CONSUMER RESISTANCE TO SPONSORED EWOM: THE ROLES OF INFLUENCER CREDIBILITY AND INFERENCES OF INFLUENCER MOTIVES By Mengtian Jiang A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Information and Media — Doctor of Philosophy 2018 ABSTRACT CONSUMER RESISTANCE TO SPONSORED EWOM: THE ROLES OF INFLUENCER CREDIBILITY AND INFERENCES OF INFLUENCER MOTIVES By Mengtian Jiang Sponsored content, written by social media influencers or micro-celebrities to endorse a product or service, has become a popular influential marketing strategy to help advertisers reach and influence potential consumers. However, since December 2015, the Federal Trade Commission has required influencers to use a clear and prominent disclosure to alert readers of the paid nature. Do consumers ponder the influencers’ motives for the product recommendation? Does the presence of disclosure or their knowledge of the influencer change their perceptions of the motives? How do they respond to sponsored content with the FTC-required disclosure? Guided by Persuasion Knowledge Model and Attribution Theory, this current study examines how a consumer makes inferences of influencer motives about a sponsored post, as well as how a consumer uses their prior knowledge of perceived influencer credibility and sponsorship disclosure to interpret and respond to a sponsored post. Instagram was used as the posting platform context for this study. The research was conducted in two phases. Phase I used three online surveys to examine how consumers infer the motives behind the behavior of a social media influencer product recommendation. This process identified six distinct types of influencer motives (goals that the influencer seeks through the post) that co-exist during consumer processing of sponsored content: Money motives, Selling motives, Image motives, Love motives, Sharing motives and Helping motives; and developed a scale to measure consumer perceptions of influencer product recommendation motives within the context of social media. The second phase of the research consisted of a 2 (disclosure) x 2 (influencer credibility) x 2 (product category) between-subjects experiment to examine the roles of influencer credibility and different types of influencer motives on consumer resistance to Instagram sponsored posts. The results showed that the presence of an FTC-required clear and conspicuous sponsorship disclosure generated stronger consumer perceptions of money and selling motives regardless of product categories and credibility. In addition, highly credible influencers appear to generate stronger consumer perceptions of image, love, and helping motives and are less likely to face consumer resistance to their messages than less credible influencers are, regardless of product categories or disclosure. Furthermore, findings revealed that different thoughts about influencer motives led to varying levels of resistance. Specifically, money and selling motives together, as well as image motives increased consumer resistance towards to sponsored content; while love, sharing and helping motives altogether reduced consumer resistance to persuasion. Theoretical and practical implications of the findings are discussed, and future research directions are suggested. Copyright by MENGTIAN JIANG 2018 ACKNOWLEDGEMENTS To my advisor and mentor Dr. Nora Rifon, To my committee members Drs. Robert LaRose, Esther Thorson, and John Besley, To my loving parents, and To my dearest Spartan family. v TABLE OF CONTENTS LIST OF TABLES ___________________________________________________________ ix LIST OF FIGURES __________________________________________________________ x CHAPTER 1 INTRODUCTION ________________________________________________ 1 CHAPTER 2 BACKGROUND _________________________________________________ 8 Types of Social Media Influencers_____________________________________________ 8 Sponsored Content in the Form of eWOM______________________________________ 9 Influencer Marketing and Disclosure _________________________________________ 10 Sponsorship Disclosure and Consumer Resistance ______________________________ 13 Consumer Processing of Sponsorship Disclosure _______________________________________ 13 Consumer Resistance as a Negative Consequence of Sponsorship Disclosure ____________ 14 Persuasion Knowledge Model _______________________________________________ 15 Three Types of Knowledge ___________________________________________________________ 15 Change-of-Meaning Principle _________________________________________________________ 16 Attribution Theory ________________________________________________________ 17 Two Types of Attributions of Influencer Motives ______________________________________ 17 Discounting Principle ________________________________________________________________ 18 Correspondent Inference Bias _________________________________________________________ 19 CHAPTER 3 SCALE DEVELOPMENT OF INFLUENCER MOTIVES _____________ 21 Conceptualizing Influencer Motives based on PKM _____________________________ 21 Conceptualizing Influencer Motives based on eWOM literature___________________ 23 Conceptualizing Influencer Motives based on Attribution Theory _________________ 24 Conceptualizing Influencer Motives in Social Media ____________________________ 25 Scale Development Process _________________________________________________ 26 Study 1: Initial Scale Development ___________________________________________ 26 Design and Procedure ________________________________________________________________ 26 Sample ______________________________________________________________________________ 27 Results ______________________________________________________________________________ 27 Study 2: Scale Development _________________________________________________ 31 Design and Procedure ________________________________________________________________ 31 Sample ______________________________________________________________________________ 31 Results ______________________________________________________________________________ 32 Study 3: Scale Validation ___________________________________________________ 34 Design and Procedure ________________________________________________________________ 34 Sample ______________________________________________________________________________ 34 Results ______________________________________________________________________________ 34 Scale Development Results __________________________________________________ 36 CHAPTER 4 THEORETICAL MODEL AND HYPOTHESIS DEVELOPMENT _____ 38 vi Role of Sponsorship Disclosure ______________________________________________ 38 Role of Influencer Credibility _______________________________________________ 40 Influencer Credibility as Agent Knowledge ____________________________________________ 40 Effect of Influencer Credibility on Different Influencer Motives and Resistance _________ 41 Interaction Effect of Disclosure and Influencer Credibility _____________________________ 43 Role of Inferences of Influencer Motives ______________________________________ 44 Effect of Inferences of Influencer Motives on Resistance _______________________________ 44 CHAPTER 5 MAIN STUDY METHOD ________________________________________ 48 Main Study Design ________________________________________________________ 48 Stimuli Development _______________________________________________________ 48 Product Choice ______________________________________________________________________ 48 Choice of Influencer Name, Product Name, and Image _________________________________ 49 Priming Story Design ________________________________________________________________ 50 Message Design______________________________________________________________________ 53 Disclosure Design ____________________________________________________________________ 53 Main Study Participants____________________________________________________ 56 Main Study Procedure _____________________________________________________ 57 Measures ________________________________________________________________ 58 Manipulation Check __________________________________________________________________ 58 Dependent Variables _________________________________________________________________ 59 Control Variables ____________________________________________________________________ 60 CHAPTER 6 MAIN STUDY RESULTS ________________________________________ 62 Manipulation Check _______________________________________________________ 62 Main and Interaction Effects on Influencer Motives and Consumer Resistance ______ 62 Effect of Different Influencer Motives on Consumer Resistance ___________________ 70 Measurement Model Testing __________________________________________________________ 70 Structural Model Testing _____________________________________________________________ 72 Post-Hoc Mediation Analysis ________________________________________________ 75 Post-Hoc SEM by Influencer Credibility Condition _____________________________ 76 CHAPTER 7 DISCUSSION___________________________________________________ 81 Summary of Findings ______________________________________________________ 81 Conceptualizing and Measuring Influencer Motives ____________________________________ 81 Role of Influencer Motives ___________________________________________________________ 83 Role of Sponsorship Disclosure _______________________________________________________ 84 Role of Influencer Credibility _________________________________________________________ 85 Role of Product Category _____________________________________________________________ 86 Limitations and Future Research Directions ___________________________________ 87 APPENDICES ______________________________________________________________ 91 APPENDIX A. TWO SCALES OF INFLUENCER MOTIVES ___________________ 92 APPENDIX B. CONSENT FORM ___________________________________________ 94 APPENDIX C. PRETEST QUESTIONNAIRE _________________________________ 95 APPENDIX D. MAIN STUDY QUESTIONNAIRE ____________________________ 104 vii APPENDIX E. DISCLOSURE + PROTEIN POWER SPONSORED INSTAGRAM POST __________________________________________________________________ 110 APPENDIX F. NO DISCLOSURE + PROTEIN POWER SPONSORED INSTAGRAM POST __________________________________________________________________ 111 APPENDIX G. DISCLOSURE + TRAVEL SITE SPONSORED INSTAGRAM POST ________________________________________________________________________ 112 APPENDIX H. NO DISCLOSURE + TRAVEL SITE SPONSORED INSTAGRAM POST __________________________________________________________________ 113 REFERENCES ____________________________________________________________ 114 viii LIST OF TABLES Table 1. An Overview of the Scale Development Process ___________________________ 26 Table 2. List of the Initial 64-Item Scale _________________________________________ 29 Table 3. EFA and Reliability Results of the 31-Item Scale __________________________ 33 Table 4. CFA Results: Validity Test and Factor Correlation ________________________ 35 Table 5. CFA Model Fit Indices _______________________________________________ 36 Table 6. Types of Motives In PKM, eWOM and Attribution Theory Studies __________ 37 Table 7. Priming Stories of Influencer Credibility Manipulation ____________________ 52 Table 8. Sponsored Content ___________________________________________________ 55 Table 9. Results of Power Analysis _____________________________________________ 56 Table 10. Multivariate Tests __________________________________________________ 63 Table 11. Results of Between Subject Effects _____________________________________ 65 Table 12. Descriptive Statistics by Experimental Conditions ________________________ 66 Table 13. CFA Results of Model 3: Validity Test and Factor Correlation _____________ 70 Table 14. CFA Results of Model 4: Validity Test and Factor Correlation _____________ 71 Table 15. SEM Hypothesis Testing Results ______________________________________ 75 Table 16. Post-Hoc SEM Results by Influencer Credibility _________________________ 77 Table 17. Hypothesis Testing Results ___________________________________________ 80 Table A. Two Scales of Influencer Motives 91 ix LIST OF FIGURES Figure 1. Proposed Theoretical Model __________________________________________ 47 Figure 2. Path Diagram of The Hypothesized Model (Model 3)______________________ 72 Figure 3. Path Diagram of The Revised Model (Model 4) __________________________ 74 Figure 4. Post-Hoc Mediation Test _____________________________________________ 76 Figure 5. Post-Hoc SEM Model for Highly Credible Influencers (Model 5) ____________ 78 Figure 6. Post-Hoc SEM Model for Less Credible Influencers (Model 6) ______________ 79 x CHAPTER 1 INTRODUCTION Advertisers and marketers use covert persuasive messages to break through advertising clutter, overcome consumers’ ad avoidance behavior, and influence consumer behavior (Litvin, Goldsmith, & Pan, 2008). For decades, consumers have encountered and become familiar with numerous types of covert persuasive messages, such as advertorials in the print media, TV infomercials, product placements in movies or other entertainment content on television or the internet, or advergames (brands integrated into digital games). With the rapid growth of digital media, covert persuasive messages are woven into editorial content using various innovative formats. Examples include native advertising, which is constructed to appear as editorial content, and not advertising, on a website; paid product reviews or recommendations, which look like authentic, consumer generated word of mouth messages, endorsements or testimonials, but in actuality are not independent of commercial entities; paid search results or promoted listings, which are not organic search results but in fact are paid by the marketer to be positioned first. In this study, we label these different types of covert persuasive messages as sponsored content, defined as “the purposeful integration of brands or branded persuasive messages into editorial media content in exchange for compensation from a sponsor (Boerman & van Reijmersdal, 2016).” Sponsored content, like other advertising messages, is protected commercial speech under the First Amendment to the U.S. Constitution. However, sponsored content is under criticism for being potentially misleading or deceptive when it successfully makes consumers believe it is not paid content. Founded in 1914, the Federal Trade Commission (FTC) issues rules pursuant to Section 5 of the FTC Act (15 U.S. Code § 57a(a)(1)(B)) to “present 1 anticompetitive, deceptive, and unfair business practices deceptive and unfair business practices, … without unduly burdening legitimate business activity (Federal Trade Commission, n.d.).” Recently, the FTC has released guidelines to address how to apply the Section 5 rule to advertising practices that fall under the umbrella of sponsored advertising across different contexts, such as endorsements and testimonials (Federal Trade Commission, 2009), digital advertising (Federal Trade Commission, 2013), and native advertising (Federal Trade Commission, 2015). In these guidelines, the FTC recommends the use of a clear and conspicuous disclosure to label sponsored content and inform consumers that the content is indeed paid for by a marketer. For instance, the FTC suggests using words and hashtags such as “#ad,” “#sponsored,” “#promotion” to label sponsored Twitter or Instagram posts; or a sentence like “Company X gave me this product to try…” in sponsored blog posts/reviews. Also, the disclosures should be “close to the claims to which they relate; in a font that is easy to read; in a shade that stands out against the background (Snyder, 2016, p.93).” A rich body of previous research has examined the disclosure effects in the contexts of video news releases (e.g. Nelson, Wood, & Paek, 2009), branded TV program placements (e.g., Boerman, van Reijmersdal, & Neijens, 2014; Boerman, van Reijmersdal, & Neijens, 2012, 2015), advergames (e.g. Van Reijmersdal, Lammers, Rozendaal, & Buijzen, 2015), and native ads in online news websites (e.g. Wojdynski & Evans, 2016). However, few studies have examined disclosure effects in the context of social media. To date, only six published studies have so far examined effects of disclosure in sponsored content in social media, such as electronic word of mouth messages (Carl, 2008; Nekmat & Gower, 2012), sponsored blogs (Campbell, Mohr, & Verlegh, 2013; Hwang & Jeong, 2016), sponsored Facebook posts (Boerman, Willemsen, & Van Der Aa, 2017), and sponsored twitter posts (Boerman & 2 Kruikemeier, 2016). This study adds to the current disclosure literature by studying consumer response to disclosure in the social media context. Social media have seen an explosion of sponsored content in the form of sponsored posts created and shared by social media influencers. Social media influencers are opinion leaders with a substantial number of social media followers who have expertise in a certain area and have the power to influence other’s opinions. The practice whereby brands or marketers provide financial or material compensation to social media influencers, who create and share brand-related sponsored posts, images or videos with their audience, is called Influencer Marketing. For instance, Felix Arvid Ulf Kjellberg, or best known as PewDiePie, is a Swedish social media influencer owns the most subscribed YouTube channel since 2013, with over 55 million subscribers as of June 2017. He creates and publishes online videos showing himself playing and commenting on video games, and is sponsored by companies to promote video games. Time and Forbes named him as one of “The World's 100 Most Influential People” (TIME, 2016) and was the highest earning YouTube star in 2016 (Hamedy, 2016). Instagram is a major hub for influencer marketing, making it a perfect context for this study. Instagram has become one of the world’s most popular social networks since the launch in 2010. On April 26, 2017, Instagram announced 700 million monthly active users (Statista, 2017), twice the size of Twitter (Constine, 2017). According to Mediakix, an influencer marketing agency, there are currently 9.7 million brand-sponsored influencer posts in 2016, which is projected to grow to 32.3 million in 2019. Mediakix also calculates that advertisers spend over $1 billion on Instagram influencer marketing in 2017 and predicts this number to increase to nearly $2.4 billion by 2019 (Mediakix, 2017). 3 Previous research often uses Friestad and Wright's Persuasion Knowledge Model (PKM, Friestad & Wright, 1994) as the theoretical framework to explain the effects of disclosure on consumer message processing and its outcomes. PKM explains how persuasion targets (e.g., a consumer) use persuasion knowledge, agent knowledge and topic knowledge to interpret and cope with an influence agent’s (e.g., a social media influencer) persuasion attempt (e.g., an ad). The “change of meaning” principle in PKM suggests once consumers recognize the persuasive intent when the disclosure is accessible in the persuasive message, they will disengage from the persuasion interaction and discount what the influence agent says, mitigating the message persuasiveness. Most empirical research has found that disclosure activates persuasion knowledge, which leads to resistance to persuasion, through the creation of consumer perceptions such as advertising skepticism (e.g., Brown & Krishna, 2004), decreased ad credibility (e.g., Wojdynski & Evans, 2016), negative brand attitudes and evaluation (e.g., Van Reijmersdal et al., 2015), reduced recommendation or forwarding intentions (e.g., Eisend, 2015) and less likelihood to click through keyword search ads (Yoo, 2009). However, a recent interview study has shown that consumers reported different levels of resistance, from skepticism to indifference to enjoyment, after the recognition of native advertisements in news websites as actual ads (Jiang, McKay, Richards, & Snyder, 2017).How will consumers respond to sponsored content with the FTC-required disclosure? The underlying process of consumer resistance to sponsored content has not been thoroughly examined. As Campbell and Kirmani (2008) and Lorenzon and Russell (2012) pointed out, most PKM studies conceptualize and measure persuasion knowledge as consumer recognition of the persuasive intent, or the recognition that a message is an ad. Advertising recognition and 4 understanding persuasive and selling intent are important components of persuasion knowledge (Ham, Nelson, & Das, 2015). However, consumer inferences of influencer motives are also an important part of persuasion knowledge. Though rarely studied, it is reasonable to posit that when a consumer reads a brand-related post from a social media influencer’s account, s/he will try to understand why the social media influencer makes these persuasive statements. During the process of making inferences, a consumer may not only realize the social media influencer wants to make money and sell products, but also make other motive assessments. Could it be their love of the product, their intention to gain fame, or intention to be useful and helpful? Together, those motive assessments will influence consumer response to the post, which may not necessarily increase consumer resistance. The Discounting principle in Attribution Theory (Jones & Davis, 1965; Kelley, 1973) states that people discount the role of one possible cause for the behavior if other plausible causes are present. Therefore, it can provide an alternative explanation to PKM and explain the different levels of consumer responses by studying how people make attributions of product recommendations to different types of influencer motives. To date, very few PKM studies have investigated how consumers make inferences of influencer motives, especially in the context of social media, and how those inferences affect the persuasiveness of the message. Consumers may possess different breadths of persuasion knowledge resulting in the inference of a range of different possible motives as an explanation for the influencer to post a recommendation. For instance, does the social media influencer makes these positive recommendation statements because s/he likes the product, wants to help others make better decisions (e.g., informing others), or because the influencer wants to get benefits from you (e.g. self-enhancement and making a profit)? Therefore, influencer motives may be a missing piece in PKM explanations of consumer resistance to persuasion and it is 5 necessary to develop and test a scale of consumer inferences of influencer product recommendation motives. One unique characteristic of influencer marketing is that it is built on the authentic relationship between an influencer and her or his followers. In other words, consumers who follow a digital influencer are likely to already possess some knowledge about this influencer’s traits and motives, based on their previous observations or interactions with the influencer. According to PKM, agent knowledge, which is the beliefs about the traits, competencies, and goals of the persuasion agent, can also serve as the information cue to help make a causal inference. However, Campbell & Kirmani (2008) reviewed PKM studies published prior to 2008 and found that few PKM studies focused on agent knowledge, possibly due to the prolific source effect literature, or the interaction between persuasion knowledge and agent knowledge. Moreover, in sponsored content, the disclosure comes from the influencer. As a result, the effectiveness of disclosure in influencer marketing may depend on the source effect. Therefore, it is important to take into account consumers’ prior knowledge about the influencer, specifically perceived influencer credibility, and to examine how this perception would interact with the sponsorship disclosure created by the influencers to influence consumer response to sponsored content. In sum, this current study examines how a consumer makes inferences of influencer motive about a sponsored post, as well as uses their prior agent knowledge (influencer credibility) to interpret and cope with an Instagram influencer’s sponsored post. Guided by the Persuasion Knowledge Model and Attribution Theory, this study developed a scale of influencer product recommendation motives and tested a research model of consumer resistance to Instagram sponsored posts. Specifically, this model incorporated the construct of consumer 6 inference of a social media influencer’s motive for presenting the message, a missing link in previous studies, with the expectation that some types of influencer motives could mitigate consumer resistance to persuasion. The model also examined the role of influencer credibility on inferences of influencer motives. Furthermore, this study provided the implications and recommendations for advertisers and endorsers on how to make effective persuasion in compliance with the FTC regulation. The paper is organized as follows. Chapter 1 introduces current covert persuasive practices in social media, discusses the prevalence of the use of sponsored content in social media and the importance of persuasion knowledge. This is followed by identifying the research gap and study contributions. Chapter 2 begins with a brief overview of study background and then introduces two theoretical frameworks, PKM and Attribution Theory. Chapter 3 focuses on the scale development of influencer motives and comes up with a sufficient measure of influencer product recommendation motives. Chapter 4 focuses on developing a theoretical model and hypotheses to explain consumer responses to sponsored electronic word of mouth (eWOM) in social media. Chapter 5 discusses the experimental design, including the stimuli development, sample, procedure, as well as measures. Chapter 6 provides data analysis and results. Chapter 7 discusses the general findings and concludes with academic, practical and policy implications, limitations and future research directions. 