Graduate School of Communication
Master’s Programme Communication Science Master’s Thesis
THE BLOGGER-PRODUCT MATCH
The Effect of Blogger Expertise and Product Type on Consumer Attitudes toward Blog Sponsoring and Purchase Intention
Student Jana Haus
Student Number 10696962
Track Corporate Communication
Supervisor Dr James Slevin
Abstract
Blogs have become an important part of consumers’ decision making process and are thought to be an effective marketing tool, however, the understanding of blog sponsoring is still vague. Sponsored recommendation blog posts are articles written by bloggers who receive benefits from organizations to review and promote products on their personal blogs. The purpose of this study is to examine whether the blogger’s level of expertise and the product type influence reader attitudes toward the sponsored post and their intention to purchase the product. This study uses a 2 (high vs. low expertise) x 2 (high vs. low involvement product) experimental research design and a total of 213 valid samples to examine consumer attitudes toward the sponsored post and purchase intention. Moreover, it tests whether propensity to trust moderates the relationship between attitude toward the sponsored recommendation post and purchase intention. The results show that attitude toward sponsored post is a crucial determinant of the reader’s purchase intention. Furthermore, it reveals that blogger expertise and the product type have to match: For low blogger expertise the attitude toward the sponsored recommendation post is more positive when the product involvement is low. In turn, for high blogger expertise the attitude toward the sponsored post is more positive when the product involvement is high. Also, the readers’ propensity to trust moderates the relationship between attitude toward sponsored post and purchase intention. Based on these findings, some implications for future research and practice are given.
Keywords: Blog sponsoring, blog advertising, expertise, product involvement, purchase intention, propensity to trust
First mentioned in the late 90s Weblogs (short: blogs) quickly turned into a „remarkable artifact of the Web“ (Kumar, Novak, Raghavan, & Tomkins, 2004, p. 35). A decade later they captured the media world’s attention as there are now millions of different blogs covering a variety of topics (Chu & Kamal, 2010; Garden, 2012). The blogosphere offers a platform for various features to the reader: it can be an independent source from traditional journalism, a participatory media where the citizens voices are heard, a personal space for interaction with relatives and friends, but also a successful marketing tool and distribution channel. Even though blogs are often described in terms of their democratic, nonprofit and participatory potentials, it is crucial to note that many of them are colonized by the market (Cammaerts, 2008). This implies naturally to corporate blogs, because they are written by employees or persons otherwise affiliated with a company, but also to blogs that are seemingly independent.
Marketers have yet realized that blogs provide a powerful source of information for consumers and enable them to customize their content to the particular needs of a specific target group in a cost-efficient manner (Chu & Kamal, 2010). The sources for the delivery of information have considerably changed with the rise of social media: direct marketing messages from organizations are no longer as effective as they used to be (Lee & Koo, 2012); consumers more likely trust their peer groups than organizations and use information provided by fellow consumers to evaluate products or services before they decide to make a purchase (Lu, Chang, & Chang, 2014). Furthermore, blogs have an enormous reach due to the reader coverage and persuasive impact at a relatively low price compared to traditional advertising (Zhu & Tan, 2007). Organization related information can appear on blogs in different forms: descriptions of personal experiences, blog advertising where the brand information is separated from the blog post and finally blog sponsoring where the brand information is in the actual post (Chu & Kamal, 2010). Sponsored recommendation blog
posts consequently are articles written by bloggers who receive products from companies to review and promote them on their personal blogs and can also be described as consumer-generated advertising (Lawrence, Fournier, & Brunel, 2013).
Even though blog sponsoring is relatively low priced, the media research firm PQ Media estimated that (Jarvis, 2006) organizations spent $20,4 million on blog advertising in 2005 and reported an increase of 145% to $49,8 million in 2006. Furthermore PQ Media (2014) found that social media sponsorship grew 13,9% to $46 million in 2009. This rapid growth demands the attention of marketers and researchers. Moreover companies like Blogsvertise 1
are emerging, which connect organizations and bloggers as official blog advertising agents. Forrester Research acknowledge blog sponsoring as sponsored conversation and states that the benefits are genuineness because the message is in the bloggers own voice (Corcoran, 2009). However, the credibility of the blogger might be questioned if an affiliation with a company is visible. It needs to be determined which factors positively influence the attitude of consumers toward sponsored posts and what leads to purchase intention. More and more companies send out products to bloggers hoping to increase sales, but without having specific knowledge which bloggers to approach and for which product blog sponsoring is useful. Randomly selecting bloggers can be cost intense and not every blogger might be suitable to review a certain product and obtain the desired outcome of convincing the reader to purchase the product.
Most blog-related studies examined factors influencing users’ acceptance of or intention to use blogs (Du & Wagner, 2006; Hsu & Lin, 2008; Keng & Ting, 2009). Only few studies focused on blog sponsoring and corresponding factors of success. Zhu and Tan (2007) examined the effectiveness of blog advertising by exploring the relationship between
! www.blogsvertise.com 1
communicator expertise, advertising intent and product involvement. Their results suggest that for involvement products, there is better advertising effectiveness when low-expertise communicators are explicit about the advertising intent or when high-low-expertise communicators are implicit about the advertising intent. However, for high-involvement products, the results showed that when low-expertise communicators are explicit about the advertising intent, the outcome is lesser advertising effectiveness. In December 2009 the Federal Trade Commission (FTC) formally instituted new guidelines concerning the use of endorsements and testimonials in advertising which require bloggers to disclose any material connections they share with marketers of products they recommend (Forrest & Cao, 2010). For example, if a person decides to purchase a new shampoo and writes about it on her blog saying that it made her hair shinier and softer, and this is her honest opinions, the post is not an endorsement under the guidelines. Nevertheless, if the person is part of a consumer marketing program under which she receives the shampoo and then writes about, the positive review is an endorsement under the guidelines. This means that any affiliations between bloggers and companies such as sponsorship or simply receiving free products must be clearly stated in the post and therefore the marketing intent can no longer be concealed.
Regarding the study by Zhu and Tan (2007) this implies that it is now required to unfold any affiliations with brands and accordingly the differentiation between explicit and implicit advertising intent is no longer necessary due to legal reasons. This creates a challenge for organizations because even though consumers believe more in other individuals, the visibility of an affiliation of the blogger with a company might hinder the effect and negatively affect readers’ attitude toward the post. However, the credibility of a blog also depends on the credentials of bloggers themselves, because they endorse the products they write about. When consumers receive information presented in a sponsored post, their perception of the bloggers expertise becomes a crucial factor that will impact their attitude toward the post as well (Zhu
& Tan, 2007). In addition to the previously mentioned study, Lu, Chang, and Chang (2014) examined the effect of sponsorship type, product type, and brand awareness on consumers’ attitudes toward sponsored posts. They tested the different effects between experience and search goods and recommended use the involvement type of a product as a variable. Both studies (Lu et al., 2014; Zhu & Tan, 2007) were conducted in Asia and cultural differences therefore limit the generalizability and a conduction of the study in another cultural setting can add new insights to the topic (Bryman, 2012).
