high costs (Agrawal & Kamakura, 1995).

1.4. Outline In this chapter, the rising trend of bloggers as endorsers was introduced, as well as the current gap in literature on the influence of the medium on endorsements. The theoretical framework following, Chapters 2 to 4, will cover the relevant extant literature and theories on secondary brand associations, celebrity endorsement, bloggers, and the different advertising media. In Chapter 5, the conceptual framework and hypotheses of this study are presented. Subsequently, the methodology chapter (Chapter 6) will describe the design of the experiment, which will be conducted by exposing participants to either a blogger or celebrity endorser in a print or social media advertisement. The results will then be analyzed and the hypotheses tested in Chapter 7. A discussion of the findings in regard to the main and sub-questions will follow in Chapter 8, along with their theoretical and managerial implications. Finally, the conclusion (Chapter 9) will summarize the present study and discuss the limitations of this research.

Chapter 2: Brands and Secondary Associations The following three chapters will establish a theoretical framework and provide an overview of the existing literature on the different types of endorsers and advertising media. Before discussing these topics, though, the concept of brands will first be introduced and the importance of external sources, such as endorsers and the channel, to consumers’ knowledge of a brand. Subsequently, the significant theories regarding endorsers, specifically, the use of celebrity endorsers in advertising and the research on their effectiveness will be presented in Chapter 3. Bloggers as today’s new digital influencers will then be discussed, as well as the current literature on sponsored blog posts. Lastly, Chapter 4 will be dedicated to the shift toward social media as an advertising channel and the impact that different channels have on the communication of a message. In order to provide a thorough understanding of the significance of endorsers from a branding perspective, what a brand constitutes first has to be discussed. Kotler (1991) defines a brand as “a name, term, sign, symbol, or design, or combination of them which is intended to identify the goods and services of one seller or group of sellers and to differentiate them from those of competitors” (p. 442). In line with this, companies invest substantially in marketing programs to build their brand and to ensure that its associations stand out in the minds of consumers (Keller & Lehmann, 2003). One of the tactics often used to achieve this is utilizing endorsers in their marketing communication. By means of exposure through advertising, consumers are expected to gradually link a particular endorser together with a brand, thus helping distinguish that brand. 2.1. Brand Knowledge Looking at how brands and knowledge in general are stored in consumers’ minds, the most commonly used model is the associative network memory model from cognitive psychology (Henderson, Iacobucci, & Calder, 1998; Keller, 1993). This network is constructed of a series of nodes and links, where the nodes are pieces of stored information and the links connect the nodes to each other (Keller, 1993). It is in this manner that a brand node, such as L’Oréal, exists in a person’s memory to which many brand associations, such as cosmetics, French, and quality, can be linked. When it comes to endorsers, a brand and the endorser each represent individual nodes that become associated as a link is built over time (Till & Shimp, 1998). Through a process termed “spreading activation,” an activated node can become a source of activation for other linked nodes, causing a spreading effect (Collins & Loftus, 1975). In this way, when a consumer sees and is reminded of an endorser, the brand may automatically come to mind or vice versa. It should be noted that repetition plays a key role in establishing and strengthening this association (Anderson, 1976). Repeated pairings of a brand and the endorser increase the likelihood

that the presence of one will be a good predictor of the other in memory (Till, 1998). For example, by seeing Michael Jordan in Nike advertisements repeatedly over the last thirty years, this link has been continually reinforced so that many consumers are aware of this connection and it can be easily triggered when prompted with either one of the concepts. Based on this associative network structure, Keller (1993) points out two key dimensions when considering a consumer’s knowledge of a brand: brand awareness and brand image. First and for all, the consumer has to be able to recognize and recall the brand, which is conceptualized as brand awareness. Brand image then consists of all the linked associations, reflecting the consumer’s perceptions of the brand. While this can be constructed of many different types of associations, such as the attributes and benefits of a product, the favorability, strength, and uniqueness of them are what distinguishes a brand in memory (Keller, 1993). It is, therefore, important to build these components since they affect customer-based brand equity, meaning how consumers respond differently to a marketing stimulus that they encounter in their environment because of their knowledge of that brand (Keller, 1993). 2.2. Leveraging Secondary Sources of Brand Knowledge One way in which brand equity can be created is through leveraging secondary associations. This constitutes linking a brand to an external entity such as a person, place, thing, or other brand (see Figure 1). As Keller (2003) points out, in today’s competitive markets, leveraging these indirect associations as a source of secondary knowledge is becoming increasingly important for brands. They serve as a means to enhance consumers’ brand knowledge by essentially borrowing associations that the brand may not possess itself. As they become attached to the brand, either new associations are created or existing ones are updated (Keller, 2003).

