#Sponsored
An experiment on the effects of sponsorship disclosures on Instagram on brand
attitude and the role of cognitive resistance strategies
Myrthe Timmermans Student number: 10550615 Master’s Thesis
Graduate School of Communication Master’s program Communication Science University of Amsterdam
Academic year: 2017-2018 Supervisor: Stephan Winter
Abstract
Influencer marketing blurs the lines between a traditional advertisement and the native environment of social media. Therefore, sponsorship disclosures are necessary means to ensure transparency and avoid misleading consumers. This study aims to research the effects of Instagram-based sponsorship disclosures on brand attitude of the sponsored brand. A between-subjects experimental design with one factor (sponsorship disclosure) with five conditions: 1) simple disclosure, 2) paid partnership disclosure, 3) gifted disclosure, 4) honesty disclosure and 5) no disclosure was used. Data was collected from 214 female Instagram users, who participated via an online between-subjects experiment. Respondents saw a manipulated post of a social influencer promoting a product of a brand, accompanied by one of the disclosures. The results showed a significant difference between the simple disclosure and gifted disclosure on brand attitude, suggesting that the use of the simple disclosure has a more negative effect on brand attitude. The hypothesized mediating effect of counterarguing could not be supported. The hypothesized mediating effect of source derogation could be supported for one of the comparisons. Namely, respondents in the gifted disclosure came up with more positive source derogations than respondents in the simple disclosure. The findings of this research can help marketers and influencers to choose an appropriate sponsorship disclosure to inform consumers about the persuasion attempt, hereby avoiding deception. Future research should look at sponsorship disclosure use of micro influencers, other types of (low or high involvement) products and on other social media channels such as Facebook, Twitter and YouTube.
Keywords; influencer marketing, Instagram, sponsorship disclosures, cognitive resistance strategies, source derogation, counterarguing
Table of contents
INTRODUCTION 4
THEORETICAL FRAMEWORK 7
SPONSORSHIP DISCLOSURE AND ITS EFFECTS ON BRAND ATTITUDE 7
TYPES OF SPONSORSHIP DISCLOSURES 9
THE EFFECT OF DISCLOSURE LANGUAGE ON BRAND ATTITUDE 11
MEDIATION VIA COGNITIVE RESISTANCE STRATEGIES 13
MEDIATION VIA COGNITIVE RESISTANCE STRATEGY: COUNTERARGUING 14
MEDIATION VIA COGNITIVE RESISTANCE STRATEGIES: SOURCE DEROGATION 15
CONCEPTUAL FRAMEWORK 17 METHODS 18 SETTING 18 SUBJECTS 18 DATA COLLECTION 19 PROCEDURE 20
STIMULUS MATERIAL: SPONSORSHIP DISCLOSURE 20
MEASURES 21
RESULTS 23
MANIPULATION CHECK 23
CORRELATION MATRIX 23
TESTING THE HYPOTHESES 25
H1(A-D): THE EFFECT OF SPONSORSHIP DISCLOSURE ON BRAND ATTITUDE 25
H2(A-D): MEDIATION OF COUNTERARGUING 26
H3(A-D): MEDIATION OF SOURCE DEROGATION 28
DISCUSSION 30
THEORETICAL IMPLICATIONS 32
MANAGERIAL IMPLICATIONS 33
LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH 34
CONCLUSION 35
REFERENCES 36
APPENDIX 42
APPENDIX A-SURVEY 42
Introduction
As social media platforms are constantly changing, advertisers continue to search for new, relevant ways to reach consumers. Therefore, brands are increasingly using opinion leaders, who are paid for placing posts in which a brand/product is promoted (Scott, 2015), nowadays labelled as ‘social influencers’. Influencer marketing is defined as ‘’the identification and use of specific key individuals who hold influence over potential buyers of a brand or product to aid in the marketing activities of the brand’’ (Brown & Hayes, 2008, p. 50). Influencer marketing could be a powerful tool, as it is blurring the lines between a genuine endorsement and a paid advertisement.
The use of influencers, such as bloggers and vloggers, has been found to significantly increase consumers’ brand attitude and purchase intentions (Hwang & Jeong, 2016; Lee & Watkins, 2016). Influencer marketing thus seems like a golden ticket for advertising in a new era, but there remains some tension however. Namely, endorsements should be truthful and should reflect honest opinions and beliefs of the endorser, to prevent misleading and misinforming consumers. To avoid deception, the Federal Trade Commission (FTC) proposed the use of sponsorship disclosures. A sponsorship disclosure is defined as ‘’a notification in editorial content, which mentions that the content is sponsored, so consumers are able to recognize the persuasive intent of the content’’ (Boerman, van Reijmersdal & Neijens, 2014, p. 215).
As influencer marketing is seen as a covert advertising strategy (Campbell, Mohr & Verlegh, 2013), it is hard to recognize what is sponsored and what not. Therefore, sponsorship disclosures are necessary labels, designed to make it identifiable for a consumer that it is a persuasion attempt of an advertisement, hereby warning and protecting consumers from being misled or deceived (Hoy & Andrews, 2004). In 2015, the FTC has come up with endorsement
guidelines for influencer marketing. Failing to comply with these rules can lead to fines up to 250,000 dollars, as well as losing credibility from followers (Rodriguez, 2015).
Research shows mixed results regarding the influence of sponsorship disclosures on brand attitude (Boerman & van Reijmersdal, 2016). Namely, some researches show a significant negative effect of disclosures (Evans, Phua, Lim & Jun, 2017) whereas others do not find an significant effect at all (van Reijmersdal et al., 2013). Nevertheless, disclosing the persuasive attempt of the message activates people’s understanding of the persuasive nature of the sponsored content and could mitigate the persuasion (van Reijmersdal et al., 2016).
Research on sponsorship disclosures tends to focus on disclosures versus no disclosures, even though there are many types of disclosures available for marketers and influencers. Language characteristics of disclosures could have differential effects, such as the use of vague terms (Evans et al., 2017). Therefore, the main aim of the study is to research the effects of using different types of sponsorship disclosure languages on brand attitude of the sponsored brand in an Instagram context. Different types of sponsorship disclosures from a social influencer will be compared by means of a between-subjects experiment. There will also be made use of a control condition, making it possible to also compare disclosed sponsorship versus non-disclosed sponsorship effects on brand attitude.