From preceding discussion of blog sponsoring, blogger expertise and product type were identified as crucial factors influencing consumers’ attitude and purchase intention. The attitude toward the sponsored recommendation post and purchase intention are likely to be affected by the involvement with the sponsored product, because a consumer’s involvement determines their information process. The present study thus aims to answer the following research question: Which effect do blogger expertise and product type have on consumer attitudes toward blog sponsoring and purchase intention? A 2 (high vs. low expertise) x ︎ 2 (high vs. low involvement product) ︎experimental design is presented to carry out the hypotheses testing. A covariate, product familiarity, is also included in order to control unexpected variation between variables. In addition, the moderating effect of consumers’ propensity to trust on the relationship between attitude toward sponsored post and purchase intention is examined because trust plays a vital role in the process of acknowledging and accepting the recommendation of a blogger and the decision to purchase a product. Based on the findings of this study, theoretical and managerial implications for academics and marketers will be presented that can help organizations to determine which products and which bloggers they should approach in order to increase sales. It will reveal where there is a „perfect match“ of product involvement and the blogger expertise which leads to a more
positive attitude toward the sponsored post. The results can help organizations to cut costs and make blog marketing more efficient.
Theoretical Framework
Blogs, unlike website, are typically personal and serving as a channel for bloggers to express their opinions. They provide a platform for self-expression which includes background information on the authors and their real life identity (Gilly & Schau, 2003). Sponsored recommendation posts are a form of online consumer review (Forrest & Cao, 2010) which often includes pictures and information about the product or hyperlinks to the sponsoring company (Zhu & Tan, 2007). To distinguish sponsored recommendation posts from online consumer review one has to look at the publishing platform. Consumer reviews are generally posted on a product or seller website while sponsored recommendation posts appear on personal blogs (Lu et al., 2014; Park, Lee, & Han, 2007). Furthermore the tone of sponsored recommendations is more positive because the message is motivated by the compensation the blogger receives from the company and the bloggers high satisfaction with the product. A sponsored post in the present study thus is defined as a blog article written by a blogger who is sponsored by a company for reviewing and promoting their products or services on their blog. The interpretation of the sponsored message however varies from individual to individual and also depends on the type of product and the credentials of the blogger.
Stuart Hall (1980) suggested that meaning is not simply determined by the sender. A message is never transparent and the audience is not solely a passive recipient of meaning. He found that messages are decoded or interpreted in different ways depending on the recipients social background and other personal factors. Hall proposed the encoding-decoding model where communication is defined as a social process depending on the individual production and consumption of meaning. For example, blog sponsoring can have multiple layers of
meaning which can be decoded in various ways depending on the individual. Consequently, the same sponsored post can have different meanings to different readers. Hall criticized linear communication models because of their concentration on message exchange and the lack of structured conception of the different stages in the process. According his work there are four stages of the encoding-decoding process (Hall, 1980) which are explained in the following in the context of blog sponsoring:
1. Production. The message is encoded by bloggers when they write the sponsored recommendation post. The message is influenced by the bloggers ideology, beliefs and values.
2. Circulation. How is the message perceived? Bloggers post the written message on their blogs. The format of the message, hence whether it is written, spoken or otherwise articulated influences the how the audience perceived the post.
3. Distribution/consumption (also: use). This means the interpretation or decoding of the message. It requires an active audience, so the individual reading the sponsored post has to be aware of the message and decode its meaning.
4. Reproduction. The last stage of the process deals with what the reader does with the message after they interpreted it based on their own values and beliefs. It determines whether the message influenced the reader or not.
How the message is produced and circulated by the blogger can also establish a certain level of trust from the audience, hence it „makes“ the reader trust in the blogger because they assume they have the ability and knowledge to endorse the product, and in terms of circulation, the blogger uses suitable language and has proficient writing skills.
The decoding is constructed by using three hypothetical positions: dominant-hegemonic positions, negotiated position and lastly the oppositional position (Hall, 1980; Kim, Jeong, Khang, & Kim, 2011). The dominant position implies that the recipient shares the encoded message of the sender, accepts and reproduces the anticipated meaning (also called „preferred reading“ as seen from the point of view of the sender). Recipients take a negotiated position when they partly share the code of the message and overall accept the preferred reading but sometimes resist and modify the message in a way that is more related to their own values, experiences and opinions. This position is based on personal conditions and involves contradictions because of a mix of adaptive and oppositional elements. Lastly, the oppositional position refers to recipients that completely understand the preferred reading but decode the message in a contrary way and therefore fully reject the meaning because of their social situation. For example the recommendation post is viewed negatively because readers know that it is intended to influence their purchase intention and therefore they do not buy the product.
Halls encoding-decoding model is used in this study to define the process the readers are experience after reading the sponsored blog post. Based on their use and reproduction of the message, they develop different attitudes toward the sponsored recommendation post and may or may not be persuaded to purchase the featured product. Because the purchase intention is influenced by how the reader decodes information, they can adapt one of the three positions explained above. So they might either completely believe the recommendation (dominant position), therefore have a positive attitude and purchase the product, partly believe the message and then depending on their personal attitude purchase the product or decide against it (negotiated position) or develop a negative attitude toward the recommendation and therefore do the opposite of what is expected of them and decide not to
buy the product (oppositional position). The following sections aim to determine how expertise and product type can possible influence these different processes.
Attitude toward Sponsored Post and Purchase Intention
From a marketers perspective, the ultimate goal of blog sponsoring or any persuasive message is to positively influence the attitude of a consumer toward the product. This persuasion does not only occur from top-down but it is also shaped by family, friends and acquaintances which present influential sources of opinion and information on products and thereupon effectively influence an individuals opinions and behavior (Brooks, 1957). Lu et al. (2014) define consumers’ attitude toward sponsored recommendation posts as psychological tendency toward the credibility of the source of a blog post during a particular exposure occasion. If a blog reader reads the specific sponsored post and claims the message is true, their attitude toward the sponsored post is positive.
