• No results found

How do digital influencers differ from regular consumers in regard to their credibility toward their online product reviews? Is their credibility affected by the role of product-source congruency?

N/A
N/A
Protected

Academic year: 2021

Share "How do digital influencers differ from regular consumers in regard to their credibility toward their online product reviews? Is their credibility affected by the role of product-source congruency?"

Copied!
104
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

1

How do digital influencers differ from regular consumers in regard to

their credibility toward their online product reviews? Is their credibility

affected by the role of product-source congruency?

by

Diamantoula Papadopoulou

(2)

2

How do digital influencers differ from regular consumers in regard to

their credibility toward their online product reviews? Is their credibility

affected by the role of product-source congruency?

by

Diamantoula Papadopoulou

University of Groningen Faculty of Economics and Business

MSc Marketing Management June 2018 Berlageweg 65 9731 LK Groningen 00306982021652 d.papadopoulou@student.rug.nl Student number S3488098

First supervisor: Dr. J. A. (Liane) Voerman Second supervisor: J.A (Jan) Koch, MSc

(3)

3

ACKNOWLEDGEMENT

The current Master thesis marks the end of a challenging academic year. During this journey a lot of people stood by my side and supported me unconditionally.

Firstly and more importantly, I would like to thank my supervisor Dr. J.A. (Liane) Voerman for her continuous support, patience and guidance during the last four months of this journey. Her motivation and valuable contributions were extremely beneficial for the completion of this project.

In addition, I would like to express my gratitude to all my friends who helped me collect the required data, as well as all the peer fellows who participated in the survey.

Moreover, I would like to thank my two best friends, Eleni and Kiki, who continuously supported and encouraged me throughout the Master thesis period. Last but not least, I would like to acknowledge the unconditional support, love and inspiration I received from my parents, Kiki and Michalis.

(4)

4

EXECUTIVE SUMMARY

In the last decade, the rise of the Internet has substantially affected the lives of consumers in different ways. The Internet, along with the tremendous popularity of social media has revolutionized the way people communicate, interact and receive information on a global scale. This digital revolution has transferred the conventional of-mouth into a new online form of communication, known as electronic word-of-mouth communication.

This new form of communication among the consumers facilitates a generation of online content by fellow individuals. In the current digital era, consumers share their opinions and experiences about products and services with fellow individuals on different online platforms. These online product reviews are perceived as trusted sources of information that aid consumers’ purchase decisions.

The relatively new phenomenon of digital influencers has also dominated the field of eWOM and online product reviews. Digital influencers have arisen in different online platforms, by sharing their product evaluations based on their experiences with them. They are considered as decision makers and the center of online conversations regarding product evaluations.

However, there is no existing literature regarding the differential effect of these different sources of online product reviews toward their perceived credibility. Therefore, the current research focuses on identifying how digital influencers differ from regular consumers regarding their perceived credibility. It also investigates the effect of product-source congruency on digital influencers’ perceived credibility.

The findings depicted that digital influencers are partially considered as more expert sources of online product reviews than regular consumers. However, regarding the perceived trustworthiness no differential effect was found. Both the digital influencers and regular consumers were equally evaluated. On the other hand, a significant effect of the role of product-source congruence on both source expertness and trustworthiness was identified.

(5)

5

TABLE OF CONTENTS

1. INTRODUCTION ... 7

1.1 EWOM- User-Generated content ... 8

1.2 Online reviews by fellow consumers ... 8

1.3 Digital Influencers ... 9 1.3.1 Instagram as a platform………..………..……..10 1.3.2 YouTube as a platform…………..……….11 1.4 Product-source congruency ... 11 1.4 Source Credibility ... 12 1.5 Problem statement ... 13

1.6 Structure of the research ... 14

2. THEORETICAL FRAMEWORK ... 15

2.1 Conceptual framework ... 15

2.2 Type of source and expertness ... 16

2.3 Type of source and trustworthiness ... 17

2.4 YouTube versus Instagram ... 19

2.5 Source Involvement/Attitude toward the source ... 20

2.6 The role of product-source congruency on Digital influencers credibility ... 21

2.7 Covariates ... 22

2.7.1 Consumer scepticism………..22

2.7.2 Product involvement………..……….23

3. RESEARCH DESIGN ... 24

3.1 Type of research design ... 24

3.3.1 Participants & Design………...24

3.1.2 Stimulus………..26

3.1.3 Sampling design………..26

3.1.4 Procedure………27

3.2 Data Collection-Factor analysis ... 27

3.2.1 Perceived source credibility………..………..29

3.2.2 Source involvement………30

3.2.3 Product-source congruency………30

3.2.4 Covariates………...30

3.2.5 Control variables……….31

3.3 Manipulation check results ... 31

3.3.1 Additional survey………32

3.4 Plan of the analysis... 33

3.4.1 Study 1………33

(6)

6 3.4.1.2 Linear Regression……….………..34 3.4.2 Study 2………....35 3.4.2.1 Two-way ANOVA……….……….35 3.4.2.2 Linear regression………35 4. RESULTS ... 37 4.1 Study 1 ... 37 4.1.1 One-way ANOVA………..37 4.1.2 Linear regression………39 4.2 Study 2 ... 42 4.2.1 Two-way ANOVA………..42 4.2.2 Linear regression………44 4.3 Multicollinearity results ... 47 5. DISCUSSION ... 49 5.1 Conclusion ... 49 5.2 Limitations ... 51

5.3 Managerial Implications & Further research ... 53

6. REFERENCES ... 55

7.APPENDICES ... 67

Appendix 1: Survey ... 67

APPENDIX 2: Additional survey for manipulation check ... 73

APPENDIX 3: Manipulation check results-ANOVA & POST-HOC TEST ... 75

APPENDIX 4: Repeated-Measures ANOVA for additional survey ... 76

APPENDIX 5: ANOVA & POST-HOC STUDY 1: EXPERTNESS ... 78

APPENDIX 6: ANOVA STUDY 1: TRUSTWORTHINESS ... 79

APPENDIX 7: REGRESSION ANALYSIS STUDY 1: EXPERTNESS ... 80

APPENDIX 8: REGRESSION ANALYSIS STUDY 1: TRUSTWORTHINESS ... 82

APPENDIX 9: ANOVA STUDY 2: EXPERTNES ... 85

APPENDIX 10: ANOVA STUDY 2: TRUSTWORTHINESS ... 88

APPENDIX 11: REGRESSION ANALYSIS STUDY 1: EXPERTNESS ... 91

APPENDIX 12: REGRESSION ANALYSIS STUDY 2: TRUSTWORTHINESS ... 95

APPENDIX 13 : MULTICOLLINEARITY CHECK STUDY 1 ... 99

(7)

7

1. INTRODUCTION

Over the last decade, the Internet’s growing popularity has significantly changed the sources of information delivery from the offline environment to a new dominant online ecosystem (Zhu & Zhang, 2010). Social media as interactive Internet-based applications have grown rapidly in importance and migrated into the ‘’mainstream’’, by introducing substantial changes to communication and collaboration among people (Kietzmann, Hermkens, McCarthy & Silvestre, 2011). In this digital era, there is considerable diversity across the ways people tend to express themselves and communicate with others through social networking sites and content communities (Kaplan & Haenlein, 2010).