Furthermore, insights in the underlying effects of sponsorship disclosure on persuasion remain scarce (van Reijmersdal et al., 2016). If people know that there is a persuasive intent in the message, this may evoke resistance strategies to mitigate the persuasion attempt. When resistance of a persuasive message takes place, it is very unlikely that attitudes will be favorable (Tormala & Petty, 2002). Cognitive resistance strategies are ‘’a person’s ability to
people use resistance strategies to cope with persuasive content, by means of defending their freedom of choice to accept the content or not. Two important resistive activities are counterarguing and source derogations. Therefore, this research also aims to research if disclosure of sponsored content actives these resistance strategies, thereby also influencing brand attitude. Taken it all together, this thesis will seek to answer: What is the effect of type of sponsorship disclosure on brand attitude and is this effect mediated by cognitive resistance strategies?
Although there is already some research about sponsorship disclosure use on Instagram, it remains unclear what the most effective form of sponsorship disclosure is, benefiting consumers as well as companies and social influencers. Marketeers should choose an appropriate type of sponsorship disclosure, hereby ensuring transparency for consumers and compliance with the rules and regulations of the FTC law. Furthermore, this thesis will help to understand how the use of sponsorship disclosures can help consumers cope with this new form of online advertising. From a theoretical perspective, this research will contribute to the current state of the art on influencer marketing and the use of various sponsorship disclosures on Instagram. Moreover, it will contribute to the literature on information processing of consumers’ as well as understanding how new forms of digital advertising are executed. Lastly, this research will give insights into the underlying effects of sponsorship disclosures on Instagram on consumer attitudes, by looking at the mediating roles of two cognitive resistance strategies, namely, counterarguing and source derogation.
Theoretical Framework
Sponsorship disclosure and its effects on brand attitude
Persuasive intent in non-traditional advertising, such as social media campaigns, is less recognized than in traditional commercials (van Reijmersdal, Smit & Neijens, 2010). By minimizing interruption of the social media environment, consumers can be persuaded without triggering advertising recognition (Lee, Kim & Ham, 2016). In particular, the use of social influencers has a covert nature. Namely, it is a way to prevent consumer resistance as the content does not appear to be from a commercial source (Campbell et al., 2013). The use of sponsorship disclosures is a tool to guarantee fair communication and prevent persuasive attempts where the public is unaware of (Brusse, Fransen & Smit, 2015). Hence, a sponsorship disclosure can aid consumers to realize what the message is actually trying to accomplish (Rozendaal, Lapierre, van Reijmersdal & Buijzen, 2011). This is important for consumers, as it’s often the only piece of information that distinguishes regular social media posts from advertisements (Evans et al., 2017).
Research indicates that sponsorship disclosures can affect behavioral, cognitive and affective consumers outcomes (Evans et al., 2017). However, in the literature review of Boerman and van Reijmersdal (2016), it is stated that there are still inconclusive findings regarding the influence of disclosures on brand attitude. Namely, there are also researches where no significant effect was found (van Reijmersdal et al., 2013; van Reijmersdal et al., 2016). For example, Dekker and van Reijmersdal (2013) compared the effects of endorsement disclosures in a television program on brand attitude. In their experiment, the host of the show promoted a digital video camera of Samsung. They then compared three conditions: a no disclosure, a disclosure of source (by saying the information comes from Samsung) and a
that they are trying to influence the public). Overall, their results showed no significant differences on brand attitude across the three conditions.
However, other research shows that disclosures have a significant negative effect on brand related attitudes (Boerman et al., 2012). Namely, Evans et al. (2017) researched the use of sponsorship disclosure on Instagram. They created an Instagram post in which a Dunkin Donuts drink was promoted by a social influencer, accompanied by a sponsorship disclosure. The respondents were assigned to one of four conditions: a ‘Paid Ad’ condition, a ‘Sponsored’ condition, a ‘SP’ condition (an abbreviation of sponsored) and a no disclosure condition. Their results showed that there was an indirect negative effect of disclosure language on brand attitude. Namely, the effect was mediated through advertising recognition and disclosure memory. Only when people understood and remembered seeing a sponsorship disclosure, there was a negative impact of disclosure language on brand attitude.
Moreover, the study of Boerman et al. (2013) researched how sponsorship disclosures in TV programs affect viewer’s processing of the program content. Participants were shown a television program, in which a sponsored item about sneakers was discussed by the host of the show. Half of the respondents were shown a disclosure (‘this program contains advertising by Alive Shoes’), whereas the other half the respondents were not shown a disclosure. After seeing the program, the respondents were asked to write down their thoughts. The results showed that respondents in the disclosure conditions wrote down less positive brand related thoughts and more negative brand related thoughts than people who did not see a disclosure.
Thus, people tend to resist persuasion attempts when they are recognized (Wei, Fischer, & Main, 2008). Consequentially, when a warning or label informs consumers about the intent of
the message, it can trigger distrusting feelings towards the message and the brand (Boerman et al., 2017). Consumers tend to be critical of the message because they do not want to be manipulated (Boerman et al., 2014). Consumers may feel threatened and will try to resist the persuasion. As the content is trying to persuade them of the product or brand, consumers could counteract this and adapt their attitude negatively. Thus, reminding consumers that the post contains sponsored content could negatively impact brand attitudes (Boerman et al., 2012). Based on these researches, the following hypothesis can be compiled:
Hypothesis 1a: The use of a sponsorship disclosure has a more negative influence on brand attitude than using no sponsorship disclosure.
Types of sponsorship disclosures
As the FTC law posits, when there is a “material connection” between the publisher and anything in an Instagram post, the influencer has to disclose it (FTC, 2017). This entails payments, gifts, free products and substantial discounts. Moreover, the disclosure should be disclosed in a proper way, meaning it should stand out, it should be easy to understand for the average consumer and should not be buried under a lot of other hashtags or involve ambiguous language. When social influencers decide to disclose sponsorship on Instagram, there are different forms and languages they could use.
Hashtags are an important feature of Instagram. Using a hashtag makes it easy for other users to find your photographs and videos (Sheldon & Bryant, 2016). Often, social influencers make use of a simple disclosure, by using a hashtag to indicate their sponsorship such as ‘#collab’ ‘#ad’, ‘#sp’ or ‘#sponsored’ added to the caption of their photo.
In some cases, the social influencer is not paid for publishing the post, but she did receive the product for free from the brand. This if often made visible with a gifted disclosure, communicated via de hashtag ‘#gifted’, indicating that the product was a present from the brand. According to the FTC, this type of post also needs to be disclosed, even if the exchange is gifted products and no monetary compensation is made (FTC, 2017).