As indicated in the research of Fishbein and Ajzen (1975) consumer attitudes directly affect their purchase behavior. Their theory of reasoned action (TRA) and later the theory of planned behavior (TPB), which was proposed to improve the predictive power of the TRA by including perceived behavioral control, are widely recognized models to apply in social psychology and marketing (Lu et al., 2014; Tsai, Chin, & Chen, 2010). According to these theories, individuals have voluntary control and make decisions based on rational and available systematic information (Ajzen & Fishbein, 1977). They suggest that the determinant of an individuals’ behaviors is their intention which derives from attitude and subjective norms. In this context attitude refers to a person’s feeling of favorableness, or the opposite, toward the behavior. Subjective norms are the person’s perceived social pressure caused by people who are important in their social environment and how they evaluate the person’s behavior (Ajzen, 1991; Tsai et al., 2010). Applied to a marketing perspective the
model explains that consumers’ purchase behavior is established by their intention to purchase but also by attitudes and subjective norms which affect the behavior hence the intention of purchasing a product or service.
In this paper the TRA is used to define the attitude of consumers toward the sponsored post and the purchase intention: the subjective norm is used in the sense of consumers’ perception of the blog post which contains a recommendation of a sponsored product. The purchase intention represents the consumer being willing and intending to buy the recommended product (Spears & Singh, 2004). This is in line with the work of Bouhlel, Mzoughi, Ghachem, and Negra (2010) who found that consumers’ attitudes toward a blog positively influence consumers’ purchase intention. Furthermore, the quality of the communication can result in an overall positive attitude toward the blog, which in turn leads the consumer to purchase the product (Fiore, Jin, & Kim, 2005). Therefore, the following hypothesis is proposed:
H1: Consumers’ positive attitude toward the sponsored recommendation post leads to higher purchase intention.
Propensity to Trust
Ulrich Beck and Anthony Giddens (1994) speak of a condition of heightened risk and uncertainty as characteristics of the contemporary society. Expertise is not only understood in terms of its diagnostic role but also as a critical involvement for how risk is produced and disseminated in modern societies. Beck relies on expertise to legitimate the authority of dominant institutions while Giddens suggest trust plays a vital role in the maintenance of social stability. While in pre-modern society the elders possessed the knowledge, post-modern societies trust in expert systems. Even though these expert are usually not questioned,
there is the possibility that they are not always right which presents a risk that the individual has to be willing to take. (Healy, 2001).
Additionally, these social processes in post-modernity have caused a continual displacement and reappropriation of expertise that causes uncertainty and doubt in a world of multiple authorities. More and more risks are created by human knowledge and practice as opposed to natural processes. Problems constructed by science and technology can only be solved by more expertise which in turn leads to more manufactured uncertainty. Not expertise itself but the easy accessibility of expertise is what distinguishes post-modern societies (Tucker, 1998). Because of this multitude of authorities, people no longer just take opinions from experts because they tend to frequently disagree with each other which leads a person to make a decision in the context of conflicting and changeable information (Giddens, 1999). The decision which expert to trust has become risky which is linked to reflexivity and individualization. Individuals have to choose how they want to be and how they want to act (Arnoldi, 2013). This decision making in social life is understood as de-traditionalization and is linked to what Beck, Giddens, and Lash (1994) call institutional reflexivity.
Giddens defines trust as „the vest of confidence in persons or in abstract systems, made on the basis of a ‚leap into faith‘ which brackets ignorance or lack of information.“ (Giddens, 1991, p. 244) Individuals must have trust because they are not able to experience the problem because of the time/space distantiation or the technical knowledge upon which to make informed decisions regarding the validity of the expertise. Applied to the context of this study, apart from the perceived blogger expertise the consumers’ propensity to trust becomes crucial in the process of acknowledging and accepting the opinion and recommendation of the blogger.
The propensity to trust of an individual can influence the relationship between attitude toward the sponsored recommendation post and purchase intention as it refers to the general attitude of an individual to trust other people (Mayer, Davis, & Schoorman, 1995). Lu, Chang and Chang (2014) tested propensity to trust as a control variable in order to limit unexpected variation. However, other studies found that trust is as important to online commerce as perceived usefulness and perceived ease of use as antecedents of the widely accepted technology acceptance model (Gefen, Karahanna, & Straub, 2003). Gefen (2000) found that consumers with a high propensity to trust are more likely to trust other people and furthermore it is primarily the individuals propensity to trust that affects their trust in the vendor. Transferred to this paper it is suggested that propensity to trust improves the relationship between consumer attitudes toward sponsored recommendation posts and purchase intention. If consumers’ propensity to trust is in generally higher they are likely to trust the blogger more easily and therefore buy the product whilst consumers with a negative propensity to trust are harder to convince to purchase the product even if their attitude toward the sponsored recommendation post is overall positive. This leads to the following hypothesis:
H2: Propensity to trust moderates the relationship between attitude toward the sponsored post and purchase intention in a way that high propensity to trust leads to higher purchase intention.
Expertise and Attitude toward the Sponsored Post
The availability of increasingly more reliable free sources constructed by online communities have changed the notion of expertise in a way that knowledge sharing platforms like Wikipedia are challenging the traditional roles of experts (Sanger, 2012). Sanger (2012) further suggest that traditional certified professionals have a role in society of articulating
their knowledge in their field of expertise because they are for example asked for comments by journalists and when they speak of their areas of expertise people listen and acknowledge their credibility. This epistemic leadership has been challenged by the rise of so called online experts. Nevertheless, it will not wither away even as social networks further mature and crowdsourcing becomes increasingly accepted. One reason is the „regress of credibility“ which begins with the observation that, in every folk epistemology, a piece of information is regarded as more credible if the communicator is believed to be credible. So if a statement comes from a questionable source it will always be believed less than when it comes from a credible source. Information credibility can be defined as trust in a source, hence, whether or not the information is perceived to be believable and the individual demonstrates believe in the information (Corritore, Kracher, & Wiedenbeck, 2003; Del Guidice, 2011; Hilligoss & Rieh, 2008; Metzger, 2007).
Even though sponsored posts are an acceptable communication media used by marketers, the credibility of sponsored posts is often questionable and consumers may feel that the message has bias and thus resist the persuasive intent (Lee & Koo, 2012; Lu et al., 2014). Research suggests that persuasion partly depends on characteristics regarding the source of the message (Homer & Kahle, 1990). In addition Lau and Ng (2001) found that individual factors such as personality, attitudinal and product involvement variables are relevant to the process of persuasion through eWOM. Credibility is the most researched source characteristic and can be narrowed down to the dimensions expertise, objectivity, and trustworthiness. It is often suggested that expertise is the most dominant dimension (Kahle & Homer, 1990). Nonetheless, the conclusions regarding its impact are conflicting. While some empirical evidence suggests that source expertise can lead to a positive attitude change (Fireworker & Friedman, 1977; Kelman & Hovland, 1953), other studies showed that there is no relationship between expertise and persuasion (Scileppi & Homer, 1969; Steiner & Homer, 1968).