Of the many different sources depicted, endorsers are considered to be such a secondary source of knowledge since originally they are not directly related to the brand (Keller, 1993). An endorser encompasses his or her own network of associations that have developed over time with his or her image. As, for instance, actor Daniel Craig becomes identified with Heineken, many of his associations transfer over to the brand, such as his masculinity, daring spirit as the character of James Bond, and charm. To this extent, endorsers can be leveraged in advertising to support, but also differentiate a brand’s image from competitors. Another potential external source that is significant to brand building is the channel. While ‘channels’ in Figure 1 refers more to distribution channels, the communication channel through which a brand sends a message can also have an effect. On a theoretical basis, any associations from any type of entity can be linked to a brand (Keller, 2005). Thus, the particular associations that consumers have with print as compared to broadcast or online media channels can be expected to affect how they receive and evaluate a brand message through that channel. When determining which secondary associations to link and the degree to which they can be leveraged to a brand, Keller (2003) distinguishes three important factors. The first concerns the knowledge that consumers have of the secondary source, as some awareness or familiarity must exist in order for any associations to transfer. Secondly, the meaningfulness of this knowledge, in terms of its relevance to the brand from consumers’ perspectives, must be considered. The last factor is about how transferable this knowledge of the secondary source is. When the entity becomes linked to the brand, to what extent will it affect the brand’s existing associations? Therefore, when selecting which of the sources of secondary knowledge to draw upon, these three factors should all be taken into account. While much literature has examined these external sources individually, little research up to this point has looked at a combination of them. Similarly noted by Keller (2003), academicians have yet to take a broader perspective on this subject in which multiple external sources are simultaneously linked to a brand. One of the scant examples is a study by Dean (1999). In this research, third-party product endorsement and event sponsorship were manipulated to determine their effect on pre-purchase attitudes. The results revealed a significant main effect for both sources, but no interaction between them. Another example is by Chao, Wührer, and Werani (2005), who incorporated celebrity endorsement as a moderator of country-of-origin effects. With the few academic examples available, this present study looks to contribute theoretically by investigating both the endorser and the channel as secondary associations. Moreover, the interaction between these two external entities will be considered.

Chapter 3: Endorsers in Marketing Communication With the wide use of endorsers in marketing communication, a general overview of endorsers and the extensive literature that has been devoted to them over the past decades will be examined in this chapter. Fundamentally, an endorser is someone “who endorses or who demonstrates the product” (Batra, Myers, & Aaker, 1996, p. 417). This person acts as “a ‘source’ of the information in the ad, contributing to the acceptability of the content of the message,” which is a result of his or her credibility or attractiveness (Batra et al., 1996, p. 415). Three main elements, thus, are part of this endorsement relationship: the endorser, the product or brand, and the consumer (Friedman & Friedman, 1979). While celebrities are the most frequently employed endorser in advertising and, correspondingly, have received the most academic attention, three other prevalent types have been identified (Friedman, Termini, & Washington, 1976). A professional expert is someone who is recognized for his or her special knowledge or expertise on the product endorsed, the typical consumer is a real person whose only knowledge of the product is from regular use, and lastly there is the company president of the product being advertised. The study by Friedman et al. (1976) suggests that while having an endorser is better than no endorser in an advertisement, different types have differing impacts on consumer evaluations. This notion can similarly be applied to celebrities and bloggers. While their influence has yet to be compared in academic research, the existing literature surrounding each endorser and their effectiveness will be examined. Additionally, due to the still developing literature on bloggers, the idea of a typical consumer endorsing a product will later be revisited. 3.1. Endorsement Models Traditionally, many theories have been developed to explain the phenomena underlying endorsements and the advantages of it for brands. As mentioned in the previous chapter, endorsers act as a source of secondary knowledge and can assist in creating distinct associations for a brand. Moreover, it has been suggested that they can evoke positive emotions through classical conditioning and heavily influence the credibility of a message. These theories concerning how endorsements serve to benefit brands will be further explored in the subsequent sections. 3.1.1. Classical Conditioning The basic premise on which endorsers are theorized to work in marketing is through classical conditioning. According to this process, an unconditioned stimulus that elicits a desired response is repeatedly paired with a neutral stimulus so that the neutral stimulus eventually becomes the

conditioned stimulus and evokes the conditioned response desired (Nord & Peter, 1980). Applying this to endorsements, the endorser acts as the unconditioned stimulus and the brand initially as the neutral stimulus. Through repeated pairings and exposure in advertising, consumers come to internalize this association and it becomes possible for the brand to elicit the same positive feelings (e.g. excitement and entertainment) that the endorser might (Erdogan, 1999). In the context of celebrities, many times marketers pair them with a product in order to draw attention to the advertisement (Nord & Peter, 1980). While the chosen celebrity endorser may not have any relevance to the actual content of the advertisement, his or her presence and attention-catching abilities are expected to increase its effectiveness. Moreover, when celebrities such as Brad Pitt or Natalie Portman are repeatedly paired with a brand, previously conditioned feelings such as liking or admiration can be stimulated. This process is also referred to as affect transfer, since the positive emotions that consumers associate with the stimulus (i.e. the endorser) are used to generate similar favorable feelings toward the brand (Till, Stanley, & Priluck, 2008). 3.1.2. The Match-up Hypothesis An important tenet related to affect transfer and the effectiveness of endorsers is the belongingness or congruency between the endorser and brand (Till, 1998). Known as the match-up hypothesis, this theory suggests that an endorser will be most effective when there is congruity with the endorsed brand or product (Kamins, 1990; Kahle & Homer, 1985). This is determined by the endorser’s image and how consumers perceive its level of ‘fit’ to be with the brand (Misra & Beatty, 1990). Therefore, when a product is advertised with an endorser, a high fit should result in greater believability and effectiveness as compared to a low fit, which has been confirmed by research (Kamins & Gupta, 1994). As an example, take Beyoncé who is a singer known for her beauty, talent, and representation of African-American women. Based on this image she embodies, it is reasonable that it would be a more appropriate match for her to endorse a cosmetics brand than to be a spokesperson for a food brand. Till (1998) further points outs that the ease and likelihood of building an associative link is greater when there is congruency. In relation to classical conditioning, this helps explain why affect is greater under conditions of appropriate fit (Till et al., 2008). 3.1.3. The Source Credibility and Source Attractiveness Models One of the key reasons cited for the wide use of endorsers in marketing, and of most interest to this present study, is their ability to add credibility to a message (Munnukka, Uusitalo, & Toivonen, 2016). This is especially important when the consumer has little information other than what is