Another disclosure social influencers could use is a paid partnership disclosure. The paid partnership disclosure is a new feature of Instagram from 2017 (Ha, 2017). According to Instagram, this will indicate to followers that the user has a commercial relationship with the business partner that’s mentioned, and that they were compensated in some way for placing the post (Instagram, n.d.). Above the post, it shows as ‘Paid partnership with….’.
Alternatively, social influencers could opt for an honesty disclosure. This type of disclosure is currently not used on Instagram but is sometimes used on blogs (Hwang & Jeong, 2016). This type of disclosure highlights that fact that the influencer thinks the product is enjoyable, not only because she is paid to say that. This type of disclosure often contains an indication that the content is based on their own honest opinions (Hwang & Jeong, 2016).
The use of certain disclosure language could be considered as a stronger indicator that the post is sponsored. Clear language could have a more negative attitudinal outcome compared to a no disclosure or language that uses more vague terms (Evans et al., 2017). In this sense, the aforementioned sponsorship disclosures can be seen as a continuum with a hierarchical order, ranging from a subtle disclosure to an obvious disclosure (see Figure 1). These four types of disclosures will be used in the research as they are the most common forms used on Instagram and cover both implicit as well as more explicit strategies.
Figure 1. Taxonomy of sponsorship disclosures.
The effect of disclosure language on brand attitude
Disclosure research has focused on various types of disclosure characteristics, such as the timing, duration and position of the disclosure (Boerman et al., 2015). Moreover, language characteristics of sponsorship disclosures are also examined, providing more or less explicit information to the consumer. For example, Evans et al. (2017) researched the use of sponsorship disclosures of a social influencer on Instagram. They compared three types of disclosures: ‘Sponsored’, ‘SP’ and ‘Paid Ad’. Their results showed that there was an indirect negative effect of disclosure language, as the effect was mediated through advertising recognition and disclosure memory. If people remember the sponsorship disclosure and understand that it is advertising, a strong disclosure has a negative effect on brand attitude. On a more granular level, ‘Paid Ad’ was recognized as sponsored content, whereas ‘SP’ and ‘Sponsored’ were less recognized.
According to the Persuasion Knowledge Model (PKM) from Friestad and Wright (1994), people’s experiences with persuasive messages helps them understand how companies incorporate persuasive intent in their messages. Once consumers identify that a message has a persuasive goal, the meaning of the message is changed from a neutral message to a goal-directed message (Hwang & Jeong, 2016). Consequentially, consumers will have a lower motivation to accept the message and a larger motive to discount the message. Sponsorship
disclosures induce this change of meaning by revealing the persuasive intent in varying degrees, depending on the type of disclosure used. Therefore, it is expected that the stronger the disclosure, the larger the negative effect of sponsorship disclosure will be. Thus, explicitly stating the word ‘paid’ in an Instagram post will have a more negative influence than only stating ‘sponsored’, ‘gifted’ or ‘*this is a sponsored post, but the content is based on my honest opinions’. Therefore, the following hypothesis can be compiled:
Hypothesis 1b: A paid partnership disclosure has a more negative influence on brand attitude than a simple, gifted and honesty disclosure.
A sponsorship disclosure that may have another effect, is the honesty disclosure. As briefly mentioned before, the honesty disclosure is not typically used on Instagram. This form of sponsorship disclosure recognizes the persuasive intent but nuances it by saying that the opinions are based on honest opinions. Therefore, the expression could mitigate the negative effects of sponsorship disclosure as the expression could give the impression that the content is unbiased (Hwang & Jeong, 2016). Hwang and Jeong (2016) tested this form of sponsorship disclosure on blogposts. Their results show that compared to a no disclosure condition, message attitude was more negative when a simple disclosure was used. However, when an honesty disclosure was used, the negative effects on message attitude disappeared. As the gifted disclosure is an ambiguous, confusing form of a simple disclosure, it can be expected that the honesty disclosure will have a less negative effect on brand attitude than these two sponsorship disclosures. Therefore, the following hypothesis can be compiled:
Hypothesis 1c: A simple and gifted disclosure has a more negative influence on brand attitude than an honesty disclosure.
Furthermore, a simple disclosure (#sponsored) and the gifted disclosure (#gifted) both make use of hashtags. However, a simple disclosure is expected to be more recognized as sponsored content, hence the negative influence is expected to be bigger than for the gifted disclosure. Namely, a simple disclosure indicates that it is sponsored, whereas a gifted disclosure only indicates that the influencer received the product for free. Therefore, the following hypothesis can be formulated:
Hypothesis 1d: A simple disclosure has a more negative effect on brand attitude than a gifted disclosure.
Mediation via cognitive resistance strategies
Any type of persuasive appeal evokes some form of resistance within receivers. Namely, people have a fundamental need for autonomy and when someone feels that his or her freedom is being limited, this causes reactance (Jenkins & Dragojevic, 2011). As persuasive messages by definition try to persuade people to hold a certain opinion or perform a behavior, these messages are a threat to autonomy. The Cognitive Response Approach (CRA) posits that people’s response to a counter attitudinal message will determine whether a message is accepted (resulting in persuasion) or rejected (resulting in resistance) (van Reijmersdal et al., 2016). When resulting in resistance, people could use resistance strategies to cope with the persuasive message. As resistance to persuasion contains negative cognitions, this ultimately leads to negative responses to the content as well.
In addition, the Persuasion Knowledge Model (PKM) posits that people have knowledge regarding persuasion and use it when they encounter persuasion attempts in advertising (Hwang & Joeng, 2016). In time, consumers will get a better understanding how to carry out
resistance strategies to defend themselves to persuasion. Consequentially, the use of such strategies has the potential to negatively influence consumers behaviors, such as purchase intention and brand attitude (Nelson, Wood & Paek, 2009; van Reijmersdal et al., 2016). Moreover, the PKM acknowledges celebrity endorsements as a persuasion tactic that people eventually learn to recognize (Boerman, Willemsen & Van der Aa, 2017). To conclude, both the CRA and the PKM acknowledge that the recognition of sponsored content can lead consumers to refute arguments, which could lead to the use of cognitive resistance strategies.
