''This blog post is sponsored''
Effects of sponsorship disclosures on message processing, through persuasion
knowledge and the role of perceived parasocial relationship in this.
Melanie de Looper 10113703
Research Master Communication Science Master's thesis
Dr. G. van Noort
Graduate School of Communication University of Amsterdam
2
Abstract
This study investigates the presence of disclosures in Dutch sponsored blog posts and the effects of disclosures and disclosure location in a sponsored blog post on the processing of the blog post, trough persuasion knowledge. It was expected that a disclosure (vs. no disclosure) and a disclosure above (vs. beneath) the sponsored blog post result in more persuasion knowledge, which subsequently leads to more critical processing. Besides, it was expected that a disclosure above the sponsored blog post is more effective for blog readers that
experience a weaker relationship with the blogger and a disclosure beneath the sponsored blog post is more effective for blog readers that experience a stronger relationship with the blogger. The results show an effect of a disclosure on the critical processing of the blog posts via persuasion knowledge, but disclosure location did not influence this effect. In addition, the results show that this did not differ for blog readers with different levels of perceived parasocial relationship with the blogger. This study shows that a disclosure activates persuasion knowledge and critical processing and stresses the importance of disclosures in sponsored blog posts. The results are useful for legislators to improve the guidelines for disclosures.
Central concepts: blog, disclosure, critical processing, persuasion knowledge, perceived
persuasive intent, parasocial relationship, blogger, blogvertorial, sponsored content
Introduction
When considering the purchase of a product or service, a large amount of consumers searches for product information posted on the internet by other consumers (Chiu & Ho, 2009). Blogs,
3
or ''interactive websites with frequently updated posts that are archived in reverse
chronological order'' (Mutum & Wang, 2011, p. 249), are an important source to provide this kind of product information (Armstrong, 2006). One of the reasons is that bloggers frequently share opinions about brands or products in their blog posts (Armstrong, 2006). The
experiences, recommendations and opinions of influential bloggers about products, services or brands are, thus, useful information for consumers when thinking about purchasing a product or service.
However, sometimes companies or advertising agencies offer bloggers a compensation in exchange for posting a blog post about a specific brand or product (Armstrong, 2006). These so called 'blogvertorials' are a form of sponsored content or ''the intentional inclusion of a brand or product in editorial content'' (van Reijmersdal, Neijens & Smit, 2007, p 403). It is not always clear for the blog reader that the blog post is sponsored. Sponsored messages are easily perceived as editorial content by the reader and the commercial intent is, therefore, less clear (Buijzen, van Reijmersdal & Owen, 2010; Calvert, 2008; Russel, 2002; Wright,
Friestad, & Boush, 2005). For sponsored blog posts it is even more difficult to notice the commercial intent because sponsored blog posts are a relative new way of advertising.
Because it is hard for the receiver to distinguish between commercial and editorial content, guidelines for sponsored blog posts in the Netherlands are established in February last year (The Dutch Advertising Code, 2014). According to these guidelines the commercial motives of sponsored blog posts should be disclosed. However, it is unclear whether bloggers actually use disclosures and if the disclosures they use are in line with the guidelines.
Therefore, the first aim of this study is to investigate whether Dutch bloggers disclose their sponsored blog posts and what kind of disclosures they use in their sponsored blog posts.
4
RQ: Are Dutch bloggers using disclosures and what type of disclosures do they use in their sponsored blog posts?
It is argued that bloggers became increasingly more influential the last couple of years (Ip & Wagner, 2008) and that the power of sponsored blog posts is increasing (Armstrong, 2006). One of the explanations for this increasing power is that in sponsored blog posts the brand or product in question is endorsed by the blogger, who is perceived as an autonomous person. According to Kirby and Marsden (2006) people will accept corporate information more easily when presented by such an autonomous person than when the information is presented by the organization itself. However, it is questionable whether the insertion of disclosures changes the power of sponsored blog posts.
Adding a disclosure possibly changes the power of sponsored blog posts because the purpose of a disclosure is exposing the persuasive intent of the sponsored message (the Dutch Advertising Code, 2014). Studies that investigated the effects of disclosures in traditional media show that recognition of the persuasive intent of a message leads to a more critical style of processing that message (Boerman, van Reijmersdal & Neijens, 2012; Boerman, van Reijmersdal & Neijens, 2014; Wentzel, Tomczak & Herrmann, 2010). Until now, very few studies focused on the effects of sponsorship disclosures in new media, such as blogs (Abendroth & Heyman, 2012; Campbell, Mohr & Verlegh, 2013; Carl, 2008). In addition, these studies did not focus on the effects of disclosures on underlying processes when coping with a sponsored message. Therefore, the second aim of this study is to investigate the effects of sponsorship disclosures on blogs on the recognition of the persuasive intent and the
processing of the message.
In the guidelines of the Dutch Advertising Code (2014), several recommendations are provided about what should be included in a disclosure. It is, for example, clearly stated in the
5
guidelines what the disclosing message should mention (i.e., in the disclosure the relationship between the blogger and the advertiser should be clarified, but also the brand name, product name and the receiving of a compensation). Besides, several recommendations are made according to the appearance of the disclosure. For example it is stated in the guidelines that the disclosure should be in a colour that is in contrast with the background colour. However, in the guidelines it is not clearly stated what the location of a disclosure should be.
It is surprising that the guidelines of the Dutch Advertising Code (2014) do not include information about where the disclosure should be placed, because it is arguable that the
desired effect of the disclosure is influenced by the location of the disclosure. Results of previous studies on the location or timing of disclosures are not uniform; A study from Campbell, Mohr & Verlegh (2013) showed that only a disclosure after the sponsored content had the desired effects, while a study of Boerman, van Reijmersdal and Neijens (2014) showed that only a disclosure before the sponsored content had the desired effect. Therefore, the third aim of this study is to investigate the effect of the location of a disclosure on the recognition of the persuasive intent and the processing of the message.
Previous research on persuasive messages showed that the relationship with the
communicator of the persuasive message is essential for the effectiveness of the message (van Noort, Antheunis & van Reijmersdal, 2012; Verlegh, Verkerk, Tuk & Smidts, 2004). Blogs are often very personal and blog readers easily develop a relationship with the blogger (Colliander & Dahlén, 2011). Other factors, such as regularly visiting the blog, result in a stronger experienced relationship with the blogger which makes blog readers trust the blogger more (Colliander & Dahlén, 2011). Blog readers who experience a stronger relationship with the blogger are possibly less susceptible to a disclosure before the persuasive message, than after the persuasive message because they see the disclosing message as a confirmation of their beliefs that the blogger is honest and trustworthy (Jacks & Devine, 2000). In contrast, a
6
disclosing message afterwards could make these blog readers feel deceived by the blogger because the blogger turned out to be not as honest as they thought. Therefore, the fourth aim of this study is to investigate if the strength of the perceived relationship with the blogger influences the effects of disclosure location on the recognition of the persuasive intent and the processing of the message.
RQ: What is the effect of disclosed sponsored blog posts on critical processing of the
message via the activation of persuasion knowledge and what is the role of disclosure location and perceived relationship with the blogger in this?