contained in the advertisement. In these cases, he or she will likely turn to the source, or the endorser, as a cue (Jain & Posavac, 2001). It is widely regarded by both marketers and researchers that the characteristics of the source have a significant effect on how persuasive a message is (Ohanian, 1990). In evaluating perceptions of endorsers and their credibility, two models are most commonly used: the source credibility and source attractiveness models (Erdogan, 1999; Ohanian, 1990; McCracken, 1989). Originally developed for studying communication, these models shed light on the conditions under which an endorser will be perceived as credible (McCracken, 1989). Ohanian (1990) describes source credibility as “a communicator’s positive characteristics that affect the receiver’s acceptance of a message” (p. 41). According to the source credibility model, the effectiveness of a message is dependent upon the perceived level of expertise and trustworthiness of the endorser (Hovland, Janis, & Kelley, 1953). Expertise is defined as “the extent to which a communicator is perceived to be a source of valid assertions,” while trustworthiness refers to “the degree of confidence in the communicator’s intent to communicate the assertions he considers most valid” (Hovland et al., 1953). Thus, whereas expertise involves the knowledge, experience, or skills that the endorser possesses, trustworthiness concerns how honest and believable the target audience perceives the endorser to be (Erdogan, 1999). It is said that an endorser who exhibits these two factors is credible, and thus persuasive (McCracken, 1989). In addition to the source credibility model, much literature has indicated that the attractiveness of the source also affects the effectiveness of a message. This is generally termed the source attractiveness model and is composed of the receiver’s perceptions of the endorser’s familiarity, likeability, and similarity (McGuire, 1985). Familiarity is described as knowledge of the source based on exposure, likability as affection for the source due to his or her physical appearance and behavior, and similarity as the resemblance perceived between the endorser and the receiver (McGuire, 1985). Together these factors contribute to the attractiveness and persuasiveness of the endorser. Therefore, going beyond purely physical attractiveness, this entails other characteristics such as intellect, personality, or athletic ability (Erdogan, 1999). It has been confirmed through multiple studies that these factors, in fact, have a significant effect on how persuasive an endorser is (Amos, Holmes, & Strutton, 2008). Although the importance of each differs in certain situations, Amos et al. (2008) found that trustworthiness and expertise exercise the most influence, followed by attractiveness. Furthermore, a three-dimensional scale based on these factors has been constructed by Ohanian (1990) and is often used to measure an endorser’s credibility.

3.2. Celebrity Endorsement In this section, celebrity endorsers specifically will be focused on and the research on the effect of celebrity endorsements on consumer attitudes and intentions. In contrast to the overarching definition of an endorser, McCracken (1989) describes 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” (p. 310). This includes not only television and movie stars, but also athletes, politicians, figures in business, and artists. Similarly, Friedman et al. (1976) classify the celebrity as someone “known to the public for his accomplishments in areas unrelated to the product class endorsed” (p. 22). Thus, celebrities are principally distinguished by their recognition in society, which is what marketers draw upon based on the following meaning transfer model. 3.2.1. The Meaning Transfer Model To understand why celebrities, particularly, are effective as endorsers of brands, the cultural meanings that they embody are important to be considered. This is explained by McCracken (1989) who states that the meanings a celebrity brings to the endorsement process are what makes him or her so valuable. Celebrities, through the roles they play and how the media portrays them, acquire an extensive variety of symbolic meanings in society. For instance, these distinctions may have to do with their gender, age, status, personality or lifestyle. Each celebrity possesses a special configuration of meanings and is said to have the ability to deliver this with more power, depth, and vividness than anonymous models or actors (McCracken, 1989). When a celebrity endorses a product, this set of meanings also enters into the endorsement relationship. According to the model, the meanings move on a path from the celebrity to the product and from the product to the consumer (McCracken, 1989). This is chiefly facilitated by advertising, which serves to bring the product and meanings of the culturally constituted world together in such a way that their similarity is apparent to the viewer (McCracken, 1986). In the last stage of the process, consumers must recognize and claim these meanings into their lives through contact with the product. Considering each of these steps, McCracken (1989) asserts that successful endorsements take their power from this transfer of meaning from the celebrity world to the consumer world. It is in this manner than celebrities add value to a brand. 3.2.2. Effectiveness of Celebrity Endorsers Due to the long practice of brands employing celebrity endorsers, many decades of literature have investigated how effective they are in influencing a brand’s target audience and the conditions under which this occurs. First and foremost, celebrities have been found to increase rates of recall