Mediation via cognitive resistance strategy: counterarguing
A resistance strategy often used by consumers is counterarguing (Zuwerink & Cameron, 2003). Counterarguing is ‘‘the generation of thoughts that dispute or are inconsistent with the persuasive argument’’ (Slater & Rouner, 2002, p. 180). It involves rebuttal of arguments of the persuasive message and is a cognitively effortful strategy at undermining influence (Zuwerink & Cameron, 2003). When an advertising message is combined with elements which reveal the persuasive intent of the message, counterarguments may be triggered (Moyer-Gusé & Nabi, 2011).
Van Reijmersdal et al. (2016) tested the effects of disclosing sponsored content in blogs on brand attitude and purchase intentions. Participants were shown a sponsored food blog post with a sponsorship disclosure (‘brand paid for this blog to persuade you’) or without a sponsorship disclosure. Their results showed that participants in the sponsorship disclosure condition activated persuasion knowledge in response to seeing disclosures, which in turn influences the use of counterarguing to reduce the persuasion attempt of the message (van Reijmersdal et al., 2016). Due to the use of counterarguments, the persuasion attempt was mitigated, and the brand attitude was lower than when no sponsorship disclosure was used.
Moreover, as previously explained, Boerman et al. (2013) researched the use of a sponsorship disclosure in a television program, by including a sponsored feature about sneakers. The results showed that the number of positive thoughts about the brand decreased and the number of negative thoughts increased when participants saw a sponsorship disclosure (Boerman et al., 2013). Thus, the use of more explicit disclosure strategies increased the use of counterarguments. Therefore, the following hypotheses can be compiled:
Hypothesis 2a: The use of a sponsorship disclosure will lead to the use of more counterarguing and consequently to a less positive brand attitude than using no sponsorship disclosure.
Hypothesis 2b: A paid partnership disclosure will lead to the use of more counterarguing and consequently to a less positive brand attitude than a simple, gifted and honesty disclosure.
Hypothesis 2c: A simple and gifted disclosure will lead to the use of more counterarguing and consequently to a less positive brand attitude than an honesty disclosure.
Hypothesis 2d: A simple disclosure will lead to the use of more counterarguing and consequently to a less positive brand attitude than a gifted disclosure.
Mediation via cognitive resistance strategies: source derogation
Another resistance strategy often used by consumers is source derogation (Zuwerink & Cameron, 2003). Source derogation is defined as ‘’thoughts noting reasons for discounting the message due to source reliability’’ (Wright, 1975, p. 2). Source derogation is not seen as a cognitively effortful strategy and is aimed at the source (Zuwerink & Cameron, 2003). When
a person is derogated, competence and intelligence are being questioned (Jenkins & Dragojevic, 2011). Sponsorship disclosures are likely to arouse negative source derogations, as it indicates that the influencer is paid for placing the post. Moreover, a disclosure indicates that it is not an authentic opinion of the influencer (Evans et al., 2017), which could negatively affect the credibility of the influencer.
Namely, followers can adapt and form their decisions and actions as a result of the interaction they experience with influencers, seeing their reviews and experiences (Li, Lee & Lien, 2014). These perceptions of the source in turn influence how the viewer evaluates the claim that is made (Bohner, Ruder & Erb, 2002). Thus, if a social influencer posts a positive opinion about a brand or product, most likely followers will like the product as well (Fransen & Fennis, 2014). With the presence of a sponsorship disclosure this could change, as it indicates that the opinion is fabricated. Additionally, claims made by the source are likely to be derogated as well, questioning the influencers’ opinion of a brand/product (Jenkins & Dragojevic, 2011).
Shareef, Mukerji, Dwivedi, Rana and Islam (2017) researched the effects of source derogations of advertisements for Facebook users, and compared three sources: an aspirational reference group, an associative reference group and the actual marketer. Their results showed that the advertisement was more effective when initiated by a regular member of the network than when artificially generated as a commercial statement, as the latter crated negative opinions about the advertisement. Moreover, they showed that informal source derogation was an important driver of successful persuasion. Furthermore, the more it is shown to the consumer that the post is fabricated, the more negative the effect on attitudes (Shareef et al., 2017). In the research of Jenkins & Dragojevic (2011), message derogation
was also negatively related to message attitude. Thus, the stronger the sponsorship disclosure, the more likely people will thus use negative source derogations to evaluate the post. Therefore, the following hypotheses can be compiled:
Hypothesis 3a: The use of a sponsorship disclosure will lead to the use of more negative source derogations and consequently to a less positive brand attitude than using no sponsorship disclosure.
Hypothesis 3b: A paid partnership disclosure will lead to the use of more negative source derogations and consequently to a less positive brand attitude than a simple, gifted and honesty disclosure.
Hypothesis 3c: A simple and gifted disclosure will lead to the use of more negative source derogations and consequently to a less positive brand attitude than an honesty disclosure.
Hypothesis 3d: A simple disclosure will lead to the use of more negative source derogations and consequently to a less positive brand attitude than a gifted disclosure.
Conceptual framework
The main question is visualized in the following conceptual model (see Figure 2).
Methods
Setting
To answer the research question and hypotheses, a cross-sectional experimental research design was conducted. More specifically, a between-subjects experimental design with one factor was used. The factor is ‘type of sponsorship disclosure’ and has five levels: 1) simple disclosure, 2) paid partnership disclosure, a 3) gifted disclosure, 4) honesty disclosure and 5) no disclosure (see Table 1). The sample of the study consisted of at least n = 200 participants, so that in every condition 40 participants could partake in the study. The experiment was programmed in Qualtrics, the available program via the University of Amsterdam. Participants were randomized over the five conditions. The sample consisted of only females, as the used influencer and the product only appeals to this gender. Moreover, the participants had a minimum age of eighteen years, as of that age children do not need parental approval for processing their online data (Ministerie van Justitie, 2001). Lastly, the participant had to have a registered Instagram account, so that they can recognize the Instagram post format.
IV: Type of sponsorship disclosure 1. Simple disclosure 2. Paid partnership disclosure 3. Gifted disclosure 4. Honesty disclosure 5. No disclosure - DV: brand attitude - Mediator: counterarguing - Mediator: source derogation - DV: brand attitude - Mediator: counterarguing - Mediator: source derogation - DV: brand attitude - Mediator: counterarguing - Mediator: source derogation - DV: brand attitude - Mediator: counterarguing - Mediator: source derogation - DV: brand attitude - Mediator: counterarguing - Mediator: source derogation N = 40 N = 40 N = 40 N = 40 N = 40
Table 1. Overview of the conditions and variables of the thesis.