The research questions of this study will be answered by using a multi-method
approach. To answer the first research question, a content analysis of the most popular Dutch blogs will be conducted. Answering the first research question contributes to the existing literature because the results of this study will give insight in the presence of disclosures in Dutch sponsored blog posts. Until now, there is no research done on disclosure usage by bloggers. The results will add to the knowledge about disclosure usage and, therefore, also contribute to the knowledge about the effectiveness of the guidelines of institutions such as the Dutch Advertising Code.
The second research question will be investigated by means of an experiment. Answering this research question will contribute to the existing literature in several ways. First, a couple of studies focused on the effects of disclosure characteristics in traditional media (Boerman, van Reijmersdal & Neijens, 2012; Boerman, van Reijmersdal & Neijens, 2014). In addition, a few studies have investigated the effects of sponsorship disclosures in new media, like blogs, but these studies have only focused on the effects of disclosures on consumer responses like brand attitude and purchase intention (Abendroth & Heyman, 2012;
7
Campbell, Mohr & Verlegh, 2013; Carl, 2008). However, this study focuses on the effects of disclosures in new media (i.e., blogs) on underlying mechanisms, that lead to such consumer responses. Therefore, the results of this study will contribute to the existing scientific
knowledge about disclosure effects on underlying mechanisms in new media.
Second, studies on sponsorship disclosures in new media and social media have only focused on the main effects of sponsorship disclosures (Campbell, Mohr & Verlegh, 2013; Carl, 2008). However, it is also important to investigate if and how receiver characteristics, such as perceived relationship with the communicator, influence this. Therefore, the results of the current study contribute to the literature by examining possible moderating effects of receiver characteristics on the effects of disclosures and disclosure location.
Theoretical Framework
Blogs and sponsored blog posts
For this study blogs are defined as frequently updated websites, where content about subjects of the authors interest is posted in reverse chronological order (Mutum & Wang, 2011). There are a lot of different types of blogs. For example, a blog can be established out of personal motives, but an organization can also develop a blog out of commercial motives (Primo, 2010). Besides this, blogs either have an internal focus or an external focus. Bloggers with an internal focus discuss their own activities and occupations (Primo, Zago, Oikawa & Consoni, 2013), for example a dancer that discusses the difficulties of being a professional dancer or a company that discusses its own products. The blogs with an external focus, on the other hand, are about other subjects of the bloggers interest (Primo, Zago, Oikawa & Consoni, 2013), for example about activities or products the blogger loves. This study will focus on the personal,
8
external focused blogs, because sponsored blog posts will be most likely present in personal, third-party blogs with an external focus (Kirby & Marsden, 2006).
Disclosing sponsored blog posts
In existing literature about sponsorship disclosures, disclosures are described in a variety of ways (Boerman, van Reijmersdal & Neijens, 2013; Cain, 2011; Campbell, Mohr & Verlegh, 2013; Dekker & van Reijmersdal, 2013). However, all these descriptions had in common that a sponsorship disclosure should inform receivers about the persuasive intent and commercial motives of a sponsored message with the aim to avoid misleading advertising.
According to the Dutch Advertising Code (2014), there are some rules a disclosure on social media, such as blogs, should conform to. A few examples are: 'if a distributor receives compensation in money or in kind from the advertiser, this must be explicitly communicated' and 'the nature of the relationship should be disclosed clearly and in an easily accessible manner, e.g. by means of layout and/or presentation'. In addition to these rules, the Dutch Advertising Code provides some examples. However, disclosures do not have to be exactly like these examples. Therefore, disclosures can exist in many different forms.
It is unclear if and where bloggers are placing disclosures in their blog posts. Besides, it is not clear what type of disclosures bloggers use and if these disclosures are in line with the
guidelines of The Dutch Advertising Code. No scientific studies have been conducted to investigate the usage of disclosures by bloggers. Therefore the following research question will be addressed in the this study:
RQ1: Are bloggers disclosing their sponsored blog posts and, if so, are these disclosures in line with the guidelines of The Dutch Advertising Code?
9 The effects of sponsorship disclosures
It is important to know whether bloggers are disclosing their sponsored blog posts, because a disclosure helps the receiver recognizing the persuasive intent of a message. A disclosure makes the persuasive intent of a message more clear, while this is originally hidden in
editorial content. If a person is aware of the persuasive intent, this leads to a more critical way of processing the message (Boerman, van Reijmersdal & Neijens, 2014).
Sponsored content, persuasion knowledge and critical processing
Studies on sponsored content in traditional media showed that sponsored content is confusing for the reader and the reader is influenced without noticing it, because the commercial content is interwoven with the editorial content in the post (Buijzen, van Reijmersdal & Owen, 2010). The commercial motives are less clear for the blog readers, which makes it difficult to
recognize the persuasive intent of a sponsored blog post (Boerman, van Reijmersdal & Neijens, 2014; Buijzen, van Reijmersdal & Owen, 2010).
Recognition of the persuasive intent is a component of persuasion knowledge. This is ''knowledge an individual has developed about tactics used in persuasive attempts'' (Friestad & Wright, 1994, p. 1). Persuasion knowledge encompasses knowledge about the sender of a persuasive message, the motives of the sender, the techniques used in the message and the effectiveness of these techniques (Campbell & Kirmani, 2000). Persuasion knowledge consists of two dimensions. The first dimension is a cognitive dimension, also called conceptual persuasion knowledge, that comprises the recognition and understanding of the commercial intent and the persuasive tactics. The second dimension results from the
activation of conceptual persuasion knowledge and is an attitudinal dimension that comprises a more critical way of processing the persuasive message.
10
Recognition and understanding of the persuasive intent and the used tactics are parts of the conceptual dimension, or conceptual persuasion knowledge. An example of the
activation of conceptual persuasion knowledge is that the blog reader realizes that the blogger is paid by brand 'X' to write a blog post about product 'X'. Conceptual persuasion knowledge is the first defence mechanism when confronted with a persuasive attempt (Brucks,
Armstrong & Goldberg, 1988; Kunkel et al., 2004). This knowledge and understanding of the persuasive tactics in turn leads to a more critical way of processing the persuasive message (Brucks, Armstrong, & Goldberg, 1988; Friestad & Wright, 1994). The underlying motivation to critically process the sponsored blog post when the persuasive intent is recognized is explained by Brehm's reactance theory (1966). Brehm (1966) states that a message that has the purpose to change ones behaviour or beliefs is experienced as a threat to one's freedom. Because individuals have the desire to maintain their freedom, they will try to resist
persuasive attempts (Sagarin, Cialdini, Rice & Serna, 2002). However, the receiver only experiences this threat to freedom and the tendency to resist the message when he or she is aware of the persuasive intent of the message (Friestad & Wright, 1994).
The activation of conceptual persuasion knowledge, thus, leads to the activation of attitudinal mechanisms, such as critical processing of the message (Boerman, van Reijmersdal & Neijens, 2012; Boerman, van Reijmersdal & Neijens, 2014; Rozendaal, Buijzen &
Valkenburg, 2012; Wetnzel, Tomczak & Herrmann, 2010). When the persuasive intent of a message is more salient, the receivers of the message will look at the message more
analytically and will think about the message in a more detached way (Escalas, 2007).