Along with the rise of Social Media, the Social Media Influencer Marketing has also emerged. Nowadays, more people have immersed themselves in the creation of their own online viewing content. In the last five years, new media personalities have evolved across different social media platforms and they have established a direct interaction. This contributes to an ideal environment for creating online communities with strong associations among their members (Havan, Single Grain, 2017). These media personalities have emerged in the online environment as everyday people, who have built a reputation for their knowledge about some specialist niche and their relationship with their audience (Havan, Single Grain, 2017). Over the years, these so called digital influencers have dominated the digital world and have grown from obscurity to a form of marketing (Wart, Forbes, 2017).

(8)

8

1.1 EWOM- User-Generated content

According to Babic et al. (2016) word of mouth (WOM) facilitates information diffusion about goods, services and brands among consumers. Richings and Root-Shaffer (1988) have stated that the traditional (offline) word of mouth plays a significant role for consumers’ buying decisions. However, when such information is disseminated through the Internet (e.g. reviews, tweets, blog posts, video testimonials) is defined as ‘’electronic word of mouth’’ (eWOM). User-generated content is the way this information is created and how consumers express themselves and engage with other fellow consumers in the online environment (Boyd & Ellison, 2008). Consumers generate online content in the moment of being social across different online platforms (Smith, Fischer & Yongjian, 2012). Bickart and Schindler (2001) have indicated that consumers are more interested in product information when it is generated by peer users on the Internet rather than when this information comes from corporations. In the recent years, consumers have the opportunity to gather various evaluations and opinions from other consumers as well as to provide their own consumption related advice (Hennig, Gwinner, Walsh & Gremler, 2004). Therefore, user-generated content differs across various media from tweets on Twitter and Facebook status updates to videos on YouTube as well as consumer-produced product reviews (Dhar & Chang, 2009; Muniz & Schau,2007).

1.2 Online reviews by fellow consumers

(9)

9 process (Park & Park, 2008). Hence, this significant volume of data is considered as very attractive and challenging for both corporations and consumers (Singh, Irani, Rana, Dwivedi, Saumya & Roy, 2017).

Over the last decade, many firms effectively have taken advantage of users’ online reviews as a new marketing tool to raise brand awareness and build loyalty among consumers (Dellarocas, 2003). Therefore, many firms facilitate both user-generated content and eWOM by enabling users to post and share their product evaluations followed by their personal experiences, either on their corporate websites or on intermediary review sites (Chen & Xie 2008). According to Zhu & Zhang (2010), corporations not only allow online engagement among consumers toward product evaluations, but they also proactively induce consumers to disseminate the word about their products on the Internet. According to the above, online consumer reviews act as a critical decision variable for consumers and as an important factor of product sales (Chatterjee, 2001).

1.3 Digital Influencers

(10)

10 the new market trends, indicate a certain level of trust and maintain a large social network. These online personalities act as honest and authentic sources of product evaluations they provide to their following. This growing trend is expanded into various markets such as fashion, beauty, technology and entertainment (Geppert, Convince and Convert, 2016). That is why consumers first turn to their favorite digital influencers, to the people whom they trust for receiving a recommendation when planning potential product purchases.

Many corporations understand their value and contribution to brand awareness and tend to take advantage of this tremendous growth by establishing strong long-term relationships with digital influencers (Ward, Forbes, 2017). The act of involving prominent personalities in a particular market niche to disseminate content as well as implementing marketing activities around them is called Influencer Marketing (ADWEEK,2016).

The existing diversity across the plethora of social media platforms provides the opportunity to these to be connected with millions of followers every moment (MediaKix, 2017). In this context, research has indicated that visual content can reach more social engagement than any other type of content (ADWEEK, 2016). Nowadays, Instagram and YouTube enable influencers to share their product experiences through photos and videos and are considered as the most popular influencing platforms among consumers (ADWEEK, 2016). Hence, this research is focused on both influencers on YouTube and Instagram.

1.3.1 Instagram as a platform

(11)

11 consumers and contributed to a significant growth. Therefore, it has also become one of the largest online platforms of the growing influencer trend (Mediakix, 2016)

1.3.2 YouTube as a platform

YouTube has become one of the new forms of social networks’ online community since its establishment in 2005 (Haridakis & Hanson, 2009) and it is considered as the third most visited content community worldwide (Alexa, 2014). Since YouTube resides in the Internet, it encourages consumers to actively engage in an online social environment. As stated in its slogan ‘’Broadcast yourself’’, everyone can enjoy the experience of receiving, publishing and modifying Web content (Pisani 2006). Therefore, it enables users to move harmoniously between the traditional online activity of being exposed to mediated content and the interpersonal activity of being the purveyor of the content (Holtz, 2006; Haridakis & Hanson, 2009).

Although the number of YouTube’s videos that consumers are more exposed to tend to be professionally produced (Kruitbosch & Nack, 2008), the most engaging and commented-on videos tend to be generated by fellow consumers, known as Youtubers (Burgess & Green, 2009). Regarding the content of the videos Burgess and Green (2009) have indicated that the user-generated videos include vlogs, entertainment videos, informational content and scripted performances. More specifically they often feature reviews, demonstrations, and ‘unboxing’ of new products (Blythe & Cairns, 2009). All in all, YouTube has started to gain the upper hand in establishing a massive presence of the digital influencers trend, something corporations consider as a new beneficial marketing opportunity.

1.4 Product-source congruency

(12)

12 attribute between the information source and the product, the former is considered as an effective and knowledgeable source regarding the product (Kamins, 1990). This provides the privilege to the spokesperson to deliver the information quickly, giving the opportunity to the consumers to understand the significant attributes of the product before moving to alternative sources of information (Lynch & Schuler, 1994).

According to Lynch and Schuler (1994), some sort of product-source congruency will result in high levels of information delivery effectiveness and to the enhancement of a favorable attitude toward the source. This can contribute to a potential increase in the perceived source credibility (Lynch & Schuler, 1994).

1.5 Source Credibility

(13)

13 the source credibility of an online product review indicates the degree to which the reviewer is considered as a credible source for product related information and his ability to provide an objective opinion regarding the product. Hence, when the source is considered as highly credible, recipients perceive the information provided as a significant support for the related issue and they are confident in their thoughts about the message (Clark, Evans & Wegener, 2011). On the other hand, when the source is presented as a low in credibility, it is expected to provide deceptive and unreliable information (Clark, Evans & Wegener, 2011). Therefore, consumers are expected to elicit a lack of trust and to generate doubts about the information they acquire (Clark, Evans & Wegener, 2011).