Subjects
In total, 301 people have participated in the experiment. However, 67 respondents were dropped from the analyses due to incomplete surveys. Moreover, only those cases were
older. This accounted in a loss of another 20 participants. This left a total of 214 completed surveys. From the respondents, the age varies from 18 to 49 years, with an average age of 24 years (M = 23.94, SD = 3.51). Educational level varied from a high school diploma to master’s degree. Most participants (n = 99) stated a bachelor’s degree to be their highest completed education. Two randomization checks were conducted for age and education. An ANOVA test showed that participants’ mean age did not differ significantly between the five conditions, F (4, 209) = .99, p = .415. Moreover, the Chi-Square test showed that participants’ educational level did not differ significantly between the five conditions, χ2 (16) = 13.68, p =
.623. Thus in both cases randomization was successful.
Data collection
The experiment itself was conducted as an online study (see Appendix A). An online study is a cost-effective way to reach a lot of participants, as there are no costs for renting a lab, hiring research conductors and equipment (Reips, 2000). Furthermore, it is also time-effective as it allows to let participants fill out the survey at the same time. The recruitment strategy for gathering participants for the research was convenience sampling, gathering subjects close to the researcher. Moreover, snowball sampling was also used. All private (female) contacts of the researcher were contacted via e-mail, WhatsApp and Facebook Messenger and asked whether they have an Instagram account and, if so, they would be willing to participate in the survey. Moreover, the link to the experiment was placed on social media channels such as LinkedIn and Facebook, as well as asking friends and family to distribute the survey on their own social media channels. A small-scale pretest was conducted in order to test the configuration, duration and clarity of the questionnaire. A total of five respondents participated and their feedback was implemented before the start of the main study.
Procedure
The participants firstly saw the factsheet, where the research was described and why they are asked to participate. Furthermore, an informed consent was shown and asked if people consent with the information. The questionnaire started with showing the manipulation (the Instagram post), where 20% saw the simple disclosure, 20% saw the paid partnership disclosure, 20% saw the gifted disclosure, 20% saw the honesty disclosure and 20% saw no disclosure. Thereafter, brand attitude is measured, as well as the mediators counterarguing and source derogation. The survey ends with demographic questions and the manipulation checks. Lastly, the participants are debriefed.
Stimulus material: sponsorship disclosure
Sponsorship disclosure was the manipulated variable. All participants saw an Instagram post of social influencer Song of Style, promoting earrings of the brand APM Monaco. However, whether they saw a disclosure and which type of disclosure was randomized beforehand. More specifically, the participants were randomly assigned to one of five conditions:
Simple disclosure. This is the most common form of sponsorship disclosure on Instagram, by means of using hashtags such as #sponsored, #paid and #ad. For this research, the disclosure was shown as ‘’#sponsored’’.
Paid partnership disclosure. The paid partnership disclosure is a new feature of Instagram from 2017 (Ha, 2017). The disclosure was shown in the header as ‘’Paid partnership with APM Monaco’’.
Honesty disclosure. This type of disclosure highlights that the influencer thinks the product is good, not only because she is paid to say that. The disclosure was shown as ‘’*this is a sponsored post, but the content is based on my honest opinions’’ (Hwang & Jeong, 2016).
Gifted disclosure. For this disclosure, the influencer is not paid to post about the product, but she did receive the product for free from the brand. This disclosure was shown as ‘’#gifted’’.
Control condition. For this condition, none of the above disclosures were shown, only the Instagram post.
Other than the disclosure, all the posts looked the same (see Appendix B). Namely, participants saw the same picture, the same number of likes and the same comments. Moreover, the same text was shown next to the image: ‘’Hanging out in my rose garden that my dad built me. Earrings from @apmmonaco’’.
Measures
Mediator variable: Counterarguing. This variable was measured with three items based on the study of Brusse et al. (2015), such as ‘’Seeing the Instagram post, I was critical of the message’’. The three items were measured on a 7-point scale (1 = totally disagree, 7 = totally agree). The principal component analysis (PCA) showed that the variable consisted of one component. The scale was reliable (α = .84). The higher the score on this scale, the more respondents come up with counterarguments (M = 4.32, SD = 1.39).
Mediator variable: Source derogation. This variable was measured with seven semantic differential items from the article of Jenkins and Dragojevic (2011), which are bipolar adjectives: “I think the source of the message is: stupid-smart, unknowledgeable- knowledgeable, uninformed-informed, unintelligent-intelligent, unqualified-qualified, unreliable-reliable, and inexpert-expert’’ (measured with a 7-point scale). The PCA showed
that the variable consisted of one component. The scale was reliable (α = .92). The higher the score on this scale, the more positive perceptions about the message source (M = 4.34, SD = 1.03).
Dependent variable: Brand attitude. This variable was measured with six semantic differential items from the article of Boerman, van Reijmersdal and Neijens (2015), which are bipolar adjectives: ‘’I think APM Monaco is bad-good, unpleasant-pleasant, unfavorable-favorable, negative-positive, dislike-like, poor quality-high quality’’ (measured with a 7-point scale). The PCA showed that the variable consisted of one component. The scale was reliable (α = .92). The higher the score on this scale, the more positive the attitude is about the brand (M = 5.10, SD = .94).
Demographics and covariates. Three demographic variables were measured in the questionnaire: the age, gender and educational level. Moreover, it was asked if the participant had an Instagram account. If not, the participant was excluded from the analyses. Lastly, it was asked how much time the participant uses Instagram on a typical day.
Manipulation checks. It was checked if the respondents saw specific elements in the Instagram posts, indicating that the influencer works with the brand. Therefore, respondents were asked if they saw one or more of the sponsorship disclosures: ‘#sponsored’, ‘#gifted’, ‘paid partnership’ and ‘*this is a sponsored post, but the content is based on my honest opinions’. Nevertheless, the analyses were run with all respondents, as the chance of giving only the right answers on these four questions is small, and thus would led to the loss of a lot of respondents. Moreover, it was also checked if the participant interpreted it as such, thus that the influencer is working with the brand, by asking ‘’Song of Style was compensated by
APM Monaco for creating the post’’ (1 = yes, 2 = no) and ‘To what extent do you think that the post you saw contained advertising? (1 = definitely no, 5 = definitely yes). Also, the familiarity with the influencer, Song of Style, was also asked.