Besides, the receiver will process the message in a more evaluative way and will have a more critical attitude towards the message (Boerman, van Reijmersdal & Neijens, 2014).
Several studies have investigated critical processing of a persuasive message in various ways. Previous studies have identified several indicators of critical processing
11
(Engeln, 2005; Rozendaal, Buijzen & Valkenburg, 2012). First, scepticism or suspicious feelings about honesty and credibility of the source are considered to be indicators of a critical processing style (Boerman, van Reijmersdal & Neijens, 2012; Rozendaal, Buijzen &
Valkenburg, 2012). Second, thinking of arguments opposing the central message of the ad, or counter arguing, is studied as an aspect of critical processing (Boerman, van Reijmersdal & Neijens, 2014; Engeln, 2005). Third, a less positive attitude towards the message is thought to indicate a more critical way of message processing (Boerman, van Reijmersdal & Neijens, 2012; Rozendaal, Lapierre, van Reijmersdal & Buijzen, 2011). However, in other studies the concept of critical processing is measured in a more direct way, by measuring how much the receiver criticizes the message (Boerman, van Reijmersdal & Neijens, 2014). In the current study the concept of critical processing will be measured in both a direct way as well as by measuring the above mentioned indicators of critical processing.
Previous research showed that acknowledging the persuasive intent in a short disclosure message in traditional media, activates critical processing of the message (Boerman, van Reijmersdal & Neijens, 2012; Boerman, van Reijmersdal & Neijens, 2014; Campbell & Kirmani, 2000). The disclosure functions as a cue for commercial related concepts and makes the receiver more aware of the commercial and persuasive intent of a message, which in turn leads to a more critical way of processing of the message (Boerman, van Reijmersdal & Neijens, 2012, Boerman, van Reijmersdal & Neijens, 2014). It is
therefore expected that a disclosure in a sponsored blog post has a positive effect on the activation of persuasion knowledge, which will result in a more critical way of processing the message, more scepticism and suspicious feelings, more counter arguing and a less positive attitude towards the blog post.
12
H1: A disclosure has a more positive effect on the recognition of the
persuasive intent (i.e., conceptual persuasion knowledge) than no disclosure, which in turn has a positive effect on the critical processing of the blog post (a), scepticism and suspicious feelings towards the blog post (b), counter arguing (c) and a negative effect on attitude towards the blog post (d).
Disclosure before versus after the sponsored blog post
It is argued that disclosure location, or timing, influences the effects of a disclosure on the recognition of the persuasive intent and the critical processing because it possibly influences the moment the blog reader sees the disclosure. In this study it is assumed that blog readers see a disclosure above the sponsored blog post before they read the actual blog post and that blog readers see a disclosure beneath the sponsored blog post after they read the actual blog post. In that respect, a disclosure above the sponsored blog post possibly has a different effect on the activation of persuasion knowledge than a disclosure beneath the sponsored blog post. On the one hand, a disclosure before the sponsored blog post could raise the illusion of invulnerability in the receiver. Because of the disclosure, the blog reader feels somewhat protected and feels like he or she is invulnerable for the commercial intent in the blog post (Cain, Loewenstein & Moore, 2011; Sagarin, Cialdini, Rice & Serna, 2002). The blog reader is, therefore, probably less motivated to defence oneself against the persuasive attempt, persuasion knowledge will possibly not be activated and the receiver will not process the sponsored message in a critical way (Sagarin, Cialdini, Rice & Serna, 2002). On the other hand, a disclosure before the sponsored content in the blog post could function as a prime of advertising related, commercial concepts (Bennett, Pecotich & Putrevu, 1999) and therefore the receivers are more likely recognize the persuasive intent of the blog post. Such a prime
13
increases the amount of attention for the sponsored content in the blog post, which causes the receiver recognizes the post more easily as advertising.
In contrast, a disclosure after the sponsored blog post possibly leads to the activation of persuasion knowledge in hindsight. According to Boerman, van Reijmersdal and Neijens (2014), receivers feel deceived when they realize afterwards that the message had a
commercial purpose when they thought it had an entertaining or informational purpose. Besides the change of meaning principle occurs. This phenomenon is explained by a shift in the meaning of the information presented in the message, in this the sponsored blog post, because the viewer becomes more aware of the source's intention to influence him or her (Friestad & Wright, 1994). This awareness results in the activation of persuasion knowledge. A disclosure afterwards could, thus, affect the way the receiver thinks about the message and what the receiver thinks about how the message have possibly influenced them (Petty,
Wegener & White, 1998). A study of Campbell, Mohr and Verlegh (2013) showed that when a disclosure is shown after the sponsored message, receivers infer greater influence of the message and experience more feelings of being undesirably influenced than when a disclosure was shown before the persuasive message.
However a study of Boerman, van Reijmersdal and Neijens (2014) showed that only a sponsorship disclosure before the sponsored content resulted in more recognition of the persuasive intent and a more critical way of processing the message. It is, therefore, expected in this study that a disclosure before the sponsored blog post will result in more recognition of the persuasive intent and subsequently a more critical way of processing the message than a disclosure after the sponsored blog post. Therefore it is important to know where bloggers are placing their disclosures.
14
H2: A disclosure placed before the sponsored blog post has a more positive effect on the recognition of the persuasive intent (i.e., conceptual persuasion knowledge), than a disclosure placed after the sponsored blog post which in turn has a positive effect on the critical processing of the blog post (a), scepticism and suspicious feelings towards the blog post (b), counter arguing (c) and a negative effect on attitude towards the blog post (d).
Perceived relationship with the blogger
The effects of disclosure location are possibly influenced by the way the receiver thinks about the communicator of the disclosure, in this case how the blog reader thinks about the blogger. An explanation for this is that some blog readers experience a parasocial relationship with the blogger (Colliander & Dahlén, 2011). A parasocial relationship is the illusion of an intimate and personal relationship which occurs between a person that appears in some form of media and a viewer of that medium. (Rubin & McHugh, 1989, p. 3). This parasocial relationship makes the blog reader experience some kind of connection with the blogger the blog reader would normally have with a good friend (Colliander & Dahlén, 2011). Therefore, blog readers could see the bloggers as sincere, trustworthy and credible communicators (Verlegh, Verkerk, Tuk & Smidts, 2004; van Noort, Antheunis & van Reijmersdal, 2012).
It is easy for blog readers to develop a parasocial relationship because the visitors of the blog get a lot of insight in the daily life of the blogger via the blog posts (Colliander & Erlandson, 2013). Some blog readers experience a stronger parasocial relationship with the blogger than other blog readers, because a parasocial relationship with the blogger becomes stronger if the number of encounters with the media person is higher (Auter, 1992). So
15
friend who is credible, sincere and trustworthy. In this study it is expected that the effects of disclosure location on conceptual persuasion knowledge and critical processing of the message differ for blog readers who experience a stronger parasocial relationship with the blogger and for blog readers who experience a weaker parasocial relationship with the blogger.