1.6 Problem statement

Several recent studies have attempted to examine the performance of online reviews by regular consumers and their subsequent effect on product sales according to their credibility. In addition, regarding digital influencers, previous research focuses on the identification of influential consumers, now considered as opinion leaders, due to their perceived knowledge and influential power (Uzunoglu & Kip, 2014). Over the years, online reviews by regular consumers published on different online review platforms has been considered as a credible source of product evaluations. However, with the presence and power of digital influencers skyrocketing, consumers become more engaged with these online personas to receive authentic and reliable product reviews. Hence, it can be encountered as highly significant both for scholars and corporations to demonstrate the current differentiated level of source credibility between digital influencers and regular consumers in terms of online product reviews.

How do the digital influencers differ from regular consumers in regard to their perceived expertness and trustworthiness, and how these are affected by the product-source congruency?

(14)

14 1 focuses on identifying the difference in the perceived credibility between the digital influencers and the regular consumers, whereas Study 2 aims to investigate the role of product-source congruency on the perceived credibility of digital influencers. The findings and the implications of both studies will be provided to the managers in order to incorporate them effectively to their strategies for achieving the desired behavioral response by consumers.

1.7 Structure of the research

(15)

15

2. THEORETICAL FRAMEWORK

Academic research on eWOM and brand related user-generated content on the Internet is evolving rapidly over the last couple of years. To situate both Study 1 and Study 2 of the current research within this tremendously increasing literature, relevant studies on online consumer reviews and the influencer marketing are reviewed under consideration.

In this context, when consumers perceive a source of communication as unbiased, believable, true and factual then it is considered as credible(Hass,1981). According to Hovland, Janis and Kelly (1953) credibility consists of 2 fundamental dimensions; expertness and trustworthiness. In regard to it, it has been indicated that both highly trustworthy and expert sources of information trigger more immediate attitude change than do sources in lack of these components (Hovland, Janis & Kelly, 1953).

2.1 Conceptual framework

As it was mentioned above, the current research is structured into two separate studies which aim to address the problem statement accordingly. The following figures represent the conceptual models of Study 1 and Study 2.

(16)

16

Figure 2: Conceptual model Study 2

2.2 Type of source and expertness

(17)

17 competence which expresses the extent of their expertness and knowledge on specific areas (Katz, 1957; Uzunoglu & Kip, 2014). Based on the research by Tomaszeski (2006), digital influencers are considered as more information savvy and up-to-date with the latest developments. The increasing level of enduring, the self-perceived knowledge and the exploratory behavior of digital influencers has turned them into the go-to sources of current and advanced information for consumers (Hsu & Tsou, 2011).

Hence, according to the above, in Study 1 it is hypothesized:

H1: Digital influencers are perceived as more expert sources of online product reviews than regular consumers.

2.3 Type of source and trustworthiness

(18)

18 In this context, it is notable that consumers are willing to respond to online communication derived from fellow peers and to be engaged with it by potentially disseminating it with others. According to Choi and Scott (2013), this generated social capital by consumers is found among high in frequency of communication online networks (Cabezudo & Izquierdo, 2012). The uncertain reduction theory claims that consumers feel uncertain and uncomfortable in interpersonal relationships, thus they strive to reduce it through the interpersonal communication (Berger & Calabrese,1975). The latter corresponds to the formation of online communities with common shared interests and strong trustworthy associations among the members (Liu et al.., 2015). Consumers are more inclined to trust and interact with peers who seem identical to themselves (Lee & Watkins,2016). Therefore, the perceived level of homophily between the information source and the recipient has an impact on the online para-social interaction and the perceived trustworthiness (Lee & Watkins, 2016). Lazarsfeld and Merton (1954) distinguish the external homophily from the internal one. External homophily covers the fields of age, gender and social classes whereas the internal homophily refers to similar beliefs, values and lifestyle among individuals. Prisbell and Andersen (1980) claim that perceived homophily is a factor which contributes to risk reduction, to a creation of a good atmosphere and to the reinforcement of the safety of the interpersonal relationship (Prisbell & Andersen,1980). Para social interaction depends on the degree of homophily between the source and the recipient, thus it is generated between current media personalities and media users (Frederick, Lim, Clavio & Walsh, 2012). Nowadays, consumers by exploring user-generated media are exposed to customized content which is created by their peers with similar interests (Choi & Morawitz, 2017). They seek guidance and advice from digital influencers because of their perceived similarities and reciprocal relationship (Labrecque, 2014). Digital influencers offer many opportunities to their audience to receive information about their personal daily life and personalities (Choi & Morawitz, 2017). As this relationship between the digital influencers and their audience continues to enhance the users’ trust toward them will be increased by continuing being exposed to them, seeking for their valuable advice (Rubin, Perse & Powell, 1985; Shan, 2016)

(19)

19 the source and the recipient acts as a heuristic cue which leads the consumer to understand whether the product or service complies with their interests, values and needs (Shan, 2016). Similarity also enhances the degree to which subjects seem familiar and likeable (Labrecque, 2014). The mere exposure to digital influencers reinforces their perceived similarity as well as their likeableness. Consequently, consumers appear more or less likely to identify the source with deception or persuasive intention (Shan, 2016).

In line with the above, for Study 1 the following hypothesis is formulated:

H2: Digital influencers are perceived as more trustworthy sources of online product reviews than regular consumers.

2.4 YouTube versus Instagram

People, as active users of the online world, tend to rely on media which makes them goal oriented and motivated to fulfill their goals and satisfy their needs (Choi & Behm-Morawitz, 2017). Users are actively looking for user-generated media which can meet their cognitive needs for information and entertainment and social interaction. (Katz, Gurevitch & Hass, 1973). By the same token, YouTube was the first social network that facilitated the emergence of social interactive communities between content creators and users (Smith, Fischer, Yongjian, 2012). Creators engage in a meaning making process by communicating their life experiences to their audience (Burgess & Green, 2009). YouTube provides the opportunity to content creators to reveal their identity and their daily life to their audience by generating an interactive relationship (Guo & Lee, 2013). YouTube’s community structure, culture and norms, are considered to lead to a higher intention to share and interact with the videos’ content (Gobel, Meyer, Ramaseshan & Bartsch, 2017).

(20)

20 message information to be easily shared in a richer and longer format (Li & Suh, 2015) in comparison to Instagram.

All in all, in accordance with the above arguments, the following hypotheses are formed and examined in Study 1:

H3: Digital Influencers on YouTube are perceived as more expert than those on Instagram toward their online product reviews.

H4: Digital Influencers on YouTube are perceived as more trustworthy than those on Instagram toward their online product reviews.