7 CHAPTER 2 BACKGROUND Types of Social Media Influencers According to Mavrck, an influencer marketing agency, it is useful to think of three types of social media influencers: mega-influencers, macro-influencers, and micro-influencers (Gottbrecht, 2016). The categorization is based on three criteria: reach (“ability to deliver content to target audience”), relevance (“connection to a brand or topic”), and resonance (“ability to drive a desired behavior from an audience”). Mega-influencers are those celebrities or social media stars who have over 1 million followers on social media, such as Kim Kardashian West. They have the highest reach, but the lowest resonance. Macro-influencers are those professionals, bloggers, or experts who have 10,000 to 1 million followers. They have the highest relevance, due to their expertise and influence in specific categories such as lifestyle, fashion, or business. Micro-influencers are those everyday consumers who have between 500 and 10, 000 followers. Although they have the lowest reach, they have the highest resonance. Dhanik (2016) states that micro-influencers are marketer’s favorites due to their affordable costs, niche target audience, and high levels of trust and engagement with followers. A study by Simply Measured shows that 92% of consumers trust word of mouth recommendations from their personal connections, while only 33% trust ads (Carlson, 2016). Another study shows that 49% of Twitter users rely on recommendations from Twitter influencers, almost as much as from their friends (56%). Forty percent made an online purchase after seeing the item used by a Twitter influencer (Swant, 2016). Sharing or talking about a brand 8 on social media is significantly correlated with purchase behavior, according to a Nielson survey (Kapadia, 2016). Not only do social media influencers bring value to the partnered brands, but they also benefit from the partnership. Influencers gain money from the sponsorship, and many make a living out of it. According to Socialyte, a social media influencer could charge $250 to $100,000 per social media post based on the number of their followers (Chafkin, 2016). The well-known American teen pop star Selena Gomez is currently the number one on the influencer rates list. She charges $550,000 per social media post across Facebook, Twitter, and Instagram (Heine, 2016). Furthermore, when they provide useful information and economic values (e.g., giveaways, discount) to their audience, it also helps them gain reputation and exposure by increasing the number of their followers, likes, comments, or shares of their posts. Sponsored Content in the Form of eWOM Brand-related social media posts created by social media influencers are electronic word of mouth (eWOM) messages, defined as “any positive or negative statement made by potential, actual, or former customers about a product or company, which is made available to a multitude of people and institutions via the Internet (Hennig-Thurau, Gwinner, Walsh, & Gremler, 2004, p. 39).” The definition of eWOM develops from the original WOM, which is “oral, person to person communication between a receiver and a communicator whom the received perceives as non-commercial, concerning a brand, a product, or a service. (Arndt, 1967b, p. 3)” In the original definition, Arndt emphasized that the word of mouth communicator is perceived as independent of the manufacturer or the commercial entities. Thus, WOM is informal communication and 9 considered to have more credibility than formal, commercial content; the speaker is perceived by the receiver as having nothing to gain by offering a disingenuous opinion. Unlike traditional offline WOM which is normally two-way, asynchronous communication and takes place in small groups with close social ties, an eWOM communicator can make an asynchronous communication to a large scale of the audience at a fast speed (Cheung & Thadani, 2010, 2012; Hung & Li, 2007). Furthermore, the electronic nature undermines the receiver’s abilities to evaluate the credibility of the communicator and message. Empirical studies have shown that that eWOM influences product judgements (Lee & Youn, 2009), product choice (Huang & Chen, 2006), brand image (Sandes & Urdan, 2013), attitudes (Lee, Park, & Han, 2008; Lee, Rodgers, & Kim, 2009), usefulness and social presence (Kumar & Benbasat, 2006), purchase intentions (Bickart & Schindler, 2001; Park & Lee, 2008; Park, Lee, & Han, 2007; Sandes & Urdan, 2013; Xia & Bechwati, 2008) and product sales (Babić, Sotgiu, & Bijmolt, 2015; Chevalier & Mayzlin, 2006; Forman, Ghose, & Wiesenfeld, 2008; Hu, Liu, & Zhang, 2008; Liu, 2006). To encourage the creation and dissemination of positive eWOM and take advantage of its positive impact on consumer behavior, marketers and manufacturers provide incentives for people to write positive product recommendations on the review site or product website. As a result, Influencer Marketing is a popular marketing strategy. Influencer Marketing and Disclosure The rapid growth of social media has given rise to influencer marketing to reach and persuade potential consumers and increase brand awareness, image or sales in the advertising clutter. Brands first identify and partner with relevant social media influencers to showcase and promote their products to their audience across social media platforms. They then send the 10 influencers free product samples, a free trial/subscription, early access to use the product/service, and/or pay hundreds to millions of dollars per post. In return, social media influencers are expected to create brand-related content showcasing the product or service, reviewing the performance and make a recommendation, or sharing their experiences with their followers. According to a 2016 study from content marketing agency Linqia, 86% of marketers surveyed have been working with influencers, and 94% reported that influencer marketing was an effective part of their overall marketing strategy (Linqia, 2016). Industry statistics also show that influencer marketing demonstrates high ROI and cost efficiency. On average, brand marketers earn $6.85 in media value for every $1 they spent on influencer marketing campaigns (Buyer, 2016). Influencer marketing is projected to be a $5-10 billion global market by 2020 (Mediakix, 2015). In the 2008 holiday season, Kmart launched an influencer marketing campaign by giving six famous bloggers $500 gift certificates to shop at their stores. In return, each blogger wrote a sponsored post about their Kmart shopping experiences without Kmart’s censorship and gave away one $500 gift certificate to their followers to go on the Kmart shopping spree. According to their agency Izea, this campaign successfully created online buzz, resulting in over 2,000 related blog comments, over 2,500 twitter contest entries, and approximately 500,000 reaches via blogs and Twitter. Kmart’s Social Media Index (SMI), the performance metric that tracks brand’s “share of social voice,” also increased from 14.49% in November to 23.21% in December (Lukovit, 2008). If a social media influencer posts about products or services in partnership with a brand, regardless of the valence of the content, the sponsored post becomes an ad or an endorsement and is therefore subject to the FTC Act and the FTC’s Endorsement Guides (Federal Trade 11 Commission, 2015). In other words, digital influencers should make a truthful and honest statement, and fully disclose their material connections with sponsors by using disclosure in influencer marketing. The material connections include “a business or family relationship, monetary payment, or the provision of free products (Federal Trade Commission, 2017).” When Warner Brothers' advertising agency promoted a video game Middle-Earth: Shadow of Mordor, they hired influencers to post positive gameplay videos on social media and their YouTube channels. During the campaign period, over 5.5 million visits watched the sponsored videos. However, the marketer and influencers did not adequately disclose the sponsorship, such that disclosures in YouTube videos were visible only if people clicked on the “Show More” button in the description box, and Facebook or Twitter posts promoting these videos did not contain any disclosures. The FTC fined and ordered the company to fully disclose the sponsorship of future influencer marketing practices (Federal Trade Commission, 2016a, 2016b). The way marketers or digital influencers disclose sponsorship varies widely. In the recent Warner Brothers’ case, the YouTube star PewDiePie disclosed his partnership with Warner Brothers in his sponsored video, but it was limited. There was one line in the description box, “This video was sponsored by Warner Brother,” which was buried in the middle of other links. Even many journalists missed the disclosure and quickly blamed PewDiePie for not disclosing the sponsorship in news articles. On the other hand, in Kmart’s 2018 campaign mentioned in the introduction, Chris Brogan, a famous social media marketer, disclosed his relationship with Kmart and its agency Izea in several ways. He used the words “sponsored post” in the title, a disclosure statement – “This post is a sponsored post on behalf of Kmart via Izea. The opinions are mine” – in the first line of his article, and another disclosure statement with a link to the 12 agency in the last line – “The preceding was a sponsored post. For more information about the sponsor (Kmart), see Izea” (Brogan, 2008). Sponsorship Disclosure and Consumer Resistance Consumer Processing of Sponsorship Disclosure Theories of Information processing explains that people acquire, process, and use information through several stages of the cognitive process: exposure to sensory information; attention; comprehension; retention/retrieval of knowledge in memory; and decision making based on the formation and change of product beliefs, brand evaluations, and attitudes (Mazis & Staelin, 1982; McGuire, 1976). In the case of consumer cognitive processing of sponsored content, a consumer is first exposed to the sponsored content (exposure), and then pays attention to the sponsorship disclosure and sponsored content (attention). After that, the consumer will need to make sense of the content and the source’s behavior (comprehension). During this stage, s/he will recognize the sponsorship and further wonder the underlying reasons and motives for the influencer’s behavior. Finally, the consumer will evaluate the content and make a decision (evaluation and decision making). Wojdynski (2016) concluded that consumers must engage in two sequential cognitive processes attention and comprehension – for disclosure to have its intended effect. First, a consumer needs to focus attention on the disclosure itself, thereby increasing the likelihood of message elaboration and subsequent understanding or comprehension of the message’s meaning. Under these conditions, a consumer can understand the information conveyed in the disclosure. Empirical studies have shown that a large portion of participants did not pay attention, recognize 13 or recall disclosure in sponsored content (e.g. Boerman et al., 2012, 2017; Wojdynski & Evans, 2016). Consumer Resistance as a Negative Consequence of Sponsorship Disclosure Consumers’ cognitive and affective responses to a persuasive message are an important factor for understanding persuasion effects. This study focuses on one particular type of consumer response: resistance to persuasion, defined by McGuire (1964) as “the ability to withstand a persuasive attack” (Knowles & Linn, 2004, p. 4). Knowles & Linn (2004) identified four faces of resistance: reactance, distrust, scrutiny, and inertia. Psychological Reactance Theory (Brehm, 1966) states that people want to maintain their choice and autonomy and do not want to be manipulated. When consumers experience a threat to their freedom, they will evoke resistance strategies to cope with an unwanted persuasive attempt. Jacks and Cameron (2003) identified seven resistance strategies: 1) attitude bolstering, 2) negative affect, 3) assertion of confidence, 4) selective exposure, 5) counter-arguing, 6) source derogation, and 7) social validation. Furthermore, van Reijmersdal et al. (2016) demonstrated that resistance could be both cognitive and affective. When people are motivated to resistance persuasion, they are likely to experience cognitions such as counter-arguing and source derogation(e.g. Tannenbaum, Macauley, & Norris, 1966), or experiencing negative affects such as anger, upset, and irritations (e.g. Jacks & Cameron, 2003; van Reijmersdal et al., 2016). When people do pay attention to and understand the disclosure, common findings are that a disclosure resulted in consumer resistance for the brand marketer, such as decreased perceived message credibility (e.g. Nekmat & Gower, 2012; Wojdynski & Evans, 2016), skepticism (e.g. Boerman et al., 2012), negatives brand attitudes (e,g, Wei, Fischer, & Main, 2008) and less 14 purchase intention (e.g. Nekmat & Gower, 2012). Furthermore, Boerman, van Reijmersdal, & Neijens (2013) found disclosure of sponsored content in a television program decreased the number of positive thoughts about the brand. However, disclosure did not always lead to consumer resistance or negative consequences. Some studies found that disclosure enhanced brand recall and brand recognition (Boerman et al., 2012; van Reijmersdal, Lammers, Rozendaal, & Buijzen, 2015)), and increased perceived credibility of the communicator of a sponsored eWOM (e.g. Carl, 2008). Boerman & van Reijmersdal (2016) summarized the boundary conditions that moderate the effectiveness of disclosure: 1) the characteristics of disclosure, such as timing, duration, language and modality, and 2) the receiver characteristics such as cognitive capacity and consumer skepticism. Persuasion Knowledge Model Three Types of Knowledge According to PKM, a consumer uses three types of knowledge – specifically knowledge of persuasion, the influence agent, and the topic - to interpret and cope with persuasion attempts (Friestad & Wright, 1994). Persuasion knowledge consists of a set of beliefs about psychological mediators (e.g. attention), marketers’ tactics, the effectiveness and appropriateness of marketers' tactics, marketers' persuasive intent and goals, and one's own coping tactics and goals. Agent knowledge consists of “beliefs about the traits, competencies, and goals of the persuasion agent.” Topic knowledge consists of “beliefs about the topic of the message” (Friestad & Wright, 1994, pp.3). As Campbell & Kirmani (2008) point out, the lines between three types of knowledge are blurry, and they are not independent and separate constructs. For instance, if a consumer tries to understand the social influencer’s motives to recommend a product, which is persuasion 15 knowledge based on the Friestad & Wright (1994)’s definition, is this consumer drawing upon their knowledge about the persuasion activated by disclosure, or their agent knowledge (e.g. this influencer is a sincere person and often makes unbiased statements), or their topic knowledge (e.g. this product is of good quality), or all of them? Therefore, Campbell and Kirmani (2008) recommended to redefine persuasion knowledge as “all knowledge related to persuasion, including persuasion-related knowledge of an agent or topic,” agent knowledge as “all nonpersuasion-related knowledge having to do with characteristics of the agent”, and topic knowledge as “all non-persuasion-related knowledge about the topic or content of the persuasion attempt” (Campbell & Kirmani, 2008, p. 552). Change-of-Meaning Principle The Change-of-Meaning Principle in PKM states that when consumers learn about persuasion strategies or tactics and possess certain persuasion knowledge (e.g. recognition of sponsorship), they will disengage from the persuasion interaction (called “the detachment effect”), and discount what the influence agent (e.g. the influencer) says (Friestad & Wright, 1994), mitigating the message persuasiveness. This principle is useful to explain resistance to persuasion caused by persuasion knowledge activation. In the context of Instagram sponsored content, this principle indicates that when a clear and distinctive disclosure is accessible to a consumer, the consumer recognizes the persuasive and selling intent of a sponsored Instagram post. Then the sponsored posts will take on a change of meaning. As a result, they will disengage from the post and discount what the influencer says in the post, and are more likely to resist the persuasion than being persuaded than when this knowledge is not activated. 16 Attribution Theory Attribution Theory consists of a family of theories and can be used to explain how receivers make inferences about reasons why communicators present a message. Attribution theory assumes that the receiver is a rational, logical person and focuses on the receiver’s attribution of the communicator’s motivation. Previous scholars have used theories of attribution to explain how consumers draw the inferences of the motives of a celebrity endorser (e.g. Choi & Rifon, 2012; Kamins, 1990), the motives behind a corporate sponsorship of a cause (e.g. Rifon, Choi, Trimble, & Li, 2004), and the motives of a WOM communicator (e.g. Curren & Folkes, 1987). Therefore, Attribution Theory can provide an additional explanation for the missing piece that PMK needs to examine: how people make inferences of influencer motives. Two Types of Attributions of Influencer Motives In advertising and marketing research, the most commonly used theories of attribution are Correspondent Inference Theory (Fiske & Taylor, 1991; Jones & Davis, 1965) and Kelley (1967, 1973)’s Attribution Theory. Correspondent inference theory “systematically accounts for a perceiver's inferences about what an actor was trying to achieve by a particular action (Jones & Davis, 1965, p. 222)”. In other words, people draw a correspondent inference of one’s behavior. Kelley (1967, 1973)’s Attribution Theory explains how receivers make inferences of the communicator’s motives. It posits that receivers attribute a communicator’s action with either an external/situational reason or an internal/dispositional reason. The attribution to the external reasons (e.g. the environment, the situation) is called situational or external attribution. The attribution of cause to internal reasons (e.g. the characteristics of the person), is called dispositional or internal attribution. 17 Settle and Golden (1974) stated that when exposed to an ad, a consumer would attribute the advertising message either to the advertiser’s desire to sell the product, or to the characteristics of the product being advertised. They further postulated that if the advertising message is attributed to the advertiser’s desire to sell, the consumer would be uncertain about the product characteristics and the likelihood to purchase would decrease. On the other hand, if the advertising message is attributed to actual product characteristics, the consumer would have higher certainty and higher likelihood to purchase the product. In the sponsored content context, a consumer will make two types of attributions about an influencer's motives for writing a sponsored post: either inferring inference of influencer recommendation motives to the external circumstances (e.g. making money, selling a product, or enhancing one’s image); or inferring influencer recommendation motives to disposition of the product and the communicator (e.g. loving the product, wanting to share information or help others). Discounting Principle The discounting principle (Jones & Davis, 1965; Kelley, 1973) explains how people make inferences of other’s behavior from a single observation. Generally speaking, the discounting principle refers to “the role of a given cause in producing an effect is discounted if other plausible causes are also presented (Kelley, 1971, p. 8).” It can be used to explain the different levels of consumer resistance to messages due to different attributions of the inferred motives. Several empirical study findings have shown evidence for the discounting principle. For instance, Brandt, Vonk, & van Knippenberg (2011) found that consumers will discount a third 18 party’s product recommendation when they are aware that the recommendation is caused by other salient motives rather than by product properties. If one recommends this product, s/he must believe this is a good product. However, when consumers believe the influencer’s opinion is caused by circumstances (e.g. get paid, or enhance self-image), they will discount the content and perceived the content as biased, and thus be less persuaded by the review (Lee & Youn, 2009). Sen and Lerman (2007) explored how consumers understand and make inferences of negative reviews. Their results showed that consumers were more likely to attribute the negative product review which was judged on hedonic criteria (e.g., consumer experience) to the self-serving or non-product related reasons. As a result, they were less likely to find these negative hedonic reviews useful. On the other hand, consumers were found to be more likely to attribute the negative product review which was judged on utilitarian criteria (e.g., functions) to the product related reasons. As a result, they were more likely to find negative utilitarian reviews useful. Correspondent Inference Bias Attribution process is subject to the correspondent inference bias (Sen & Lerman, 2007), when the discounting principle does not apply. Correspondent inference bias indicates that people tend to make attributions of a person’s behavior to dispositional causes rather than situational causes, even if they know other situational factors exist. Correspondence bias is often used to explain the advertising effectiveness of celebrity endorsement. One of the most cited studies was conducted by Cronley, Kardes, Goddard, and Houghton (1999), whose findings supported consumer correspondent inferences bias in evaluating celebrity endorsed ads. In this study, American student participants made correspondent inferences of Cindy Crawford’s product endorsement, even if other situational 19 causes are present. Specifically, they believed Cindy Crawford actually liked the product, frequently used the brand and viewed the brand as a good product, even if they know the celebrity received $6 million dollars to endorse the product. These correspondent inferences are subsequently positively related to attitudes toward the ad, the brand, and the endorser. 