On the one hand, a disclosure before the sponsored message is possibly not effective for blog readers who experience a stronger relationship with the blogger and will not activate persuasion knowledge in the blog reader. This is explained in two ways. First, blog readers who experience a stronger relationship with the blogger will be more resistant to counter-attitudinal messages, in this case a disclosure, than blog readers who experience a weaker relationship with the blogger (Jacks & Devine, 2000). A disclosure which exposes the ulterior motives of the blogger conflicts with the receiver's beliefs, since the receiver is convinced the blogger is credible, sincere and trustworthy. The disclosure states that the message was not written as a result of the bloggers own intention, but as a result of a company's request. According to Jacks and Devine (2000) blog readers who experience a stronger relationship with the blogger will try to resist the counter-attitudinal message. As a result, if the disclosure is positioned before the sponsored message, the blog readers with a stronger relationship with the blogger will pay more attention to resisting the counter-attitudinal disclosure message. They will affirm their initial beliefs instead of thinking about the ulterior, commercial motives behind the message. Therefore, persuasion knowledge will less likely be activated for this group of blog readers.
In addition, it could be that blog readers who experience a stronger relationship with the blogger see the disclosure as a confirmation of their beliefs that the blogger is credible, honest and sincere. By showing a disclosure before the sponsored blog post, the blogger does not only show the ulterior motives of the blog post, but also that the blogger is honest about
16
the sponsorship. (O'sullivan, 2003). A disclosure could, thus, contribute to the perceived sincerity and credibility of the blogger for blog readers who already held this belief. As a result, instead of thinking about the ulterior motives and the persuasive intent while reading the blog post, the receiver will pay attention to the confirmation of his or her beliefs and persuasion knowledge will not be activated.
In contrast, a blog reader who experiences a weaker relationship with the blogger will perceive the blogger as less credible, trustworthy and sincere (Verlegh, Verkerk, Tuk & Smidts, 2004; van Noort, Antheunis & van Reijmersdal, 2012). Therefore, these blog readers are probably less resistant to counter-attitudinal messages than blog readers who do
experience a stronger relationship with the blogger (Jacks & Devine, 2000). Because blog readers who experience a weaker relationship will feel less resistance towards the disclosure, they will accept the disclosure more easily. It is, thus, argued that a disclosure before the sponsored message will have a positive effect on the activation of conceptual persuasion knowledge for blog readers who do not experience a stronger relationship with the blogger. Consequently the activation of persuasion knowledge leads to a more critical way of processing the blog post.
On the other hand, a disclosure after the sponsored message could have a stronger effect for blog readers who perceive a stronger relationship with the blogger. The blogger is perceived as more credible by blog readers who experience a stronger relationship with the blogger. Therefore, it is possibly extra disappointing if, after reading the sponsored blog post, it turns out that the blogger was not honest or sincere. Especially because receivers feel undesirably influenced when a disclosure is placed after the sponsored message (Campbell, Mohr & Verlegh, 2013), blog readers who experience a stronger relationship with the blogger will probably not appreciate a disclosure after the sponsored blog post. Blog readers who experience a stronger relationship with the blogger feel deceived by the blogger when they
17
realize that the message they just read had the purpose to persuade them instead of
entertaining or informing them. The change of meaning principle occurs and blog readers realize that the content had a commercial motive. In that way, a disclosure after the sponsored blog post results in the recognition of the persuasive intent for blog readers with a stronger relationship with the blogger.
In contrast, a blog reader who experience a weaker relationship with the blogger will perceive the blogger as less credible, trustworthy and sincere (Verlegh, Verkerk, Tuk & Smidts, 2004; van Noort, Antheunis & van Reijmersdal, 2012). Therefore, these blog readers are possibly already more suspicious without seeing a disclosure and will probably recognize the persuasive intent more easily without a disclosure. It is argued that a disclosure after the sponsored message will have no or little effect, because the disclosure does not tell the blog reader anything new.
H3a-b: The effect of disclosure location on recognition of the persuasive intent (i.e., conceptual persuasion knowledge) is moderated by perceived parasocial relationship, such that (a) a disclosure placed above the sponsored blog post will result in more recognition of the persuasive intent for receivers with a weaker perceived parasocial relationship with the blogger than for receivers with a stronger perceived parasocial relationship with the blogger, whereas (b) a disclosure placed beneath the sponsored blog post will result in more
recognition of the persuasive intent for receivers with a stronger perceived parasocial relationship than receivers with a lower perceived parasocial relationship.
18
Figure 1: Conceptual moderated mediation model
H1 & H2 H1& H2
H3a-b
Method
This study consists of two phases. The first phase includes a content analysis of Dutch blog posts to test RQ1 and RQ2. Blog posts of the most popular blogs within the four biggest blog categories (i.e., fashion, beauty, food and tech blogs) in the Netherlands are coded for several variables. This content analysis should explore whether bloggers are placing disclosures, where bloggers place their disclosures and whether the used disclosures are in line with the guidelines. Besides, the results of this content analysis are used as input to create the stimuli for second phase. The second phase of this study includes a pre-test and a main experiment. The results of the pre-test are used to check for which blogger participants are most divided on the parasocial relationship scale. This is because, ideally, the experiment should include both participants who experience a stronger parasocial relationship with the blogger as well as
Disclosure - Above - Beneath - Control (no disclosure) - Critical Processing
- Scepticism & Suspicious Feelings - Counter Arguing
- Attitude towards the Blog Post Recognition of Persuasive
Intent (i.e., Conceptual Persuasion Knowledge)
Perceived Parasocial Relationship
19
participants who experience a weaker or no parasocial relationship with the blogger. The main experiment was conducted to test H1, H2 and H3.
Content analysis
Data collection. To further explore if and how bloggers are disclosing sponsored blog
posts a content analysis was conducted. To answer the research question, blog posts
originating from blogs belonging to one of the four main blog categories (i.e., food, fashion, beauty and tech) were analyzed. The main categories were determined by using the website www.technorati.com. This website shows the biggest blog categories per country (categories with the highest number of blogs). Subsequently, five of the most popular blogs within each category (with the most unique visitors per month) were selected. Per blog the five most recent blog posts were used for analysis (n=100). Blog posts were collected from June 2014 till October 2014 and each blog post was coded by the researcher of this study. The blog posts were coded for seven main areas: general information about the blog and blogger, general information about the blog post, brands mentioned in the blog post, products mentioned in the blog post, disclaimer, disclosure in the blog post and comments beneath the blog post. Not all information is used for this study, because the content analysis was part of a bigger study for which some other information was needed. The results of this content analysis were used to answer RQ1: Are bloggers placing disclosures in their blog posts and, if so, are they in line with the guidelines? and RQ2: where are bloggers placing disclosures in their sponsored blog posts? Besides, the results were used to create the stimulus material for the experiment. A blog from the category in which sponsorship disclosures are most common was used as stimuli for the experiment.
Results. One of the purposes of the content analysis was to answer RQ1: Are bloggers
20
results from the content analysis show that only 11.4% of all the blog posts (N=100) contain a disclosure, while 50.9% of all the blog posts contains at least one brand name. From the blog posts that contain at least one brand name, 22.4% is disclosed. Although it cannot be said with certainty that all the blog posts that included a brand name are sponsored, the difference is noteworthy.