Results
Manipulation check
In total, 45 participants saw the simple disclosure (21%), 44 saw the paid partnership disclosure (20,6%), 39 saw the honesty disclosure (18,2%), 44 saw the gifted disclosure (20,6%) and 42 saw the no disclosure condition (19,6%). In the end of the survey, most participants indicated that the research was about the evaluation of sponsored content by social influencers on Instagram, but none of the respondents indicated that it was about comparing different forms of sponsorship disclosures. Almost all respondents (90,7%) thought Song of Style was compensated for the post and were on fairly sure that the post contained advertising (M = 4.47, SD = .80), regardless of the condition they were in. Only 153 people correctly recalled the sponsorship disclosure they were assigned to (71,5%). However, since this would account in the loss of a third of all data, all analyses are conducted with all the respondents (n = 214). Later, the same analyses are run again with only the respondents who correctly recalled their condition (n = 153).
Correlation matrix
Prior testing the hypotheses, a correlation analysis was run to check if there are variables that correlate with one another (see Table 2). Significant results are in bold.
Table 2. Results of Pearson’s correlations including means and standard deviations (n = 214).
Note. * Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed).
Variables M SD 1 2 3 4 5 6 7 8 9 10
1. Sponsorship disclosure conditions 2.97 1.43
_ 2. Counterarguing 4.32 1.39 -.11 _ 3. Source derogation 4.34 1.03 -.01 -.43** _ 4. Brand attitude 5.10 .94 .12 -.39** .47** _ 5. Age 23.95 3.51 .06 -.03 .04 -.01 _ 6. Educational level 4.70 1.15 -.03 0.01 .00 -.02 .24 _ 7. Familiarity with influencer .12 .33
.01 -.11 .21** .13 .01 .01 _
8. Compensated .98 .14
.03 .08 -.04 -.01 -.04* .10 -.05 _
9. Advertising certainty 4.47 .80
-.19 .26** -.10 -.16* -.04 .16 .04 -.45** _ 10. Time spent on Instagram 2.93 1.42
Testing the hypotheses
H1(a-d): the effect of sponsorship disclosure on brand attitude
To see if there are differences between the types of sponsorship disclosures and their effects on brand attitude, a series of contrast analyses was conducted. This test was chosen as the independent variable, sponsorship disclosure is categorical and the dependent variable, brand attitude is continuous. For hypothesis 1a, the control group was compared with the four experimental groups (paid partnership, simple, gifted and honesty disclosures) (contrast 1). In the second contrast, the paid partnership disclosure was compared with the simple, gifted and honesty disclosure (contrast 2). In the third contrast, the honesty disclosure was compared with the simple and gifted disclosure (contrast 3). Lastly, in contrast 4, the simple disclosure was compared with the gifted disclosure. The results are reported in Table 3.
Table 3. Contrast analyses for hypotheses 1(a-d) with all of the respondents (n = 214).
IV df t p
Sponsorship disclosure vs. no sponsorship disclosure 209 .60 .552 Paid partnership disclosure vs simple, gifted and honesty disclosure 209 .73 .467 Honesty disclosure vs. simple and gifted disclosure 209 .77 .443 Simple disclosure vs. gifted disclosure 209 1.66 .098
Note. * significant at p < 0.05 level.
There are no significant effects for the contrast analyses when testing the effects of sponsorship disclosures on brand attitude. The same contrasts analyses were run again with only participants who correctly recalled their condition. The results are reported in Table 4.
Table 4. Contrast analyses for hypotheses 1(a-d) with respondents who passed the manipulation check (n = 153).
IV df t p
Sponsorship disclosure vs. no sponsorship disclosure 153 .40 .687 Paid partnership disclosure vs simple, gifted and honesty disclosure 153 .52 .602 Honesty disclosure vs. simple and gifted disclosure 153 .05 .959 Simple disclosure vs. gifted disclosure 153 2.00 .046*
The fourth contrast analysis shows a significant difference between the conditions (t (209) = 2.00; p = .046 (one-tailed)). Participants who saw #gifted in the Instagram post (M = 5.31, SD = .85) had a significant higher brand attitude about APM Monaco than participants who saw #sponsored in the Instagram post (M = 4.83, SD = 1.02). To conclude, hypothesis 1a, 1b and 1c are rejected. Hypothesis 1d, which stated that a simple disclosure has a more negative effect on brand attitude than a gifted disclosure, is supported. This effect only became significant during the analysis with respondents who passed the manipulation check.
H2(a-d): mediation of counterarguing
Furthermore, to see if counterarguing is a mediator in the relation between sponsorship disclosure and brand attitude, a series of mediation analyses using PROCESS from Hayes (2012) was conducted. This test is chosen as the independent variable, sponsorship disclosure, is categorial and the mediator, counterarguing is continuous, and the dependent variable, brand attitude is continuous. The results are reported in Table 5.
Table 5. Summary of mediation analysis via PROCESS for hypothesis testing 2(a-d) when using all respondents.
Sponsorship conditions B SE CI p
Sponsorship vs control (n = 214)
Sponsorship on brand attitude .10 .16 -.21, .42 .533 Sponsorship on counterarguing -.31 .24 -.79, .16 .190 Counterarguing on brand attitude
Indirect effect
Simple, gifted & honesty vs paid partnership (n = 172) Sponsorship on brand attitude
Sponsorship on counterarguing Counterarguing on brand attitude Indirect effect
Honesty vs simple and gifted (n = 128) Sponsorship on brand attitude
Sponsorship on counterarguing Counterarguing on brand attitude Indirect effect
Simple vs gifted (n = 89) Sponsorship on brand attitude Sponsorship on counterarguing Counterarguing on brand attitude Indirect effect -.26 .02 .11 .01 -.25 .12 -.14 -.48 -.28 -.28 -.33 .41 -.26 -.22 .05 .15 .16 .26 .05 .16 .18 .27 .05 .16 .20 .29 .07 .18 -.35, -.18 -.27, .31 -.21, .44 -.48, .50 . -.34, -.16 -.18, .41 -.49, .21 -1.01, -.06 -.39, -.18 -.60, .05 -.72, .06 -.17, .99 -.39, -.13 -.59, .14 < .001* .901 .483 .970 < .001* .440 .438 .079 < .001* .101 .096 .166 < .001* .227
In every model, counterarguing has a significant effect on brand attitude (b = -.26, SE = .05, CI -.35 to -.18, p < .001). Thus, the more someone comes up with counterarguments about the Instagram post, the less positive the brand attitude becomes about the sponsored brand, APM Monaco. No other significant effects were found. The same analyses were run again with only participants who correctly recalled their condition. The results are reported in Table 6.