Expertise in this paper refers to how capable the blogger is of making right assumptions about a product or service by possessing the appropriate skills, competency, and knowledge towards it (Homer & Kahle, 1990; Zhu & Tan, 2007). Credibility and reputation are interlinked with the characteristics of the bloggers because they do not simply contribute a sponsored post but also endorse the products they write about. Also high credibility can lead to consumers’ adoption of eWOM recommendations and expertise can positively affect an attitude change (Cheung, Luo, Sia, & Chen, 2009). The readers process the recommendations given on the blog and their perception of the bloggers become a crucial factor that will impact their responses to the content of the blog post (Zhu & Tan, 2007). Thus, the following hypothesis is proposed:
H3: When the perceived blogger expertise is high the attitude toward the sponsored recommendation post is more positive.
Expertise and Product Involvement
Involvement is a motivational construct (Celsi & Olson, 1988) and while product involvement is defined as the perceived relevance of a product based on an individuals needs, values, and interests (Zaichkowsky, 1985). Involvement can be either enduring or situational (Zhu & Tan, 2007). However, research suggests that product involvement is an enduring involvement, hence, it is relatively stable and the involvement levels with a specific stimulus are sustainable (Andrews, Durvasula, & Akhter, 2001; Havits & Howard, 1995; Zaichkowsky, 1985). The two main consequences of product involvement are information processing and persuasion (Andrews et al., 2001). Petty, Cacioppo and Schumann (1983) found, based on the elaboration likelihood model (ELM), that different features of an advertisement may be more or less effective depending on the involvement level. For low involvement product, peripheral cues are more important than issue-relevant arguments, but
for high involvement, the opposite is true. The level of product involvement has been widely recognized as major factor affecting an individuals response to marketing messages (Danaher & Mullarkey, 2003; Petty et al., 1983; Zhu & Tan, 2007).
According to Wang (2005)consumers could be influenced by endorsement of third-parties differently based on the product type. Consumers often seek experts’ opinions for high involvement products because high-expertise endorsers are supposed to be a credible source of information that may help the consumer to evaluate the product (Zhu & Tan, 2007). Endorsers with low expertise are generally effective by drawing from their similarity with consumers. Thus, high expertise would be more suited for products that present high financial or performance risks, while low expertise is more effective for low-risk products. Furthermore the „match-up hypothesis“ suggests that if there is a fit between the endorsers and the endorsed product, the endorsement will be more effective and the product receives a higher perceived believability, favorable attitude and purchase intent (Kamins, 1990; Kamins & Gupta, 1994). In turn a mismatch between the endorser’s expertise and the product type may lower the persuasion (Kamins & Gupta, 1994). This leads to the following hypotheses:
H4a: For low communicator expertise the attitude toward the sponsored post will be more positive when the product involvement is low than when it is high.
H4b: For high communicator expertise the attitude toward the sponsored post will be more positive when the product involvement is high than when it is low.
Model
Based on the literature review a the following model was build (see figure 1). The expertise of the blogger and the product type featured in the blog post influence the readers’ attitude toward the sponsored post. This attitude has an effect on the purchase intention. The
relationship between the variable attitude toward sponsored post und purchase intention is moderated by the consumer’s propensity to trust.
Methodology
A 2 (high vs. low involvement products) x 2 (high vs. low blogger expertise) factorial between-subjects experimental design is chosen to answer the research question and test how product type and expertise of the blogger influence readers’ attitude toward sponsored recommendation posts and purchase intention. Experimental research methodology gives the researcher more control of the situation and has a high internal validity which can provide a precision in measurement as well as support for causal claims (Bryman, 2012). Prior to the formal experiment a pretest was conducted to identify the level of involvement of different product categories. Based on the results the product categories with the highest and the lowest involvement were chosen and blog posts featuring these products were designed. For each product category a high expertise and a low expertise blog was created which lead to a total of four different blog interfaces. The respondents were randomly assigned to one of the four treatments in this experiment. In the following the design of the stimulus material and the variables will be explained.
Figure 1
Blog Development
Blogs vary in style, design, format and popularity which makes it difficult to isolate the independent variable and diminish alternative explanations by using already existing blogs. Because of this, in line with Zhu and Tan (2007) and Lu et al (2014), fictitious blogs are created to avoid any prior familiarity with the blog or disturbing side factors that do not matter in the context of this experiment. In order to make the blog posts look realistic, the website Wordpress was used to set it up. Wordpress is one of the most popular tools to create a blog and offers different templates that differ in design, layout, format, and background complexity. To avoid bias, a simple and minimalistic template was chosen.
Four different blog interface were created in English language to observe the reaction to the posts. In order to rule out interactive effects that may influence the respondents perception no hyperlinks to other websites were provided in the post and no comments by other users were shown. The design displayed a body of text describing the product, the blog title as well as the bloggers nickname and basic information (depending on their level of expertise, for low expertise no detailed information was given) on the side. The headline of the post stated that the post was sponsored by a company and this affiliation of the blogger with the company was also mentioned in the post itself. The independent variables - blogger expertise and product involvement - were operationalized as follows in the blog post.
Blogger Expertise
The blogger’s level of expertise was tested by manipulating the blogger profile and the content of the blog post. For high expertise manipulation it was stated in the blogger profile that the blogger has a profession that is related to the product, this is in line with the suggestion of Homer and Kahle (1990). In addition, to emphasize the blogger’s expertise
more professional opinions, technical terms and formal language were presented in the sponsored recommendation post. For blogs with a low expertise endorser, no information was provided regarding the bloggers background, more casual descriptions of the product were given and the language used was rather informal.
Even though the language was different, both the high and the low expertise blogs were written equally well in terms of grammar, sentence structures and expressions. The difference of the product-relevant expertise is the presence or consequently absence of professional opinions, technical terms and formality. Furthermore, all sponsored recommendation posts were formulated as one-sided messages instead of two-sided messages, so only the positive features of the product were described. Two-sided messages have different effects on the readers and were therefore avoided (Crowley & Hoyer, 1994). In the final experiment the manipulation of the variable blogger expertise was checked using a five-item scale based on Ohanian (1990) and Homer and Kahle (1990). This 7-point semantic differential scale was used to assess the level of expertise.
Product Involvement
In order to find suitable product categories, a pretest was conducted among ten students of the University of Amsterdam. The students rated six product categories on an 8-item, 7-point semantic differential scale based on the Personal Involvement Inventory (PII) introduced by Zaichkowsky (1985) (see Appendix B). Based on the results of this pretest, mineral water represents a low involvement and camera presents a high involvement product. The same scale used in the pretest was used to validate the manipulation of the product involvement in the actual experiment. In order to avoid that prior knowledge of the brand would create bias in the results (Campbell & Keller, 1995) fictive brands were created. In order to find out how
familiar the respondents were with the product category, the product familiarity of the respondent was also tested as a control variable in the survey.