According to the content analysis, disclosing blog posts is most common on fashion blogs. Of all blog posts on fashion blogs in which at least one brand name is mentioned 27.8% contain a disclosure whereas for beauty blogs this is only 23.5%. For tech blogs 18.8% of the blog posts that include at least one brand name contain a disclosure and only 14.3% of blog posts on food blogs in which at least one brand name is mentioned contain a disclosure. In general, it is concluded that sponsorship disclosures on blogs are not (yet) widely used by bloggers in the Netherlands.
Besides, the disclosures that are used by bloggers are not completely in line with the requirements of the Dutch Advertising Code (2014). According to the Dutch Advertising Code, (2014) a disclosure should contain the brand and product name and in the disclosure the receiving of a compensation should be explicitly mentioned. However, the results of the content analysis show that only 46.2% of the disclosures contain the brand name and only 23.1% contain the product name. In 15.4% of the disclosures the blogger explicitly mentions that he or she received a compensation for the blog post. Disclosures used by bloggers could be improved in order to meet the guidelines of the Dutch Advertising Code (2014). To answer RQ1, it is concluded that sponsorship disclosures are seldom used by bloggers. In some blog categories disclosures are used more than in other categories, but in all categories disclosures are not frequently used by bloggers. Besides, the disclosures that are used by the bloggers are not in line with the guidelines of the Dutch Advertising Code.
21
bloggers placing disclosures in their sponsored blog posts? The results show that in 53.8% of the blog posts that contain a disclosure, this disclosure is placed beneath the text of the sponsored blog post. Only 15.4% of the disclosed blog posts contain a disclosure above the text of the sponsored blog post. The remaining part of the disclosed blog posts contain disclosures elsewhere, for example interwoven in the text of the blog post, beneath a picture or between sentences. To answer RQ2, if a blog post is disclosed this is mostly done with a disclosure beneath the sponsored blog post.
Experimental study
Pre-test. For the experiment it is desired to include both participants who experience a
stronger parasocial relationship with the blogger as well as participants who experience a weaker or no parasocial relationship with the blogger. Because a parasocial relationship most likely exists when the familiarity with a media person is higher (Auter, 1992), it is arguable that the more familiar one is with a blogger, the more likely one has developed a parasocial relationship with that blogger. Therefore the five most popular fashion bloggers in the Netherlands are included in a pre-test where the familiarity with the bloggers is measured. In total 28 participants took part in the pre-test. Responses of the participants could range from 1(very unfamiliar) to 5 (very familiar). As shown in Table 1, the variance on the familiarity with the blogger was highest for the blogger Chloe Sterk (variance = 2.94) which means the scores on familiarity with this blogger are the most scattered in comparison with the scores of familiarity with the other bloggers. It is, thus, more likely that the participants will also be more divided on the parasocial relationship scale for blogger Chloe Sterk, than for the other bloggers. Therefore, a blog post of this blogger will be used as stimuli in the experiment.
22 Table 1
Variance of familiarity with five of the most popular Dutch fashion bloggers
Mean SD Variance
Familiarity Chloe Sterk 1.97 1.71 2.94
Familiarity Diana Leeflang 1.71 1.49 2.21
Familiarity Linda Tol 1.50 1.31 1.71
Familiarity Sabrina Meijer 1.82 1.53 2.33
Familiarity Yara Michels 1.76 1.54 2.37
Design and participants. To test the hypotheses, an online experiment was conducted.
An online experiment is suitable because it fits the target audience of this study, namely individuals who read blogs online. This experiment concerned a one-factor (no disclosure vs. disclosure above the blog post vs. disclosure beneath the blog post) between-subjects design. Participants of this experiment were randomly assigned to the condition without a disclosure, with a disclosure above the sponsored blog post or with a disclosure beneath the sponsored blog post. For this experiment a total of 188 (n = 188) participants were recruited by mail and via social media. The age from the participants ranged from 18 to 63 years old (M = 31.30,
SD = 11.79), and only females could participate in this study because that is the target
audience of fashion blogs. The majority of the participants was highly educated, 68.6% completed a higher professional education or a study at university level (M = 5.73, SD = 1.30). Participating in the experiment took approximately ten to fifteen minutes. The participants that did not complete the online survey were excluded for further analysis.
Procedure. The participants were recruited through an online invitation to participate
in a study on blog posts which included an internet hyperlink. Individuals could take part in this experiment by clicking on the internet link in the invitation. The participants were randomly assigned to one of the three conditions. In the no disclosure condition the
23
participants only saw a print screen of the sponsored blog post from www.chloesterk.nl. In the disclosure above condition the participants saw a print screen of the sponsored blog post including a disclosure above the sponsored blog post. In the disclosure beneath condition the participants saw a print screen of the sponsored blog post including a disclosure beneath the sponsored blog post. The participants were told to read the blog post and click the next button when finished reading. Subsequently, the participants filled out a questionnaire with questions concerning conceptual persuasion knowledge, critical processing of the message and the perceived relationship with the blogger. The questionnaire was identical in all conditions, except in the two conditions with a disclosure a questions was included to control if the participants could recall the disclosure. Lastly, several control questions and demographic variables were asked in all the conditions. Finally, the participants were thanked for their participation.
Stimuli. The stimuli used for this experiment was derived from the blog
'www.chloesterk.nl', that is founded by Chloe Sterk. On this blog she writes about fashion and posts pictures of her outfits and other fashion items. This blog is selected for this experiment based on the results of the content analysis and the pre-test. The blog post used in the
experiment of this study includes a short text about a new denim clothing line of the brand Hennes & Maurits and two pictures of blogger Chloe Sterk wearing some items of the clothing line.
The blog post used for this study is fictional and was created by adjusting an existing blog post of the blog www.chloesterk.nl. In that way, prior exposure to the stimuli was avoided, but the fictional blog post looked like an actual blog posts of www.chloesterk.nl. It was important to avoid prior exposure to the stimuli, because if blog readers have seen the blog post before, it could be that they remembered whether the message was sponsored, which could perhaps influence the possible effects of a disclosure. The original blog post is
24
about a new clothing line of retailer C&A. This post included more pictures of Chloe Sterk wearing the clothes and more text than the fictional blog post created for this study. The fictional blog post is shortened so the participants would not lose their attention during the experiment. The original blog post did not include a disclosure, so this text was added in the fictional blog post. The brand C&A is changed into Hennes & Maurits to make the blog post more dissimilar with the original blog post. Hennes & Maurits was chosen because this is a well known, international brand. The content analysis showed that 76.9 percent of the brands in sponsored blog posts with a disclosure contained a well-known brand, so this should be a realistic representation of the real world sponsored blog posts.