20 CHAPTER 3 SCALE DEVELOPMENT OF INFLUENCER MOTIVES It is not uncommon for consumers to read social media posts that feature a brand or recommend a product. How do consumers interpret social media influencer’s motives to recommend a product? What does a consumer think about why the social media influencer makes these persuasive statements? Are those influencer motives dichotomous? Or could multiple motives co-exist? This chapter examines how consumers infer the motives behind the behavior of social media influencers making product recommendations in social media. The goal of this chapter is to conceptualize and develop an efficient scale to measure consumer perceptions of influencer product recommendation motives, one that could be used in the following experiment to examine consumer resistance to sponsored Instagram posts. Conceptualizing Influencer Motives based on PKM Persuasion knowledge is the key concept used to understand how consumers interpret and cope with sponsored content. Ham et al. (2015) categorized two types of persuasion knowledge: dispositional and situational persuasion knowledge. Dispositional persuasion knowledge reflects “the culmination of consumers’ knowledge, skills, abilities, exposure to, and experience with persuasion and advertising (Wojdynski, Evans, & Hoy, 2017, p. 5);” and situational persuasion knowledge refers to “the evaluations and behaviors consumers carry out in response to the recognition of a persuasive communication or advertisement (Wojdynski et al., 2017, p. 5).” Beliefs about influencer motives can be seen as personal experience as a consumer and with consumer interpretation of the specific communication, a combination of both dispositional and 21 situational persuasion knowledge, which are the centerpiece of persuasion knowledge (Campbell & Kirmani, 2008). Research using persuasion knowledge usually considers whether persuasion agents have perceived ulterior motives (Campbell & Kirmani, 2000). If a consumer infers that there is an ultimate motive, she is likely to use persuasion knowledge to cope with the interaction. However, the ulterior motive is not clearly defined in previous literature. General use of the term “ulterior motive” has been equated with the hidden intention to persuade people. Sponsored content, like all advertising messages, is created to persuade. But different consumers have different interpretations of the intent, and in different contexts. For instance, Tutaj & van Reijmersdal (2012) found that consumers perceived three different intentions associated with sponsored content and banners ads: 1) the selling intent, which aims to sell the products/services, 2) the persuasive intent, which aims to influence others’ opinion and 3) the informational intent, which aims to provide information to others. Jeong and Lee (2013) generalized two types of “ultimate motives” of a website: 1) customer-oriented motives, for instance, the website tries to help and satisfy a customer, or 2) firm-serving motives, for instance, the website tries to sell as much as it can. It is worth noting that among PKM studies which examine the motives of influencer agents (e.g. an ad, a salesperson, an influencer), they either broadly focus on “ultimate motives”, which are vague and too general or merely focus on the “selling intent” or “persuasive intent,” and are overly simplistic and narrow-sighted, based mostly on past research that has dichotomized motive attributions. When a consumer sees a social media influencer post sponsored content, would it be possible for a consumer to think the influencer wants to sell the product and persuade you, but also other motives that do not intent to sell? Unfortunately, PKM 22 literature did not discuss or explain non-selling motives much. Therefore, we examined literature from other field to understand such as intentions to inform or entertain others, engage and connect with the audience, or enhance public image. Conceptualizing Influencer Motives based on eWOM literature Sponsored posts created and shared by social media influencers can be considered as eWOM messages. A rich body of eWOM literature has studied what motivates people to generate word of mouth messages about brands and products on social media. Important factors such as consumer evaluations (i.e. satisfaction, commitment, perceived value, quality, trust and loyalty), personality traits (i.e. innovativeness and altruism), instrumentalism (i.e. selfenhancement), involvement (i.e. product involvement), as well as social function (i.e. social connection and belonging) have been identified to influence the generation of eWOM (Anderson, 1998; Arndt, 1967a; Brown, Barry, Dacin, & Gunst, 2005; Cheung & Thadani, 2010; Daugherty, Eastin, & Bright, 2008; Dichter, 1966; Hennig-Thurau et al., 2004; Matos & Rossi, 2008; Sun, Youn, Wu, & Kuntaraporn, 2006; Sundaram, Mitra, & Webster, 1998). Based on the eWOM literature, we summarize that when a consumer sees a social media influencer posted sponsored content, s/he may think this influencer wants to promote the products (selling motives), to share useful information with others (sharing motives), to engage with the audience (socializing motives), to enhance own public image (image motives), and/or to help others (helping moitves). It is worth noting that previous eWOM research fails to recognize financial incentives, as another reason to generate eWOM. This could be due to the fact that Arndt’s (1967b) original definition of word of mouth, “oral, person to person communication between a receiver and a communicator whom the received perceives as non-commercial, 23 concerning a brand, a product, or a service, (pp.3)” emphasized that word of mouth (WOM) communicator is perceived as independent of the manufacturer or the commercial entities. However, Influencer Marketing is an attempt to disguise the sponsored poster as someone who is engaging in pure eWOM, and this increasing amount of eWOM messages that blend branded information naturally into the online editorial environment may be detected by the consumer. Therefore, economic benefits are also another important motive of the influencer to generate and share eWOM, which eWOM literature failed to address. Conceptualizing Influencer Motives based on Attribution Theory Used in the study of celebrity endorser and sponsorship effects, Attribution Theory may provide a missing link for understanding consumer perceptions of influencer motives. Kelley (1967, 1973)’s Attribution Theory explains how receivers make inferences of the communicators. It posits that receivers attribute a communicator’s action with either an external/situational reason (e.g., the environment, the situation) or an internal/dispositional reason (e.g., the characteristics of the person). Celebrity endorsement studies using Attribution Theory have shown that consumers are likely to infer one of two types of motives for a celebrity endorsement: extrinsic or intrinsic motives. For instance, Choi (2002) summarized three plausible causes for celebrity endorsements of a product: 1) product attribution, the influencer’s belief in the product qualities and/or genuine affection for the product; 2) money attribution, the compensation the influencer received for the endorsement, 3) and image management attribution, the motivation to maintain and enhance positive images. Among these attributions, money and image management attributions were extrinsic, and product attributions is intrinsic. Similarly, Rifon, Choi, Trimble and Li (2004) also 24 identified four motives for a company to sponsor a CSR website: 1) altruistic motives, for instance, the company cares about getting health information to their consumers; 2) profit motives, for instance, the company wants to persuade people to purchase their products; 3) public image motives, for instance, CSR sponsorship creates a positive corporate image; and 4) the ethical or moral reasons. The altruistic motives are intrinsic, while the product motives and public image motives are extrinsic. It is reasonable to expect that the same process would occur for consumers viewing a micro-celebrity post. In the social media context, the influencer motives for posting a sponsored post could be either extrinsic, such as profit or image enhancement driven motives, or intrinsic, such as a genuine belief that the product has merit, or a personality trait that wants to others. It is worth noting here that Rifon, et al. (2004) empirically showed that intrinsic and extrinsic motives are not mutually exclusive and perceptions of both types of motives can coexist with each other, which means that people can process different motives at the same time. Conceptualizing Influencer Motives in Social Media In sum, none of previous litearture fully explore the types of influencer motives for posting and sharing sponsored content in social media. PKM merely focused on the selling intents and neglected all the other possible non-selling motives. eWOM literature addressed all the possible non-selling motives but failed to recognize financial incentives that also drive influencer to make product recommendations. Attribution Theory studies explained many motives but failed to take into account the intentions to sell, or the intention for social connection and social bonding. This current chapter will combine these three literature and explore all the 25 possible influencer motives, including motives and whether a consumer would infer multiples motives a consumer when they see a social media influencer posts sponsored content. Scale Development Process We followed the scale development procedures suggested by Churchill (1979) and a few recent consumer and marketing research (Babin, Darden, & Griffin, 1994; Goldsmith & Hofacker, 1991; Homburg, Schwemmle, & Kuehnl, 2015; McNally & Griffin, 2007). The procedure is shown in Table 1. Table 1. An Overview of the Scale Development Process Process Initial Items Generation Scale Development and Items Reduction Scale Validation Data and Methods Surveyed 100 participants with a thought-listing approach (Study 1) Two expert judges coded the responses and write the items Surveyed 335 participants (Study 2) Statistical procedures: EFA to assess dimensionality and reduce items Surveyed 461 participants (Study 3) Statistical procedures: CFA to assess reliability & validity Results Generated an initial scale of 67 items. The EFA model generated a 31-item scale of six factors with good validity and reliability. The CFA model confirmed the 31-item scale of six factors with a good model fit and developed a concise measure of 19-items with good validity and reliability. Study 1: Initial Scale Development Design and Procedure We posted a Human Intelligence Task (HIT) on Amazon MTurk and only allowed MTurk workers who lived in the United States, aged 18 or older and had at least 95% HIT 26 approval rate to participate in the study. Qualified participants received $1 in exchange of their completion of a 10-minute survey. After indicating their consent, participants were asked to read a sponsored post recommending Optimum Nutrition protein powder. Then, we used the thought-list procedures to collect thoughts about influencer motives, by asking “Why do you think the person posted and shared his/her opinion of the Optimum Nutrition protein powder on social media?” Participants were given at least two minutes to write down as many reasons as they could come up with for why they think someone would recommend a product in social media. After that, participants were asked to report their demographic information. Sample We recruited a total of 100 adults from Amazon Mturk. Among them, 40 were female (40%), 58 were male (58%), and two people preferred not to disclose. The majority of the participants were Caucasian (72%), followed by African American (11%), Hispanic or Latino (6%), and Asian American (6%). Their ages ranged from 22 to 74 years old, with an average age of 40 (SD=10.6). They reported that they spent an average of 14.4 hours on social media every week (SD=11.8), and spent an average 170.8 seconds (SD=119) on writing the reasons or influencer motives during the study. Results Two coders used a line-by-line analysis to code the thoughts of influencer motives. As this was an exploratory practice attempting to identify different types of motives, the coders discussed each of their decisions to come to some consensus on the range of motives that 27 appeared in the data. The coders identified six major types of influencer motives for posting sponsored content: money, selling, fame, love, sharing, and helping. Money motives are those responses that mentioned the writer had a material connection with the firm, or received an incentive or free product or financial compensation, or wanted to promote and sell the products or to persuade others to buy or use the product. Image motives are those responses that mentioned the writer wanted to enhance and maintain their self-image, feel good about themselves, seek positive evaluations from others. Love motives are responses that mentioned the writer liked the product; are satisfied with the product and/or product performance; or perceived value and good quality in the product. Sharing motives are those responses that mentioned the writer wanted to bond with others and maintain relationships, or to enjoy sharing on social media. Helping motives are those responses that mentioned the writer wanted to help others with their purchase decision or to save others from negative experiences. As a result, two experts generated an initial list of 64 items, as shown in Table 2. This initial list not only combined previous PKM, eWOM and Attribution Theory literature to create a comprehensive measure of influencer motives, but also contributed new and unique items to the social media context, such that “s/he wants to gain followers/likes/shares.” 28 Table 2. List of the Initial 64-Item Scale Type of Motives Label Money_1 Money_2 Money_3 Money_4 Money_5 Money Motives Selling Motives Image Motives Money_6 Money_7 Money_8 Money_9 Money_10 Selling_1 Selling_2 Selling_3 Selling_4 Selling_5 Selling_6 Selling_7 Image_1 Image_2 Image_3 Image_4 Image_5 Image_6 Image_7 Image_8 Image_9 Image_10 Image_11 Image_12 Image_13 Image_14 Image_15 Image_16 Item S/he benefits by this sponsorship. S/he is paid to recommend the product. S/he wants to make money. S/he is promoting the product for the benefits of the brand. S/he wants to monetize their relationship with their followers in the future. S/he receives a free product. S/he receives promotional coupons. S/he works for the company. S/he wants to receive future sponsorships. S/he gets commission. S/he wants to persuade people to buy the product. S/he wants to promote the product. S/he wants to sell the product. S/he wants to encourage others to buy the product. S/he wants to help attract new customers. S/he wants to increase product sales. S/he wants to increase company profits. S/he wants to create a positive public image on social media. S/he wants to get exposure. S/he wants to maintain their reputation. S/he wants sponsors to notice their social influence. S/he wants to gain followers. S/he wants to gain likes. S/he wants to gain shares. S/he wants to show off their lifestyle. S/he wants to brag about his or her new recipe. S/he wants to impress others with their knowledge about the product. S/he wants to impress others with their expertise. S/he wants to feel like an expert. S/he wants to promote himself or herself as the guru on social media. S/he wants to appear successful. S/he wants to brag about being sponsored. S/he is seeking attention. 29 Table 2. (cont’d) Love Motives Image_17 Love_1 Love_2 Love_3 Love_4 Love_5 Love_6 Love_7 Love_8 Love_9 Love_10 Social_1 Social_2 Social_3 Sharing motives Social_4 Social_5 Social_6 Social_7 Social_8 Social_9 Helping_1 Helping_2 Helping_3 Helping_4 Helping_5 Helping Motives Helping_6 Helping_7 Helping_8 Helping_9 Helping_10 Helping_11 S/he loves attention. S/he frequently uses the product. S/he views the brand as a good product. S/he likes the product. S/he is excited about the product performance. S/he expresses their enjoyment with the product. S/he is satisfied with the product. S/he thinks this product works for this person. S/he thinks this product does a lot of good for this person. S/he believes the product appears superior to other competitors. S/he thinks the product exceeds their expectation. S/he wants to entertain others. S/he wants to engage with the audience. S/he wants to provide a topic for further discussion with the readers. S/he wants to share the product they use with others. S/he wants to express their own opinion of the product. S/he wants to share their recipe with others. S/he enjoys sharing on social media. S/he is bored. S/he wants to connect with others on social media. S/he cares about the followers. S/he has a genuine concern for the welfare of the followers. S/he cares about getting useful information to the followers. S/he wants to give information about how to make healthy food. S/he wants to let people know how to use the product. S/he wants to help others to make better purchase decisions. S/he wants to help others get the information they want. S/he wants to inform interested groups about the product. S/he wants to encourage others to take advantage of the product. S/he wants other to benefit from their experience. S/he knows their followers are interested. 30 Study 2: Scale Development Design and Procedure This study tests the 61-item scale with an online survey. After participants were asked to read a sponsored post recommending Optimum Nutrition protein powder, they were then asked to state their agreement with 61 possible reasons why someone would recommend a product in social media. All items were measured on a traditional five-point Likert scale (5=strongly agree, 1= strongly disagree). After that, participants reported their demographic information. Sample 341 Amazon Mechanic Turk workers who lived in the United States and had at least a 90% HIT approval rate participated in the online survey. They received $1.5 as an incentive after their completion. The sample size is determined based on several considerations. Hinkin, Tracey, and Enz (1997) recommends a minimum sample size for a EFA of 150 observations. Recent literature suggests a minimum item-to-response ratio of 1:5 for principal component analysis (Osborne & Costello, 2004). Therefore, to factor analyze 41 items, this study needs at least 204 participants. Furthermore, Comrey and Lee (1992) suggests that “the adequacy of sample size might be evaluated very roughly on the following scale: 50 – very poor; 100 – poor; 200 – fair; 300 – good; 500 – very good; 1000 or more – excellent” (p. 217). Therefore, a sample size of 341 is sufficient for EFA. Among the participants, 53% were female and 46% were male. Their age ranged from 18 to 65 years old (Mean=37, SD=11). They were mostly white (80.1%), had a college degree (49.6%), and had a household income between $30,000 and $59,999 (47.6%). 31 Results Exploratory factor analysis (EFA) was used to identify the dimensionality of initial items. We performed the principal component factor analysis with Promax rotation as the factors were assumed to be correlated, according to Hinkin, Tracey, and Enz (1997) using SPSS 24. Eight factors with eigenvalues greater than 1 was extracted, explaining a total of 66.8% variance. However, not every item loaded on the intended dimension and there were large cross loadings. Therefore, 14 items that had either factor loadings lower than .60 or cross loadings were removed. The two experts also decided to remove 16 similar items to reduce the number of items, dropping down to 31 items. As shown in Table 3, A principal component factor analysis with Promax rotation analysis of this 31-item scale revealed six factors with eigenvalues greater than 1, and explained a total of 73.9% variance. All items loaded on the factor they were expected to load on, indicating good discriminant validity. All the factor loadings were higher than .50, indicating good convergent validity. The Cronbach’s Alpha reliability of each type of motives are above .84, indicating great internal reliability. Table 3 presents the factor loadings in Exploratory Factor Analysis and the reliability of each factor. 32 Table 3. EFA and Reliability Results of the 31-Item Scale Item Money_1 Money_2 Money_5 Money_6 Money_7 Money_9 Selling_3 Selling_5 Selling_6 Selling_7 Image_5 Image_6 Image_7 Image_13 Image_14 Image_16 Image_17 Love_2 Love_3 Love_5 Love_6 Love_7 Sharing_4 Sharing_5 Sharing_7 Helping_1 Helping_2 Helping_3 Helping_6 Helping_7 Helping_10 1 0.77 0.74 0.58 0.94 0.94 0.82 2 3 4 5 6 0.93 0.80 0.94 0.95 0.73 0.75 0.70 0.81 0.72 0.86 0.78 0.90 0.95 0.86 0.89 0.90 0.80 0.77 0.85 0.90 0.95 0.82 0.77 0.81 0.67 33 Mean 4.24 4.06 4.10 3.98 3.92 4.19 4.19 4.31 4.22 4.01 4.16 4.09 4.00 3.86 3.83 3.85 3.81 3.96 4.07 4.04 4.11 3.93 4.10 3.89 4.09 3.27 3.07 3.41 3.43 3.51 3.63 SD Cronbach’s ⍺ 1.05 1.13 0.99 0.91 1.09 1.07 1.03 1.10 0.97 0.93 1.03 1.11 0.97 0.99 1.05 1.12 0.89 1.11 1.07 1.09 1.09 1.02 1.05 0.95 1.00 1.00 1.10 1.19 0.84 1.06 1.12 1.15 1.18 0.93 1.14 1.14 1.10 Study 3: Scale Validation Design and Procedure Another round of data is collected to validate the 31-item in Study 2. All the design and procedure in Study 3 are the same as in Study 2. Sample 461 US adults participated in the online survey. 66 of them were recruited from Mturk, and received $1 as an incentive for their participation. 395 of them were recruited from a large Midwestern University, in exchange for extra credits. The sample size meets the minimum requirement of 200 observations for a CFA (Hinkin et al., 1997). The majority of the participants were female (62.5%) and Caucasian (66.6%). Their ages ranged from 18 to 65 years old, with an average age of 24 (SD=7). Results We conducted a confirmatory factor analysis (CFA) with the 31 items using AMOS 23. A maximum likelihood factor analysis using the Promax with Kaiser Normalization rotation methods confirmed the six factors. We used Fornell and Larcker (1981)’ s criterion to assess the convergent validity and discriminant validity of the construct. As shown in Table 4, a good model fit (χ2/d.f.=2.65, p <.001; NFI=.89, IFI=.93, TLI= .91, CFI=.92, RMSEA=.06, AIC=1327.93, BCC=1344.08), all above the thresholds suggested by Hooper, Coughlan, and Mullen (2008), except for the 2 statistic. Jöreskog and Sörbom (1993) has shown χ2 value tends to be significant when the sample size is large. 34 Table 4. CFA Results: Validity Test and Factor Correlation Model 1 Money Selling Image Love Sharing Helping Model 2 Money Selling Image Love Sharing Helping CR AVE MSV Money 0.91 0.91 0.88 0.94 0.75 0.87 0.63 0.72 0.52 0.76 0.5 0.53 0.39 0.39 0.25 0.47 0.47 0.34 1 0.63 0.50 0.03 0.02 -0.14 0.89 0.9 0.85 0.92 0.74 0.83 0.73 0.76 0.58 0.78 0.5 0.62 0.4 0.4 0.27 0.46 0.46 0.38 1 0.63 0.52 0.03 0.03 -0.14 Selling Image Love Sharing Helping 0.22 0.18 0.06 1 0.13 0.25 0.06 1 0.68 0.51 1 0.58 1 1 0.36 0.17 0.15 0.06 1 0.13 0.24 0.05 1 0.68 0.55 1 0.62 1 1 Table 4 presents the results of the reliability and validity test. Results in Table 5 revealed that the composite reliability of each factor was above the .7 threshold (Hair, Black, Babin, Anderson, & Tatham, 2006). Also, the AVE of each factor were above. 50 (Malhotra & Dash, 2011), demonstrating good convergent validity. Furthermore, the factor correlations were smaller than each factor’s square root of AVEs and thus demonstrated good discriminant validity (Fornell & Larcker, 1981). Furthermore, we tested an efficient scale with 19 items (Model 2) selected from the 31 items based on factor loadings and face validity. CFA results in Table 4 and 5 confirmed that Model 2 has an adequate model fit (χ2/d.f.