for both the message and the endorsed brand (Friedman & Friedman, 1979; Atkin & Block, 1983). Following the match-up hypothesis, Misra and Beatty (1990) observe that recall of brand information is enhanced when a fit is perceived between the celebrity and brand. Moreover, there is evidence to suggest that celebrity endorsers can also aid in brand recognition (Petty et al., 1983). Previous research regarding these cognitions, as well as attitude formation, has looked at responses from the context of the consumer’s level of involvement (Petty & Cacioppo, 1983). It is stated that consumers process information differently depending on their involvement with the message or product. In a study, Petty et al. (1983) find that famous endorsers increased product category recall but reduced brand name recognition of the tested product under low involvement, while under high involvement, brand name recognition was enhanced. In terms of product attitudes, the status of the endorser (i.e. celebrity or ordinary) had a significant impact under low involvement, but not high. Thus, the impact of a celebrity endorser in an advertisement may differ in accordance with the viewer’s own involvement (Petty et al., 1983). Expanding on consumer attitudes, it is generally agreed that celebrity endorsers are able to generate positive attitudes toward an advertisement and the endorsed brand (Atkin & Block, 1983; Kamins, 1989; Ohanian, 1991; Goldsmith, Lafferty, & Newell, 2000). For instance, Atkin and Block (1983) discovered that the use of a celebrity figure led to more favorable product evaluations in alcohol advertising with the brand being perceived as more enjoyable and pleasant. Results from Till et al. (2008) also show that celebrity endorsements positively affect brand attitude and can be strengthened through repeated exposure. Various studies have further examined the conditions under which attitudes can be enhanced. In relation to physical attractiveness, Kahle and Homer (1985) found that participants who saw an attractive celebrity liked the product more than those who were exposed to a less attractive source. Similarly, endorsers rated with higher credibility were determined to be associated with more positive attitudes toward the advertisement and subsequently attitude towards the brand (Goldsmith et al., 2000). Research has also examined celebrity endorsement with regard to two-sided messages, meaning those that present both positive and negative aspects of the advertised product. Ratings of the advertisement were found to be higher in the two-sided as opposed to one-sided situation (Kamins, Brand, Hoeke, & Moe, 1989). The effect of celebrity endorsements on purchase intentions, though, is somewhat more ambiguous. Friedman et al. (1976) find in their study a marginally higher purchase intent when a celebrity endorser was used in comparison to no endorser. In contrast, Petty et al. (1983) observe that advertisements featuring famous endorsers had no significant effect. Several studies also indicate that although consumers generally have a favorable attitude towards celebrity

endorsements, this does not necessarily translate into purchase intentions (O’Mahony & Meenaghan, 1997; Kamins et al., 1989). 3.2.3. Negative Aspects of Celebrity Endorsement While many studies indicate that celebrity endorsements are relatively effective, there are also instances when utilizing celebrities have been shown to have no significant effect. For example, a study by Mehta (1994) found that between the celebrity and non-celebrity (i.e. an unknown model) conditions, there were no significant differences in the measures of persuasion. Moreover, the cognitive responses recorded did differ in that viewers paid more attention to the celebrity in the commercial featuring a celebrity, while they focused more on the advertised product and brand in the non-celebrity situations. Therefore, it is suggested that celebrity endorsers draw attention away from the brand and towards the endorser, decreasing the advertisement’s effectiveness (Mehta, 1994). This is similar to the findings of Petty et al. (1983) mentioned, where under low involvement, increased attention toward the famous endorser resulted in lower brand name recognition. Furthermore, measures of believability are not always found to be enhanced by celebrity endorsers. While Friedman and Friedman (1979) contend that this depends on the product type, the majority of studies find insignificant or no impact on the advertisement’s believability (Friedman et al., 1976; Atkin & Block, 1983; Freiden, 1984). O’Mahony and Meenaghan (1997) also indicate that celebrity endorsers are not regarded by consumers to be particularly believable. Therefore, the possibility of another more credible or more effective type of endorser is worth investigating in today’s advertising industry. 3.3. Bloggers As an alternative to celebrities, bloggers have appeared as a new type of endorser with growing power over the digital space. While a concrete definition of bloggers does not yet exist, they are deemed as the authors or writers of blogs (Zhu & Tan, 2007). Comparable to an online journal, blogs are personal websites where entries are posted in reverse chronological order (Wright, 2006). These entries are referred to as blog posts and typically contain text, images, links to other Web pages, or a combination of these (Agarwal et al., 2012). By this definition, vloggers are considered to be a subset of bloggers, where instead of purely written content, videos are used to disseminate information. Based on a series of in-depth interviews, Uzunoglu and Kip (2014) identified bloggers as the “content generators of digital platforms” (p. 595). Thus, it is clear that bloggers are closely tied to online media. Through the information that they share, bloggers are able to bring readers of similar interests together, forming virtual communities (Agarwal et al., 2012). From the perspective of