In the fictional blog post that was created for this study the blogger writes about the new clothing line from Hennes & Maurits and about how she went to the store to buy items from the clothing line. Besides, she tells about the outfit she wears inspired on this new clothing line and two pictures are included in the blog post where the blogger wears this outfit. In the first condition without the disclosure this was all the respondent saw. In the second condition a disclosing message was added above the two pictures of the blogger (but beneath the title of the blog post). In the third condition a disclosing message was added beneath the text of the blog post. The disclosing message was the same in both conditions namely: ''This blog post is sponsored by Hennes & Maurits''. This text was chosen because the content analysis revealed that bloggers include the brand name in their disclosures almost half of the times (46.2%) but usually do not include the product name in the disclosure (23.1%) and do not explicitly mention the receiving of a compensation for writing the blog post (15.4%). According to this results, the disclosures used by bloggers in real-life are not completely in line with the guidelines of the Dutch Advertising Code as they advise bloggers to include all the above mentioned elements in their disclosures. Nevertheless, only brand name was included in the disclosure used for this study, because in this study the effects of
25
real-world disclosures were investigated. The look, feel and content of the blog post was in all three conditions the same. In all three conditions the blog post was shown in the context of the blog, so that it was clear for the participants that they were reading a blog post from
www.chloesterk.nl.
Measures
Conceptual persuasion knowledge. According to Rozendaal et al. (2011), conceptual
persuasion knowledge consists of several components. Based on these components four items are used in this study to measure conceptual persuasion knowledge: ''The blog post contained advertising'', ''Hennes & Maurits is behind the blog post'', ''The purpose of the blog post is selling a product'' and '' The purpose of the blog post is to persuade me''. The fifth item
originates from a study of Boerman, van Reijmersal and Neijens (2012): '' The blog post about Hennes & Maurits is advertising''. Participants could answer the questions on a range from 1 (strongly disagree) and 7 (strongly agree; M = 6.21, SD = 1.32). Factor analysis showed the items load on one factor (eigenvalue = 4.26; explained variance = 85.0%; Cronbach's alpha = .95)
Critical processing. Critical processing is measured with two statements that are used
in a study of Boerman, van Reijmersdal and Neijens (2014): ''While reading the message, I criticized what was said about brand X'' and ''While reading the message, I was sceptical toward what was said about brand X''. Participants indicated to what extent they agreed with the statements on a scale from 1(strongly disagree) to 7 (strongly agree; M = 4.41, SD = 1.34). Factor analysis showed the items load on one factor (eigenvalue = 2.42; explained variance = 60.4%; Cronbach's alpha = .77)
Scepticism and suspicious feelings. To measure scepticism and suspicious feelings,
26
trustworthy, convincing, prejudiced. Participants could answer on a scale from 1 (strongly disagree) to 7 (strongly agree; M = 3.54, SD = 1.21). The items originate from a study of Boerman, Tutaj and van Reijmersdal (2013). Factor analysis showed the items load on one factor (eigenvalue = 2.36; explained variance = 59.0%; Cronbach's alpha = .59)
Counter arguing. Counter arguing is measured with a four-item scale including the
items: ''I found myself actively agreeing with the points made in the message'', ''I was looking for flaws in the authors arguments'', ''It was easy to agree with the arguments made in the message'' and ''I found myself actively disagreeing with the author'', The items originate from a study of Nabi, Moyer-Gusé and Byrne (2007) and participants could answer on a scale from 1 (strongly disagree) to 7 (strongly agree; M = 4.26, SD = 1.21). Factor analysis showed the items load on one factor (eigenvalue = 2.19; explained variance = 54.6%; Cronbach's alpha = .71)
Attitude towards the blog post. Attitude towards the blog post is measured with three
semantic differentials: Good/bad, pleasant/unpleasant and favourable/unfavourable, based on a study of Chu and Kamal (2008) (M = 3.79, SD = 1.45). Factor analysis showed the items load on one factor (eigenvalue = 2.38; explained variance = 79.5%; Cronbach's alpha = .87)
Perceived parasocial relationship with the blogger. Perceived relationship with the
blogger is measured with a eleven-item scale. The first part of the questionnaire is a modified version of the audience persona interaction scale used in a study of Auter and Palmgreen (2000) which includes items about parasocial interaction with television program characters. For this study not all items are suitable, so six items are used. Examples of the items are: ''I have the same beliefs and thoughts as blogger X'' and ''I would like to be more like blogger X''. Besides, two items of the parasocial interaction scale were derived from a study of Rubin, Perse and Powell (1985). The items used in the current study are: ''Blogger X makes me feel comfortable, as if I am with friends'' and ''I look forward to new blog post from blogger X''.
27
Last, three items are based on the connectedness scale of Russel, Norman and Heckler (2004): ''I like the way Blogger X writes the blog posts'', ''I buy things that I have seen Blogger X with'' and ''I use terms and language of Blogger X''. Participants could answer on a scale from 1 (strongly disagree) to 7 (strongly agree; M = 3.52, SD = 1.21). Factor analysis showed the items load on one factor (eigenvalue = 6.20; explained variance = 56.4%; Cronbach's alpha = .92).
Control variables. To control for possible confounds, several control variables are
included. First, participants are asked whether they know the blog where they could answer on a scale from 1 (not at all) and 7 (very good) and how often they visit the blog where they could answer on a scale from 1 (never) to 7 (daily or more). Besides, attitude towards Hennes & Maurits was measured on a two-item scale ranging from 1 (very negative) to 7 (very
positive; M = 5.08, SD = 1.71). Factor analysis showed the items load on one factor (eigenvalue = 2.16; explained variance = 72.0%; Cronbach's alpha = .78)
Manipulation check. To check whether the participants actually saw the disclosure,
the participants in the conditions with a disclosure were asked to recall the disclosure (Did you see the following text: ''This blog post was sponsored by Hennes & Maurits'' or in Dutch: ''Deze blog post is gesponsord door Hennes & Maurits'').
Results
Recall of disclosure
Because this study focused on the effects of disclosures, some additional steps were taken to check whether participants were really exposed to the disclosure. At the end of the experiment the participants had to answer the question: ''Did you see the following text in the blog post: 'this blog post is sponsored by Hennes & Maurits'?''. Fifty-six percent of the participants (n = 64) that were exposed to a disclosure indicated that they did not recall the disclosure and four
28
percent of the participants (n = 3) that were not exposed to a disclosure indicated that they did recall a disclosure, these participants were excluded for further analyses. In that way it is more likely the results found in this study are because of differences in disclosure locations, instead of differences in recall of the disclosure. The sample that remained and was used for further analyses consisted of 121 participants (control n = 71, disclosure beneath blog post n = 32, disclosure above blog post n = 19). It was also tested whether there was a difference in disclosure recall between the above condition and the beneath condition to check whether disclosure location influenced recall of the disclosure, χ² (1) = 3.43, p = .06. In the disclosure above condition, 65% of the participants recalled the disclosure, while 48% of the participants in the disclosure beneath condition recalled the disclosure. For this study it is assumed that there was no difference between the above condition and the beneath condition, because the analysis showed only a marginal significance.