=2.06, p <.001; NFI=.95, IFI=.97, TLI= .96, CFI=.97, RMSEA=.048, AIC=426.52, BCC=433.07), according to Hu and Bentler (1999). Similarly, the composite reliability and AVE of each factor were above the thresholds, demonstrating good 35 convergent validity. Both factors’ square root of AVEs were greater than the factor correlation, demonstrating good discriminant validity. Appendix A presents the scale items in Model 2. Table 5. CFA Model Fit Indices Model 1: 31-item scale Model 2: 19-item scale *** p < .01 χ2/d.f. 2.65*** 2.06*** NFI .89 .95 IFI TLI .93 .91 .97 .96 CFI .92 .97 RMSEA 0.06 0.048 AIC 1327.93 426.52 BCC 1344.08 433.07 Scale Development Results The studies presented in this chapter emphasized on understanding how consumers perceive and make inferences of influencer product recommendation motives. This chapter first reviewed literature in Persuasion Knowledge Model, Attribution Theory, and eWOM literature, and then conceptualized consumer perception of influencer motives in the social media context as a multi-dimensional construct. Then, we described the procedure of scale development and validation that measured influencer motives, which demonstrated reliability and validity. Using a thought-listing approach, Study 1 identified six major types of influencer motives and generated an initial list of 64 items. Study 2 reduced 64 items to 31 items and tested the scale of 31 items with an online survey. Study 3 validated the 31-item scale and developed an efficient 19-item scale to measure consumer perceptions of influencer product recommendation motives in the context of social media. As shown in Table 6, This current study revealed that there are six dimensions of influencer motives: 1) money motives, 2) selling motives, 3) image motives, 4) love motives, 5) sharing motives and 6) helping motives. This six-factor model contributes to the current influencer motives literature by expanding the PKM literature, eWOM literature and Attribution theory studies to the comprehensive measure of influencer motives for sharing and posting 36 sponsored content in social media. The 19-item scale is also used to examine the role of influencer motives in consumer resistance to sponsored content in the following chapters. Table 6. Types of Motives In PKM, eWOM and Attribution Theory Studies Types of Motives Relevant Concepts Studied in PKM literature? Money Motives Money attribution; Profit Motives No Selling Motives Image Motives Love Motives Selling Intent Selfenhancement; Image enhancement attribution Customer satisfaction; Product attribution Studied in eWOM literature? Studied in Attribution Theory literature? Examples of Relevant Studies No Yes Choi (2002; Rifon et al. (2004) Yes No No No Yes Yes No Yes Yes Sharing Motives Social connection and belonging; Intention to share social information No Yes No Helping Motives Altruism No Yes Yes 37 Jeong & Lee (2013); Xie, Boush, & Liu (2013); Rodgers (2007); Tutaj & van Reijmersdal (2012) Choi (2002); Rifon et al. (2004); Yoon, Gürhan-Canli, & Schwarz (2006) Choi (2002); Cronley et al. (1999) Alexandrov, Lilly, & Babakus, (2013); Walsh, Gwinner, & Swanson (2004) Alexandrov et al. (2013); Jeong & Lee (2013); Yang (2013); Yoon et al. (2006) CHAPTER 4 THEORETICAL MODEL AND HYPOTHESIS DEVELOPMENT Role of Sponsorship Disclosure Numerous studies using PKM provide evidence that the FTC-required clear and conspicuous disclosure activates persuasion knowledge. For instance, Boerman et al. (2017) found that a sponsorship disclosure can help consumers to recognize celebrity endorsements on Facebook as advertising. Therefore, an FTC-required clear and conspicuous disclosure in the sponsored post will serve to inform readers that the influencer, in fact, has a material relationship with the brand. As a result, we would expect consumers make stronger attributions of extrinsic motives such as money motives, selling motives, and image motives. Discounting principle in Attribution Theory can provide an additional explanation for how the presence of disclosure affects inferences of influencer motives. Discounting principle refers to “the role of a given cause in producing an effect is discounted if other plausible causes are also presented (Kelley, 1971, p. 8).” Using this theory, Brandt, Vonk, and van Knippenberg (2011) found that consumers will discount a third party's product recommendations when they are aware that the recommendation is caused by other salient motives rather than by product properties. If one recommends this product, s/he must believe this is a good product (e.g. love motives). However, when consumers believe the influencer's opinion is caused by circumstances (e.g., to get paid, or to enhance self-image), they will discount the content and perceived the content as biased, and thus be less persuaded by the review (Lee & Youn, 2009). In other words, without seeing the disclosure or recognizing the message is sponsored content, a consumer would attribute the influencer’s recommendation to the intrinsic motive such as genuine love for 38 the product (love motives). However, when the disclosure is present and accessible, a consumer would perceive the second plausible reason such as profit gains. As discussed earlier, when a blogger posts a product recommendation, there are six plausible reasons: 1) this blogger receives compensation and wants to make money out of it (money motives); 2) this blogger wants to sell the product (selling motives); 3) this blogger wants to increase his or her social influencer and reputation (images motives); 4) this blogger loves the product (product motives), 5) this blogger wants to share a good product with others (sharing motives), and 6) this blogger wants to provide helpful information (helping motives). Without seeing the disclosure or recognizing the message is sponsored content, a consumer normally would attribute the bloggers’ recommendation to the intrinsic reasons, such as product motives, sharing motives and helping motives, and thus show less resistance to the sponsored post. However, when the disclosure is present and accessible, a consumer would perceive the plausible extrinsic reasons, such as the money, selling and image motives, and thus demonstrate resistance to the sponsored post. Based on the above discussion, we propose: H1. A sponsored post with disclosure will make readers generate stronger perceptions of a) money motives, b) selling motives, and c) image motives, than a sponsored post without disclosure. H2. A sponsored post with disclosure will make readers generate weaker perceptions of a) love motives, b) sharing motives, and c) helping motives, than a sponsored post without disclosure. H3. A sponsored post with disclosure will make readers show more resistance towards the sponsored post than a sponsored post without disclosure. 39 Role of Influencer Credibility Influencer Credibility as Agent Knowledge As brands hire social media influencers to create sponsored content, they become the brand ambassadors or the endorsers of the brand. Depending on a consumer’s knowledge of this influencer, the audience may perceive the influencer as a celebrity endorser (i.e. a megainfluencer), as an expert endorser (i.e. a micro- or macro-influencer), or an average consumer endorser for the brand (i.e. an unknown micro-influencer). According to McCracken (1989)’s meaning transfer model, endorsers are effective when their positive images or characteristics are transferred to the endorsed product or the corporation. Therefore, the effectiveness of the influencer depends on the desirable image of the endorser and consumer perceptions of that positive image. Perhaps the most studied concept that creates the desirable image of the endorser is source credibility, specifically influencer credibility in this study. Eisend (2006) summarized there are three dimensions of source credibility in marketing communications: the inclination toward truth (“the source will tell the truth”), the potential of truth (“the source knows the truth”), and the presentation of truth (providing an intensifying function for the source credibility perception, “appears to tell the truth”). In the field of advertising, source credibility is often seen as a function of expertise, trustworthiness and attractiveness (e.g. Eisend, 2006; Goldsmith, Lafferty, & Newell, 2000; Ohanian, 1990). Expertise is defined as “the extent to which a communicator is perceived to be a source of valid assertions (Hovland, Janis, & Kelley, 1953)”. It means that an endorser has special knowledge or experience with a topic or a product category. Trustworthiness is defined as “the degree of confidence in the communicator’s intent to communicate the assertion s/he considers most valid (Hovland et al., 1953).” Attractiveness 40 refers to how attractive the source is to the receiver (Ohanian, 1990). It is associated with physical beauty and to some extent likability. Some studies have also shown that a credible celebrity has a positive effect on ad effectiveness, such as increasing consumers’ attitudes towards the ad, towards the brand, and purchase intention (e.g. Choi & Rifon, 2012; Goldsmith et al., 2000). Effect of Influencer Credibility on Different Influencer Motives and Resistance As previously discussed in Chapter 2, social media influencers are considered to be “celebrities” with either a macro- or micro- influence, which comes from the content that s/he posts as his or real self. The effectiveness of celebrity endorsers has been studied using Attribution Theory, and research shows that consumers are likely to infer one of two types of motives for a celebrity endorsement, one that is driven by extrinsic reasons (such as money, selling, and image) and the second that is driven by intrinsic reasons (such as a genuine belief that the product has merits). It is reasonable to expect that the same attribution process of celebrities would occur for consumers viewing a social media influencer post. Unlike celebrities who acquire fame and recognition from appearing as characters in entertainment content, social media influencers generate their own fame for being an authentic self. Therefore, consumer knowledge of the influencer is also an information cue to help people make a correspondent inference of the influencer’s behavior. In that case, influencer credibility could serve as an additional information cue for making attributions of influencer motives. Discounting principle in Attribution Theory suggests that consumers will discount the arguments in the message if extrinsic motives such as money, selling and image motives are present. 41 Hovland and his colleagues’ source credibility model (Hovland et al., 1953; Hovland & Weiss, 1951) and McGuire (1985)’s source attractiveness model has shown the effectiveness of a credible source on liking. When consumers possess or retrieve positive prior knowledge about the influencers, for instance, the halo effect would carry over the positive attitudes towards influencers to the sponsored brand that the influencer creates content for. Regan, Straus, & Fazio (1974) has also found that when participants like someone, they are going to make more internal attributions than external attributions of that person’s behavior. Therefore, it is likely that correspondent inference bias will take place when influencer credibility is high (Jones & Harris, 1967; Sen & Lerman, 2007). To demonstrate correspondence inference bias, it is necessary to show that, despite knowing that the influencer is being paid to recommend a product, a consumer would assume that the endorser actually meant what s/he was saying. Kardes (1993) has argued that correspondence bias makes people disregard the extrinsic reasons of endorsements (e.g., money motives) and focus on the intrinsic reasons of endorsements (e.g. love motives), and thus contributes to the effectiveness of endorsement advertisements. As a result, we expect people make stronger attributions to the extrinsic reasons and demonstrate more resistance when the influencer credibility is low vs high as the discounting principle predicts; and make stronger attributions to the intrinsic reasons and generated less resistance when the influencer credibility is high vs low as the correspondence inference bias and source effect predicts. H4. Compared to a high credible influencer, a low credible influencer will make readers generate stronger perceptions of a) money motives, b) selling motives, and c) image motives. 42 H5. Compared to a low credible influencer, a highly credible influencer will make readers generate stronger perceptions of a) love motives, b) sharing motives, and c) helping motives. H6. Compared to a low credible influencer, a highly credible influencer will make readers generate less resistance towards the sponsored post. Interaction Effect of Disclosure and Influencer Credibility According to the PKM, agent knowledge (AK) interacts with persuasion knowledge in influencing consumer resistance. For instance, Hardesty and colleagues (2002) found brand familiarity (an example of agent knowledge) moderated the effect of high or low invoice prices (which activated persuasion knowledge) on consumer evaluations of advertised offers among highly skeptical consumers. Specifically, when skeptical consumers who are more familiar with the brand (high AK), their evaluations were greater for high invoice prices (versus low invoice prices). However, when skeptical consumers who are less familiar with the brand (low AK), their evaluations of high versus low invoice prices did not differ much. Similarly, Ahluwalia and Burnkrant (2004) found an interaction effect between source favorability (or previous agent attitudes) and the consumer’s persuasion knowledge on message persuasion. The difference in message attitudes between consumers with high vs. low persuasion knowledge was smaller if the message came from favorable sources (positive AK), than unfavorable sources (negative AK). Within the context of Influencer Marketing, agent knowledge is likely to be an important source of information that can influence a consumer’s inferences of influencer motives and can potentially influencer their response towards the message. We expect influencer credibility (agent knowledge) may mitigate consumer resistance to sponsored posts with disclosure (which 43 activates persuasion knowledge). Therefore, we propose an interaction effect between disclosure and influence credibility on influencer motives and consumer resistance. H7. The effect of the sponsorship disclosure will be weaker on consumer perceptions of a) money motives, b) selling motives, c) image motives or a highly credible influencer than a low credible influencer. H8. The effect of the sponsorship disclosure will be stronger on consumer perceptions a) love motives, b) sharing motives, c) helping motives for a highly credible influencer than a low credible influencer. H9. The effect of the sponsorship disclosure on consumer resistance will be weaker for a highly credible influencer than a low credible influencer. Role of Inferences of Influencer Motives Effect of Inferences of Influencer Motives on Resistance According to PKM, when consumers make inferences that the influence agent have an ulterior motive, as the change of meaning principle predicts, they will demonstrate resistance to persuasion. For instance, Hass and Grady (1975) showed that when people were forewarned of the communicator's persuasive intent, they would tend to resist the message. Similarly, Brown and Krishna (2004) found that when a consumer saw a firm offered a more expensive option by default, s/he is more likely to make counterarguments, compared to viewing a firm offered a less expensive option. This is probably because forewarning and seeing a firm offered a more expensive option will make people make inferences about the behavior of money motives or selling motives. 44 Many studies have shown evidence that persuasion knowledge could lead to resistance to persuasion. Morales (2005) found that for a firm that placed extra efforts in making or displaying products, when consumers thought the motive was to persuade people, they no longer felt gratitude toward the efforts and were not motivated to reward the firms. In this case, it is either money or selling motives prevented consumers from responding positively to firms. Other research also found the similar negative relationship between selling motives and resistance. For instance, Campbell and Kirmani (2000) found that when the ulterior persuasion motive of selling a product of a salesperson was highly accessible, consumers evaluated the salesperson as insincere. Similarly, Bambauer-Sachse & Mangold (2013) found that when consumers are aware that marketers can manipulate online reviews, such persuasion knowledge will make them less influenced by online reviews. Furthermore, Wojdynski and Evans (2016) found that recognizing a native ad, which is situational persuasion knowledge activated by disclosure location and language, resulted in increased ad skepticism and negative perceptions of credibility of that native ad. On the contrary, when consumers are less accessible to the persuasive motives, they respond favorably to persuasion. Brown and Krishna (2004) showed that activation of persuasion knowledge about a retailer’s motives influences consumer cognitive processing. In their experiment, participants imagined themselves making online purchases. One example is to purchase a monitor on a shopping website. The website offered a 21-inch monitor option by default (the more expensive option), but the participant has the choice to downgrade it to a cheaper 17-inch monitor. The results showed that when seeing the firm offering a more expensive option by default, consumers with more knowledge about this tactic were more likely to report counterarguments than those 45 who have less knowledge about this tactic. Although the authors did not directly measure consumer’s inference of the marketer motives, they postulated that consumer’s inference of the marketer motives is a mediating factor. In other words, the high default choice would activate consumer’s inferences of the marketer’s motives. Specifically, a consumer would think marketers were acting out of self-interest rather than in the interests of the consumer. As a result, the consumer would defend against or discount the persuasion. However, as discussed in the earlier section, PKM limits the motives to the selling and persuasive intent, and fails to explain consumer respond to other non-selling and non-persuasive intents, such as genuine love of the product, or pure sharing or helping intention. Discounting principle in attribution provides an alternative explanation why making attributions to extrinsic reasons lead to resistance. As discounting principle predicts, the attribution to the extrinsic motives such as money motives or image motives will discount the persuasiveness of the message. As a result, consumers will defend against the message. For instance, Yoon et al. (2006) found that when a consumer attributes the company’s corporate social responsibility (CSR) activities to insincere motives (e.g. money), the company image is hurt. However, making attributions to intrinsic reasons may be inversely related to resistance. For instance, Yoon et al. (2006) found that when a consumer makes the attribution of a company’s corporate social responsibility (CSR) activities to sincere motives (intrinsic), the company image is actually improved. Similarly, Sørum, Grape, and Silvera (2003) found making attributions to extrinsic reasons such that the person truly likes the product (love motives) had a positive effect on evaluations of the advertisement. Based on change of meaning, the discounting principle and previous findings, we propose: 46 H10. Making inferences of influencer’s motives to a) money motives, b) selling motives, c) image motives will be positively related to resistance to sponsored post. H11. Making inferences of influencer motives to a) love motives, b) sharing motives, c) helping motives will be negatively related to resistance to sponsored post. Figure 1 presents the final proposed theoretical model. Figure 1. Proposed Theoretical Model 47 CHAPTER 5 MAIN STUDY METHOD Main Study Design This study uses a 2 (disclosure: presence vs. absence) x 2 (influencer credibility: high vs. low) x 2 (product category: protein supplements vs. travel website) post-test only betweensubjects experimental design. Disclosure is manipulated as whether or not the Instagram blogger discloses that s/he had a partnership with a brand and whether or not labels the post with “#sponsored” in the end. Influencer credibility is manipulated with a priming method inspired by Wyer, Srull, and Gordon (1984). Participants will read a short introduction about the influencer, where personality traits such as expertise, attractiveness, trustworthiness and character traits are positively primed in the high credibility condition, but negatively primed in the low credibility condition. Product category is manipulated as an Instagram post either promoting a dietary supplement product as an example of credence goods or promoting a travel website as an example of experience goods. Stimuli Development Product Choice This study uses two product categories to enhance the ecological validity. Generally speaking, there are three types of products: search, experience and credence goods. Nelson (1970, 1974) first brought up the ideas of experiences goods, which are products that can be evaluated after purchase or use, in comparison to search goods whose quality can be evaluated prior to purchase. Later, Darby and Karni (1973) introduced the credence goods which are 48 difficult for consumers to evaluate and be certain of product quality even after purchase. We decided to use credence goods and experience goods to test consumer judgment of products based on electronic word of mouth, so that people need to rely on recommendations or prior experience to make a purchase decision. We then decided to use travel website as an example of experience goods and protein powder supplements as an example of credence goods for the following reasons. Travel website has the experience qualities and protein powder is considered as dietary supplements, which has the credence qualities. Furthermore, travel is a popular topic on Instagram, with over 230 million posts using #travel hashtags. And protein powder is closely related to the popular topics such as health, fitness and nutrition on Instagram. According to a recent survey of dietary supplements knowledge, protein is a less commonly used supplements (5.1%), compared to vitamin/mineral supplements (83.1%) and fish oil (53%) (Owens, Toone, & Steed-Ivie, 2014), which can minimize existing product knowledge and familiarity. Last but not least, both travel websites and protein powders appeal to both genders. Choice of Influencer Name, Product Name, and Image Fictitious brand names and gender-neutral influencer names were used to minimize the confounding effect of pre-existing attitudes and gender stereotypes and to enhance the congruence effect. A fictitious blogger whose expertise is congruent with the endorsed product is chosen for the main study to better transfer the meanings of the product to the consumers, based on the McCracken (1989)’s Meaning Transfer Model. A pretest of 137 US adults was conducted to determine the appropriate fictitious brand names, gender-neutral influencer name. They were mostly female (63.5%), Caucasian (73%) and 49 has an average age of 29 years old (SD = 12.6). Participants indicated their perceived gender image and their likability of ten gender-neutral names (e.g. “Pat”) for influencers. They also rated their familiarity and likability of nine fictitious brand names for protein powder (e.g. “Nutritionlab”) created by random Instagram name generator websites, as well as nine fictitious names for travel websites (e.g. “AdventureTours.com”) created by a made-up-words website. Participants also indicated their how much they like ten images featuring protein drinks and a travel scene. Based on the pretest results, we chose Alex as the influencer name, because this name was perceived by most participants (59.1%) with a high likability rating (mean = 3.59, SD = 1.03). We also chose Vital Nutrition as the fictitious brand name for protein products as it was rated as the most likable name (mean = 2.77, SD = 1.33) and was rated relatively low on familiarity (mean =2.69, SD = 1.18). Similarly, we chose vacations.com as the fictitious brand name for the travel website as it has the likability rating (mean = 2.89, SD = 1.28) and a relatively low rating on familiarity (mean = 2.9, SD = 1.23). Two most liked product images were chosen to be used in the sponsored post. Independent Samples T-test results showed that participants showed a similar level of likings of for the chosen protein product image (mean = 4.25, SD = 1.1) and the chosen travel image (mean = 4.35, SD = .82), t (1, 260) = .78, p =. 436. Priming Story Design To manipulate previous knowledge of the influencer credibility, participants read a short profile of the influencer before they read the sponsored post, as shown in Table 7. This story primes people to perceive the influencer as highly credible or not credible in the following aspects. First, the story manipulated the expertise of the influencer by showing the influencer has 50 a relevant advanced degree or years of working experiences, or showing the influencer is an amateur. Second, Jin & Phua (2014) demonstrated that the number of Twitter followers had a positive effect on physical attraction, trustworthiness and competence. Therefore, the story manipulated the number of followers as 3 million followers in the high credibility condition, or as 500 followers in the low credibility condition. The number of the followers was within the range of a social media influencer’s requirement. Third, the story manipulated the level of trustworthiness by telling the influencer is sincere or insincere. The story also manipulated the influencer’s character traits, as Rifon, Jiang, and Kim (2016) demonstrated that character traits are more important than physical attractiveness in predicting trustworthiness and expertise. As a result, we manipulated the influencer as down to earth and caring in the high credibility condition, or as arrogant and uncaring in the low credibility condition. The rest of the influencer profile remained the same. The designed priming story was pre-tested with 79 MTurk workers. Independent Samples T-test results confirmed that participants in the high influencer credibility condition perceived Alex as more credible (mean = 3.34, SD = .82) than those in the low influencer credibility condition (mean = 2.90, SD = .51), t (1, 61.187) = -2.89, p < .01. In sum, the manipulation of the influencer credibility was considered as successful. 51 Table 7. Priming Stories of Influencer Credibility Manipulation Product Protein Powder High Influencer Credibility Condition As a registered dietitian nutritionist (RDN) with a master’s degree1, Alex currently has over 3 million2 followers on Instagram as a food blogger. Alex posts pictures of food as well as easy and healthy recipes every day, and sometimes gives opinions on how to eat well and feel amazing. You have followed Alex on Instagram for a while, and have decided that Alex is a decent person and appears to be sincere3, down to earth4, and caring4. Travel site As a journalist who has worked for National Geographic Traveler for 8 years1, Alex currently has over 3 million2 followers on Instagram as a travel blogger. Alex posts fascinating travel pictures and tells adventure stories. Sometimes, Alex also shares budget travel tips, food recommendations, and details on what to do or where to stay, etc. You have followed Alex on Instagram for a while, and have decided that Alex is a decent person and appears to be sincere3, down to earth4, and caring4. 