Table 6. Summary of mediation analysis via PROCESS for hypothesis testing 2(a-d) when only using
respondents who passed the manipulation check.
Sponsorship conditions b SE CI p
Sponsorship vs control (n = 153)
Sponsorship on brand attitude .07 .19 -.30, .44 .711 Sponsorship on counterarguing -.48 .26 -1.01, .04 .068 Counterarguing on brand attitude
Indirect effect
Simple, gifted and honesty vs paid partnership (n = 121)
Sponsorship on brand attitude Sponsorship on counterarguing Counterarguing on brand attitude Indirect effect
Honesty vs simple and gifted (n = 95) Sponsorship on brand attitude Sponsorship on counterarguing Counterarguing on brand attitude Indirect effect
Simple vs gifted (n = 63) Sponsorship on brand attitude Sponsorship on counterarguing Counterarguing on brand attitude Indirect effect -.28 -.07 .11 -.07 -.27 .10 -.00 -.28 -.31 -.10 -.48 .78 -.26 -.28 .05 .18 .21 .30 .06 .20 .21 .28 .07 .19 .23 .30 .10 .24 -.39, -.18 -.42, .28 -.30, .52 -.64, .51 -.38, -.15 -.29, .47 -.42, .41 -.84, .28 -.44, -.17 -.48, .28 -.95, -.01 .18, 1.39 -.44, -.06 -.76, .19 < .001* .705 .587 .825 < .001* .621 .974 .327 < .001* .627 .046* .011* .011* .240
Note. * significant at p < 0.05 level.
Again, the significant effect of counterarguing on brand attitude is seen in every model. Additionally, in the simple versus gifted model, there is a significant direct effect of the conditions on brand attitude, which was also supported for hypothesis 1d when respondents were tested who passed the manipulation check (b = -.48, SE = .23, CI -.95 to -.01, p = .046). Moreover, there is a significant direct effect of the conditions on the mediator, counterarguing (b = .78, SE = .30, CI .18 to 1.39, p = .011). Thus, respondents in the simple disclosure came
up with more counterarguments than respondents in the gifted disclosure. Furthermore, in this model, the significant effect of counterarguing on brand attitude is also present when controlling for the conditions (b = -.26, SE = -.10, CI -.44 to -.06, p = .011). Lastly, when controlling for the mediator counterarguing, the conditions are not significant anymore (b = -.28, SE = .24, CI -.76 to .19, p = .240). As the indirect effect is not significant, the mediation did not take place. To conclude, hypothesis 2a, 2b, 2c and 2d are all rejected.
H3(a-d): mediation of source derogation
Furthermore, to see if source derogation is a mediator in the relation between sponsorship disclosure and brand attitude, again a series of mediation analyses using PROCESS from Hayes was conducted. This test is chosen as the independent variable, sponsorship disclosure, is categorial and the mediator, source derogation is continuous, and the dependent variable, brand attitude is continuous. The results are reported in Table 7.
Table 7. Summary of mediation analysis via PROCESS for hypothesis testing 3(a-d) when using all respondents.
Sponsorship conditions b SE CI p
Sponsorship vs control (n = 214)
Sponsorship on brand attitude .10 .16 .21, .42 .533 Sponsorship on source derogation -.10 .18 -.45, .24 .567 Source derogation on brand attitude
Indirect effect
Simple, gifted & honesty vs paid partnership (n = 172) Sponsorship on brand attitude
Sponsorship on source derogation Source derogation on brand attitude Indirect effect
Honesty vs simple and gifted (n = 128) Sponsorship on brand attitude
Sponsorship on source derogation Source derogation on brand attitude Indirect effect
Simple vs gifted (n = 89) Sponsorship on brand attitude Sponsorship on source derogation Source derogation on brand attitude Indirect effect .43 .15 .11 .15 .43 .05 -.14 -.21 .40 -.06 -.33 -.01 .42 -.33 .06 .14 .16 .18 .06 .15 .18 .20 .08 .16 .14 .22 .09 .18 -.32, .54 -.14, .43 -.21, .43 -.20, .50 .31, .55 -.24, .33 -.50, .22 -.61, .19 -.26, .54 -.37, .27 -.72, .06 -.45, .44 .25, .59 -.68, .02 < .001* .311 .483 .392 < .001* .738 .438 .297 < .001* .732 .096 .984 < .001* .062
As can be seen, in every model source derogation has a significant effect on brand attitude (b = .43, SE = .06, CI -.32 to .54, p < .001). This means that the more someone comes up with positive perceptions about the source (Song of Style), the more positive the brand attitude will become about the sponsored brand, APM Monaco. No other significant effects were found. Again, the same analyses are run again with only participants who correctly recalled their condition. The results are reported in Table 8.
Table 8. Summary of mediation analysis via PROCESS for hypothesis testing 3(a-d) when only using
respondents who passed the manipulation check.
Sponsorship conditions b SE CI p
Sponsorship vs control (n = 153)
Sponsorship on brand attitude .07 .19 -.30, .44 .711 Sponsorship on source derogation -.27 .20 -.66, .13 .182 Source derogation on brand attitude
Indirect effect
Simple, gifted & honesty vs paid partnership (n = 121)
Sponsorship on brand attitude Sponsorship on source derogation Source derogation on brand attitude Indirect effect
Honesty vs simple and gifted (n = 95) Sponsorship on brand attitude Sponsorship on source derogation Source derogation on brand attitude Indirect effect
Simple vs gifted (n = 63) Sponsorship on brand attitude Sponsorship on source derogation Source derogation on brand attitude Indirect effect .40 .18 .11 .07 .38 .09 -.01 -.21 .38 .07 -.48 -.06 .36 -.46 .08 .17 .21 .22 .08 .19 .21 .22 .09 .19 .25 -.11 .22 .24 .26, .54 -.17, .52 -.30, .52 -.36, .50 .22, .54 -.29, .46 -.42, .41 -.63, .22 .19, .56 -.31, .45 -.57, .45 .14, .58 -.89, -.02 -.95, -.01 < .001* .312 .587 .753 < .001* .651 .974 .346 < .001* .718 .046* .811 .002* .041*
Note. * significant at p < 0.05 level.