brands, bloggers can therefore be an effective means by which to reach specific categories of consumers (Uzunoglu & Kip, 2014). 3.3.1. Bloggers as Opinion Leaders Much of the literature on bloggers today regards them as online opinion leaders, who introduce new information, ideas, and opinions, and then disseminate this to their connected network (Song, Chi, Hino, & Tseng, 2007; Uzunoglu & Kip, 2014; Mendoza, 2010). Thus, they play an important role in bringing together and mediating the flow of information, adding influence as it is passed on (Uzunoglu & Kip, 2014). Opinion leaders are originally described by Katz and Lazarsfeld (1955) as “the individuals who were likely to influence other persons in their immediate environment” (p.3). Similarly, bloggers are said to have the ability to affect the attitudes and behavior of their surrounding online community (Song et al., 2007). This includes, in essence, any reader or follower of the blog. Also referred to as digital influencers, bloggers have been able to extend their reach even further due to the integration of blogs with known microblogs such as Facebook, Twitter, and Instagram (Uzunoglu & Kip, 2014). In assessing the influential role of bloggers in connection with brands, Uzunoglu and Kip (2014) develop a multi-step communication flow model. First, the brand communicates with the target blogger via activities such as product samples or events. The blogger then decodes and shares the brand’s message through his or her personal blog or microblogs. Once the message has been transmitted to readers and followers, they can, in turn, recode it by commenting or sharing it with their own network. This process of message dissemination continues further as followers share it among themselves or, potentially, respond back to the blogger or brand. Each of these steps, therefore, constitutes a two-way communication process with the blogger serving as the key mediator in spreading the brand’s message. In this way, the messages shared by bloggers can have wide influence and come to be reference points for readers and consumers (Uzunoglu & Kip, 2014). However, it is also noted that with the immense number of blogs in existence, not all bloggers are influential or opinion leaders (Agarwal et al., 2012). Determining which individuals hold influence is important for companies when selecting endorsers who can offer an effective presence. Thus, many recent studies have focused on developing metrics or algorithms with the purpose of identifying the influential bloggers (Song et al., 2007; Agarwal et al., 2012; Akritidis, Katsaros, & Bozanis, 2011). Labrecque et al. (2013) state that to be influential in the context of social media, an individual must have content that is relevant and have a sufficient reach.

3.3.2. Bloggers as Peer Endorsers One of the main reasons for the influence that bloggers exert over their followers stems from their inherent identity as ordinary consumers. It is found that Internet users typically view bloggers as ‘one of us’ (Uzunoglu & Kip, 2014). Considering that essentially any individual with access to the Internet can start a blog and post information on products or services, bloggers can be likened to peer consumers (Lu, Chang, & Chang, 2014). Lee and Youn (2009) indicate that when evaluating products, consumers place more trust in peers than information from companies and are more likely to use this information in their decisions. Thus, the fact that bloggers are often perceived as peers may give them a certain edge as endorsers. Following the definition of an endorser, a peer endorser is “a typical satisfied customer who endorses or demonstrates a product or service and acts as a source of information to influence the acceptability of the message” (Munnukka et al., 2016, p. 182). In recent years, the use of peers as endorsers has increased in advertising as a means to improve the credibility of a message (Munnukka et al., 2016). In examining the credibility of peer endorsers, Munnukka et al. (2016) explain that it is constructed of four dimensions: trustworthiness, expertise, attractiveness, and similarity. Prior research has suggested that the effectiveness of this type of endorser is the result of the person’s similarity to the audience (Friedman & Friedman, 1979). Therefore, in extending the three credibility constructs typically used for research on celebrity endorsers (Ohanian, 1990), Munnukka et al. (2016) also include perceived similarity in their study. The findings show that it is a significant additional dimension of the credibility of peer endorsers and has one of the strongest effects, along with the endorser’s trustworthiness, on attitudes toward the advertisement. In situations where the main goal is to have the audience identify with the endorser, peer endorsers can be especially valuable (Batra et al., 1996). It is proposed that their similarity with the consumer enhances identification, and hence the effectiveness of an advertisement (Friedman & Friedman, 1979). Particularly when non-complex and low-risk products are involved, peer endorsers have been shown to have a greater influence than celebrity or expert endorsers (Friedman & Friedman, 1979). Thus, the nature of bloggers as “empowered consumers” may make them more effective endorsers in certain cases (Mendoza, 2010). 3.3.3. Sponsored Recommendation Posts With bloggers being a relatively recent phenomenon, the few studies on their effectiveness as endorsers have been investigated in the context of sponsored blog posts. Characterized as a type of online consumer review, which provides product information based on experience, a sponsored