1 Manipulation of expertise. Manipulation of the number of followers. 3 Manipulation of trustworthiness 4 Manipulation of character traits. 2 52 Low Influencer Credibility Condition Never attended college1, Alex currently has over 500 followers2 on Instagram as a food blogger. Alex posts pictures of food as well as easy and healthy recipes every day, and sometimes gives opinions on how to eat well and feel amazing. You have followed Alex on Instagram for a while, and have decided that Alex is a decent person but sometimes appears to be insincere3, arrogant4 and uncaring4. As a travel enthusiast1, Alex currently has over 500 followers2 on Instagram as a travel blogger. Alex posts fascinating travel pictures and tells adventure stories. Sometimes, Alex also shares budget travel tips, food recommendations, and details on what to do or where to stay, etc. You have followed Alex on Instagram for a while, and have decided that Alex is a decent person but sometimes appears to be insincere3, arrogant4 and uncaring4. Message Design The sponsored content is adapted from real social media posts about a protein supplement and a travel website. The brand information was integrated into the content to mimic the sponsored post. The study only contained positive product information with the intention to promote the brand. Only one-sided (positive) information, instead of two-sided (both positive and negative but slightly positive) information is included in the sponsored post, as a recent study tests the moderating effect of message sidedness but does not find differences in the effect of sponsorship disclosure between one-sided message and two-sided message on source credibility, message attitude, brand attitude and behavioral intention (Hwang & Jeong, 2016). Disclosure Design This study designed sponsorship disclosure based on the FTC’s recommendation of using clear and conspicuous disclosure and the real practices to enhance ecological validity. Recently, the FTC staff sent out 90 letters reminding influencers and marketers to clearly disclose sponsorship relationship in their paid promotion or social media endorsements (Federal Trade Commission, 2017). The FTC letters pointed out that many consumers will not understand vague disclosures like "Thanks [brand]," "#sp," or "#partner" (Federal Trade Commission, 2017) and recommends to use “#sponsored”, “#ad” or “#promotion” in sponsored Instagram posts. A quick search on Instagram generated 836,332 posts with “#sponsored”, over 4.06 million posts with “#ad”, over 4.3 million posts with “#promotion”, and over 11 million posts with “#sp”, as of July 2017. Along with the previous findings on characteristics of effective disclosure (e.g. Boerman et al., 2014, 2015; Wojdynski & Evans, 2016), this study uses simple and straightforward 53 disclosure language, and places multiples disclosure at the beginning and in the middle of the sponsored content to make sure the disclosure is clear and conspicuous. Specifically, this study uses one sentence to disclose the sponsorship at the beginning of the post, “So excited to partner with @ [brand name]," and “#sponsored” at the end of the post. In the disclosure condition, participants read a disclosure at the beginning of the article, stating “partner with [brand]” and at the end with “#sponsored." “@ [brand name]” and “#sponsored” were in the hyperlink blue color. In the no disclosure condition, participants read a disclosure at the beginning of the article, stating “partner with [brand]” and at the end with “#sponsored." “@ [brand name]” and “#sponsored” were in the hyperlink blue color. The designed sponsored post was pre-tested with 79 MTurk workers. Independent Samples T-test results confirmed that participants in the disclosure condition was more confident recalling seeing a disclosure for sponsorship (mean = 3.3 vs. 2.3, SD = 1.2 vs 1.1, t (1, 77) = 3.85, p < .01) and seeing and #sponsored (mean = 3.5 vs 2.49, SD = 1.31 vs 1.28, t (1, 77) = 3.46, p < .01) than those in the low influencer credibility condition. Although participants in the disclosure conditions did not report significantly more confidence in recalling a sentence indicating the material connection between Alex and the brand (mean = 3.4 vs 3.16, SD = 1.1 vs 1.55, t (1, 64) = -.789, p = .423) due to the small sample size, the trend was still consistent with the earlier two findings. Therefore, the manipulation of the message and disclosure design was considered as successful. Table 8 presents the message content of the four different sponsored posts. 54 Table 8. Sponsored Content Product Disclosure No Disclosure So excited to partner with @Vital Nutrition to share my new recipe with you – a super creamy chocolate smoothie bowl! This smoothie bowl is a fusion between a super healthy nutritious smoothie and the flavor of the most sinful chocolate cake. My new protein powder from @Vital Nutrition Protein tastes extremely good and really smells like Powder chocolate (this is super seldom with other vegan protein powders!) I mixed 2 tbsp of it with ¼ avocado, ¾ cup almond milk, 2 tbsp cocoa powder, and 2 frozen bananas. It turned out very, very creamy! I also decorated with fruit toppings of choice. I LOVE it! #Sponsored So excited to share my new recipe with you – a super creamy chocolate smoothie bowl! This smoothie bowl is a fusion between a super healthy nutritious smoothie and the flavor of the most sinful chocolate cake. My new protein powder from @Vital Nutrition tastes extremely good and really smells like chocolate (this is super seldom with other vegan protein powders!) I mixed 2 tbsp of it with ¼ avocado, ¾ cup almond milk, 2 tbsp cocoa powder, and 2 frozen bananas. It turned out very, very creamy! I also decorated with fruit toppings of choice. I LOVE it! So excited to partner with @Vacations.com to explore Key West in the next few days! Key West is famous for its cocktail-fueled sunset celebrations on Mallory Square, but I recommend you spend one night nightboarding instead. it's an excursion where you go either paddle boarding or kayaking in the dark, only there are lights around your craft and the kayaks have glass bottoms. You can see so much more at night. I even saw a lobster scurrying through the sea grass! It's a nice break from all the partying. I used @Vacations.com to book my trips and it gave me a discounted hassle-free getaway. You should definitely give it a try! #Sponsored So excited to explore Key West in the next few days! Key West is famous for its cocktail-fueled sunset celebrations on Mallory Square, but I recommend you spend one night nightboarding instead. it's an excursion where you go either paddle boarding or kayaking in the dark, only there are lights around your craft and the kayaks have glass bottoms. You can see so much more at night. It's a nice break from all the partying. I used @Vacations.com to book my trips and it gave me a discounted hassle-free getaway. You should definitely give it a try! Travel site 55 Main Study Participants A power analysis with a three-group ANCOVA main and interaction effects is used to determine the sample size. Results in Table 9 suggested a sample size of 333 is at least required for the main study, with an effect size f=.21 (as in Campbell and Kirmani (2000)), power (1-β) =.80 (as in Banerjee, Chitnis, Jadhav, Bhawalkar, and Chaudhury (2009)), and at a significant level of α = 0.05. Table 9. Results of Power Analysis F tests - ANCOVA: Fixed effects, special, main effects and interactions Analysis: A priori: Compute required sample size Input: Effect size f = 0.21 α err prob = 0.05 Power (1-β err prob) = 0.8 Numerator df = 7 Number of groups = 8 Number of covariates = 5 Output: Noncentrality parameter λ = 14.68 Critical F = 2.04 Denominator df = 320 Total sample size = 333 Actual power = .80 A total of 360 U.S. adults were recruited from Amazon Mechanic Turk, who 1) currently live in the United States, 2) are at least 18 years old, and 3) have at least 95% approval rate. Participants who do not qualify for the requirements were not invited to complete the study. They received $1 as an incentive for their participation. According to an Forbes article, the average-adult reading speed is 300 words per minute (Nelson, 2012). Therefore, we excluded thirteen participants who spend fewer than 13 seconds before submitting their responses on the influencer bio page (which contains 67-83 words), or spent fewer than 19 seconds before 56 submitting their responses on the sponsored post page (which contains 97-112 words) from analysis. This leaves us a final sample of 347 for data analysis. Among the participants, 55% were female and 45% were male. Their age ranged from 18 to 75 years old (Mean=37, SD=11). The majority of the participants were Caucasian (76.4%), followed by African American (11.2%), Hispanic or Latino (5.2%), and Asian American (4.9%). Half of the participants had a household income between $30,000 and $69,999 (49.6%). Furthermore, they reported an average of 15 years of education (SD =2.7). Main Study Procedure Participants were told that the study is interested in their evaluation of an Instagram post. This Instagram post was sponsored by a fictitious brand, but the sponsorship was not revealed to the participants explicitly. They were randomly assigned to one of the eight experimental conditions. Upon completing the consent form, participants were exposed to the influencer credibility priming manipulation. Half of the participants were randomly assigned to read an introduction priming a highly credible digital influencer who shows expertise in the area, was highly trustworthy, and is very likable. And the other half read an introduction priming a low credibility of the influencer, who has a low expertise, low trustworthiness and is unlikable. The study used a writing task as an attention check to make sure participants read through the content of the stimuli. After participants read the short bio of the influencer, they were asked to briefly describe their impression of the influencer. After that, participants were randomly assigned to one of four sponsored posts and then read a sponsored post featuring either a protein powder supplement or a travel site. Half of the participants were assigned to a sponsored post with disclosure, and the rest of them read a sponsored post without disclosure. 57 See Appendix E-H for the screenshots of four sponsored Instagram posts. After reading the sponsored post, participants were asked to answer a questionnaire. Participants responded to items, presented in a randomized order, measuring the strength of their inferences about influencer motives, their resistance toward and involvement in the sponsored post. After that, they answered the manipulation check questions and the control measures of perceived appropriateness of sponsored content. Lastly, they reported their demographic information. Upon completing the questionnaire, participants received $1 for their participation. Measures For the ease of reading, this section will not discuss measures in the sequence of survey items, but in the sequence of independent variables, dependent variables, control variables, and demographic questions. Manipulation Check To assess the manipulation of sponsorship disclosure, we measured both participant recall of the disclosure and recognition of the sponsored post as an ad. To measure disclosure recall, one question adapted from (Boerman et al., 2014) asked participants “In the post you just read, how confident are you in recalling a disclosure for sponsorship? (1=extremely unconfident,5=extremely confident)”. Two new items were also included on the five-point Likert scale: “In the post you just read, how confident are you in recalling #sponsored?” “In the post you just read, how confident are you in recalling a sentence stating the author partnered with a brand?” To measure ad recognition, one statement adapted from Campbell and Kirmani (2000) 58 asked participants to rate on a on a five-point Likert scale, 1=Strongly Disagree, 5=Strongly Agree: “While I read the Instagram post, it was pretty obvious that it was an ad.” Influencer credibility was measured with fifteen items adapted from Ohanian (1990) on a five-point semantic differential scale (Cronbach’s α = .96). The attractiveness dimension was measured with five items: attractive/unattractive, classy/not classy, beautiful/ugly, elegant/plain, sexy/not sexy. The trustworthiness dimension was measured with five items: dependable/undependable, honest/dishonest, reliable/unreliable, sincere/insincere, trustworthy/untrustworthy. And the expertise dimension was measured with five items: expert/not an expert, experienced/inexperienced, knowledgeable/unknowledgeable, qualified/unqualified, skilled/unskilled. Dependent Variables Influencer motives were measured with the scale developed in Chapter 3 in a five-point Likert scale, 1=Strongly Disagree, 5=Strongly Agree. All reliabilities of the constructs are greater than .8, indicating excellent internal consistencies (Landis & Koch, 1977). Money motives were measured with three items (Cronbach’s α = .90): “benefits by this sponsorship;” “is paid to recommend the product;” and “wants to receive future sponsorships.” Selling motives were measured with three items (Cronbach’s α = .86): “wants to sell the product;” “wants to increase product sales;” and “wants to increase company profits.” Image motives were measured with four items (Cronbach’s α = .88): “wants to gain followers;” “wants to gain likes;” “wants to gain shares;” and “wants to promote himself or herself as the guru on social media.” 59 Love motives were measured with three items (Cronbach’s α = .89): “views the brand as a good product;” “likes the product;” and “is satisfied with the product.” Sharing motives were measured with three items (Cronbach’s α = .82): “wants to share the product they use with others;” “wants to express their own opinion of the product;” and “enjoys sharing on social media.” Helping motives were measured with three items (Cronbach’s α = .89): “cares about getting useful information to the followers;” “wants to help others to make better purchase decisions;” and “wants to help others get the information they want.” Resistance was measured with seven statements adapted from van Reijmersdal et al. (2016) on a five-point Likert type scale, 1=Strongly Disagree, 5=Strongly Agree: “While reading the Instagram post, I contested/refuted /doubted /countered the information in the content;” “While reading the Instagram post, I felt angry/irritated/annoyed.” These items showed a great internal consistency (Cronbach’s α = .90). Control Variables Participants' age, gender, perceived appropriateness of the sponsored post, involvement in the sponsored post were included as control variables. Perceive appropriateness of the sponsored content was measured with three adapted items on a five-point Likert type scale, 1=strongly disagree, 5=strongly agree (Cronbach’s α = .83): “It is appropriate that an Instagram blogger gets paid to post some sponsored content;” “Instagram bloggers can post content for pay, because they provide information to us for free;” “It is acceptable to me if the brands pay to place their messages in an Instagram post” (Nelson et al., 2009; Yoo, 2009). 60 Involvement in the sponsored post was measured on a 10-item five-point semantic scale (Cronbach’s α = .96): “The content I read is important/unimportant, of irrelevant/relevant, means a lot/nothing, valuable/worthless to me, interesting/boring, exciting/unexciting, appealing/unappealing, needed/not needed, mundane/fascinating, involving/uninvolving to me (Zaichkowsky, 1994).” 61 CHAPTER 6 MAIN STUDY RESULTS Manipulation Check An Independent Samples T-test was used to test the manipulation of sponsorship disclosure. Results of disclosure recall showed that participants in the disclosure condition were more confident recalling a disclosure for sponsorship (mean = 3.50 vs. 2.45, SD = 1.41 vs.1.31, t (1, 345) = -7.191, p < .001), recalling #sponsored (mean = 3.97 vs 2.63, SD = 1.36 vs 1.43, t (1, 345) = -8.963, p < .001), and recalling a sentence indicating the material connection between Alex and the brand (mean = 3.99 vs 3.46, SD = 1.16 vs 1.35, t (1,342.707) = -3.93, p < .001), compared to those in the no disclosure condition. Results of advertising recognition confirmed that participants in the disclosure condition (mean = 4.2, SD = .97) were more likely to recognize the sponsored post as an ad than those in the no disclosure condition (mean = 3.63, SD = 1.19), t (1, 337.759) = -5.21, p < .001. Therefore, the manipulation of the sponsorship disclosure was successful. An Independent Samples T-test was used to test the manipulation of influencer credibility. Results revealed that participants in the relatively high influencer credibility condition perceived Alex as more credible (mean = 3.74, SD = .76) than those in the relatively low influencer credibility condition (mean = 3.14, SD = .77), t (1, 345) =- 7.328, p < .001. In sum, the manipulation of the influencer credibility was considered as successful. Main and Interaction Effects on Influencer Motives and Consumer Resistance To test H1-9, multivariate analysis of covariance was used to examine the effects of sponsorship disclosure and influencer credibility in the two product categories on various types 62 of influencers motives and consumer resistance. Age, gender, years of education, involvement and perceived appropriateness of sponsored content were entered as covariates. Table 10 presents the results of between subject effects. Table 11 presents the descriptive statistics of each influencer motive and consumer resistance by experimental condition. Table 10. Multivariate Tests Wilk's Λ F (7, 329) Partial η2 Intercept .596 31.787*** .404 Age Gender Education Appropriateness Involvement .963 .984 .941 .850 .686 1.788 .780 2.918** 8.283*** 21.465*** .037 .016 .059 .150 .314 Sponsorship Disclosure Influencer Credibility Product Category Credibility * Disclosure .903 .898 .958 .992 5.060*** 5.349*** 2.039* .396 .097 .102 .042 .008 .985 .978 .982 .708 1.060 .860 .015 .022 .018 Disclosure * Product Category Credibility * Product Category Credibility * Disclosure * Product Category Note: * p < .05; ** p < .01, *** p < .001 As shown in Table 10, results of multivariate tests showed that the covariates, age (F (7, 329) =1.78, p = .089; Wilk's Λ = .963, partial η2 = .037) and gender (F (7, 329) = .792, p = .604; Wilk's Λ = .98, partial η2 = .017) did not significantly influence the six types of influencer motives. However, years of education (F (7, 329) = 2.92, p < .01; Wilk's Λ = .941, partial η2 = .59), perceived appropriateness of sponsored content (F (7, 329) = 8.28, p < .001; Wilk's Λ = .85, partial η2 = .15), and involvement (F (7, 329) =21.465, p < .001; Wilk's Λ = .686, partial η2 = .31) and were significant covariates. 63 Multivariate test results also showed that there were statistically significant main effects of sponsorship disclosure (F (7, 329) = 5.06, p < .001; Wilk's Λ = .903, partial η2 = .097), influencer credibility (F (7, 329) = 5.349, p < .001; Wilk's Λ = . 898, partial η2 = .102), and product category (F (7, 329) = 2.039, p < .05; Wilk's Λ = .958, partial η2 = .042) on the combined dependent variables of six influencer motives and consumer resistance after controlling for all covariates. However, multivariate test results did not find a significant two-way interaction effect, disclosure  influencer credibility (F (7, 329) = .396, p = .905; Wilk's Λ = .992, partial η2 = .008), or a significant three-way interaction effect, disclosure  influencer credibility  product category (F (7, 329) = .86, p = .539; Wilk's Λ = .982, partial η2 = .018) after controlling for the covariates. Therefore, H7-9 were rejected. Results in Table 11 and Table 12 showed that after controlling for covariates, participants who read a sponsored post about a protein supplement condition perceived stronger selling motives (Mean = 4.26 vs. 3.91, F (1, 334) = 11.1, p < .001, partial η2 = .032) and image motives (Mean = 4.24 vs. 4.01, F (1, 334) = 5.76, p < .05, partial η2 = .017) than those who read a sponsored post about a travel site. However, consumer perceptions of money motives, love motives, sharing motives and helping motives, as well as consumer resistance did not differ by product category. 64 Table 11. Results of Between Subject Effects F (1, 334) Partial η2 Money Motives 2.92 0.009 Selling Motives 11.10*** 0.032 Image Motives 5.76* 0.017 Love Motives 2.09 0.006 Sharing Motives 0.58 0.002 Helping Motives 0.34 0.001 Resistance 0.06 0.000 Money Motives 24.31*** 0.068 Selling Motives 9.17*** 0.027 Image Motives 0.30 0.001 Love Motives 0.22 0.001 Sharing Motives 0.023 0.000 Helping Motives 0.004 0.000 Resistance 0.006 0.000 Money Motives 0.52 0.002 Selling Motives 0.00028 0.000 Image Motives 8.03** 0.023 Love Motives 6.34* 0.019 Sharing Motives 4.66* 0.014 Helping Motives 12.99*** 0.037 5.32* 0.016 IV Product Category Disclosure Credibility DV Resistance Note: * p < .05; ** p < .01, *** p < .001 65 Table 12. Descriptive Statistics by Experimental Conditions Credibility Low Money Motives High Total Low Selling Motives High Total Low Image Motives High Total Love Motives Low Disclosure Protein Travel Total Product Supplement Site Category M SD No disclosure 4.13 0.86 Disclosure 4.68 Total M SD 4.09 0.92 4.11 0.89 0.52 4.48 0.97 4.59 0.77 4.39 0.77 4.27 0.96 4.33 0.87 No disclosure 4.20 0.80 3.77 1.15 3.98 1.00 Disclosure 4.57 0.70 4.38 0.84 4.47 0.78 Total 4.38 0.77 4.08 1.04 4.23 0.93 No disclosure 4.16 0.83 3.93 1.05 4.05 0.95 Disclosure 4.62 0.62 4.43 0.90 4.53 0.77 Total 4.38 0.77 4.17 1.01 4.28 0.90 No disclosure 3.97 0.98 3.95 1.05 3.96 1.01 Disclosure 4.51 0.61 4.04 0.82 4.29 0.75 Total 4.22 0.87 4.00 0.94 4.11 0.91 No disclosure 4.17 0.85 3.66 1.04 3.92 0.98 Disclosure 4.45 0.57 3.98 0.86 4.21 0.76 Total 4.31 0.73 3.82 0.96 4.06 0.89 No disclosure 4.07 0.92 3.81 1.05 3.94 0.99 Disclosure 4.48 0.59 4.01 0.83 4.25 0.75 Total 4.26 0.80 3.91 0.95 4.09 0.90 No disclosure 4.35 0.76 4.10 0.90 4.23 0.84 Disclosure 4.46 0.64 4.13 0.95 4.30 0.81 Total 4.40 0.70 4.11 0.92 4.26 0.82 No disclosure 4.12 0.79 3.78 1.03 3.95 0.93 Disclosure 4.02 0.85 4.03 0.70 4.03 0.77 Total 4.07 0.82 3.91 0.88 3.99 0.85 No disclosure 4.24 0.78 3.94 0.98 4.10 0.89 Disclosure 4.24 0.78 4.08 0.82 4.16 0.80 Total 4.24 0.78 4.01 0.90 4.13 0.85 No disclosure 3.91 1.06 3.86 0.89 3.88 0.97 Disclosure 3.67 0.94 3.67 0.92 3.67 0.92 Total 3.80 1.01 3.77 0.90 3.78 0.95 66 M SD Table 12. (cont’d) High Total Low Sharing Motives High Total Low Helping Motives High Total Low Resistance High Total No disclosure 4.08 0.91 3.99 0.82 4.03 0.86 Disclosure 4.09 0.94 4.14 0.69 4.12 0.82 Total 4.08 0.92 4.07 0.76 4.08 0.84 No disclosure 3.99 0.99 3.92 0.85 3.96 0.92 Disclosure 3.89 0.96 3.92 0.83 3.90 0.90 Total 3.94 0.97 3.92 0.84 3.93 0.91 No disclosure 3.90 1.00 3.83 0.98 3.86 0.98 Disclosure 3.75 0.85 3.71 1.01 3.73 0.93 Total 3.83 0.93 3.77 0.99 3.80 0.96 No disclosure 4.08 0.92 3.96 0.88 4.02 0.90 Disclosure 3.95 1.04 4.25 0.62 4.10 0.86 Total 4.02 0.98 4.11 0.77 4.06 0.88 No disclosure 3.99 0.96 3.89 0.93 3.94 0.94 Disclosure 3.85 0.95 4.00 0.86 3.92 0.91 Total 3.92 0.96 3.95 0.90 3.93 0.93 No disclosure 3.44 1.15 3.45 0.85 3.45 1.01 Disclosure 3.25 0.92 3.39 1.00 3.32 0.96 Total 3.36 1.04 3.43 0.92 3.39 0.98 No disclosure 3.80 0.97 3.78 1.11 3.79 1.04 Disclosure 3.74 1.09 3.97 0.75 3.86 0.93 Total 3.77 1.03 3.88 0.94 3.82 0.98 No disclosure 3.62 1.07 3.62 1.00 3.62 1.03 Disclosure 3.50 1.03 3.71 0.92 3.60 0.98 Total 3.56 1.05 3.66 0.96 3.61 1.01 No disclosure 2.46 1.17 2.20 1.02 2.34 1.10 Disclosure 2.49 1.09 2.53 1.18 2.51 1.13 Total 2.48 1.13 2.35 1.10 2.42 1.11 No disclosure 2.15 1.20 2.15 1.22 2.15 1.20 Disclosure 2.18 1.06 1.80 0.76 1.99 0.93 Total 2.17 1.13 1.97 1.02 2.07 1.08 No disclosure 2.31 1.19 2.18 1.12 2.25 1.15 Disclosure 2.34 1.08 2.14 1.03 2.24 1.06 Total 2.32 1.13 2.16 1.07 2.24 1.11 67 H1a-c predicted that the presence of disclosure in the sponsored post would positively affect money motives, selling motives, and image motives. Results in Table 11 and Table 12 showed that after controlling for covariates and product category, participants in the disclosure condition reported stronger perceptions of money motives (F (1, 334) = 24.31, p < .001, partial η2 = .068) and selling motives (F (1, 334) = 9.17, p < .001, partial η2 = .027) than those in the non-disclosure condition. However, participants in the disclosure and no disclosure groups did not differ in their perceptions of image motives, F (1, 334) = .03, p = .59, partial η2 = .001. Therefore, H1a-b were supported and H1c was rejected. H2a-c predicted a negative effect of sponsorship disclosure on love motives, sharing motives and helping motives. Results in Table 11 and Table 12 showed that after controlling for covariates and product category, participants in the disclosure and no disclosure groups did not differ in their perceptions of love motives (F (1, 334) =.22, p = .64, partial η2 = .001), sharing motives (F (1, 334) = .023, p = .88, partial η2 = .000) or helping motives (F (1, 334) = .004, p = .95, partial η2 = .000). In other words, sponsorship disclosure did not significantly influence people’s perceptions of love motives, sharing motives or helping motives. Therefore, H2a-c were rejected. H3 predicted that the presence of disclosure in the sponsored post would positively affect consumer resistance. Results in Table 11 and Table 12 showed that after controlling for covariates and product category, participants in the disclosure condition (mean = 2.25, SD = 1.15) did not generate more resistance towards the sponsored post than those in the nondisclosure condition (mean = 2.24, SD = 1.06), (F (1, 334) = .006, p = .094, partial η2 = .000), rejecting H3. 