2.5 Source Involvement/Attitude toward the source

In today’s world, consumer’s experience is strongly correlated with the notion of engagement (Mollen and Wilson, 2010). According to previous research, the concept of involvement with a topic is considered as one over the various notions that conceptualize consumer engagement (Uzunoglu & Kip, 2014). In the context of product endorsements, the level of involvement influences the way that consumers assess endorsers’ characteristics (Munnukka, Uusitalo, Toivonen, 2016).Therefore, consumers may evaluate endorsers differently under low and high involvement situations (Chaiken, 1980; Petty & Goldman, 1981). The Elaboration Likelihood Model indicates that there are two different routes that consumers follow when processing information, the central or the peripheral route (Petty & Cacioppo, 1986). Consumers are highly involved with a topic when the personal importance and interest in the topic increases (Petty & Cacioppo, 1986). Hence, they prefer the central route of processing information and they are more motivated to activate all their cognitive resources to process the delivered information (Petty & Cacioppo, 1986). Consequently, the degree of involvement is perceived as an important factor of determining the trustworthiness and the expertness of the source, because of consumers’ high involvement which may lead to a higher willingness to assess source trustworthiness and expertness (Munnukka, Uusitalo, Toivonen, 2016).

This leads to the following hypotheses examined in both Study 1 and Study 2: H5: A high level of consumers’ source involvement has a positive effect on the

perceived expertness of the source.

(21)

21

2.6 The role of product-source congruency on Digital influencers

credibility

Kamins (1990) claims that a product endorsement is more effective when there is a fit between the endorser and the endorsed product. The associative learning can contribute to the establishment of potential links or relationships between two or more concepts (Klein, 1991; Martindale, 1991). Furthermore, in the context of endorsements, the experiences people create both with the product and the endorser may lead to the establishment of specific connections (Till & Busler, 2000). These connections constitute people’s the association set for the product and the endorser (Till & Busler, 2000). More generally, Kamins & Gupta (1994) stated that the matching link between the spokesperson with the product derives from a broader degree of congruency between the former and the latter.

According to the above, the degree of congruency between a product and the information source has a significant effect on consumers’ attitude and perceived credibility toward the source (Kamins, 1990; Kamins & Gupta, 1994). According to McCracken (1989), product-source congruency also consists of cultural and symbolic meanings associated with the information source which are transferred onto the products and then consequently to their audience (Ilicic & Webster,2011). Kirmani and Shiv (1998) indicated that the high extent of product-source congruency can be perceived as a strength of source’s persuasiveness and thus credibility. Previous research by Kamins & Gupta (1994) has indicated that the most important dimension which leads to congruency is the perceived expertness of the source. Prior studies have stated that sources high in expertness tend to generate a more favorable attitude toward the issues advocated (Kamins & Gupta, 1994) as well as to a better fit between the spokesperson and the product (Kamins & Gupta, 1994).

(22)

22 As mentioned above, YouTube as a channel holds the privilege of providing video content in a broader format than Instagram. The length of Youtuber’s video content and the argumentation on it generate a more favorable attitude toward them (Cacioppo, Petty & Morris, 1983). The elaborated video content they provide to their audience contributes significantly to their perceived expertness and trustworthiness.

According to the above, it is hypothesized examined in Study 2:

H7: An inclusion of low product-source congruency products into digital influencers’ product portfolio has a negative effect on the perceived expertness by

consumers.

H8: An inclusion of low product-source congruency products into digital influencers’ product portfolio has a negative effect on the perceived

trustworthiness by consumers.

H9: An inclusion of low product-source congruency products into Youtubers’ product portfolio has a less negative effect on the perceived expertness than this of

Instagramers.

H10: An inclusion of low product-source congruency products into Youtubers’ product portfolio has a less negative effect on the perceived trustworthiness than

this of Instagramers.

2.7 Covariates

Two other variables that might affect the perceived credibility are also investigated in the current research. The degree of scepticism toward online product reviews and the level of the involvement with the product both in Study 1 and Study 2. Hence, their impact is also examined.

2.7.1 Consumer scepticism toward online product reviews

(23)

23 also become sceptical against them (Sher & Lee, 2009). Consumers, based on their prior experiences are characterized by a different level of scepticism toward online recommendations (Friestad & Wright, 1994,). In other words, highly sceptical consumers tend to respond more negatively to online product reviews as they are considered to generate stereotypes for certain categories of product information (online product reviews) (Sher & Lee, 2009). Hence, in line with Reimer, consumers with a higher level of scepticism holds a significantly higher level of suspicion toward the delivered information. In consequence, the observed level of a lower initial trust leads to a decrease in the trustworthiness of the online product reviews (Reimer & Benkenstein, 2016) as well as the expertness, which directly influence the perceived credibility of the information source.

H11: If consumers’ scepticism toward online product reviews goes up the perceived expertness of the source goes down.

H12: If consumers’ scepticism toward online product reviews goes up the perceived trustworthiness of the source goes down.

2.7.2 Product involvement

According to Traylor (1981), product involvement is defined as the degree up to which a consumer understands and recognizes a specific product. The higher the degree of the understanding and the recognition, the higher the level of the involvement with the product (Zaichkowsky, 1985). In line with the above, Engel (1995) defined the involvement with a product as the condition where the consumer showcases a personal interest in a specific product. Cacioppo, Petty & Morris, (1983) adopt the Elaboration Likelihood Model, which indicates that individuals who are highly motivated to process the delivered message as well as those who are able to, are the ones who will process it more effortfully and thoughtfully. In line with the findings of Neese and Taylor (1994), a high level of involvement with a product may have a higher positive effect on the perceived source credibility.

According to the above, the following hypotheses for Study 1 and Study 2 are formulated:

H13: A high product involvement level will have a positive influence on consumers’ evaluation of the perceived source expertness.

(24)

24

3. RESEARCH DESIGN

According to the theories and findings reviewed, the first study of this research focuses on measuring the perceived credibility of digital influencers versus regular consumers toward their online product reviews. The following study aims to examine the effect of the product-source congruency on the perceived credibility of digital influencers. In this section, the methodology for data collection, the measurement tools and the analysis plan will be described.

3.1 Type of research design

In order to address the above mentioned problem statement, an online survey has been conducted as a conclusive research in a descriptive research design, subjected to a quantitative analysis (Malhotra, 2010). The survey was undertaken to determine the factors underlying the perceived credibility of digital influencers and regular consumers toward their online product reviews. Respondents were asked to express their degree of agreement on a variety of questions on a 7-point Likert scales (1=strongly agree, 7=strongly disagree). Likert scales were used due to their advantage of being easily understandable by respondents (Malhotra, 2010).

The survey was administered in English and distributed worldwide. The link of the survey was distributed to the participants via social network sites; Facebook posts and messages as well as Instagram stories.

3.1.1 Participants and Design

(25)

25 influencers in order to test their effect on influencers’ perceived trustworthiness and expertness.

Table 1 provides a summary of which stimulus was assigned to which condition.

Product Consistent/Foundation Less

consistent/Running shoes

Inconsistent/Printer

Type of source Youtuber Condition 1 Condition 2 Condition 3 Instagramer Condition 4 Condition 5 Condition 6

Regular consumer

Condition 7

Table 1: Research Design

The following conditions are used for Study 1

Condition 1: A description of the influencer on YouTube, followed by the foundation review.