recommendation post is described as “a blog article written by a blogger who is sponsored by the producers of any product” (Lu et al., 2014, p. 259). Zhu and Tan (2007) further state that it involves mentioning, reviewing, and/or promoting products on a personal blog, and should be considered an advertisement. As a paid sponsorship, the blogger receives monetary compensation or certain benefits, such as free samples or discounts, from the sponsoring marketer in exchange for posting about the product (Forrest & Cao, 2010). Various studies have researched the effect of this new form of advertising under different conditions. For example, the type of product recommended by a blogger is shown to have a significant relationship with attitudes toward the sponsored post (Lu et al., 2014). When the product is a search good, meaning its attributes are more objective and can be easily evaluated without interacting with the product (e.g. a cell phone), consumers are found to have a more positive attitude than when the recommended product is an experience good (e.g. a package tour). Even in the latter case, though, Lu et al. (2014) still observed consumer attitudes toward the blog post to be slightly favorable. Another condition that has received attention is whether a blogger reveals the advertising intent of a post. Laws in many countries now stipulate that bloggers must disclose if a post is sponsored to avoid any deception (Martin & Smith, 2008). Zhu and Tan (2007) indicate that explicitly mentioning a post is sponsored does not always result in lower effectiveness, depending on the consumer’s product involvement and the perceived expertise of the blogger. The majority of the studies on sponsored blog posts thus far, therefore, point to a relatively positive influence on viewers.

Chapter 4: Advertising Media In the previous chapter, the existing literature on celebrities and bloggers as endorsers and the effect that each has on consumers was discussed. As mentioned, a factor that has yet to be investigated with endorsements, but may be a significant contributor to an endorser’s influence is the type of medium involved. With the medium being the means by which an endorsement is communicated, Chen and Hsieh (2012) state that different advertising media create different results with consumers. Traditional media, such as print and television, has long been established as the primary form of advertising, but the Internet represents a powerful new medium for marketers to engage with and more effectively deliver advertisements to consumers (Li et al., 2012). Differences in the characteristics of these media have been shown to affect how consumers view an advertisement (Yoon & Kim, 2001). Thus, this final chapter of the theoretical framework will focus on the literature concerning the types of traditional and online media used today and the specific characteristics of each. Moreover, the differences in credibility and impact that different medium types have on the communication of a message will be expanded upon. As this study also plans to look at the sender of a message within a medium, the existing literature on this will be discussed. 4.1. Traditional vs. Online Media For decades, traditional media, which includes newspapers, magazines, television, and radio, have dominated advertising. All of these communication forms consist of a one-way, linear information flow where the consumer is a passive recipient of the message (Bezjian-Avery, Calder, & Iacobucci, 1998). Consumers have little control over what they are exposed to, as whatever appears on a TV screen or in a magazine is what is viewed. In this way, advertising is pushed out from firms whose purpose is to inform, persuade, or remind customers about the brand or its offerings (Campbell, Pitt, Parent, & Berthon, 2011). With the emergence of the Internet and, especially, social media, the way in which marketers communicate with consumers has drastically changed (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004). Compared to traditional media, the Internet has allowed for more personalized advertising targeted at different groups and more direct engagement between marketers and consumers (Yoon & Kim, 2001). Rather than being mere receivers of messages, individuals have gained greater power over the information they encounter (Labrecque et al., 2013). Moreover, due to the features of Web 2.0 and social media platforms such as Facebook, Twitter, and Instagram, consumers are able to easily create their own content and enter into two-way communication with brands (Campbell et al., 2011). As a result of the networks that are formed on social media,

advertisements are diffused and delivered through more human connections, rather than being pushed directly from a company to its target audience (Li et al., 2012). It is precisely this interactive nature that constitutes the core difference between the Internet and traditional media (Bezjian-Avery et al., 1998; Yoon & Kim, 2001). La Ferle, Edwards, and Lee (2000) further point out that online media is unique in that it combines the elements of other media into one: print, visual, and sound. This array of attributes, along with the wide availability and lower cost of online channels, has made advertising through the Internet a popular means (Uzunoglu & Kip, 2014). With regard to product endorsements, Chang, Chen, and Tan (2012) discuss several unique characteristics within the context of social media. First, endorsers have the ability to construct an online representation of him or herself through a profile page, providing viewers with a more personal idea of who he or she is. Second, various built-in tools such as commenting and tagging allow endorsers to respond to and have a discussion with readers. These features help provide additional transparency and a stronger feeling of connectedness between the endorser and consumer, which may in turn affect an advertisement’s effectiveness (Chang et al., 2012). 4.2. The Impact of Different Media Given the unique characteristics of each medium, an advertisement delivered through one channel will have a different impact than through another. Each is known to have its strengths and weaknesses that affect how different messages are transmitted and received (Batra et al., 1996). For instance, quick messages tend to be better suited for television where advertisements are devised to maintain familiarity and reinforce a brand. In contrast, print media provides a more effective learning environment for when time is needed to absorb the message and compare brand options (Kim, Hwang, & Fesenmaier, 2005). Yoon and Kim (2001) further note that advertising effectiveness differs depending on the kind of consumer and product. The results from their study suggest that for highly involved products, such as automobiles and luxury watches, the Internet is a better channel for advertising. Contrastingly, for low involvement products, television is more appropriate. It is proposed that a possible reason for this finding is that the Internet allows consumers to better fulfill their need for information. In an experiment designed to measure the effectiveness of print, radio, and television, recall was found to be significantly higher for those exposed to an advertisement through print or radio than television (Klein, 1981). A different study, focusing on tourism advertising, observed that television was the most ‘powerful’ channel in influencing destination choice, with newspapers ranking second and magazines closely in third (Kim et al., 2005). Further analysis showed that television was most useful in building destination and advertisement awareness, whereas print