68 H4a-c predicted that influencer credibility would negatively affect money motives, selling motives and image motives. Results in Table 11 and Table 12 showed that after controlling for covariates, participants generated weaker perceptions of images motives for a highly credible influencer, compared to a low credible influencer, F (1, 334) = 8.03, p < .01, partial η2 = .023). However, their perceptions of money motives (F (1, 334) = .52, p = .47, partial η2 = .002) or selling motives (F (1, 334) = .00028, p = .98, partial η2 = .000) did not differ in the high vs low influencer credibility condition. Therefore, H4ab were rejected but H4c was supported. H5a-c predicted that influencer credibility would positively affect love motives, sharing motives, and helping motives. Results in Table 11 and Table 12 showed that after controlling for covariates, participants generated stronger perceptions of love motives (F (1, 334) = 6.34, p < .05, partial η2 = .019), sharing motives (F (1, 334) = 4.66, p < .05, partial η2 = .014), and helping motives (F (1, 334) = 12.99, p < .001, partial η2 = .037) for a highly credible influencer, compared to a low credible influencer. However, their perceptions of did not differ in the high vs low influencer credibility condition. Therefore, H5a-c were supported. H6 predicted that the influencer credibility would negatively affect consumer resistance. Results in Table 11 and Table 12 showed after controlling for covariates, participants demonstrated significantly less resistance to a sponsored post shared by a highly credible influencer (mean = 2.07, SD = 1.08), compared to a low credible influencer (mean = 2.42, SD = 1.11), F (1, 334) = 5.32, p < .05, partial η2 = .016). Therefore, H6 was supported. 69 Effect of Different Influencer Motives on Consumer Resistance Measurement Model Testing To test the measurement model, a confirmatory factor analysis with maximum likelihood estimation with Promax rotation was run on the correlation matrix of the latent variables. The model fit index of proposed theoretical model (Model 3) demonstrated a good model fit based on the recommended cutoff values (Hu & Bentler, 1999), χ2/d.f.=2.698, p <.001; NFI=.90, IFI=.94, TLI= .93, CFI=.94, RMSEA=.07, AIC=946.087, BCC=963.354. Table 13. CFA Results of Model 3: Validity Test and Factor Correlation CR AVE MSV 1 2 3 4 5 6 1. Money 0.90 0.76 0.70 1 2. Selling 0.86 0.68 0.70 0.84 1 3. Image 0.88 0.65 0.42 0.65 0.61 1 4. Love 0.89 0.73 0.89 0.10 0.06 0.38 1 5. Sharing 0.82 0.61 0.89 0.15 0.06 0.43 0.95 1 6. Helping 0.90 0.74 0.80 -0.06 -0.10 0.23 0.85 0.89 1 7. Resistance 0.93 0.66 0.20 0.07 0.15 0.03 -0.38 -0.38 -0.44 7 1 The validity test results in Table 13 showed that the composite reliability of each factor was above the .7 threshold (Hair et al., 2006) and the AVE of each factor were above. 50 (Malhotra & Dash, 2011), demonstrating good convergent validity. However, the AVEs of selling motives, love motives, sharing motives and helping motives were less than the MSV; and the square root of AVEs of selling motives, love motives, sharing motives and helping motives were less than the factor correlations, showing discriminant validity issues (Fornell & Larcker, 1981). Discriminant validity issues indicate that the variables that measure selling motives, love 70 motives, sharing motives and helping motives were highly correlated with other variables outside the parent factor than with the variables within the parent factors. As shown in Table 13, love motives were highly correlated with sharing motives (r = .95) and helping motives (r = .85); and selling motives and money motives were also highly correlated (r = .84), which explained why these factors lack discriminant validity. Table 14. CFA Results of Model 4: Validity Test and Factor Correlation AVE 0.84 MSV 0.47 MS 1 Image MS CR 0.91 Image 0.88 0.65 0.47 0.69 1 LSH 0.96 0.90 0.17 0.06 0.38 0.93 0.66 0.17 0.11 0.03 Resistance Note: LSH: A second-level factor of love, sharing and helping motives MS: A second-level factor of money and selling motives LSH Resistance 1 -0.42 1 To solve the discriminant validity issues, we added two second-level constructs in the CFA model (Model 4) by combining selling and money motives into a second-level factor called SM motives, and combining love, sharing and helping motives into a second-level factor called LSH motives. CFA results of this re-specified model showed that this model has a great model fit (χ2/d.f.=2.75, p <.001; GFI = .84, NFI=.90, IFI=.93, TLI= .92, CFI=.93, RMSEA=.072, AIC=955.328, BCC=985.9). To further test the validity of this revised CFA model, results in Table 14 showed that the composite reliability and AVE of each factor were above the thresholds, demonstrating good convergent validity. Furthermore, the AVEs of each factor were less than the MSV of each factor; and the factors’ square root of AVEs were greater than the factor correlation, demonstrating good discriminant validity. 71 Structural Model Testing To test the hypothesized relationships in H7-8, we performed a structural equational modeling for the original hypothesized model. Figure 2 presents the full path diagram of the structural equation model with respective path coefficients. The model fit indexes in the hypothesized model appears to have a satisfactory fit to the data with the exception of the 2 statistic (χ2/d.f.=2.71, p <.001; NFI=.90, IFI=.94, TLI= .93, CFI=.94, RMSEA=.07, AIC=955.243, BCC= 971.3), according to Hooper, Coughlan, & Mullen (2008). Figure 2. Path Diagram of The Hypothesized Model (Model 3) H7a-c predicted the positive relationships between money, selling and image motives and consumer resistance to sponsored post. Results of the hypothesized theoretical model showed that selling motives (β= .26, p < .05) positively predicted consumer resistance. However, image motives (β= .17, p = .054) did not positively predicted consumer resistance at a .05 significance 72 level. Furthermore, money motives negatively predicted consumer resistance (β= -.25, p < .05), which was in the opposite direction of the hypothesized relationship. Therefore, H7a and H7c were rejected while H7b were supported. H8a-c predicted the negative relationships between love, sharing and helping motives and consumer resistance to sponsored post. Results of the hypothesized theoretical model revealed that neither love motives (β= -.18, p = .554) nor sharing motives (β= .08, p = .810) was significantly related to consumer resistance. However, results found that helping motives negatively predicted consumer resistance to sponsored post (β= -.40, p < .05). Therefore, H8a-b were rejected while H8c was supported. We also performed a structural equational modeling for the revised model with the second-level factors, as shown in Figure 3. The model fit indices revealed an acceptable model fit for the revised model (χ2/d.f.=2.635 , p <.001; NFI=.90, IFI=.94, TLI= .93, CFI=.94, RMSEA=.069, AIC=933.662, BCC=949.574). Results of this revised structural model confirmed that image motives positively predicted consumer resistance to sponsored post (β= .17, p < .05), in support of H7c; and love, sharing and helping motive altogether negatively predicted consumer resistance to sponsored post (β= -.49, p < .001), in support of H8a-c. However, money and selling motives together did not significantly predict consumer resistance to sponsored post (β= .03, p = .707), rejecting H7a-b. 73 Figure 3. Path Diagram of the Revised Model (Model 4) 74 The SEM results of the hypothesized and revised model are summarized in Table 15. Table 15. SEM Hypothesis Testing Results Hypothesis From To Hypothesized Model (Model 3) β Results Revised Model (Model 4) β Results Money Resistance -.25 Rejected Motives Rejected .03 Selling H10b Resistance .26* Supported Motives Image H10c Resistance .17 Rejected .17* Supported Motives Love H11a Resistance -.18 Rejected Motives Sharing H11b Resistance .08 Rejected -.49*** Supported Motives Helping H11c Resistance -.40* Supported Motives Note: Hypothesized Model: CFI (comparative fit index) = .94; NFI (normed fit index) = .91; TLI (Tucker-Lewis index) = .93, RMSEA (root mean square error of approximation) = .07; 2/df=2.71, p <0.001. Revised Model: CFI = .94; NFI= .90; TLI = .93, RMSEA = .079; 2/df=2.635, p <0.001. One indicator of all constructs is set to 1 to standardize the measurement scale. *p <.05, ** p <.05, ***p <.001 H10a Post-Hoc Mediation Analysis As several advertising scholars used ad recognition as a measure of persuasion knowledge (e.g. Boerman, van Reijmersdal, & Neijens, 2014; Tutaj & van Reijmersdal, 2012), it is likely to predict that the effect of money and selling motives on consumer resistance was mediated by ad recognition. This current study measured ad recognition as part of the manipulation check of disclosure condition. Therefore, we performed a post-hoc mediation analyses with Haye’s PROCESS path-analysis macro for SPSS 24 (Hayes, 2013; Model 4), using a bootstrap estimation approach with 5000 samples. Ad recognition was the mediator, money and selling motives altogether was used as the independent variable, and consumer resistance was used as the dependent variable. 75 The mediation results revealed that money and selling motives together significantly positively predicted ad recognition, b = .52, SE = .07, t = 7.75, p < .001, and that ad recognition was a significant predictor of resistance, b = .29, SE = .05, t = 5.27, p < .001. Furthermore, money and selling motives together was no longer a significant predictor of consumer resistance after controlling for the mediator, ad recognition, b = -.004, SE = .07, p = .95. Results also showed that the direct effect of money motives on resistance was not significant, b = -.004, SE = .07, 95% CI = [-.1489, .1406]; whereas the indirect effect from money motives to ad recognition and then to resistance was significant, b = .15, SE = .03, 95% CI = [.0887, .2236]. These results indicated a full mediation, as shown in Figure 4, that ad recognition fully mediated the relationship between money motives and consumer resistance to sponsored content. Figure 4. Post-Hoc Mediation Test Post-Hoc SEM by Influencer Credibility Condition As the current study found a significant main effect of influencer credibility on four types of influencer motives and consumer resistance, we performed two sets of post-hoc structural equational modeling using AMOS 23 to how consumers used influencer motives to sponsored content differed in different influencer credibility conditions. One SEM model used data from participants who read a sponsored post shared by a highly credible influencer, and the other one used data from participants who read a sponsored post shared by a less credible influencer. 76 Table 16. Post-Hoc SEM Results by Influencer Credibility Model 5 Model 6 High Low From To β β Money and Selling Motives Ad Recognition .33*** .48*** Ad Recognition Resistance .08 .22** Image Motives Resistance .16* .07 Love, Sharing and Helping Motives Resistance -.53*** -.31*** Note: Model 5: CFI (comparative fit index) = .92; NFI (normed fit index) = .87; TLI (TuckerLewis index) = .91, RMSEA (root mean square error of approximation) = . 043; 2/df=2.285, p <0.001. Model 6: CFI = .91; NFI= .84; TLI = .90, RMSEA = .082; 2/df=2.16, p <0.001. One indicator of all constructs is set to 1 to standardize the measurement scale. *p <.05, ** p <.05, ***p <.001 First, we tested how consumers use their inferred motives to respond to sponsored content shared by a highly credible influencer. The SEM model had an adequate model fit (χ2/d.f.= 2.29, p <.001; NFI=.87, IFI=.92, TLI= .91, CFI=.92, RMSEA=.043, AIC=2693.151, BCC=2785.791). Results of this model in Figure 5 and Table 16 confirmed that money and selling motives together positively predicted ad recognition (β= .33, p < .001). However, ad recognition did not significantly predict consumer resistance to sponsored post (β= .07, p = .29). As a result, money and selling motives together did not influence consumer resistance when the influencer credibility was high. Also, image motives did not predict consumer resistance to sponsored post (β= .16, p < .05); and love, sharing and helping motive altogether negatively predicted consumer resistance to sponsored post (β= -.53, p < .001), when the influencer credibility was high. 77 Figure 5. Post-Hoc SEM Model for Highly Credible Influencers (Model 5) Then, we examined how consumers use their inferred motives to respond to sponsored content posted by a less credible influencer. This SEM model had an adequate model fit (χ2/d.f.= 2.16, p <.001; NFI=.84, IFI=.91, TLI= .90, CFI=.91, RMSEA=.082, AIC=885.881, BCC=896.867). As shown in Figure 6 and Table 16, results of this structural model shown in Figure 5 and Table 16 confirmed that money and selling motives together positively predicted ad recognition (β= .48, p < .001), which in turn positively predicted consumer resistance to sponsored post (β= .22, p < 0.01). Thus, money and selling motives together increased consumer resistance for less credible influencers. Furthermore, love, sharing and helping motive altogether negatively predicted consumer resistance (β= -.31, p < .001); but image motives did not significantly influence consumer resistance (β= .07, p = .385) for less credible influencers. 78 Figure 6. Post-Hoc SEM Model for Less Credible Influencers (Model 6) 79 The results of this study are summarized in Table 17. Table 17. Hypothesis Testing Results Hypothesis IV DV Results H1a Sponsorship Disclosure Money Motives Supported H1b Sponsorship Disclosure Selling Motives Supported H1c Sponsorship Disclosure Image Motives Rejected H2a Sponsorship Disclosure Love Motives Rejected H2b Sponsorship Disclosure Sharing Motives Rejected H2c Sponsorship Disclosure Helping Motives Rejected H3 Sponsorship Disclosure Resistance Rejected H4a Influencer Credibility Money Motives Rejected H4b Influencer Credibility Selling Motives Rejected H4c Influencer Credibility Image Motives Supported H5a Influencer Credibility Love Motives Supported H5b Influencer Credibility Sharing Motives Supported H5c Influencer Credibility Helping Motives Supported H6 Influencer Credibility Resistance Supported H7a Disclosure  Credibility Money Motives Rejected H7b Disclosure  Credibility Selling Motives Rejected H7c Disclosure  Credibility Image Motives Rejected H8a Disclosure  Credibility Love Motives Rejected H8b Disclosure  Credibility Sharing Motives Rejected H8c Disclosure  Credibility Helping Motives Rejected H9 Disclosure  Credibility Resistance Rejected H10a Money Motives Resistance H10b Selling Motives Resistance H10c Image Motives Resistance H11a Love Motives Resistance H11b Sharing Motives Resistance H11c Helping Motives Resistance Rejected* Supported Supported Note: * Post hoc analysis revealed that this effect was fully mediated by advertising recognition. 80 CHAPTER 7 DISCUSSION Summary of Findings Sponsored content, written by social media influencers or micro-celebrities to endorse a product or service, has become a popular influential marketing strategy for advertisers to reach and influence potential consumers. However, the FTC has mandated that influencers must disclose sponsorship in a clear and prominent way (FTC, 2015). The objective of this current research is to understand how consumers making inferences of Instagram influencer motives for posting product recommendations, and how consumers use their knowledge of influencer credibility and sponsorship disclosure to interpret and respond to sponsored posts. Guided by the Persuasion Knowledge Model and Attribution Theory, this study first used a thought listing approach and three online surveys to develop a scale of influencer product recommendation motives in the Instagram influencer marketing context. This study then conducted a 2 x 2 x 2 online experiment design focused on disclosure, influencer credibility, and product category to investigate the roles of influencer credibility and sponsorship disclosure on consumer inferences of influencer motives and consumer resistance, as well as the role of different types of influencer motives on consumer resistance. This chapter summarizes the findings of the current study, provides the academic and practical implications, and discusses the study’s limitations as well as future research directions. Conceptualizing and Measuring Influencer Motives Understanding what motivates social media influencers to post sponsored content in social media is an important piece of consumer persuasion knowledge. This study highlighted the 81 importance of consumer inferences of influencer motives as a part of persuasion knowledge, which has rarely been studied in the previous literature. Therefore, the first part of this research explored and measured how consumers infer the motives behind social media influencers’ product recommendations. Furthermore, this study contributed to the current literature by developing and validating a 31-item long scale and a 19-item short scale of influencer motives in the context of social media with a good model fit, reliability, and validity. The findings indicate that the inference of influencer motives was a multi-dimensional construct with six distinct types: Money, Selling, Image, Love, Sharing, and Helping. The findings expand previous PKM research’s definitions of influencer motives from the broad and vague definition of “ultimate motives” (e.g. DeCarlo, 2005) and the overly simplistic definitions of “selling motive” (e.g. Xie, Boush, & Liu, 2013) or “persuasive intent”(e.g. Tutaj & van Reijmersdal, 2012) in the retail sales and advertising context to a comprehensive definition of six-dimensional influencer motives in the social media influencer marketing context. It is worth noting that different types of influencer motives can co-exist with each other. When a consumer sees a social media influencer post sponsored content, he or she would have multiple thoughts about the social media influencer’s underlying motives, instead of only perceiving a single motive or the dichotomized motives. In other words, a consumer is likely to think an influencer is motivated to post a recommendation on social media for various reasons, such as making money, selling products, enhancing his or her public images, a genuine love for the product, sharing information, and helping other consumers to make a decision. 82 Role of Influencer Motives This study provides an important theoretical implication by developing and testing a conceptual model of consumer resistance to sponsored Instagram posts. This conceptual model incorporates the construct of consumer inferences of a social media influencer’s motive for recommending the product on social media, a missing link in previous studies. As expected, image motives directly increased consumer resistance towards sponsored content; however, love, sharing, and helping motives altogether reduced consumer resistance to persuasion. These findings suggest that marketers and influencers might emphasize on creating content that encourages readers to perceive the motives of influencers’ recommendations as a genuine approval of the product, an enjoyment of sharing, and an intention to help others to reduce resistance to sponsored posts. Surprisingly, the findings showed that money and selling motives do not directly increase consumer resistance towards a sponsored post, but rather indirectly influenced consumer resistance. The post hoc mediation analysis revealed that the effect of money and selling motives on consumer resistance was fully mediated by advertising recognition. In other words, when a consumer realizes that a social media influencer wants to sell something or benefit financially from the sponsorship, the consumer will be more likely to recognize the sponsored content as an advertisement, and thus generates greater resistance to the content. This result confirms the change of meaning principle in PKM (Friestad & Wright, 1994), which states that once consumers activate persuasion knowledge — in this case, making inferences to money and selling motives and recognizing the sponsored posts as ads —they will demonstrate resistance to persuasion. 83 Role of Sponsorship Disclosure This study also examined the role of sponsorship disclosure on different types of influencer motives as well as consumer resistance. Consistent with the PKM literature, this study found that a sponsored post with sponsorship disclosure would leads to stronger consumer inferences of money motives and selling motives, than a post without a sponsorship disclosure does. This tells us the FTC-required distinct and conspicuous disclosure in sponsored content has its intended effect of making consumers aware of the influencers’ financial motives. However, contrary to expectations, our findings do not show a significant effect of sponsorship disclosure on consumer resistance. Consumers who read a sponsored post with disclosure do not demonstrate significantly more resistance than those who read than a sponsored post without disclosure. This result contradicts to previous empirical findings that sponsorship disclosure would often leads to resistance to persuasion. Our study shows that consumer resistance is a result of complex consumer cognitive appraisals of all kinds of influencer motives. Being exposed to a sponsored post with disclosure, consumers would still make inferences about non-monetary motives (e.g. helping motives), based on their knowledge of the influencer and their understanding of the sponsored content, which possibly mitigates consumer resistance caused by money motives and selling motives. This finding is important for advertisers and marketers who worry that displaying disclosure in a sponsored post leads to more resistance to persuasion. Furthermore, the study did not find any main effect of sponsorship disclosure on image motives, love motives, sharing motives, or helping motives. This finding suggests that nonselling and persuasive intents, such as image, love, sharing or helping motives are independent of sponsorship disclosure. Future research might explore the relationships between disclosure and 84 non-financial motives and investigate the antecedents that would possibly influence these motives. Role of Influencer Credibility Influencer credibility has frequently been studied as an outcome of persuasion knowledge activation in previous studies. For instance, Campbell and Kirmani (2000) measured the perceived credibility of a salesperson as the consumers’ evaluations of the salesperson after exposure to the salesperson’s flattering comments about how the customer looked in the jacket being sold. This current research contributes to the above study by considering influencer credibility as an additional information cue for attributing influencer motives as well as by providing evidence that influencer credibility is a critical factor that affects different types of influencer motives and consumer resistance. The results of multivariate tests showed that a highly credible influencer would lead a consumer to make stronger inferences that the product recommendation was made to enhance the influencer’s public image, communicate his or her genuine love of the product, or express his or her intention to help others. Consumers are also less likely to resist a credible influencer’s message than they are for a less credible influencer. Post-hoc SEMs with good model fits revealed that influencer credibility affected how consumers used influencer motives to resist sponsored content. When influencer credibility is relatively high, consumer resistance to sponsored content is negatively influenced by love, sharing, and helping motives; increased by image motives; and is not influenced by money and selling motives via ad recognition. When influencer credibility is relatively low, consumer resistance to sponsored content is positively influenced by money and selling motives. Inferences about love, sharing, and helping motives 85 generated