Condition 4: A description of the influencer on Instagram, followed by the foundation review.

Condition 7: A description of the regular consumer, followed by the foundation review.

The following conditions are used for Study 2

Condition 1: A description of the influencer on YouTube, followed by the foundation review.

Condition 2: A description of the influencer on YouTube, followed by the running shoes review.

Condition 3: A description of the influencer on YouTube, followed by the printer review.

Condition 4: A description of the influencer on Instagram, followed by the foundation review.

Condition 5: A description of the influencer on Instagram, followed by the running shoes review.

(26)

26 3.1.2 Stimulus

The current research focuses on the beauty industry, which seems to be continuously developing with the rise of digital influencers. According to a survey conducted by PR WEEK (2018), 98 per cent of the respondents agreed that the beauty industry works more effectively with digital influencers in comparison to other industries. Therefore, as the target digital influencer, a beauty influencer both on YouTube and Instagram was employed. As the target product, a foundation was selected because it is perceived as one of the most important and essential beauty product women use and look for online reviews. Furthermore, in order to draw the participants’ attention to the product-source congruence, a small survey among fellow peers was conducted. In this survey, a pair of running shoes and a printer were used to represent the two categories, meaning one less product-source congruent item and one product-source incongruent item respectively. In addition, in order to minimize the brand familiarity effects, a fictitious brand name and a fictitious digital influencer were used. This guarantees that the fictitious name cannot be related to an existing brand, hence it cannot influence respondents’ answers. The name for all the products was Brand.

Regarding the source of the online product review, Jenny Smith was created, described either as an Instagramer, Youtuber or as a regular consumer. The fictitious review source minimizes any potential variation in respondents’ knowledge and attitude toward the familiar source (Till & Busler, 2000).

3.1.3 Sampling design

(27)

27 3.1.4 Procedure

Respondents, as mentioned above were randomly exposed to one of the seven in total conditions.

● In the introductory page, participants were informed that they would assist in the completion of the current master thesis by participating in the following survey followed by high anonymity and confidentiality.

● Following the introductory page, respondents were assigned to one of the seven conditions and they were instructed to carefully read the informational text. They were provided by relevant information regarding the characteristics of the online product review source, by being exposed to a stimulus featuring Jenny Smith either as a digital influencer (Youtuber or Instagramer) or a regular consumer. The purpose of the stimulus was to establish Jenny as representative online product review source of each category.

● The main stimulus was designed to visually pair Jenny with a picture of the reviewed product. In the conditions Jenny was featuring solely as a Youtuber or an Instagramer, she was also paired with the picture of the running shoes and the printer apart from the target product; the foundation. Each condition had a small background information text regarding the source along with her product review. The structure of the stimulus as well as the product review were identical apart from the background information.

● After their exposure to the respective stimulus, the respondents were asked to assess the perceived credibility of the review source by answering a variety of questions.

A copy of the questionnaire is included in the Appendix 1.

3.2 Data Collection-Factor analysis

(28)

28 analysis is appropriate, the Eigenvalues above 1 indicate the extracted number of factors (Malhotra, 2010).

The next step of the analysis aimed to examine the internal consistency of each set of items forming particular scales. Cronbach’s alpha was computed in order to measure this internal consistency. A value of Cronbach’s alpha above 0.6 indicates a satisfactory internal consistency (Malhotra, 2010). In the current study, the value of Cronbach’s alpha was above 0.6 for each set of items. In Table 2 all the constructs, related items, used scales as well as the Eigenvalues and the Cronbach’s Alpha are presented.

Variable Source Items Measureme

nt

EV Cronbach

’s Alpha Trustworthiness Ohanian

(1990)

I perceive Jenny as a reliable source of information

I believe Jenny is a trustworthy source of information

I would like to have a friendly chat with Jenny

Jenny makes me feel comfortable

● I perceive Jenny as a dishonest source of information 7-point Likert scale→ From 1=Strongly agree to 7= Strongly disagree. 2.575 0.814 Expertness Ohanian (1990) ● I perceive Jenny as an experience source of (reviewed

category) products I think Jenny is knowledgeable

about (reviewed category) products

● Jenny is an expert for (reviewed category) products ● I think Jenny doesn’t have much

experience of (reviewed category) products 7-point Likert scale→ From 1=Strongly agree to 7= Strongly disagree. 3.134 0.907 Involvement/Atti tude toward the sources of online

product reviews

Lee, Park & Han (2008)

● When I think of buying a product I do read online reviews

by (source)

● I rely on (source’s) reviews when considering a new product ● Online reviews by (source) are

(29)

29

Product-source congruency

Rifon(20 06)

I think Jenny fits well with the product

I perceive Jenny as a believable source of information

Jenny doesn’t match up with the product

● Jenny is an appropriate source of information for the product

● There is a congruency between

Jenny and the product

7-point Likert scale→ From 1=Strongly agree to 7= Strongly disagree. 3,008 0.889 Consumer scepticism toward online product reviews Skarmeas & Leonidou (2013)

● I am generally uncertain about online reviews

● I am not doubtful about online reviews

● I am generally sceptical about online reviews 7-point Likert scale→ From 1=Strongly agree to 7= Strongly disagree. 1.545 0.660 Product involvement Chandras ekaran (2004) ● I am particularly interested in the (reviewed product) ● (The reviewed product) is not

relevant to me

● I am quite involved with (the reviewed product) 7-point Likert scale→ From 1=Strongly agree to 7= Strongly disagree. 2.077 0.777

Table 2: Operationalization table

*Excluded items from FA

3.2.1 Perceived source credibility

(30)

30 In addition to it, the source expertness was measured by asking four questions in relation to its perceive experience and knowledge toward the reviewed product. All the items were correlated to each other, having an Eigenvalue of 3.134. Following the factor analysis, the reliability analysis indicated a high value of the Cronbach’s Alpha at 0.907.

3.2.2 Source involvement

Three different items were used in order to measure respondent’s involvement with the depicted source. Respondents were also asked to determine their involvement on a 7-point Likert scale developed by Lee, Park & Han (2008). According to factor analysis the Eigenvalue of the source involvement factor is 2.369 followed by a Cronbach’s Alpha of 0.864.

3.2.3 Product-source congruency

Five different questions regarding the congruency between the information source and the reviewed product were stated. The main aim of these questions was to identify the effect of the product-source congruency on digital influencers’ perceived credibility. The respondents had to rank these statements on a 7-point Likert scale from agree’’ to ‘’strongly disagree’’. The applied scale was based on Rifon (2006). Based on the factor analysis the item ‘’ There is a congruency between Jenny and the product’’ was excluded due to its communality rate of 0.095. After the exclusion of the item the Eigenvalue of the extracted factor is 3.008 followed by a Cronbach’s Alpha of 0.889. 3.2.4 Covariates

Next, the respondents were asked to answer three statements regarding both their degree of scepticism toward online product reviews as well as the level of their involvement with the product. The scale for consumers’ scepticism was based on Skarmeas & Leonidou (2013). All the items lead to one factor with an Eigenvalue of 1.545 and a Cronbach’s Alpha of 0.660. These low values can be attributed to a low communality, just above the threshold of 0.4, of the second item of the scale.