Measurements
The experiment was conducted among blog readers, therefore it was crucial to examine how they - as the recipients of the message - perceive the sponsored recommendation post. Existing scales were used to measure attitude toward sponsored post (ATP), purchase intention and propensity to trust (PTT) (see Appendix B). These scale were already used by Lu, Chang, and Chang (2007) in the context of blog sponsoring and found to measure the right constructs. The scale used for ATP was a four item scale of Boush, Friestad, and Rose (1994) and Obermiller and Spangenberg (1998). If respondents score high on each item of this scale they have a positive attitude toward the sponsored post. For purchase intention a five item scale was used by Putrevu and Lord (1994) and Taylor and Baker (1994). The higher the respondents agree with the statements, the higher their intention to purchase the product is.
These two dependent variables examine the readers’ perception of the sponsored recommendation post, however, their perception of the blogger and the blog post also depends on how much an individual is willing to trust other people in general. Thus, the moderating variable PPT was measured based on a five-item scale developed by Laurent, Kapferer and Roussel (1995). Participants who score high on this scale have a higher propensity to trust. The respondents of the survey were asked to rate each item of these three variables on a seven-point Likert scale reaching from 1 (strongly disagree) to 7 (strongly agree).
Pilot Test
A pilot test was conducted with ten students of the University of Amsterdam in order to check the manipulations and other possible problems which could occur in the actual experiment. Each respondent read one randomly assigned experimental blog and completed the questionnaire. Afterwards the students were asked for feedback and suggestions. Once the respondents found that the treatments have been successfully operationalized the final experiment was conducted. Some suggestions were given regarding the phrasing of questions and the design of the experimental blogs and were integrated in the final version.
Procedures and Sample
An electronic survey was posted via an online survey platform Qualtrics in order to sample 2
from the mass population of online users above the age of 18 (see Appendix A & B for treatments, constructs, scales and measures). Even people who were not reading blogs regularly could participate in the experiment because the treatments showed a blog post that did not require any previous experience or knowledge of blogs and the questions were based on the blog interface that was presented to them.
The link was shared on popular social networking sites like Facebook and Twitter on the wall of relevant groups and pages that were connected to blogs and blog reading in general. Furthermore the experiment was posted in blog specific forums like Blogger. This way of distribution was chosen because it is assumed that blog readers are generally experienced internet users and therefore browse social networks regularly (Lu et al., 2014).
After accessing the experiment, the respondents were randomly assigned to one of the four experimental situations. First, a description and introduction about the survey and the
www.qualtrics.com
experimental blog post was given and afterwards the respondents viewed the blog post and completed the questionnaire which measured their behavioral intents toward the treatment, captured demographic information and checked for manipulation and covariate control. It took about 8 minutes to complete the survey per average.
The sample comprised of 213 respondents. The sample number of each experiment version was between 101 and 112 responses due to the random assignment of one of the four treatments. 127 women (59,6%) and 84 men (39,4%) and 2 „others“ (0,9%) participated in the experiment. This is consistent with blog usage population since, according to previous studies more females use blogs than males (Hsu & Lin, 2008). The majority of respondents was between 21 and 30 years old (72,8%) and 48,4 per cent were highly educated (bachelor or college degree). These findings are in line with previous studies which found that e-customers are generally younger and higher educated tha conventional e-customers (Frost, Goode, & Hart, 2010; Lu et al., 2014; Yang, Chandlrees, Lin, & Chao, 2009). 43,2 per cent of respondents stated that they read blogs 1-5 times per month and 30,5 per cent claimed that they never read blogs. The majority of respondents were from the Netherlands and Germany.
Results
Validity and Reliability Tests of Scales
First the items „I don’t believe in what the blogger wrote in this article“ and „I have no intention to buy this product“ which are negatively worded were recoded so that a high value indicates the same type of response on every item on the scale. Then a principal-component factor analysis with varimax rotation was conducted, yielding in 6 factors. All items loaded on the intended factors respectively, indicating satisfactory convergent and discriminant
validity. The constructs have a Cronbach’s Alpha above 0,8 indicating adequate reliability (see Table 1) (Nunnally & Bernstein, 1994).
Manipulation Checks
The manipulation of the independent variables was verified using manipulation checks. An independent-sample t-test was conducted to compare the score of blogger expertise and product involvement for high versus low expertise bloggers and high versus low involvement products. For both expertise and product involvement, the results show a significant difference between the means for different levels of treatment (see Table 2). The magnitude of the differences in the means of blogger expertise was moderate (eta squared = .08). For product involvement the magnitude of differences in means was large (eta squared = .13) (Cohen, 1988). In addition, a one-way between-groups analysis of covariance was conducted to compare the effect of the four different combination of treatments HEHI , HELI , LEHI 3 4 5
Table 1
Constructs & Reliability.
Construct Number of Items Cronbach’s Alpha
ATP 4 .85 Purchase Intention 4 .84 Product Familiarity 2 .88 PTT 5 .95 Blogger Expertise 5 .92 Product Involvement 6 .92 Note. N = 213.
HEHI = High Expertise + High Involvement Product
3
HELI = High Expertise + Low Involvement Product
4
LEHI = Low Expertise + High Involvement Product
and LELI on participants ATP. The independent variable was the type of treatment, and the 6 dependent variable consisted of scores on the ATP scale. Participants’ scores on the product familiarity scale were used as the covariate in this analysis. Preliminary checks were conducted to ensure that there was no violation of the assumptions of normality, linearity, homogeneity of variances, homogeneity of regression slopes, and reliable measurement of the covariate. There was a significant difference between the four treatment groups on attitude towards sponsored post [F(3,208) = 9.01, p = .000, partial eta squared = .12]. This shows that the combined treatment groups vary from each other and can therefore be compared.
There was no significant effect of the covariate product familiarity on ATP [F(1,208) = .84, p = .36, partial eta squared = .01]. How familiar the respondents were with the product prior to reading the sponsored post hence does not influence their attitude toward the sponsored post.
Hypotheses Tests
All tests of hypotheses were conducted at a 5% level of significance. In order to answer H1, H2 and H3 regression analysis was performed. The regression model with purchase intention as dependent variable and ATP, age, sex, education, frequency and length of blog use as
Table 2
Manipulation Checks. Independent
Variable
Treatments N Mean Std Dev t P
Blogger Expertise High 103 4.14 1.37 5.44 0.01**
Low 110 3.18 1.21 Product Involvement High 112 4.53 1.23 4.26 0.01** Low 101 3.72 1.52 Note. N = 213. *p < 0.05, **p < 0.01.