Randomization
To check if random assignment to the different groups was successful, analyses were
conducted for the control variables age, education level, familiarity with the blog and attitude towards Hennes and Maurits. The experimental groups did not differ with respect to
educational level, χ² (10) = 10.96, p = .36, familiarity with the blog 'www.chloesterk.nl', F(2, 118) = .22, p = .80 and familiarity with the blogger Chloe Sterk, F(2, 118) = .08, p = .92. However, all three conditions did significantly differ with respect to age, F(2, 118) < 1, ns. In the condition with a disclosure above the sponsored blog post the participants were the
youngest (M = 26.58, SD = 8.42) followed by the condition with no disclosure (M = 30.97, SD = 11.32) and in the condition with a disclosure beneath the sponsored blog post participants were the oldest (M = 35.1, SD = 13.62). The attitude towards Hennes & Maurits did also significantly differ between all three conditions F(2, 118) = 3.69, p = .03. Participants in the
29
condition with a disclosure beneath the sponsored blog post had the most negative attitude towards Hennes & Maurits (M = 4.65, SD = 1.83) followed by participants in de condition without a disclosure (M = 5.23, SD = 1.72). Participants in the condition with a disclosure above the sponsored blog post had the most positive attitude (M = 5.95, SD = .94). Based on these results, it was decided to include attitude towards Hennes & Maurits and age in all analyses as covariates to make sure they do not confound the effects.
Hypotheses testing
To test the two mediation hypotheses (H1 and H2), the Hayes' (2012) PROCESS macro in combination with model 4 was used. This method estimated a simple mediation and generated 95% bootstrap confidence intervals for direct and indirect effects. In this estimation method 5,000 bootstrap samples were used to estimate the bias corrected confidence intervals. Two dummy variables were constructed, the first dummy variable represented the disclosure condition (0 = no disclosure, 1 = disclosure) and the second dummy variable represented the disclosure location condition (0 = beneath, 1 = above).The PROCESS macro in combination with model 4 was used for disclosure and disclosure location independently. Besides, for each dependent variable the PROCESS macro was run separately, so a total of eight analyses were conducted to test the effects of the disclosure (H1) and disclosure location (H2) on the
dependent variables through conceptual persuasion knowledge.
Effects of disclosure. With regard to H1, in which it was proposed that a disclosure
has a positive effect on the conceptual persuasion knowledge, which subsequently will result in more critical processing of the message, more scepticism and suspicious feelings, more counter arguing and a less positive attitude towards the blog post, the PROCESS macro was run four times. The results show that, controlling for age and attitude towards Hennes & Maurits, sponsorship disclosure had a positive effect on conceptual persuasion knowledge (b
30
= .78, p = .001, 95% bias corrected interval [.32, 1.24]). Participants who were exposed to a disclosure had significantly more conceptual persuasion knowledge than participants who were not exposed to a disclosure. The results show that conceptual persuasion knowledge, in turn, had a positive significant effect on critical processing of the blog post (b = .37, p < .001, 95% bias corrected interval [.18, .56]). Conceptual persuasion knowledge also had a positive effect on scepticism and suspicious feelings (b = .34, p < .001, 95% bias corrected interval [.20, .49]), counter arguing (b = .33, p < .001, 95% bias corrected interval [.17, .48]) and a negative effect on attitude towards the blog post (b = -.36, p < .001, 95% bias corrected interval [-.55, -.17]). More conceptual persuasion knowledge thus leads to a more critical way of processing the blog post, more scepticism and suspicious feelings, more counter arguing and a less positive attitude towards the blog post.
Sponsorship disclosure had no direct effect on critical processing of the message (b = .07, p = .78, 95% bias corrected interval [.43, .57]), scepticism and suspicious feelings (b = -.02, p = .92, 95% bias corrected interval [-.40, .36]), counter arguing (b = -.06, p = .76, 95% bias corrected interval [-.47, .34]) and attitude towards the blog post(b = .09, p = .23, 95% bias corrected interval [-.41, .59]). However, sponsorship disclosure had a significant indirect effect through conceptual persuasion knowledge on critical processing (effect = .29, SE = 0.15, 95% bias corrected interval [.07, .69]), scepticism and suspicious feelings (effect = -.27,
SE = .13, 95% bias corrected interval [.07, .58]), counter arguing (effect = .25, SE = .13, 95%
bias corrected interval [.07, .60]) and attitude towards the blog post (effect = -.28, SE = .18, 95% bias corrected interval [-.72, -.04]). The data, thus, support H1 and it is concluded that a disclosure of sponsored content in a blog post increases conceptual persuasion knowledge, which in turn increases the critical processing of the message, scepticism and suspicious feelings, counter arguing and decreases attitude towards the blog post.
31
Effects of disclosure location. With regard to H2, in which it was expected that a
disclosure above the sponsored blog post results in more conceptual persuasion knowledge, which in turn will lead to more critical processing of the message, more scepticism and suspicious feelings, more counter arguing and a less positive attitude towards the blog post, the PROCESS macro was run four times. The results show that disclosure location did not have a significant effect on conceptual persuasion knowledge (b = -.25, p = .56, 95% bias corrected interval [-1.12, .62]). Participants exposed to a disclosure above the sponsored content and participants exposed to a disclosure beneath the sponsored content did not differ in conceptual persuasion knowledge. Because disclosure location had no effect on conceptual persuasion knowledge there is no indirect effect of disclosure location on critical processing of the blog post (effect = -.10, SE = .17, 95% bias corrected interval [-.47, .27]), scepticism and suspicious feelings (effect = -.08, SE = .14, 95% bias corrected interval [-.38, .23]), counter arguing (effect = -.07, SE = .13, 95% bias corrected interval [-.34, .23]) and attitude towards the blog post (effect = -.07, SE = .17, 95% bias corrected interval [-.13, .58]).
Disclosure location also had no significant direct effect on critical processing of the blog post (b = .08, p = .83, 95% bias corrected interval [-.72, .89]), scepticism and suspicious feelings (b = -.13, p = .68, 95% bias corrected interval [-.77, .51]), counter arguing (b = -.24, p = .47, 95% bias corrected interval [-.90, .43]) and attitude towards the blog post (b = -.03, p = .95, 95% bias corrected interval [-.94, .88]). These findings suggest that disclosure location has no effect on critical processing, scepticism and suspicious feelings, counter arguing and attitude towards the blog post. Therefore, the H2 is not supported by the results, and it is concluded that disclosure location has no effect on conceptual persuasion knowledge or critical processing of the blog post.
Perceived parasocial relationship with the blogger. To test the moderated mediation
32
the dependent variables through conceptual persuasion knowledge was moderated by perceived parasocial relationship, the Hayes' (2012) PROCESS macro in combination with model 7 was used. This method estimated moderation and mediation at once and generated 95% bootstrap confidence intervals for direct and indirect effects. In this estimation method 5,000 bootstrap samples were used to estimate the bias corrected confidence intervals. The PROCESS macro in combination with model 7 was run for each independent variable separately, so a total of four analyses were conducted to test H3. The results show that the effect of disclosure location on conceptual persuasion knowledge was not moderated by perceived parasocial relationship with the blogger (interaction effect: b = .24, p = .13, 95% bias corrected interval [-.07, .54]). The effect of a disclosure on conceptual persuasion knowledge did not differ for respondents with higher levels of perceived parasocial relationship with the blogger and respondents with lower levels of perceived parasocial relationship with the blogger. It was also tested whether the indirect effects of disclosure location on the dependent variables were moderated by perceived parasocial relationship with the blogger. The results show that the indirect effect of disclosure location on critical
processing, through conceptual persuasion knowledge, was not moderated by perceived parasocial relationship (effect = .09, SE = .06, 95% bias corrected interval [-.11, .21]), neither was the indirect effect of disclosure location on scepticism (effect = .05, SE = .04, 95% bias corrected interval [-.01, .13]), counter arguing (effect = .04, SE = .03, 95% bias corrected interval [-.07, .12]) and attitude towards the blog post (effect = -.04, SE = .03, 95% bias corrected interval [-.13, .01]). The indirect effects of disclosure location did not differ for respondents with higher levels of perceived parasocial relationship with the blogger and respondents with lower levels of perceived parasocial relationship with the blogger. Therefore, H3 is not supported by the results of this study.