(31)

31 Chandrasekaran (2004). All the items lead to one factor with an Eigenvalue of 2.077 and a quite high Cronbach’s Alpha of 0.777 as depicted in Table 2.

3.2.5 Control variables

At the end of the survey, respondents were asked to provide some demographic information like their gender and age. The collection of the gender was a filter question as only females could participate in the survey.

3.3 Manipulation check results

Following the factor and the reliability analysis, the results of the manipulation check were examined. For the purpose of Study 2, the manipulation check questions were only included in the conditions where the type of source was either a Youtuber or an Instagramer. The participants were also asked to answer questions regarding the perceived congruency between the presented source and the featured product in the condition they were assigned to. The aim of the questions was to check whether the respondents did perceive a significant difference between the product-source congruency of the foundation, the running shoes and the printer. The first step for examining the effectiveness of the manipulation check was an ANOVA in order to compare the three different product types and identify if there any significant difference among them.

According to the ANOVA test, there is a significant difference somewhere among the three types of products as depicted in Table 3.

Source F Sig. Mean

Square

Product 18.074 0.000 6.777

Note:*** p-value<0.01; **p-value<0.05; *p-value<0.10 Table 3: Anova test for manipulation check results

(32)

32 that the participants perceived the printer and the running shoes the same whereas they identified a difference between the foundation and the other two. For a detailed view, see Appendix 3.

Product Mean Std. Error

Foundation 3.758 0.078

Running shoes 4.307 0.084

Printer 4.357 0.074

Table 4: Descriptives for product type

A possible explanation for this is the fact that the participants were only assigned to one condition presenting only one out of the three products. They did not have the opportunity to be exposed to all of them in order to evaluate the difference between the printer and the running shoes (as expected).

3.3.1 Additional survey for the manipulation check

As mentioned above the manipulation check was not successful as the respondents did not interpret any significant difference between the running shoes and the printer in terms of their product-source congruency. This may have happened due to the unilateral exposure to only one featured product. For this reason, an additional survey was conducted in order to determine how respondents react when they are exposed to not only one product but to all three. In total, 30 female participants between the age of 18-35 were asked to rank the three products based on their fit with the source on a Likert scale from 1 (Strongly agree) to 7 (Strongly disagree).

(33)

33

Source F Sig. Mean Square

Product Sphericity Assumed

513.907*** 0.000 203.200

Note:*** p-value<0.01; **p-value<0.05; *p-value<0.10 Table 5: Within-Subjects effects

Product N Minimum Maximum Mean Std.Deviati

on Variance Foundation 30 1 2 1.27 0.450 0.202 Running shoes 30 3 6 3.67 0.711 0.506 Printer 30 5 7 6.47 0.571 0.326

Table 6: Additional survey for manipulation

3.4 Plan of the analysis

In this section, the plan of analyzing the data is presented. Firstly, an ANOVA was performed for both studies, followed by a regression analysis using dummy variables. The section describes the steps for the above mentioned analyses as well as the difficulties derived from them.

3.4.1 Study 1

3.4.1.1 One-Way ANOVA of a possible difference between the sources on both expertness and trustworthiness.

In order to examine the differences in the mean values of both expertness and trustworthiness, for the three types of online product review sources, a one-way analysis of variance was conducted.

(34)

34

3.4.1.2 Linear Regression

The next step after the analysis of variance was to identify the nature and the degree of association between the variables by conducting a regression analysis (Malhotra,2010). In the regression analysis, two dummy variables were computed for the type of online product reviews source. These dummy variables were used in three different models of the regression analysis following below.

Dummy variables

In the current study, the independent variable is the type of online product reviews source, which has three different levels; the Instagramer, the Youtuber and the regular consumer. Due to the difficulty of testing the hypothesis using a three-level variable, two dummy variables were created for the ease of the analysis. Since the main purpose of this study is to examine the difference between the digital influencers and the regular consumer toward both their perceived expertness and trustworthiness, the regular consumer was set as the reference group of the two dummy variables. The first dummy variable(Dummy1_Insta) is 1 when the participants were exposed to a condition depicting Jenny Smith as an Instagramer and 0 to the conditions presenting Jenny both as a Youtuber and regular consumer. The second dummy variable (Dummy2_Youtube) is 1 when the respondents saw Jenny Smith as a Youtuber and 0 when she was depicted both as an Instagramer and a regular consumer.

The following models will be examined with both the perceived expertness and trustworthiness as a dependent variable.

DV: Expertness/Trustworthiness Model 1:

DV=βο +β1*Dummy1_Insta +β2*Dummy2_Youtube +ε Model 2:

DV=βο+β1*Dummy1_Insta +β2*Dummy2_Youtube + β3*Sourceinvo + β4*Sourceinvo*Dummy1_Insta +β5*Sourceinvo*Dummy2_Youtube +ε Model 3:

(35)

35 3.4.2 Study 2

3.4.2.1 Two-Way ANOVA of a possible difference between the sources and the different products on both expertness and trustworthiness.

The plan of the analysis for Study 2 follows the one of Study 1. In the beginning, in order to examine the differences in the mean values of both expertness and trustworthiness, for the two types of online product review sources as well as for the three different products, a two-way analysis of variance was conducted.

3.4.2.2 Linear Regression

In the regression analysis of Study 2, three dummy variables were computed. The first dummy variable indicates the two different online product review sources and the other two the three different types of products. For the purpose of the current study, four different models were estimated for the analysis.

Dummy variables

In comparison to Study 1, in Study 2 there are two independent variables. The first one is the type of online product reviews source with two levels; the Instagramer and the Youtuber. The second one is the type of products which represents the product fit and the consistency with the source (Foundation; Consistent, Running shoes; Less consistent, Printer; Inconsistent). Regarding the creation of the first dummy variable the Youtuber was selected as the reference group. The first dummy variable (Dummy1_YoutubeS) is 0 when the participants were assigned to a condition presenting Jenny as a Youtuber and 1 as an Instagramer. For the second and the third dummy variables the consistent product, the foundation, was set as the reference group. The second dummy variable (Dummy1_Lessconsistent) is 1 when the respondents were exposed to the condition with the running shoes as the featured product and 0 to the conditions of both the foundation and the printer. Following the second dummy variable, the third one (Dummy2_Inconsistent) is 1 when the participants saw the printer as the featured product and 0 when they saw either the running shoes or the foundation.

The following models will be examined with both the perceived expertness and trustworthiness as a dependent variable.