less resistance, and consumer resistance is not influenced by image motives for lowcredibility influencers. Based on the findings on the importance of influencer credibility, our recommendation is that advertisers and marketers use social media influencers who have good reputations and credible histories. It is important that influencers present themselves in a highly authentic and down-to-earth manner and ask for likes, shares, or comments from their readers when appropriate to elicit more thoughts about helping motives and fewer thoughts about image motives and to generate less consumer resistance. In addition, this study generated some unexpected results regarding the sponsorship disclosure × influencer credibility interaction effect on different types of influencer motives and consumer resistance. Our findings did not find any significant interaction effect between sponsorship disclosure and influencer credibility, meaning that the effect of influencer credibility on influencer motives and consumer resistance did not differ by the disclosure condition. This contradicts PKM, which posits that agent knowledge interacts with persuasion knowledge to affect how consumers cope with persuasive messages. Future research could study other types of agent knowledge, such as the pre-existing agent attitudes, or other types of persuasion knowledge, such as consumer knowledge of a specific type persuasion tactic, to further explore the interaction effect. Role of Product Category This study used two types of goods—one credence good (i.e., protein supplement powder) and one experience good (i.e., a travel site) — to enhance the ecological validity of the online experiment. The results do not indicate any significant two-way interaction effects or any significant three-way interaction effect. However, there were significant main effects of product 86 category on selling motives and image motives. Specifically, the sponsored post recommending the protein supplement made people infer stronger selling motives and stronger image motives than the sponsored post recommending the travel site. This finding implies that marketers might consider using influencer marketing for experience goods, as readers are less likely to perceive the selling and image motives of the influencer who recommends experience goods, which may lead to decreased resistance. Still, future research could use other credence and experience goods to further investigate the differences between these two product categories. Limitations and Future Research Directions In addition to the above discussion, this study has some limitations, which provide directions for future research. To begin, during the first stage of the scale development of influencer motives, a few cross-loading items together with the low factor loading items were excluded to purify the dimensionality of the scale. For instance, the Image_4 item “the person wants sponsors to notice their social influence” loaded not only on the image motives dimension but also on the money motives dimension. In addition, the difference between the two factorloadings was smaller than 0.10. Therefore, we removed this item from the scale. It is possible that the deleted cross-loading items could influence consumer resistance. However, the study failed to test these cross-loading motives since these items were excluded from the scale in the early stages of the scale development. Second, the generalization of the experiment findings is limited to the use of fictitious influencers and followers, the two specific products used in the study, Instagram, and the type of sponsorship. To enhance the external validity of the study, future research could use real microcelebrities as influencers and their real followers as study participants to test the model in an authentic relationship between the influencer and their readers. Future research could also test 87 the scale of influencer motives and the theoretical model across different products and media platforms to see if the results can be replicated. Moreover, this experiment used a sponsored post that recommends a product. There are other types of sponsored content on social media, such as those that provide offers and promotion codes, contain affiliate shopping links, offer giveaways, and promote good causes. It would be interesting to test if the results can be replicated across different types of sponsorship. This study also has limitations in its manipulation of credibility. Although consumers’ perception of a highly credible influencer was significantly stronger than those of low-credibility influencers, the average score of the perceived credibility in the low-credibility category is around the midpoint of the scale. Also, participants overall did not show much resistance to the sponsored posts in the experiment, with an average score that is lower than the midpoint of the scale. This could be due to the participants’ highly positive attitudes toward sponsored content (mean = 3.78, SD = 1.16). Further studies could revise the low-credibility manipulation and the message content to create variances. Moreover, this study only examined the effects of disclosure and influencer credibility on influencers’ motives, which further influenced consumer resistance. Since the study results indicate that love, sharing, and helping motives reduce consumer resistance, future studies could investigate other antecedents to love, sharing, and helping motives in addition to disclosure and influencer credibility. Furthermore, this study did not take individual differences into account. Future studies could investigate the boundary conditions and possible moderating factors such as age, education, the need for cognition, and perceived appropriateness of sponsored content, which were found to be significant covariates in this study. For instance, to explore the role of 88 involvement, future research could expand the theoretical model by considering the information dual-processing models, such as elaboration likelihood model (e.g. Cheung, Sia, & Kuan, 2012; Lee et al., 2008; Park & Kim, 2009; Park et al., 2007) or the heuristic-systematic model (e.g. Chaiken, 1980). 89 90 APPENDICES 91 APPENDIX A. TWO SCALES OF INFLUENCER MOTIVES Table A. Two Scales of Influencer Motives Full Scale: 31-item Money Money_1 Motives Money_2 Money_5 Money_6 Money_7 Money_9 Selling Selling_3 Motives Selling_5 Selling_6 Selling_7 Image Image_5 Motives Image_6 Image_7 Image_13 Image_14 Image_16 Image_17 Love Love_2 Motives Love_3 Love_5 Love_6 Love_7 Social Social_4 Motives Social_5 Social_7 Helping Helping_1 Motives Helping_2 Helping_3 Helping_6 Helping_7 Helping_10 Efficient Scale: 19-item Money Money_1 Motives Money_2 Money_9 benefits by this sponsorship. is paid to recommend the product. wants to monetize their relationship with their followers in the future. receives a free product. receives promotional coupons. wants to receive future sponsorships. wants to sell the product. wants to help attract new customers. wants to increase product sales. wants to increase company profits. wants to gain followers. wants to gain likes. wants to gain shares. wants to promote himself or herself as the guru on social media. wants to appear successful. is seeking attention. loves attention. views the brand as a good product. likes the product. expresses their enjoyment with the product. is satisfied with the product. thinks this product works for him/her. wants to share the product they use with others. wants to express their own opinion of the product. enjoys sharing on social media. cares about the followers. has a genuine concern for the welfare of the followers. cares about getting useful information to the followers. wants to help others to make better purchase decisions. wants to help others get the information they want. knows their followers are interested. benefits by this sponsorship. is paid to recommend the product. wants to receive future sponsorships. 92 APPENDIX A. (cont’d) Selling Motives Selling_3 Selling_6 Selling_7 Image Motives Image_5 Image_6 Image_7 Image_13 Love Motives Love_2 Love_3 Love_6 Social Motives Social_4 Social_5 Social_7 Helping Helping_3 Motives Helping_6 Helping_7 wants to sell the product. wants to increase product sales. wants to increase company profits. wants to gain followers. wants to gain likes. wants to gain shares. wants to promote himself or herself as the guru on social media. views the brand as a good product. likes the product. is satisfied with the product. wants to share the product they use with others. wants to express their own opinion of the product. enjoys sharing on social media. cares about getting useful information to the followers. wants to help others to make better purchase decisions. wants to help others get the information they want. 93 APPENDIX B. CONSENT FORM 1. EXPLANATION OF THE RESEARCH and WHAT YOU WILL DO: • You are being asked to participate in a research study of social media posts. • During this study, we will ask you to read a social media post and answer a few questions. • The duration of this online survey is approximately 10 minutes. Please allow sufficient time to provide thoughtful, fair and honest responses. • You must be at least 18 years old to participate in this research. 2. YOUR RIGHTS TO PARTICIPATE, SAY NO, OR WITHDRAW: • Participation in this research project is completely voluntary. You have the right to say no. You may change your mind at any time and withdraw. You may choose not to answer specific questions or to stop participating at any time. 3. COSTS AND COMPENSATION FOR BEING IN THE STUDY: • There are no costs associated with completing this research study. • You will be awarded $1 after your completion of the study. 4. CONTACT INFORMATION FOR QUESTIONS AND CONCERNS: • If you have concerns or questions about this study, such as scientific issues, how to do any part of it, or to report an injury, please contact the researcher (Mengtian Jiang, Room 309, 404 Wilson Road, East Lansing, MI 48823, jiangme2@msu.edu, 517-515-9687). • If you have questions or concerns about your role and rights as a research participant, would like to obtain information or offer input, or would like to register a complaint about this study, you may contact, anonymously if you wish, the Michigan State University’s Human Research Protection Program at 517-355-2180, Fax 517-432-4503, or e-mail irb@msu.edu or regular mail at 4000 Collins Rd, Suite 136, Lansing, MI 48910. 5. DOCUMENTATION OF INFORMED CONSENT:  By clicking on this button, I agree to participate in this online survey. 94 APPENDIX C. PRETEST QUESTIONNAIRE Please indicate the gender image that comes to your mind when you see the following names: For instance, if you see the name "Alice" and think this is a name for a girl, then you should select "Female". Male Female Unisex Alex    Andy    Dakota    Jesse    Jordan    Kendall    Max    Pat    Ray    Taylor    Please indicate how much you like the following names. Dislike a lot Dislike a Neither like little nor dislike Alex    Like a little Like a lot                         Max      Pat      Ray      Taylor      Andy   Dakota  Jesse Jordan Kendall 95 Below are the suggested new brand names for a dietary supplement product. Please indicate how much you like the following new brand names. Dislike very Dislike a Neither like Like a little Like very much little nor dislike much Greensland      Lesupplement      Vital nutrition      Supergreens      Vitaminos      The super elixir Vera           Greensmania      Nutritionlab      Below are the suggested new brand names for a dietary supplement product. Please indicate how familiar you are with the following new brand names. Very unfamiliar Neither Familiar Very familiar unfamiliar unfamiliar nor familiar Greensland      Lesupplement      Vital nutrition      Supergreens      Vitaminos      The super elixir Vera           Greensmania      Nutritionlab      96 Below are the suggested new brand names for a travel agency's website. Please indicate how much you like the following new brand names. Dislike Dislike a Neither like Like a Like very very much little nor dislike little much AdventureTours.com      GoTours.com      Tourism.com      Vacations.com      GreatEscapesTravel.com      FloridaAdventure.com      FloridaVacations.com      KeyWestExperience.com      VacationKeyWest.com      Below are the suggested new brand names for a travel agency's website. Please indicate how familiar you are with the following new brand names. Very unfamiliar Neither Familiar Very unfamiliar unfamiliar familiar nor familiar AdventureTours.com      GoTours.com      Tourism.com      Vacations.com      GreatEscapesTravel.com      FloridaAdventure.com      FloridaVacations.com      KeyWestExperience.com      VacationKeyWest.com      97 Please indicate how much you like the following images. Dislike very much Dislike a little Neither like nor dislike Like a little Like very much      Dislike very much Dislike a little Neither like nor dislike Like a little Like very much      98 Dislike very much  Dislike very much  Dislike a little  Dislike a little  Neither like nor dislike  Neither like nor dislike  99 Like a little Like very much   Like a little Like very much   Dislike very much  Dislike very much  Dislike a little  Dislike a little  Neither like nor dislike  Neither like nor dislike  100 Like a little Like very much   Like a little Like very much   Dislike very much Dislike a little Neither like nor dislike Like a little Like very much      Dislike very much Dislike a little Neither like nor dislike Like a little Like very much      101 Dislike very much  Dislike very much  Dislike a little  Dislike a little  Neither like nor dislike  Neither like nor dislike  102 Like a little Like very much   Like a little Like very much   What year were you born? 19__ ________________________________________________________________ What is your gender?  Female  Male What is your race?  African-American  Asian  Caucasian  Hawaiian Nation or Pacific Islander  Hispanic, Latino or Spanish origin  Native American or Alaskan native  Multiracial  Other ________________________________________________ Thank you very much for your time and cooperation. Please enter your M-Turk Worker ID below. After you enter the ID, please click the ">>" button. ________________________________________________________________ You will receive a completion code on the next page. Please copy and paste the code on the HIT page to receive your incentive. Thank you very much for your time and effort. Your responses are very important to us. 103 APPENDIX D. MAIN STUDY QUESTIONNAIRE We use social media to follow people we like or we find helpful. Today we are going to ask you to review a social media post written by someone named Alex. We would like you to imagine that you are currently following Alex and are familiar with Alex’s posts. First, we would like you to read a profile of Alex. Then, we will show you one of Alex’s posts and we would you to review it and offer your opinions about the post. Based on the short bio, please briefly describe your impression of Alex. ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ Now, you are going to see a screenshot of Alex’s latest Instagram posts. We would like to ask you please read the post carefully. We will then ask you to review it and offer your opinions about it. Please briefly summarize the content of the post that you just saw. ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ 104 The followings are the common reasons why someone would recommend a product in social media. Think about the reasons that are related to the Alex’s post that you just read. To what extent do you agree or disagree with these statements? Alex writes this post because s/he... Strongly Somewhat Neither Somewhat Strongly agree agree agree disagree disagree nor disagree benefits by this sponsorship. is paid to recommend the product. wants to receive future sponsorships. wants to sell the product. wants to increase product sales. wants to increase company profits. wants to gain followers. wants to gain likes. wants to gain shares. wants to promote himself or herself as the guru on social media. wants others to know what s/he values about himself or herself. wants others to learn more about himself or herself. wants others to understand what is important to him or her. views the brand as a good product. likes the product. is satisfied with the product. wants to share the product they use with others. wants to express their own opinion of the product. enjoys sharing on social media. cares about getting useful information to the followers. wants to help others to make better purchase decisions. wants to help others get the information they want. 105 While reading the Instagram post, Strongly agree Somewhat Neither Somewhat Strongly agree agree disagree disagree nor disagree I contested the information in the content. I refuted the information in the content. I doubted the information in the content. I countered the information in the content. I felt angry. I felt irritated. I felt annoyed. The content that I read in the Instagram post is ... 1 2 3 4 5 Unimportant Important irrelevant relevant means nothing means a lot worthless to me valuable to me boring interesting unexciting exciting unappealing appealing not needed needed mundane fascinating uninvolving to me involving to me 106 In your opinion, Alex, the person who writes the Instagram post, is ... 1 2 3 4 5 Attractive Unattractive Classy Not classy Beautiful Ugly Elegant Plain Sexy Not sexy Dependable Undependable Honest Dishonest Reliable Unreliable Sincere Insincere Trustworthy Untrustworthy Expert Not an expert Experienced Inexperienced Knowledgeable Unknowledgeable Qualified Unqualified Skilled Unskilled In the post you just read, how confident are you in recalling seeing the following items? I am I am Neither Somewhat Extremely extremely somewhat confident unconfident unconfident confident confident nor unconfident a disclosure for sponsorship #sponsored a sentence indicating the material connection between Alex and the brand. 107 Please indicate to what extent you agree with the following statements. Strongly Somewhat Neither Somewhat Strongly agree agree agree disagree disagree nor disagree It is highly unethical for an Instagram blogger to post sponsored content without identifying the sponsored brands. It is appropriate that an Instagram blogger gets paid to post some sponsored content. Instagram bloggers can post content for pay, because they provide information to us for free. It is acceptable to me if the brands pay to place their messages in an Instagram post. What year were you born? 19__ (PLEASE ENTER THE LAST TWO DIGITS OF YOUR BIRTH YEAR IN THE BOX BELOW) ________________________________________________________________ What is your gender? Female Male What is your race? African-American Asian Caucasian Hawaiian Nation or Pacific Islander Hispanic, Latino or Spanish origin Native American or Alaskan native Multiracial Other ________________________________________________ Please indicate the answer that includes your entire household income in (previous year) before taxes. Less than $10,000 $10,000 to $19,999 $20,000 to $29,999 $30,000 to $39,999 $40,000 to $49,999 $50,000 to $59,999 108 $60,000 to $69,999 $70,000 to $79,999 $80,000 to $89,999 $90,000 to $99,999 $100,000 to $149,999 $150,000 or more Not including kindergarten, how many years of formal education have you completed? (PLEASE ENTER THE NUMBER OF YEARS OF FORMAL EDUCATION IN THE BOX BELOW) ________________________________________________________________ Thank you very much for your time and cooperation. Please enter your MTurk ID below. After you enter the ID, please click the ">>" button. ________________________________________________________________ Thank you very much for your time and effort. Your responses are very important to us. 109 APPENDIX E. DISCLOSURE + PROTEIN POWER SPONSORED INSTAGRAM POST 110 APPENDIX F. NO DISCLOSURE + PROTEIN POWER SPONSORED INSTAGRAM POST 111 APPENDIX G. DISCLOSURE + TRAVEL SITE SPONSORED INSTAGRAM POST 112 APPENDIX H. NO DISCLOSURE + TRAVEL SITE SPONSORED INSTAGRAM POST 113 REFERENCES 114 REFERENCES Alexandrov, A., Lilly, B., & Babakus, E. (2013). The effects of social- and self-motives on the intentions to share positive and negative word of mouth. Journal of the Academy of Marketing Science, 41(5), 531–546. https://doi.org/10.1007/s11747-012-0323-4. Anderson, E. W. (1998). Customer Satisfaction and Word of Mouth. Journal of Service Research, 1(1), 5–17. https://doi.org/10.1177/109467059800100102 Arndt, J. (1967a). Role of product-related conversations in the diffusion of a new product. Journal of Marketing Research, 291–295. Arndt, J. (1967b). Word of mouth advertising: A review of the literature. Advertising Research Foundation. Babić, A. R., Sotgiu, F., & Bijmolt, T. H. A. (2015). The Effect of Electronic Word of Mouth on Sales: A Meta-Analytic Review of Platform, Product, and Metric Factors. Journal of Marketing Research, 53(3), 297–318. https://doi.org/10.1509/jmr.14.0380 Babin, B. J., Darden, W. R., & Griffin, M. (1994). Work and/or Fun: Measuring Hedonic and Utilitarian Shopping Value. Journal of Consumer Research, 20(4), 644. https://doi.org/10.1086/209376 Bambauer-Sachse, S., & Mangold, S. (2013). Do consumers still believe what is said in online product reviews? A persuasion knowledge approach. Journal of Retailing and Consumer Services, 20(4), 373–381. https://doi.org/10.1016/j.jretconser.2013.03.004 Banerjee, A., Chitnis, U. B., Jadhav, S. L., Bhawalkar, J. S., & Chaudhury, S. (2009). Hypothesis testing, type I and type II errors. Industrial Psychiatry Journal, 18(2), 127–131. https://doi.org/10.4103/0972-6748.62274 Bickart, B., & Schindler, R. M. (2001). Internet forums as influential sources of consumer information. Journal of Interactive Marketing, 15(3), 31–40. Boerman, S. C., & Kruikemeier, S. (2016). Consumer responses to promoted tweets sent by brands and political parties. Computers in Human Behavior, 65, 285–294. https://doi.org/10.1016/j.chb.2016.08.033 Boerman, S. C., & van Reijmersdal, E. A. (2016). Informing Consumers about ’Hidden’Advertising A Literature Review of the Effects of Disclosing Sponsored Content. In Advertising in new formats and media: Current research and implications for marketers (pp. 115–146). London, UK: Emerald Group Publishing. Boerman, S. C., van Reijmersdal, E. A., & Neijens, P. C. (2012). Sponsorship Disclosure: Effects of Duration on Persuasion Knowledge and Brand Responses. Journal of Communication, 62(6), 1047–1064. https://doi.org/10.1111/j.1460-2466.2012.01677.x 115 Boerman, S. C., van Reijmersdal, E. A., & Neijens, P. C. (2013). Appreciation and effects of sponsorship disclosure. In Advances in Advertising Research (Vol. IV) (pp. 273–284). Springer. Retrieved from http://link.springer.com/chapter/10.1007/978-3-658-023652_21 Boerman, S. C., van Reijmersdal, E. A., & Neijens, P. C. (2014). Effects of sponsorship disclosure timing on the processing of sponsored content: A study on the effectiveness of European disclosure regulations. Psychology & Marketing, 31(3), 214–224. Boerman, S. C., van Reijmersdal, E. A., & Neijens, P. C. (2015). How audience and disclosure characteristics influence memory of sponsorship disclosures. International Journal of Advertising, 34(4), 576–592. Boerman, S. C., Willemsen, L. M., & Van Der Aa, E. P. (2017). “This Post Is Sponsored”: Effects of Sponsorship Disclosure on Persuasion Knowledge and Electronic Word of Mouth in the Context of Facebook. Journal of Interactive Marketing, 38, 82–92. https://doi.org/10.1016/j.intmar.2016.12.002 Brandt, A. C., Vonk, R., & van Knippenberg, A. (2011). Augmentation and Discounting in Impressions of Targets Described by Third Parties with Ulterior Motives. Social Cognition; New York, 29(2), 210–220. https://doi.org/http://dx.doi.org.proxy2.cl.msu.edu/10.1521/soco.2011.29.2.210 Brehm, J. W. (1966). A theory of psychological reactance. Retrieved from http://doi.apa.org/psycinfo/1967-08061-000 Brogan, C. (2008, December 2). Sponsored Post-Kmart Holiday Shopping Dad Style. Retrieved from http://dadomatic.com/sponsored-post-kmart-holiday-shopping-dad-style/ Brown, C. L., & Krishna, A. (2004). The Skeptical Shopper: A Metacognitive Account for the Effects of Default Options on Choice. Journal of Consumer Research, 31(3), 529–539. https://doi.org/10.1086/425087 Brown, T. J., Barry, T. E., Dacin, P. A., & Gunst, R. F. (2005). Spreading the word: Investigating antecedents of consumers’ positive word-of-mouth intentions and behaviors in a retailing context. Journal of the Academy of Marketing Science, 33(2), 123–138. Buyer, L. (2016, January 18). PR Under the Influence: Why Influencer Marketing is all the Buzz. Retrieved from http://socialprchat.com/pr-under-the-influence-why-influencermarketing-is-all-the-buzz/ Campbell, M. C., & Kirmani, A. (2000). Consumers’ Use of Persuasion Knowledge: The Effects of Accessibility and Cognitive Capacity on Perceptions of an Influence Agent. Journal of Consumer Research, 27(1), 69–83. https://doi.org/10.1086/314309 116 Campbell, M. C., & Kirmani, A. (2008). I know what you’re doing and why you’re doing it: the use of the persuasion knowledge model in consumer research. Handbook of Consumer Psychology, 549–575. Campbell, M. C., Mohr, G. S., & Verlegh, P. W. J. (2013). Can disclosures lead consumers to resist covert persuasion? The important roles of disclosure timing and type of response. Journal of Consumer Psychology, 23(4), 483–495. https://doi.org/10.1016/j.jcps.2012.10.012 Carl, W. J. (2008). The role of disclosure in organized word‐of‐mouth marketing programs. Journal of Marketing Communications, 14(3), 225–241. https://doi.org/10.1080/13527260701833839 Carlson, K. (2016, July 1). We Trust Influencers Just as Much as Friends. Retrieved from https://business.experticity.com/we-trust-influencers-just-as-much-as-friends/ Chafkin, M. (2016, November 30). Confessions of an Instagram Influencer. Bloomberg.Com. Retrieved from https://www.bloomberg.com/news/features/2016-11-30/confessions-ofan-instagram-influencer Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39(5), 752–766. http://dx.doi.org /10.1037/0022-3514.39.5.752 Cheung, C. M. K., & Thadani, D. R. (2010). The effectiveness of electronic word-of-mouth communication: A literature analysis. Proceedings of the 23rd Bled EConference ETrust: Implications for the Individual, Enterprises and Society, 329–345. Cheung, C. M. K., & Thadani, D. R. (2012). The impact of electronic word-of-mouth communication: A literature analysis and integrative model. Decision Support Systems, 54(1), 461–470. https://doi.org/10.1016/j.dss.2012.06.008 Cheung, C. M.