LELI = Low Expertise + Low Involvement Product
independent variables is significant, F(6, 204) = 13,23, p < .001. The regression model can therefore be used to predict the purchase intention of the participant, the strength of the prediction is moderate: 28 per cent of the dependent variable can be predicted on the basis of independent variables. Only two independent variables are significant. Age has a significant, weak association with purchase intention, b* = .18, t = 2,82, p <.05, 95% CI [.05, .32]. Per year of age, the purchase intention increases by .05. ATP has a significant, moderately strong association with purchase intention, b* = .49, t = 8.14, p < .001, 95% CI [.33, 5.54]. Thus, H1 is supported: consumers’ positive ATP leads to higher purchase intention.
Table 3
Regression model to predict Purchase Intention.
Standardized Coefficients (Execpt Constant)
H1 H2
Constant 2.01*** 3.77***
ATP .49***
ZScore (ATP) .44***
ZScore (PTT) .19**
ZScore (ATP) X ZScore (PTT)
.04*
Age .18** .17**
Sex -.03 -.02
Education -.06 -.04
Frequency of Blog Use -.01 -.03
Length of Blog Use .04 .01
R2 .28*** .33***
F 13.23*** 12.29***
Note. N = 213.
Dependent variable: Purchase Intention. *p < .05 ** p < .01 ***p <.001.
In order to test the effect of PTT on the relationship between ATP and purchase intention a multiple regression analysis was conducted with the standardized variables ATP and PTT and the interaction variable (ZScore (ATP) X ZScore (PTT)) and the control variables age, sex, education, frequency and length of blog as independent and purchase intention as dependent variable.
The variables ATP and PTT are first changed from continuous to standardized variables. Those standardized variables are multiplied and then create the interaction term. 33 per cent (R2 = .33) of the variation in purchase intention can be predicted on the basis of the
independent variables. The model is significant, F(8, 202) = 12.29, p < .001. The effect of the interaction term is significant with p < .05 thus there is moderation.
Figure 2 shows the direction of the moderating variable: PTT promotes the relationship between the dependent and the independent variable, hence it strengthens the relationship. Thus, H2 is confirmed.
Figure 2
Next a regression model with ATP as dependent variable and age, sex, education, frequency and length of blog as independent and expertise as independent variables is significant, F(6, 204) = 10.72, p < .001 (see Table 4) . The model can therefore be used to predict ATP, the strength of the prediction is moderate: 24 per cent of the variation in work performance can be predicted on the basis of age, sex, education, frequency and length of blog as independent and expertise( R2 = .24). Expertise, b* = .46, t = 7.41, p < .001, 95% CI [.24, .41], has a
significant, moderate association with ATP. H3 is supported by the results of this study.
A one-way between-groups analysis of variance was conducted to explore the impact of the treatments on ATP. Subjects were divided into four groups according to the combination of expertise and product type presented to them. There was a statistically significant difference at the p < .05 level in ATP for the four groups [F(3, 209)=8,74, p < .001]. Despite reaching statistical significance, the actual difference in mean scores between the groups was moderate
Table 4
Regression model to predict ATP.
Standardized Coefficients (Execpt Constant) H3 Constant 2,08*** Expertise .46 Age .06 Sex -.04 Education .09
Frequency of Blog Use .06
Length of Blog Use .01
R2 .24***
F 10.71***
Note. N = 213.
Dependent variable: ATP. *p < .05 ** p < .01 ***p <.001.
(Cohen, 1988). The effect size, calculated using eta squared, was .11.Post-hoc comparisons using the Tukey HSD test indicated that the mean scores of the four treatment groups were significantly different from each other. For low communicator expertise the ATP is more positive for low involvement products (MLELI 3.86, SD .97) than for high involvement
products (MLEHI 3.29, SD .85). H4a is therefore confirmed. For high communicator expertise
the attitude toward the sponsored post is more positive when the product involvement is high (MHEHI 4.17, SD .93) than when it is low (MHELI 3.81, SD .97). This result supports H4b.
Conclusion & Discussion
Sponsored recommendation posts are often thought to be an effective marketing tool, therefore an examination of the reader’s attitude toward these posts written by bloggers who receive benefits from a sponsor is important. This study conducted a 2 (high vs. low expertise) x 2 (high vs. low involvement product) experimental research design. Moreover, it tested whether propensity to trust moderates the relationship between attitude toward sponsored post and purchase intention. The findings are consistent with those of earlier research concluding that the recommender and informant of a product are both positively related to the consumer’s purchase intention (Park et al., 2007). Furthermore, the results reveal that attitude toward sponsored recommendation post is a crucial determinant of the readers’ behavioral intention, which implies that increasing a user’s positive attitude is likely to result in a higher purchase intention. Also, readers’ propensity to trust plays a vital role and significantly influences the relationship between attitude toward sponsored post and purchase intention. Finally, product type and bloggers’ expertise have to match in order to reach a high attitude toward sponsored post.
The results show that consumers’ positive attitude toward the sponsored recommendation posts leads to higher purchase intention, which was tested in H1. If consumers believe in the
sponsored post and develop a positive attitude toward the sponsored post, they will form the intention to purchase the product recommended by the blogger. This is in line with the findings of previous studies: a credible recommendation blog post can affect consumers’ willingness to buy products (Bouhlel et al., 2010; Lu et al., 2014).
The second hypothesis assumed that an individuals’ propensity to trust moderates the relationship between attitude toward sponsored post and purchase intention. This hypothesis was supported. If readers displayed a high propensity to trust they are more likely to buy the product featured in the blog post than when their propensity to trust was low. As explained in H1, high attitude toward the sponsored posts leads to higher purchase intention. H2 further reveals that if the attitude toward the sponsored post in low, but the individual’s propensity to trust is high, their purchase intention is higher than the purchase intention of those with a lower propensity to trust. Similar effects were found in previous studies regarding online commerce were trust is a major antecedent of the widely accepted technology acceptance model and consumers with a higher propensity to trust are more likely to trust the vendor of a product (Gefen, 2000; Gefen et al., 2003). The present study shows that those results are applicable in the context of blog sponsorship as well and that trust is a crucial factor in the process of accepting the recommendation posts of bloggers.
As hypothesized in H3, when the perceived blogger expertise is high, the attitude toward the sponsored post is more positive. A possible explanation for this finding lies in readers’ prior expectation of an expert. Experts who have a higher expertise in the product category are expected to analyze and review products as part of their personal or work interest and publish their opinions on their blogs (Wang, 2005). Due to consumers’ previous experiences with advertisements they might be more likely to accept that an expert is paid to endorse a product on which consumers seek advice. These findings are opposed to the results of Homer
et al which found no significant relationship between expertise and persuasion (1969; 1968). However, the following hypotheses reveal, that the appropriate combination of expertise and product type also has an impact on the attitude toward the sponsored post.