(36)

36 Model 1: DV=βο +β1*Dummy1_YoutubeS +β2*Dummy1_Lessconsistent + β3*Dummy2_Inconsistent + ε Model 2: DV=βο +β1*Dummy1_YoutubeS +β2*Dummy1_Lessconsistent + β3*Dummy2_Inconsistent +β4*Dummy1_YoutubeS*Dummy1_Lessconsistent +β5*Dummy1_YoutubeS*Dummy2_Inconsistent + ε Model 3: DV=βο +β1*Dummy1_YoutubeS +β2*Dummy1_Lessconsistent + β3*Dummy2_Inconsistent +β4*Dummy1_YoutubeS*Dummy1_Lessconsistent +β5*Dummy1_YoutubeS*Dummy2_Inconsistent +β6*Sourceinvo+ β7*Dummy1_YoutubeS*Sourceinvo + ε Model 4: DV=βο +β1*Dummy1_YoutubeS +β2*Dummy1_Lessconsistent + β3*Dummy2_Inconsistent +β4*Dummy1_YoutubeS*Dummy1_Lessconsistent +β5*Dummy1_YoutubeS*Dummy2_Inconsistent +β6*Sourceinvo +

(37)

37

4. RESULTS

In this chapter, the results from the analysis of variance and the regression analysis will be explained and discussed both for Study 1 and Study 2.

4.1 Study 1

4.1.1 One-Way ANOVA of a possible difference between the sources on both expertness and trustworthiness

Expertness

The results of the analysis of variance among the Instagramer, the Youtuber and the Consumer on the perceived expertness showed that there is a difference between the three sources. The p-value is significant at the alpha level of 0.1 as seen in Table 8. As it is depicted in Table 7, there is a convergence of the means of the Youtuber and the Instagramer. According to Table 7, there is an indication that the respondents seemed to perceive both the Instagramer and the Youtuber as more expert sources than the regular consumer.

On the other hand, following the analysis of variance, the pair comparisons did not indicate any significant difference between the three sources. However, the pair comparison of the Youtuber with the regular consumer almost reached the significant level of 0.1 with a p-value of 0.133. Therefore, this raises concerns of accepting hypotheses 1 and 3. For a detailed view, see Appendix 5.

Trustworthiness

(38)

38

Type of source N Mean

Expertness Std. Deviation Mean Trustworthines s Std. Deviation Instagramer 29 3.6379 0.56135 3.7672 0.58984 Youtuber 32 3.6016 0.54573 3.8516 0.58150 Regular consumer 31 3.9032 0.65408 3.7661 0.52415

Table 7: Mean per condition

Df Sum of squares Mean square F Sig. Expertnes s Sum of squares Mean square F Sig. Trustworthiness Between Groups 2 1.682 0.841 2.42 4 0.094 0.150 0.075 0.235 0.791 Within Groups 89 30.890 0.347 28.466 0.320 Total 91 32.573 38.616 Table 8: ANOVA

(39)

39 4.1.2 Regression analysis Study 1

In the following Table 9, the results of the regression analysis for Study 1 are depicted.

Expertness/ Trustworthi

ness

Model 1 Model 2 Model 3

Main effects Dummy1_I nsta -0.265* 0.001 -0.338 -0.102 -0.359 -0.123 Dummy2_Y outube -0.302** 0.085 -0.376 -0.002 -0.382 -0.014 Source Involvemen t 0.076 0.085 0.076 0.081 Interaction effects Dummy1_I nsta*Source Invo -0.050 -0.014 -0.065 -0.024 Dummy2_Y outube*Sou rceInvo -0.020 -0.007 -0.032 -0.015 Covariates Product involvement 0.074 0.062 Consumer scepticism -0.018 -0.046 R2 0.052 0.005 0.067 0.041 0.074 0.050 Adjusted R2 0.030 -0.017 0.012 -0.014 -0.004 -0.029 F-value 2.424* 0.235 1.227 0.744 0.954 0.634

Note:*** p-value<0.01; **p-value<0.05; *p-value<0.10 Table 9: Regression results

Dependent variable: Expertness Model 1

(40)

40 Instagramer, something which is partially in line with the preliminary expectations. Table 9 shows that the Youtuber, as a review source, leads to an increase of 0.302 in the perceived expertness compared to the Instagramer with a lower increase of 0.265.

However, due to the previous contradiction mentioned above, hypothesis 1 and hypothesis 3 are partially supported.

Model 2

The aim of the analysis of the second model was identify the interaction effect of the respondents’ involvement with the source on the perceived expertness of the source. The overall model is not significant as are the p-values of all the variables. The interaction effects do not indicate any effect on the perceived expertness of the online product reviews sources. Participants did not evaluate differently the perceived source expertness depending on their level of involvement with the featured source. Hence, hypothesis 5 is not accepted.

Model 3

In the last model, the variable ‘’Product involvement’’ as well as the ‘’Consumer scepticism’’ were added. In this case, their effect on the perceived source expertness was examined. The overall model is not significant. The p-value of the Consumer scepticism as well as the Product involvement do not have any effect on the perceived source expertness, as it is depicted in Table 9. Therefore, the results suggest that the respondents did not evaluate differently the perceived expertness of the source under high or low conditions of both scepticism toward the online product reviews and involvement with the product. All in all, hypothesis 11 and hypothesis 13 are not supported.

Dependent variable: Trustworthiness Model 1

(41)

41 respondents did not interpret any difference between the three sources regarding their trustworthiness. Therefore, hypothesis 2 and hypothesis 4 are not accepted.

Model 2

Once again, model 2 was estimated to examine the interaction with the degree of source involvement. In line with the above, the overall model is not significant. There is no effect of the interaction variables on the perceived trustworthiness. Respondents did not differentiate between the sources in regard to their trustworthiness. Hence, the level of involvement did not indicate any effect on their evaluations. Hypothesis 6 is not accepted.

Model 3

Finally, the scepticism toward the online product reviews as well as the product involvement were also added in model 3. The regression results identified no effect of both variables on the perceived trustworthiness of the source, as the overall model is not significant. It is concluded that both hypotheses 12 and 14 are not accepted. For a detailed view, see Table 9.

According to the above, there is an indication of a significant difference between the digital influencers and the regular consumers regarding their perceived expertness. However, there is partial support for hypothesis 1 and hypothesis 3 as partially significant results were established by the ANOVA. The regression analysis determined a differential effect of the three sources on the perceived expertness. More specifically, digital influencers scored higher than the regular consumer in the perceived expertness. In addition, among the Youtuber and the Instagramer, the former indicated a higher increase in the perceived expertness than the latter.

(42)

42

4.2 Study 2

4.2.1 Two-Way ANOVA of a possible difference between the sources and the different products on both expertness and trustworthiness.