-Y., Sia, C.-L., & Kuan, K. K. (2012). Is this review believable? A study of factors affecting the credibility of online consumer reviews from an ELM perspective. Journal of the Association for Information Systems, 13(8), 618. Chevalier, J. A., & Mayzlin, D. (2006). The Effect of Word of Mouth on Sales: Online Book Reviews. Journal of Marketing Research, 43(3), 345–354. Choi, S. M. (2002). Attributional approach to understanding celebrity /product congruence effects: Role of perceived expertise (Ph.D.). Michigan State University, United States -Michigan. Choi, S. M., & Rifon, N. J. (2012). It Is a Match: The Impact of Congruence between Celebrity Image and Consumer Ideal Self on Endorsement Effectiveness. Psychology & Marketing, 29(9), 639–650. https://doi.org/10.1002/mar.20550 117 Churchill, G. A. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 64–73. Comrey, A. L., & Lee, H. B. (1992). A first course in factor analysis. NY: Psychology Press. Constine, J. (2017, April 26). Instagram’s growth speeds up as it hits 700 million users. Retrieved from http://social.techcrunch.com/2017/04/26/instagram-700-million-users/ Cronley, M. L., Kardes, F. R., Goddard, P., & Houghton, D. C. (1999). Endorsing products for the money: The role of the correspondence bias in celebrity advertising. NA-Advances in Consumer Research Volume 26. Retrieved from http://acrwebsite.org/volumes/8333/volumes/v26/NA-26 Curren, M. T., & Folkes, V. S. (1987). Attributional influences on consumers’ desires to communicate about products. Psychology and Marketing, 4(1), 31. Darby, M. R., & Karni, E. (1973). Free competition and the optimal amount of fraud. The Journal of Law and Economics, 16(1), 67–88. Daugherty, T., Eastin, M. S., & Bright, L. (2008). Exploring Consumer Motivations for Creating User-Generated Content. Journal of Interactive Advertising, 8(2), 16–25. https://doi.org/10.1080/15252019.2008.10722139 DeCarlo, T. E. (2005). The effects of sales message and suspicion of ulterior motives on salesperson evaluation. Journal of Consumer Psychology, 15(3), 238–249. Dhanik, T. (2016, December 12). Micro, Not Macro: Rethinking Influencer Marketing. Retrieved from http://adage.com/article/digitalnext/micro-macro-influencer-marketingkim-kardashian/307118/ Dichter, E. (1966). How Word-of-Mouth Advertising Works. Harvard Business Review, 44(6), 147. Eisend, M. (2006). Source credibility dimensions in marketing communication–A generalized solution. Journal of Empirical Generalizations in Marketing, 10(2), 1–33. Eisend, M. (2015). Persuasion knowledge and third-person perceptions in advertising: the moderating effect of regulatory competence. International Journal of Advertising, 34(1), 54–69. https://doi.org/10.1080/02650487.2014.993792 Federal Trade Commission. (2009, October 5). FTC Publishes Final Guides Governing Endorsements, Testimonials | Federal Trade Commission. Retrieved April 24, 2017, from https://www.ftc.gov/news-events/press-releases/2009/10/ftc-publishes-final-guidesgoverning-endorsements-testimonials 118 Federal Trade Commission. (2013, March). .com Disclosures: How to Make Effective Disclosures in Digital Advertising | Federal Trade Commission. Retrieved April 24, 2017, from https://www.ftc.gov/tips-advice/business-center/guidance/com-disclosureshow-make-effective-disclosures-digital Federal Trade Commission. (2015, May). The FTC’s Endorsement Guides: What People Are Asking | Federal Trade Commission. Retrieved March 13, 2016, from https://www.ftc.gov/tips-advice/business-center/guidance/ftcs-endorsement-guides-whatpeople-are-asking#product Federal Trade Commission. (2016a, July 11). Warner Bros. Settles FTC Charges It Failed to Adequately Disclose It Paid Online Influencers to Post Gameplay Videos. Retrieved from https://www.ftc.gov/news-events/press-releases/2016/07/warner-bros-settles-ftc-chargesit-failed-adequately-disclose-it Federal Trade Commission. (2016b, November 21). FTC Approves Final Order Requiring Warner Bros. to Disclose Payments to Online Influencers. Retrieved from https://www.ftc.gov/news-events/press-releases/2016/11/ftc-approves-final-orderrequiring-warner-bros-disclose-payments Federal Trade Commission. (2017, April 18). FTC Staff Reminds Influencers and Brands to Clearly Disclose Relationship. Retrieved from https://www.ftc.gov/news-events/pressreleases/2017/04/ftc-staff-reminds-influencers-brands-clearly-disclose Federal Trade Commission. (n.d.). About the FTC. Retrieved from https://www.ftc.gov/about-ftc Fiske, S. T., & Taylor, S. E. (1991). Social cognition, 2nd. NY: McGraw-Hill. Forman, C., Ghose, A., & Wiesenfeld, B. (2008). Examining the relationship between reviews and sales: the role of reviewer identity disclosure in electronic markets. Information Systems Research, 19(3), 291. Fornell, C., & Larcker, D. F. (1981). Structural equation models with unobservable variables and measurement error: Algebra and statistics. Journal of Marketing Research, 382–388. Friestad, M., & Wright, P. (1994). The Persuasion Knowledge Model: How People Cope with Persuasion Attempts. Journal of Consumer Research, 21(1), 1–31. Goldsmith, R. E., & Hofacker, C. F. (1991). Measuring consumer innovativeness. Journal of the Academy of Marketing Science, 19(3), 209–221. https://doi.org/10.1007/BF02726497 Goldsmith, R. E., Lafferty, B. A., & Newell, S. J. (2000). The impact of corporate credibility and celebrity credibility on consumer reaction to advertisements and brands. Journal of Advertising, 29(3), 43–54. 119 Gottbrecht, L. (2016, October 18). The Three Types of Influencers All Marketers Should Know [Infographic]. Retrieved from http://www.mavrck.co/the-three-types-of-influencers-allmarketers-should-know-infographic/ Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate data analysis (sixth). Pearson. Ham, C.-D., Nelson, M. R., & Das, S. (2015). How to Measure Persuasion Knowledge. International Journal of Advertising, 34(1), 17–53. https://doi.org/10.1080/02650487.2014.994730 Hamedy, S. (2016, December 5). Forbes 2016 list of highest paid YouTubers will make you weep. Retrieved from http://mashable.com/2016/12/05/forbes-highest-paid-youtubestars-2016/#1902dwFaJ8qd. Hardesty, D. M., Carlson, J. P., & Bearden, W. (2002). Brand familiarity and invoice price effects on consumer evaluations: the moderating role of skepticism toward advertising. Journal of Advertising, 31(2), 1–15. Hass, R. G., & Grady, K. (1975). Temporal delay, type of forewarning, and resistance to influence. Journal of Experimental Social Psychology, 11(5), 459–469. Hayes, A. F. (2013). Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach (1 edition). New York: The Guilford Press. Heine, C. (2016, July 19). Selena Gomez’s Social Media Posts Are Evidently Worth $550,000 Apiece – Adweek. Retrieved from http://www.adweek.com/digital/selena-gomezs-socialmedia-posts-are-evidently-worth-550000-apiece-172552/ Hennig-Thurau, T., Gwinner, K. P., Walsh, G., & Gremler, D. D. (2004). Electronic word-ofmouth via consumer-opinion platforms: What motivates consumers to articulate themselves on the Internet? Journal of Interactive Marketing, 18(1), 38–52. Hinkin, T. R., Tracey, J. B., & Enz, C. A. (1997). Scale construction: Developing reliable and valid measurement instruments. Journal of Hospitality & Tourism Research, 21(1), 100– 120. Homburg, C., Schwemmle, M., & Kuehnl, C. (2015). New Product Design: Concept, Measurement, and Consequences. Journal of Marketing, 79(3), 41–56. https://doi.org/10.1509/jm.14.0199 Hooper, D., Coughlan, J., & Mullen, M. (2008). Structural equation modelling: Guidelines for determining model fit. Articles, 2. 120 Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion; psychological studies of opinion change. Retrieved from http://doi.apa.org/psycinfo/1954-05952-000 Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication effectiveness. Public Opinion Quarterly, 15(4), 635–650. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. Hu, N., Liu, L., & Zhang, J. J. (2008). Do online reviews affect product sales? The role of reviewer characteristics and temporal effects. Information Technology and Management, 9(3), 201–214. https://doi.org/10.1007/s10799-008-0041-2 Huang, J.-H., & Chen, Y.-F. (2006). Herding in online product choice. Psychology & Marketing, 23(5), 413–428. Hung, K. H., & Li, S. Y. (2007). The influence of eWOM on virtual consumer communities: Social capital, consumer learning, and behavioral outcomes. Journal of Advertising Research, 47(4), 485–495. Hwang, Y., & Jeong, S.-H. (2016). “This is a sponsored blog post, but all opinions are my own”: The effects of sponsorship disclosure on responses to sponsored blog posts. Computers in Human Behavior, 62, 528–535. https://doi.org/10.1016/j.chb.2016.04.026 Jacks, J. Z., & Cameron, K. A. (2003). Strategies for Resisting Persuasion. Basic and Applied Social Psychology, 25(2), 145–161. https://doi.org/10.1207/S15324834BASP2502_5 Jeong, H. J., & Lee, M. (2013). Effects of recommendation systems on consumer inferences of website motives and attitudes towards a website. International Journal of Advertising, 32(4), 539–558. Jiang, M., McKay, B. A., Richards, J. I., & Snyder, W. (2017). Now You See Me, But You Don’t Know: Consumer Processing of Native Advertisements in Online News Sites. Journal of Interactive Advertising, 17(2), 92–108. https://doi.org/10.1080/15252019.2017.1399839 Jin, S.-A. A., & Phua, J. (2014). Following Celebrities’ Tweets About Brands: The Impact of Twitter-Based Electronic Word-of-Mouth on Consumers’ Source Credibility Perception, Buying Intention, and Social Identification With Celebrities. Journal of Advertising, 43(2), 181–195. https://doi.org/10.1080/00913367.2013.827606 Jones, E. E., & Davis, K. E. (1965). From acts to dispositions the attribution process in person perception. Advances in Experimental Social Psychology, 2, 219–266. 121 Jones, E. E., & Harris, V. A. (1967). The attribution of attitudes. Journal of Experimental Social Psychology, 3(1), 1–24. Jöreskog, K. G., & Sörbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS command language. Lincolnwood, IL: Scientific Software International. Kamins, M. A. (1990). An investigation into the “match-up” hypothesis in celebrity advertising: When beauty may be only skin deep. Journal of Advertising, 19(1), 4–13. Kapadia, A. (2016, March 12). Numbers Don’t Lie: What a 2016 Nielsen Study Revealed About Referrals. Retrieved from http://www.business2community.com/marketing/numbersdont-lie-2016-nielsen-study-revealed-referrals-01477256 Kardes, F. R. (1993). Consumer inference: Determinants, consequences, and implications for advertising (Vol. 349). Hillsdale, NJ: Lawrence Erlbaum. Kelley, H. H. (1967). Attribution theory in social psychology. In Nebraska symposium on motivation. University of Nebraska Press. Retrieved from http://psycnet.apa.org/psycinfo/1968-13540-001 Kelley, H. H. (1971). Attribution in social interaction. In E. E. Jones, D. E. Kanouse, H. H. Kelley, R. E. Nisbett, S. Valins, & B. Weiner (Eds.), Attribution: Perceiving the Causes of Behavior. Morristown, N.J.: General Learning Press. Kelley, H. H. (1973). The processes of causal attribution. American Psychologist, 28(2), 107. Knowles, E. S., & Linn, J. A. (2004). Resistance and persuasion. Psychology Press. Kumar, N., & Benbasat, I. (2006). Research Note: The Influence of Recommendations and Consumer Reviews on Evaluations of Websites. Information Systems Research, 17(4), 425–439. https://doi.org/10.1287/isre.1060.0107 Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 159–174. Lee, J., Park, D.-H., & Han, I. (2008). The effect of negative online consumer reviews on product attitude: An information processing view. Electronic Commerce Research and Applications, 7(3), 341–352. https://doi.org/10.1016/j.elerap.2007.05.004 Lee, M., Rodgers, S., & Kim, M. (2009). Effects of valence and extremity of eWOM on attitude toward the brand and website. Journal of Current Issues & Research in Advertising, 31(2), 1–11. Lee, M., & Youn, S. (2009). Electronic word of mouth (eWOM) How eWOM platforms influence consumer product judgement. International Journal of Advertising, 28(3), 473– 499. https://doi.org/10.2501/S0265048709200709 122 Linqia. (2016, November 29). The State of Influencer Marketing 2017. Retrieved from http://go.pardot.com/l/153091/2016-11-29/9t6n1k Litvin, S. W., Goldsmith, R. E., & Pan, B. (2008). Electronic word-of-mouth in hospitality and tourism management. Tourism Management, 29(3), 458–468. https://doi.org/10.1016/j.tourman.2007.05.011 Liu, Y. (2006). Word of mouth for movies: Its dynamics and impact on box office revenue. Journal of Marketing, 70(3), 74–89. Lorenzon, K., & Russell, C. A. (2012). From apathy to ambivalence: How is persuasion knowledge reflected in consumers’ comments about in-game advertising? Journal of Marketing Communications, 18(1), 55–67. Lukovit, K. (2008, December 12). Kmart Enlists Bloggers For Social Net Campaign. Retrieved from https://www.mediapost.com/publications/article/96569/kmart-enlists-bloggers-forsocial-net-campaign.html Matos, C. A. de, & Rossi, C. A. V. (2008). Word-of-mouth communications in marketing: a meta-analytic review of the antecedents and moderators. Journal of the Academy of Marketing Science, 36(4), 578–596. https://doi.org/10.1007/s11747-008-0121-1 Malhotra, N. K., & Dash, S. (2009). Marketing Research: An Applied Orientation (5th ed.). New Delhi: Pearson Education. Mazis, M. B., & Staelin, R. (1982). Using Information-Processing Principles in Public Policymaking. Journal of Marketing & Public Policy, 1, 3–14. https://doi.org/10.2307/60000037 McCracken, G. (1989). Who is the Celebrity Endorser? Cultural Foundations of the Endorsement Process. Journal of Consumer Research, 16(3), 310–321. McGuire, W. J. (1964). Some contemporary approaches. Advances in Experimental Social Psychology, 1, 191–229. McGuire, W. J. (1976). Some internal psychological factors influencing consumer choice. Journal of Consumer Research, 302–319. McGuire, W. J. (1985). Attitudes and attitude change. In G. Lindzey & E. Aronson (Eds.), Handbook of Social Psychology (Vol. 2). New York: Random House. McNally, R. C., & Griffin, A. (2007). A measure and initial test of managers’ perceptions of relationship marketing in inter-organizational exchanges. Journal of the Academy of Marketing Science, 35(3), 382–397. https://doi.org/10.1007/s11747-007-0046-0. 123 Mediakix. (2015, December 21). Influencer Marketing Is A $5-$10 Billion Dollar Global Market. Retrieved from http://mediakix.com/2015/12/influencer-marketing-5-10-billiondollar-market/ Mediakix. (2017, March 29). Instagram Influencer Marketing Is Now A $1 Billion Industry. Retrieved from http://mediakix.com/2017/03/instagram-influencer-marketing-industrysize-how-big/ Morales, A. C. (2005). Giving Firms an “E” for Effort: Consumer Responses to High‐Effort Firms. Journal of Consumer Research, 31(4), 806–812. https://doi.org/10.1086/426615 Nekmat, E., & Gower, K. K. (2012). Effects of Disclosure and Message Valence in Online Word-of-Mouth (eWOM) Communication: Implications for Integrated Marketing Communication. International Journal of Integrated Marketing Communications, 4(1), 85–98. Nelson, B. (2012, June 4). Do You Read Fast Enough To Be Successful? Retrieved from https://www.forbes.com/sites/brettnelson/2012/06/04/do-you-read-fast-enough-to-besuccessful/. Nelson, M. R., Wood, M. L. M., & Paek, H.-J. (2009). Increased Persuasion Knowledge of Video News Releases: Audience Beliefs About News and Support for Source Disclosure. Journal of Mass Media Ethics, 24(4), 220–237. doi.org/10.1080/08900520903332626. Nelson, P. (1970). Information and Consumer Behavior. Journal of Political Economy, 78(2), 311–329. Nelson, P. (1974). Advertising as Information. Journal of Political Economy, 82(4), 729–754. Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers’ perceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19(3), 39–52. Osborne, J. W., & Costello, A. B. (2004). Sample size and subject to item ratio in principal components analysis. Practical Assessment, Research & Evaluation, 9(11), 8. Owens, C., Toone, T., & Steed-Ivie, M. (2014). A Survey of Dietary Supplement Knowledge, Attitudes, and Use in a Rural Population. Journal of Nutrition & Food Sciences. https://doi.org/10.4172/2155-9600.1000304 Park, D.-H., & Kim, S. (2009). The effects of consumer knowledge on message processing of electronic word-of-mouth via online consumer reviews. Electronic Commerce Research and Applications, 7(4), 399–410. 124 Park, D.-H., & Lee, J. (2008). eWOM overload and its effect on consumer behavioral intention depending on consumer involvement. Electronic Commerce Research and Applications, 7(4), 386–398. https://doi.org/10.1016/j.elerap.2007.11.004 Park, D.-H., Lee, J., & Han, I. (2007). The Effect of On-Line Consumer Reviews on Consumer Purchasing Intention: The Moderating Role of Involvement. International Journal of Electronic Commerce, 11(4), 125–148. https://doi.org/10.2753/JEC1086-4415110405 Regan, D. T., Straus, E., & Fazio, R. (1974). Liking and the attribution process. Journal of Experimental Social Psychology, 10(4), 385–397. Rifon, N. J., Choi, S. M., Trimble, C. S., & Li, H. (2004). Congruence effects in sponsorship: The mediating role of sponsor credibility and consumer attributions of sponsor motive. Journal of Advertising, 33(1), 30–42. Rifon, N. J., Jiang, M., & Kim, S. (2016). Don’t Hate me Because I am Beautiful: Identifying the Relative Influence of Celebrity Attractiveness and Character Traits on Credibility. In Advances in Advertising Research (Vol. VI) (pp. 125–134). Springer Gabler, Wiesbaden. https://doi.org/10.1007/978-3-658-10558-7_11 Rodgers, S. (2007). Effects of sponsorship congruity on e-sponsors and e-newspapers. Journalism & Mass Communication Quarterly, 84(1), 24–39. Sandes, F. S., & Urdan, A. T. (2013). Electronic Word-of-Mouth Impacts on Consumer Behavior: Exploratory and Experimental Studies. Journal of International Consumer Marketing, 25(3), 181–197. https://doi.org/10.1080/08961530.2013.780850 Sen, S., & Lerman, D. (2007). Why are you telling me this? An examination into negative consumer reviews on the web. Journal of Interactive Marketing, 21(4), 76–94. Settle, R. B., & Golden, L. L. (1974). Attribution Theory and advertiser credibility. JMR, Journal of Marketing Research (Pre-1986), 11(000002), 181. Snyder, W. (2016). Ethics in Advertising: Making the Case for Doing the Right Thing. New York: Routledge. Sørum, K. A., Grape, K. M., & Silvera, D. (2003). Do dispositional attributions regarding peer endorsers influence product evaluations? Scandinavian Journal of Psychology, 44(1), 39– 46. Statista. (2017). Instagram: active users 2017. Retrieved from https://www.statista.com/statistics/253577/number-of-monthly-active-instagram-users/ Sun, T., Youn, S., Wu, G., & Kuntaraporn, M. (2006). Online word-of-mouth (or mouse): An exploration of its antecedents and consequences. Journal of Computer-Mediated Communication, 11(4), 1104–1127. 125 Sundaram, D. S., Mitra, K., & Webster, C. (1998). Word-of-mouth communications: A motivational analysis. NA-Advances in Consumer Research Volume 25. Retrieved from http://acrwebsite.org/volumes/8208/volumes/v25/NA-25 Swant, M. (2016, May 10). Twitter Says Users Now Trust Influencers Nearly as Much as Their Friends. Retrieved from http://www.adweek.com/digital/twitter-says-users-now-trustinfluencers-nearly-much-their-friends-171367/ Tannenbaum, P. H., Macauley, J. R., & Norris, E. L. (1966). Principle of congruity and reduction of persuasion. Journal of Personality and Social Psychology, 3(2), 233. TIME. (2016, April 20). Felix Kjellberg (a.k.a. PewDiePie). Retrieved June 29, 2017, from http://time.com/collection-post/4302406/felix-kjellberg-pewdiepie-2016-time-100/ Tutaj, K., & van Reijmersdal, E. A. van. (2012). Effects of online advertising format and persuasion knowledge on audience reactions. Journal of Marketing Communications, 18(1), 5–18. https://doi.org/10.1080/13527266.2011.620765 van Reijmersdal, E. A., Fransen, M. L., van Noort, G., Opree, S. J., Vandeberg, L., Reusch, S., … Boerman, S. C. (2016). Effects of Disclosing Sponsored Content in Blogs. American Behavioral Scientist, 60(12), 1458–1474. https://doi.org/10.1177/0002764216660141 van Reijmersdal, E. A., Lammers, N., Rozendaal, E., & Buijzen, M. (2015). Disclosing the persuasive nature of advergames: moderation effects of mood on brand responses via persuasion knowledge. International Journal of Advertising, 34(1), 70–84. https://doi.org/10.1080/02650487.2014.993795 van Reijmersdal, E. A., Tutaj, K., & Boerman, S. C. (2013). The effects of brand placement disclosures on skepticism and brand memory. Communications-The European Journal of Communication Research, 38(2), 127–146. Walsh, G., Gwinner, K. P., & Swanson, S. R. (2004). What makes mavens tick? Exploring the motives of market mavens’ initiation of information diffusion. Journal of Consumer Marketing, 21(2), 109–122. Wei, M.-L., Fischer, E., & Main, K. J. (2008). An Examination of the Effects of Activating Persuasion Knowledge on Consumer Response to Brands Engaging in Covert Marketing. Journal of Public Policy & Marketing, 27(1), 34–44. https://doi.org/10.1509/jppm.27.1.34 Wojdynski, B. W. (2016). Native advertising: Engagement, deception and implications for theory. In R. Brown & V. Jones (Eds.), The new advertising: Branding, content and consumer relationships in a data-driven, social media era (Vol. 1, pp. 203–236). Santa Barbara, California: Praeger. 126 Wojdynski, B. W., & Evans, N. J. (2016). Going Native: Effects of Disclosure Position and Language on the Recognition and Evaluation of Online Native Advertising. Journal of Advertising, 45(2), 157–168. Wojdynski, B. W., Evans, N. J., & Hoy, M. G. (2017). Measuring Sponsorship Transparency in the Age of Native Advertising. Journal of Consumer Affairs. Retrieved from http://onlinelibrary.wiley.com/doi/10.1111/joca.12144/full Wyer, R. S., Srull, T. K., & Gordon, S. (1984). The effects of predicting a person’s behavior on subsequent trait judgments. Journal of Experimental Social Psychology, 20(1), 29–46. https://doi.org/10.1016/0022-1031(84)90010-6 Xia, L., & Bechwati, N. N. (2008). Word of Mouse: The Role of Cognitive Personalization in Online Consumer Reviews. Journal of Interactive Advertising, 9(1), 3–13. https://doi.org/10.1080/15252019.2008.10722143 Xie, G.-X., Boush, D. M., & Liu, R. R. (2013). Tactical deception in covert selling: A persuasion knowledge perspective. Journal of Marketing Communications, 21(3), 224–240. https://doi.org/10.1080/13527266.2012.754369 Yang, H. (2013). Market Mavens in Social Media: Examining Young Chinese Consumers’ Viral Marketing Attitude, eWOM Motive, and Behavior. Journal of Asia-Pacific Business, 14(2), 154–178. https://doi.org/10.1080/10599231.2013.756337 Yoo, C. Y. (2009). The Effects of Persuasion Knowledge on Click-Through of Keyword Search Ads: Moderating Role of Search Task and Perceived Fairness. Journalism and Mass Communication Quarterly, 86(2), 401–418. Yoon, Y., Gürhan-Canli, Z., & Schwarz, N. (2006). The effect of corporate social responsibility (CSR) activities on companies with bad reputations. Journal of Consumer Psychology, 16(4), 377–390. Zaichkowsky, J. L. (1994). The personal involvement inventory: Reduction, revision, and application to advertising. Journal of Advertising, 23(4), 59. 127