H4a and H4b were confirmed. First, it was found that for low communicator expertise the attitude toward the sponsored post is more positive when the product involvement is low than when it is high. An explanation for this could be that when a low-involvement product is featured in the blog post the reader is not motivated or willing to process more technical and formal information about the product and the expert opinion therefore is perceived as unnecessary. Non-experts have a greater similarity to the audience and their honesty about an affiliation with a company can enhance the blogger’s credibility (Zhu & Tan, 2007).
The results of H4b show that for high communicator expertise the attitude toward the sponsored post will be more positive when the product involvement is high than when it is low. One reason could be that reviews of high involvement products require more elaborate skills and efforts because they usually contain more information, technological details or other complex attributes that need to be explained. Thus, the positive attitude toward the sponsored post that derives from high expertise and high involvement products is because the readers acknowledge that the expert deserves the rewards resulting from the sponsorship. In turn, if the blogger’s expertise is high and the product featured is a low involvement product, this might cause doubt and a more negative attitude toward the sponsored post because the interest of the blogger is less accessible and understandable to the reader.
In terms of Hall encoding-decoding model (1980), a mismatch between the bloggers expertise and the product involvement leads to an oppositional position of the reader when decoding the message - they develop a more negative attitude toward the post and therefore do the opposite of what is expected of them and decide not to buy the product. A match
between expertise and product involvement is likely to lead to the dominant position where the reader accepts the encoded message of the sender and reproduces the anticipated meaning, hence develops a positive attitude and consequently intends to purchase the product.
Based on the above mentioned findings, the present study offers a theoretical contribution to the field of online marketing communication in the blogosphere. While previous research regarding blog sponsorship is still relatively limited and most research has been conducted in Asia, this study presents a quantitative experimental research that aims to understand readers’ attitudes toward sponsored blog posts in a western context. Blog sponsoring needs to be distinguished from online consumer reviews because bloggers receive financial benefits from the company they are affiliated with, which makes it necessary to determine which factors improve the effectiveness of blog sponsoring. The implications for practice deriving from this research will be explained in the following section.
Practical Implications
The results appear to have important implications for organizations seeking to promote their products via blogs. From a marketing perspective the ultimate goal of blog sponsoring is to create a positive attitude toward the sponsored post because, as shown in this study, this will lead to consumers’ intention to buy the featured product. In order to establish this positive attitude the blogger expertise and the product involvement have to match. The findings of this two-way interaction between blogger expertise and product type show that marketers need to consider two aspects before collaborating with bloggers: First, who recommends the sponsored product, a high or a low expertise blogger? And second, which type of product will be recommended, a high or a low involvement good?
For low involvement products, it is not recommended to use an expert as an endorser since low involvement products do not need a lot of technical or detailed information and the readers might feel that the blogger is overselling or doubt their interest. When a blogger with low expertise endorses a low involvement good, they produce a better advertising effect. This means that organizations should encourage non-expert bloggers to promote low involvement goods in an open and honest way consumers can relate to.
For high involvement products, consumers develop a more negative attitude toward the post if the product is endorsed by a non-expert. Organizations should therefore approach expert bloggers because these products need more explanation and detailed information. Also, providing free trial products to expert bloggers can be beneficial because they can eventually become opinions leaders which then informally influence the attitude or behavior of others in a desired way (Li & Du, 2010; Venkatraman, 1989).
Also, the moderating effect of propensity to trust shows that trust is crucial factor influencing the relationship between attitude toward the sponsored post and purchase intention. A high propensity to trust leads to a higher purchase intention, so organizations and bloggers should actively focus on being perceived as trustworthy or establishing trust. This is in line with Giddens argumentation that building trust in late modernity is a reflexive process which requires continuous communication and openness (Möllering, 2006). In order to develop trust it needs to be constantly reproduced and sustained. Bloggers who agree to a sponsorship are advised to carefully select the products they are recommending to avoid losing their readers’ trust.
Marketers need to be aware that not every product is suitable for every blogger in order to avoid spending money on sending out products and collaborating with bloggers that do not match the product type. Acting according to the above mentioned recommendations can make
the advertising efforts more effective and increase sales by targeting the „right“ blogger for the „right“ product type. Apart from traditional companies the results can be beneficial for companies connecting bloggers and organizations, as well as for bloggers themselves. Especially for the blogger it is desirable to achieve a positive attitude of the reader toward the blog post because this might not only lead to a higher popularity of the blog, but because a positive attitude leads to a higher purchase intention this can result in more collaborations with suitable companies. However, this might also have a negative effect. The following section will give recommendations for future research as well as limitations of this study.
Limitations and suggestions for future research
Further research should examine whether too many collaborations with similar companies have a negative impact on the overall attitude toward the blog - not the blog post itself. The present study only focused on the attitude toward one blog post, not the general attitude toward the blog. It can however be assumed, that the overall attitude toward the blog depends on the frequency of sponsored posts and the balance between sponsored posts and posts that display the personal opinions of the bloggers and are not in collaboration with a company. Furthermore variables should be included to the present model which measure not only the consumer’s propensity to trust, but their level of trust in organizations or virtual experts such as bloggers.
Moreover, the generalizability of this study is limited because the respondents were mainly from the Netherlands and Germany. Further studies should extend the sample profile by random sampling and engage in cross-cultural comparisons to extend the generalizability of the results. Also, this study used product familiarity as a control variable to limit unexpected variation, continuing research should use more control variables in the experiment.
While this research focused on blogger expertise and the involvement level of the featured product, future research should include other factors such as authenticity of the blogger or similarity between the reader and the blogger as well as blog or blogger reputation. Furthermore, other product classifications should be examined such as hedonic/utilitarian or tangible/intangible and it is recommended to use different product categories for each treatment. Future studies should create at least two experimental blogs per treatment to make the results more stable and generalizable. Apart from products, services like entertainment, travel, or restaurants should be subject to further research as consumers are more likely to need opinions and experiences from others on those, but might feel biased if the blogger has received benefits for testing them for free or was even paid to test them. In addition, while this research implied that an expert blogger has a more formal writing style than a non-expert blogger, the effect of language in sponsored recommendation posts could have limited the results of this study and present an interesting starting point for future investigations.
Lastly, the mobile photo and video-sharing social network Instagram is on the rise to become equally successful in terms of marketing as blogs by simply showing pictures and videos of products, without giving much written information or product reviews. Most bloggers also use Instagram and tag for example the brands they are wearing on their photos. However, there are also Instagram users who do not write successful blogs but have thousands of followers and are approached by companies to promote their products. Future research should therefore look at Instagram as a microblogging platform and invest which factors make Instagram users suitable to endorse different product types.
Literature
Ajzen, I. (1991). The theory of planned behavior. Organizational behavior and human decision processes, 50(2), 179-211.
Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and review of empirical research. Psychological Bulletin, 84(5), 888-918.
Andrews, J.C., Durvasula, S., & Akhter, S.H. (2001). A framework for conceptualizing and measuring the involvement construct in advertising research. Journal of Advertising, 19(4), 27-40.
Arnoldi, J. (2013). Risk. New York: John Wiley & Sons.
Beck, U., Giddens, A., & Lash, S. (1994). Reflexive Modernization: Stanford Univeristy Press.
Borstorff, P. C., Marker, M. B., & Bennett, D. S. (2007). Online Recruitment: Attitudes and Behaviors of Job Seekers. Journal of Strategic E-commerce, 5(1&2), 1-23.
Bouhlel, O., Mzoughi, N., Ghachem, M.S., & Negra, A. (2010). Online purchase intention: Understanding the blogosphere effect. International Journal of e-Business
Management, 4(2), 37-51.
Boush, D.M., Friestadt, M., & Rose, G.M. (1994). Adolescent skepticism toward TV
advertising and knowledge of advertiser tactics. The Journal of Consumer Research, 21(1), 165-175.
Brooks, R.C. (1957). "Word-of-Mouth" Advertising in Selling New Products. Journal of Marketing, 22(2), 154-161.
Bryman, A. (2012). Social research methods (Vol. 4). New York: Oxford University Press Inc.
Cammaerts, B. (2008). Critiques on the Participatory Potentials of Web 2.0. Communication, Culture & Critique, 1, 358-377.
Campbell, M.C., & Keller, K.L. (1995). When Attention-Getting Advertising Tactics Elicit Consumer Inferences of Manipulative Intent: The Importance of Balancing Benefits and Investments. Journal of consumer psychology, 4(3), 225-254.
Celsi, R.L., & Olson, J.C. (1988). The role of involvement in attention and comprehension processes. Journal of consumer research, 15(2), 210-224.
Cheung, M.Y., Luo, C., Sia, C.L., & Chen, H. (2009). Credibility of electronic word-of-mouth: informational and normative determinants of on-line consumer
recommendations. International Journal of Electronic Commerce, 13(4), 9-38. Chu, S.C., & Kamal, S. (2010). The Effect of Perceived Blogger Credibility and Argument
Quality on Message Elaboration and Brand Attitudes: an Exploratory Study. Journal of Interactive Advertising, 8(2), 26-37.
Cohen, J.W. (1988). Statistical power analysis for the behavioral sciences (Vol. 2). Hillsdale, NJ: Lawrence Erlbaum Associates.
Corcoran, S. (2009). Sponsored Conversations: When it's ok to pay bloggers to post. Retrieved 6th of October, 2014, from http://blogs.forrester.com/sean_corcoran/ 09-03-02-sponsored_conversations_when_it’s_ok_pay_bloggers_post
Corritore, C.L., Kracher, B., & Wiedenbeck, S. (2003). Online trust: Concepts, evolving themes, a model. International Journal of Human-Computer Studies, 58, 737-758. Crowley, A.W., & Hoyer, W. (1994). An Integrative Framework for Understanding
Two-Sided Persuasion. Journal of consumer research, 20, 561-574.
Danaher, P.J., & Mullarkey, G.W. (2003). Factors affecting online advertising recall: A study of students. Journal of Advertising Research, 43(3), 252-267.
Del Guidice, K. (2011). Crowdsourcing credibility: The impact of audience feedback on web page credibility. American Society for Information Science and Technology, 47(1), 1-9.
Dellarocas, C. (2003). The Digitization of Word of Mouth: Promise and Challenges of Online Feedback Mechanisms. Management Science, 49(10), 1407-1424.
Du, H.S., & Wagner, C. (2006). Weblog success: exploring the role of technology. International Journal of Human-Computer Studies, 64(9), 789-798.
Fiore, A.M., Jin, H.J., & Kim, J. (2005). For fun and profit: hedonic value from image interactivity and responses toward an online store. Psychology & Marketing, 22(8), 669-694.
Fireworker, R.B., & Friedman, H.H. (1977). The effects of endorsements on product evaluation. Decision Sciences, 8(3), 576-583.
Fishbein, M., & Ajzen, I. (1975). Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research. Reading, MA: Addison-Wesley.
Forrest, E., & Cao, Y. (2010). Opinions, Recommendations and Endorsements: The New Regulatory Framework for Social Media. Journal of Business & Policy Research, 5(2), 88-99.
Frost, D., Goode, S., & Hart, D. (2010). Individualist and collectivist factors affecting online repurchase intention. Internet Research, 20(1), 6-28.
Garden, M. (2012). Defining blog: A fool's errand or a neccessary undertaking. Journalism, 13(4), 483.
Gefen, D. (2000). E-commerce: the role of familiarity and trust. Omega - The international Journal of Management Science, 28(6), 725-737.
Gefen, D., Karahanna, E., & Straub, D.W. (2003). Trust and TAM in Online Shopping: An Integrated Model. MIS Quarterly, 27(1), 51-90.
Giddens, A. (1991). Modernity and Self-Identity: Stanford University Press.
Giddens, A. (1999). Risk. The second Reith lecture. Retrieved 16/11, 2014, from http:// news.bbc.co.uk/hi/english/static/events/reith_99/week2/lecture2.htm
Gilly, M.C., & Schau, H.J. (2003). We are what we post? Self-presentation in personal web space. Journal of consumer research, 30, 385-403.
Hall, S. (1980). Encoding/Decoding. London: Hutchinson.
Havits, M.E., & Howard, D.R. (1995). How enduring is enduring involvement? A seasonal examination of three recreational activities. Journal of consumer psychology, 4(3), 255-276.
Healy, S. (2001). Risk as social process: the end of "the age of appealing to the facts". Journal of hazardous materials, 86, 39-53.
Hilligoss, B., & Rieh, S.Y. (2008). Developing a unifying framework of credibility
assessment: Construct, heuristics, and interaction in context. Information Processing and Management, 44, 1467-1484.
Homer, P. M., & Kahle, L.R. (1990). Source Expertise, Time of Source Identification, and Involvement in Persuasion: An Elaborative Processing Perspective. Journal of Advertising, 19(1), 30-39.
Hsu, C.L., & Lin, J.C. (2008). Acceptance of blog usage: the roles of technology acceptance, social influence and knowledge sharing motivation. Information & Management, 45(1), 65-74.
Jarvis, J. (2006). Power Couple: How Bloggers and Advertisers Can Finally Hook up. Advertising Age, 77(19), 42.
Kahle, L.R, & Homer, P. (1990). Source Expertise, Time of Source Identification, and Involvement in Persuasion: An Elaborative Processing Perspective. Journal of Advertising, 19(1), 30-39.