33 Additional analyses
The results of the current study show some unexpected findings, namely that the perceived parasocial relationship with the blogger does not influence the effects of disclosure location on conceptual persuasion knowledge. However, expecting perceived parasocial relationship to not play a role at all in the activation of persuasion knowledge is contradictory to results of previous studies. Therefore, additional analyses were conducted to see whether
perceived parasocial relationship influenced conceptual persuasion knowledge and the dependent variables directly. The results show that perceived parasocial relationship had a significant direct effect on conceptual persuasion knowledge (b = -.28, p < .001, 95% bias corrected interval [-.59, -.18]). Participants with higher levels of perceived parasocial relationship with the blogger had a significant lower conceptual persuasion knowledge than participants with lower levels of perceived parasocial relationship. Perceived parasocial relationship also had a significant direct effect on critical processing of the blog post (b = -.29,
p < .001, 95% bias corrected interval [-.50, -.03]), as well as a significant direct effect on
scepticism and suspicious feelings (b = .40, p < .001, 95% bias corrected interval [.54, -.23]), counter arguing (b = -.45, p < .001, 95% bias corrected interval [-.60, -.28]) and attitude towards the blog post (b = .58, p < .001, 95% bias corrected interval [.43, .82]).
Because perceived parasocial relationship has an effect on conceptual persuasion knowledge, it was also tested whether perceived parasocial relationship has an indirect effect on the dependent variables through conceptual persuasion knowledge. The results show that perceived parasocial relationship also had significant indirect effects via conceptual
persuasion knowledge on critical processing (effect = -.08, SE = .04, 95% bias corrected interval [-.19, -.02]), scepticism and suspicious feelings (effect = -.06, SE = .03, 95% bias corrected interval [-.14, -.01]) and counter arguing (effect = -.05, SE = .03, 95% bias corrected interval [-.13, -.01]). However, perceived parasocial relationship had no significant indirect
34
effect on attitude towards the blog post (effect = .05, SE = .04, 95% bias corrected interval [-.01, .14]). It is concluded that higher levels of perceived parasocial relationship with the blogger decrease conceptual persuasion knowledge, which in turn result in less critical processing of the blog post, less scepticism and suspicious feelings and less counter arguing. Besides, higher levels of perceived parasocial relationship also directly lead to less critical processing, less scepticism and suspicious feelings, less counter arguing and a more positive attitude towards the blog post.
Conclusion and Discussion
This study focused on the effects of disclosures and disclosure location in sponsored blog posts on the recognition of the persuasive intent and the processing of the blog post. The current study is, therefore, the first to provide insight into the effects of disclosures in new media (i.e., blogs) on underlying mechanisms of consumer responses. In addition, this study is the first to give insight in the usage of sponsorship disclosures by Dutch bloggers.
The present study had four main aims. First of all, it was investigated whether bloggers in the Netherlands are using disclosures in their sponsored blog posts and if so, where they place these disclosures and what kind of disclosures they are using. Second, this study aimed to investigate the effects of sponsorship disclosures on the recognition of the persuasive intent and the processing of the message. The third aim was to investigate the effect of disclosure location on the recognition of the persuasive intent and the processing of the message. The last aim of this study was to investigate if the strength of the perceived parasocial relationship with the blogger influences the effects of disclosure location on the recognition of the persuasive intent and the processing of the message.
35
most popular Dutch blogs was conducted. The results of this content analysis show that, despite the new guidelines of the Dutch Advertising Code, bloggers in the Netherlands are using disclosures in only twenty percent of all blog posts in which brands and/or products are mentioned. In addition, the results show that most of the disclosures used by bloggers were placed beneath the sponsored blog posts and very few disclosures were placed above the sponsored blog post. Besides, it turned out that the disclosures used by bloggers are most of the time not in line with the guidelines provided by the Dutch Advertising Code.
Regarding the second, third and fourth aim of the study, an experiment was conducted. As expected, the results of this study show that a disclosure in a sponsored blog post leads to more conceptual persuasion knowledge, which in turn leads to more critical processing, more scepticism and suspicious feelings, more counter arguing and a less positive attitude towards the blog post. This is in accordance with results of previous studies (Boerman, van
Reijmersdal & Neijens, 2012; Boerman, van Reijmersdal & Neijens, 2014; Campbell & Kirmani, 2000). However, it is important to take into account an alternative explanation for these results. The indirect effect of a disclosure on critical processing could also be the result of differences between the experimental groups. In the disclosure conditions, the participants that did not recall the disclosure were not included for further analyses. It could be that there are differences between the group in which participants only saw the blog post and the group in which participants also saw a disclosure and recalled it. It is possible, for example, that participants that did recall the disclosure are more familiar with reading blog posts or maybe more critical towards blogs. In that way, in the disclosure conditions only participants that are more critical are included, while in the control condition also participants that are less critical are included. However, it is unknown whether factors, such as being more familiar with blogs or being more critical towards blogs result in better recall of disclosures. Therefore, further research is needed to investigate the effects of these factors.
36
It was expected that a disclosure above the sponsored blog post results in more conceptual persuasion knowledge and a more critical way of processing the message, than a disclosure beneath the sponsored blog post. However, in contrast to previous research (Boerman, van Reijmersdal & Neijens, 2014; Campbell, Mohr & Verlegh, 2013) and the expectations in this study, the results of this study show that disclosure location did not affect the recognition of the persuasive intent, the processing of the message, scepticism and
suspicious feelings, counter arguing and positive attitude towards the blog post. An alternative explanation for the absence of an effect of disclosure location is the exclusion of participants that did not recall the disclosure. It is unknown whether the same procedure was followed in earlier studies (Boerman, van Reijmersdal & Neijens, 2014). If in previous studies recall of the disclosure was not measured, or participants that not recall the disclosure were not excluded, the results found between disclosure location can also be ascribed to differences in visibility or noticeability of a disclosure. It is arguable that a disclosure before the sponsored content is more visible and will be noticed more easily by the receiver than a disclosure after the sponsored content. If that is the case, the differences in previous studies are possible found because the disclosure after the sponsored content was not noticed by participants. In addition, previous research focused on disclosure location in traditional media, such as television (Boerman, van Reijmersdal & Neijens, 2014). It could be that disclosure location has an effect for traditional media, such as television, because these medium types are more passive.
Therefore, the differences between a disclosure before and after the sponsored content are possibly bigger, the receiver cannot do anything about the sequence of sponsored content and the disclosure. Although it seems logical that most people read from top down, it is possible that the blog reader reads the sponsored content first even though the disclosure is placed above the sponsored content, or reads the disclosure first even though it is placed beneath the sponsored content. This could be another explanation for the absence of an effect of