media was linked to requests for more travel information. In a retail context with a large firm promoting a sale through ten different media channels, Danaher and Dagger (2013) measured advertising effectiveness in terms of purchase incidence (i.e. store visit) and purchase outcome. Their study showed that television, radio, newspaper, catalogs, sponsored search, and direct mail advertising all had a significant influence. However, magazine advertising had no influence on either measure of purchase behavior. Therefore, these varying findings serve as evidence that different media have different effects on consumer response depending on the context and should be used in line with the advertising objectives desired. 4.3. Media Credibility The importance of credibility in determining the persuasiveness of a message can be seen in the growing body of research not only with endorsements, but also concerning advertising. Moore and Rodgers (2005) contend that consumers often ignore or avoid advertisements that they view as low in credibility. Thus, in order to have an effect on the attitudes and behavior of an audience, the source and the medium through which an advertisement is delivered should both be viewed as credible (Moore & Rodgers, 2005). With source credibility being discussed earlier in regards to endorsers, media credibility refers to consumers’ perceptions of how believable and trustworthy a medium is (Kiousis, 2001). If a medium is perceived to lack either, people are less likely to pay attention to the content and, hence, the advertising (Johnson & Kaye, 1998). Several studies have examined consumers’ perceptions of advertising credibility and trust in different media. Findings from Moore and Rodgers (2005) indicate that credibility is highest for newspapers, followed by television, magazines, and radio, while the Internet scored the lowest. The credibility of each media was also found to be affected by suspicion of advertising, with consumers being the most skeptical about the Internet. Similarly, in terms of trust, it was observed that while consumers are not particularly trustful of advertising media, Internet advertising is perceived as less trustworthy than traditional advertising (Soh, Reid, & King, 2007). In contrast, a survey conducted by Kiousis (2001) found online channels to be deemed as more credible than television. Further studies report that these perceptions differ depending on consumer demographics, such as age and media use, and should be accounted for (Becker, Martino, & Towers, 1976). In focusing on young Web users, Johnson and Kaye (1998) found that they are more likely to judge online media as a credible source. Thus, the research at present on the perceived credibility of different media is mixed and warrants further investigation. With the differences in credibility associated with different types of media, this reveals that the medium through which a message is communicated plays an important role in how that

message is perceived. In relation to endorsements, the medium in which an endorsed message is seen can affect its credibility differently. Therefore, besides attempting to determine if one type of medium is more credible overall, how the credibility of a medium contributes to perceptions of an endorser will also be researched. 4.4. The Sender In considering the medium through which a message is communicated, a significant aspect to also take into account is who the sender of the message is. Looking at traditional marketing communication literature, a distinction in personal versus non-personal channels has previously been noted (Belch & Belch, 2004). Generally, personal communication between two or more individuals is expected to be more persuasive than channels without personal contact (Belch & Belch, 2004). The differing influence of the channel, or in other words, the communicator of a message has also been researched with regard to viral advertising. Cho, Huh, and Faber (2014) distinguish between the advertiser, who creates the message, and the sender, who distributes the message. The results of their study show that depending on which of the two delivers the message, there is a difference in trust and its resulting impact on receivers. Thus, it can be said that the sender has an impact on how a message is evaluated. What has yet to be addressed in the context of endorsements, though, is the endorser as the sender compared to the brand as the sender. Within online media, especially social media, an endorser now has the power to send brand-related messages through his or her personal account. As this was previously not an option for endorsers, the present study will examine the difference in impact between the endorser’s personal channel and the brand channel.

Chapter 5: Hypotheses Based on the theoretical framework, a conceptual model and hypotheses are developed in this chapter. As depicted in Figure 2, the primary focus of this research is the influence of the type of endorser, together with the type of medium, on the credibility of a message. The independent variable Endorser type is divided into celebrity versus blogger, while Medium type consists of three levels, namely personal social media, brand social media, and traditional media. Thus, besides the division between social media and traditional media, there is a further distinction made on social media between the sender of the message in terms of the endorser him/herself and the brand. While a similar distinction in the sender was attempted to be found for traditional media, it is uncommon to see an advertisement that is not brand-sponsored where an endorser communicates the message him/herself. Finally, the effect of perceived credibility on consumers’ attitudes toward the message is examined. As it can be observed in the conceptual model below, Hypothesis 3 is not shown due to the fact that it does not directly concern the effect of a variable. However, it will be described further on in this chapter and tested thereafter. Figure 2. Conceptual model with hypotheses. Looking at the extant literature, there is currently not enough evidence on bloggers and celebrities to unequivocally support that one is more credible than the other as an endorser. While several studies have shown that celebrities are not always regarded as particularly believable and do not consistently enhance the believability of a message (O’Mahony & Meenaghan, 1997; Friedman & Friedman, 1979), it has yet to be determined if they are perceived to be less credible in general. Moreover, contributing to this is the fact that research on the effectiveness of each endorser has primarily examined them in their typical medium, meaning celebrities in traditional advertisements and bloggers solely online. Without separating the medium type out, it cannot be said that one type of endorser is always seen as more credible. Thus, a main effect of endorser type on the perceived credibility of a message, regardless of the medium type, cannot be hypothesized, but will still be tested in the analysis. However, when looking at different types of media, a main effect is expected to occur as the specific characteristics of each may lend to a medium’s associated credibility. Social media, through the way in which it is designed, adds a layer of transparency unfound in traditional media (Chang et