Expertness

The results of the two-way ANOVA indicated that there is a significant difference in the mean values of expertness among the three different products; the foundation, the running shoes and the printer. As it is shown in Table 11, the overall model is significant. In line with the preliminary expectations, the perceived expertness decreases as the level of product-source congruency decreases too. However, as it was mentioned above, the manipulation check was not successful. This may have also contributed to the unexpected differences in the mean value of expertness between the running shoes and the printer as seen in Figure 5. Moreover, the analysis determined that there is no significant interaction effect between the type of source and the product-source congruency. This indicates that the respondents did evaluate negatively the perceived source expertness without interpreting any difference between the two sources.

Trustworthiness

The two-way ANOVA on the perceived trustworthiness shows that there is a difference on trustworthiness among the three types of products. The participants did evaluate the trustworthiness of the source differently among the foundation, the running shoes and the printer. There is evidence that the perceived trustworthiness decreases as the level of product-source congruency decreases respectively. However, the interaction between the type of source and the product-source congruency is insignificant.

Table 10: Mean per condition

Source/Product Foundation Running shoes Printer

Expertness Trustworthiness Expertness Trustworthiness Expertness Trustworthiness

Instagram 3.6379 3.7672 4.8393 4.3304 4.6500 4.5214

YouTube 3.6016 3.8516 4.7700 4.2100 4.7197 4.1136

(43)

43 Expertness/ Trustworth iness df Type III Sum of Squares Mean Square F Sig. Expertnes s Type III Sum of Squares Mean Square F Sig. Trustwort hiness Corrected Model 5 51.374 10.275 20.740 0.000 12.943 2.589 5.076 0.000 Intercept 1 3431.526 3431.526 6926.637 0.000 3068.835 3068.835 6018.109 0.000 Type of source 1 0.006 0.006 0.013 0.909 0.983 0.983 1.928 0.167 Type of product 2 50.959 25.479 51.431 0.000 9.680 4.840 9.491 0.000 Type of source*Typ e of product 2 0.164 0.082 0.166 0.847 1.972 0.986 1.933 0.148 Error 176 87.192 0.495 89.748 0.510 Total 182 3602.500 3218.125 Corrected Total 181 138.566 102.691

Note:*** p-value<0.01; **p-value<0.05; *p-value<0.10 Table 11: Between-subjects test

(44)

44 4.2.2 Regression analysis Study 2

In the following Table 12, the results of the regression analysis for Study 2 are depicted.

Expertness/ Trustworthin

ess

Model 1 Model 2 Model 3 Model 4

Main effects Dummy1_Y outubeS 0.006 0.159 0.036 -0.084 0.013 -0.132 -0.083 -0.197 Dummy1_L essconsistent 1.187** * 0.454** * 1.168** * 0.0358* 1.109*** 0.268 1.084*** 0.251 Dummy2_In consistent 1.065** * 0.506** * 1.118** * 0.262 1.107*** 0.245 1.023*** 0.197 Source Involvement 0.095** 0.146** 0.088* 0.130** Interaction effects Dummy1_Y outubeS*Du mmy1_Less consistent 0.033 0.205 0.060 0.270 0.134 0.307 Dummy1_Y outubeS*Du mmy2_Inco nsistent -0.106 0.492* -0.164 0.454* -0.030 0.520** Dummy1_Y outubeS*So urceinvo 0.004 -0.066 -0.034 -0.083 Covariates Product involvement 0.241*** 0.209** Consumer scepticism 0.115 0.016 R2 0.370 0.107 0.371 0.126 0.398 0.182 0.451 0.224 Adjusted R2 0.359 0.092 0.353 0.101 0.374 0.149 -0.422 0.184 F-value 34.782* ** 7.097** * 20.740* ** 5.076** * 16.459*** 5.528** * 15.713*** 5.522***

(45)

45 Dependent variable: Expertness

Model 1

Model 1 was estimated to examine the effect of product-source congruency on the perceived expertness of the digital influencers. According to the regression results, the model is significant as they are the values of Dummy1_Lessconsistent and Dummy2_Inconsistent. This indicates that both the inconsistent product (the printer) and the less consistent (running shoes) do contribute to a decrease in the perceived expertness of the source. However, as it was also explained above, the manipulation check was not successful which is something that lead the respondents not to interpret the difference on the congruency of the running shoes and the printer with the source. This may have had an impact on their evaluations of the perceived expertness. This is indicated by the higher coefficient of the Dummy1_Lessconsistent compared to the one of the Dummy2_Inconsistent. Nevertheless, based on the above hypothesis 7 is supported.

Model 2

In model 2 the two interaction effects were added to examine any potential difference between the congruency of the products with the Instagramer and the Youtuber. The overall model is significant as well as the variable Dummy1_InstaS. However, as it is depicted in Table 12, there is no significant difference among the two types of sources as the interaction effects are not significant. Based on the above, hypothesis 9 is rejected.

Model 3

(46)

46 Model 4

In the last model, the effect of the involvement with the product as well as the one of consumers’ scepticism toward the online product reviews were investigated. Based on the findings, the model is significant as it is the p-value of the product involvement variable. When the product involvement decreases by 1 unit the perceived expertness decreases by 0.241, which is something that support hypothesis 13. The p-value of consumer scepticism is not significant so that hypothesis 11 is rejected.

Dependent variable: Trustworthiness Model 1

Following the regression analysis on the perceived expertness, this model was structured for investigating the effect of product-source congruency on the perceived source trustworthiness. The overall model is significant, and it also indicates that there is a significant negative effect of the different levels of product-source congruency on the perceived source trustworthiness. The p-values of both Dummy1_Lessconsistent and Dummy2_Inconsistent are significant. Therefore, hypothesis 8 is accepted.

Model 2

The results as seen in Table 12, support that the overall model is highly significant. However, only the interaction effect between the source and the inconsistent product is significant. The p-value of the second interaction is not significant. In addition, the variable Dummy1_Lessconsistent is significant. The interaction effect shows that the inclusion of an inconsistent product by the Instagramer leads to a more negative effect on her perceived trustworthiness. For this reason, hypothesis 10 is partially accepted since there is not any evidence about the interaction with the less consistent product.

Model 3

Referenties

GERELATEERDE DOCUMENTEN

The ineffective Westphalian state system would soon be the ineffective and outdated mode of thinking, allowing the idea of glocal cosmopolitanism to grow in influence, through

In the current study it is hypothesized that the effect of the independent variables (the presence of demographic/ psychographic characteristics attached to an OCR)

Source credibility → Cognitive trust → higher eWOM adoption → More positive attitude. 05/07/2018

Next to this, unfavorable cognitive responses in an online review context — contrary to the source credibility literature regarding advertising — are negatively moderating the effects

The aim of this empirical research is to analyze the relationship between the sender’s expertise with the product and the quality of the arguments presented in an online

—   Respondents randomly assigned to each condition using Qualtrics. —  

A study by Holbrook (1999) outlines that among other criteria for taste formation the correlation between popular appeal and expert judgments is positive.

Monetary incentive Type of argument Perceived credibility of online customer review, perceived by the receiver Monetary reward proneness Consumer criticism regarding OCRs