al., 2012). For instance, built-in tools such as commenting and tagging allow users to enter into discussions with each other. This gives the poster of a message the opportunity to respond and elaborate on the original information. Moreover, the ability for the owner of an account to construct a personal profile provides viewers with insight into that user. These unique features, which are absent in traditional media, make social media seem more intimate and interactive (Yoon & Kim, 2001). As consumers are able to obtain more information about the product or brand and have more direct contact with the endorser, this may contribute to the believability and trustworthiness of the message. The differing characteristics of social media compared to traditional media lead to the first part of Hypothesis 1. H1a: Endorsed messages on social media are perceived to be more credible than endorsed messages in traditional media. Beyond the medium, though, the sender of a message should also be considered. An individual’s own social media account is generally personal in nature. It is a space where users have the freedom to share their own thoughts and experiences. Thus, serving as a form of self-representation, the content that is posted is closely linked with the credibility of the user (Zhu & Tan, 2007). In the case of endorsements, it is probable than an endorser will only recommend a product on his or her personal account if a positive experience with it prevails. Contrastingly, the content on a brand’s social media account is managed by the brand itself. Messages posted are especially curated to fit with the brand’s strategy. This may lead to lower perceptions of credibility, despite the message being the same as when it is sent from the endorser. Belch and Belch (2004) similarly note that information sent through personal communication channels are generally more influential than through non-personal channels. Taking who the sender is on social media into account leads to the second part of the hypothesis. H1b: Within social media, endorsed messages are perceived to be more credible if posted from the endorser’s personal account than from the brand’s account. While it has been proposed that social media is more credible than traditional media, this difference is expected to be moderated by the type of endorser. Due to the fame and familiarity surrounding celebrities, they have the ability to draw more attention to an advertisement than bloggers. Furthermore, the meanings that they carry with them, as based on the meaning transfer model, are more deeply rooted in the cultural world (McCracken, 1989). When celebrities are used as endorsers, they are then better able to carry their associations across different types of media. However, when it comes to bloggers, who have developed their identity entirely through personal blogs and social media, their meanings are rooted in the online space. It is expected therefore that

when they are removed from their usual online context, much of their credibility in endorsing a message will be lost. This moderation by the type of endorser is presented in two parts. H2a: The difference in the credibility of a message between social media and traditional media is larger for bloggers than for celebrity endorsers. Further focusing on social media, Hypothesis 1b contends that messages posted by the endorser will be seen as more credible than when posted by the brand. This difference in credibility when it comes to the sender of a message is also expected to differ between when the endorser is a blogger or a celebrity. For bloggers, it is believed that if they are featured on a brand’s social media account, this will greatly reduce their credibility as it goes against their identity as a blogger. However, when promoting a product on their own personal account, this is much more believable since bloggers are known for the content that they create themselves. Thus, a blogger must be the sender him/herself in order to achieve maximum credibility. In contrast, for celebrities, the sender of a message is less important. Since the large monetary compensation received by celebrities for endorsements is generally known, consumers will suspect the brand to be behind the message whether it is posted from the celebrity or the brand. H2b: Within social media, the difference in the credibility of a message between an endorser’s personal account and the brand’s account is larger for bloggers than for celebrity endorsers. In building on Hypothesis 2, the subsequent hypothesis specifies the expected interaction between the type of endorser and the type of medium. As mentioned at the beginning of this chapter, there is not believed to be a main effect of endorser type. However, taking the medium type into account, it is likely that an endorser is more effective in the context where he or she originally became known. For bloggers, this means online since their identity has been formed through their blog and social media posts. As a result, the online space is where people recognize them. Additional support for this can be drawn from the source attractiveness model, as it has been noted that bloggers are commonly viewed as peer users (Lu et al., 2014). To this extent, it has been shown that one of the factors most significant in the influence of peer endorsers is their similarity to the audience (Munnukka et al., 2016). Thus, the high perceived similarity between bloggers and social media users should contribute to their credibility. Taking these factors together, bloggers are expected to be viewed as the more credible endorser on social media. H3a: On social media, messages endorsed by bloggers are perceived to be more credible than messages endorsed by celebrities.