Again, the significant effect of source derogation on brand attitude is seen in every model. Addiotionally, in the simple versus gifted model a significant direct effect is seen of the conditions on brand attitude, which was also supported for hypothesis 1d when testing respondents who passed the manipulation check (b = -.48, SE = -.25, CI -.57 to .45, p = .046). However, there is no significant direct effect of the conditions on source derogation (b = -.06,
SE = -.11, p = .811). Furthermore, in this model the significant direct effect of source derogation on brand attitude is also present when controlling for the conditions, (b = .36, SE = .22, p = .002). Lastly, the indirect effect of the conditions on brand attitude is significant as well when controlling for source derogations (b = -.46, SE = .24, CI -.95 to -.01, p = .041). As the indirect effect is significant, mediation takes place. To conclude, hypothesis 3a, 3b, 3c are rejected, and hypothesis 3d, which stated that a simple disclosure will lead to the use of more negative source derogations and consequently to a less positive brand attitude than a gifted disclosure is supported. This effect only became significant during the analysis with respondents who passed the manipulation check.
Discussion
This research was conducted in order to answer the central research question: What is the effect of type of sponsorship disclosure on brand attitude and is this effect mediated by cognitive resistance strategies? In this research, participants were presented with a manipulated Instagram post of a social influencer (Song of Style), promoting a brand of earrings (APM Monaco). Participants were shown one of four sponsorship disclosures, namely, a simple disclosure, a paid partnership disclosure, a gifted disclosure, an honesty disclosure or no disclosure. Thereafter, brand attitude and two cognitive resistance strategies were measured – counterarguing and source derogation.
Hypothesis 1a (sponsorship vs control), 1b (paid partnership vs simple, gifted and honesty) and 1c (honesty vs simple and gifted) were all rejected. However, there was a significant effect of sponsorship disclosure on brand attitude when testing hypothesis 1d. This effect only became significant when a subsample was used with respondents who passed the
than a gifted disclosure. Only finding the significant effect for the last hypothesis could be due to that the simple disclosure is the most common form used on Instagram and thus very recognizable for respondents, whereas the gifted disclosure is less used and ambiguous. As disclosures can activate persuasion knowledge, which is shown to be an important factor in sponsorship disclosure research (Boerman & van Reijmersdal, 2016), it could be the case that the simple disclosure activated more persuasion knowledge than the gifted disclosure, consequently negatively influencing brand attitude.
Moreover, the connotation of the word ‘sponsored’ could be experienced as more negative than ‘gifted’, which indicates that someone is given something voluntarily. The results are also in line with the PKM. Namely, it states that celebrity endorsements are a persuasion tactic that people eventually learn to recognize (Boerman et al., 2017). As the simple disclosure was potentially more recognized, consumers could more easily identify the persuasive goal of the message (Hwang & Jeong, 2016). As consumers are more often confronted with the sponsored disclosure than with the gifted disclosure, they thus have a better understanding of the goal of the message and can adapt their attitude accordingly.
Moreover, the results show that the cognitive resistance strategy counterarguing was not a significant mediator in the relation between sponsorship disclosure and brand attitude for Hypothesis 2a, 2b, 2c and 2d. Furthermore, the results show that the cognitive resistance strategy source derogation was not a significant mediator in the relation between sponsorship disclosure and brand attitude for Hypothesis 3a, 3b, 3c. However, for hypothesis 3d, source derogation was found to be a significant mediator, hereby accepting hypothesis 3d. Namely, respondents in the gifted disclosure came up with more positive source derogations which consequently led to a more positive brand attitude than respondents in the simple disclosure.
Again, this effect only became significant when a subsample was used with respondents who passed the manipulation check.
An explanation for the results of the cognitive resistance strategies could lay in the nature of these two concepts. Namely, whereas counterarguing is seen as a cognitive effortful strategy, source derogation is not (Zuwerink & Cameron, 2003). Moreover, counterarguing is mainly focused on the information of the message, whereas source derogation focuses on the source of the message (Zuwerink & Cameron, 2003). As Instagram is a very visual channel, the picture is the most important part of the post and it is probable that all the focus of the respondents was on that part and less on the caption (and thus less on the sponsorship disclosure). Therefore, respondents were possible not motivated enough to critically assess the message, which is also hard to do for a picture than for text.
This is in line with the PKM, which states that consumers will have a lower motivation to process the message when they know the persuasive intent of the message (Hwang & Jeong, 2016). Whereas for source disclosure, as previously stated, the gifted disclosure could be less recognized as sponsored, and could cause for more positive source derogations as the post seems more authentic and thus the source more likeable.
Theoretical implications
This study has contributed to the current state of the art on influencer marketing and sponsorship disclosure research. Moreover, this research strengthens results from previous research of disclosure characteristics on consumer attitudes (Boerman et al., 2015). Namely, when disclosure language clearly communicates that the influencer is paid by the brand (#sponsored), consumers will use this knowledge and will more negatively derogate the
(Evans et al., 2017). Thus, as the research shows that different sponsorship disclosures can have differential effects, it highlights the importance of sponsorship research to determine what the most effective way is to disclosure sponsorships.
Moreover, most disclosure research is done on TV shows, online blogs and Facebook (Boerman et al., 2013; Hwang & Jeong, 2016; van Reijmersdal et al., 2016), thus this research contributes to the scarce amount of research on Instagram (Evans et al., 2017). Furthermore, the results of the research are in line with some of the assumptions of the Persuasion Knowledge Model (Friestad & Wright, 1994). Namely, consumers will learn defensive resistance strategies to cope with persuasion attempts, eventually negatively influencing brand attitude (Nelson et al., 2009). The thesis has also contributed to the underlying effects of sponsorship disclosures on consumer attitudes, as it was requested by van Reijmersdal et al. (2016) to research source derogation within an influencer marketing context.
Managerial implications
Influencer marketing and the use of sponsorship disclosures is still relatively new. Whereas consumer seek transparency, marketeers on the other hand seek a way to advertise that is not too obvious, hence using an organic content-based way, seamlessly fitting in the environment of other Instagram posts. The findings of this research show that using a disclosure does not have a more negative effect on brand attitude than using no sponsorship disclosure. Therefore, to protect consumers from being misled or deceived (Hoy & Andrews, 2004), when brands want to work with social influencers, the use of sponsorship disclosures is advised. As is shown that a sponsorship disclosure can do no harm, using a sponsorship disclosure can inform consumers correctly and will not downgrade the success of the advertising attempt.
This same advice goes for social influencers – to maintain their credibility and to avoid getting high fines from the FTC laws, sponsorship disclosure use is advised. More specifically, it is advised that brands and influencers use disclosures that are understandable and clear. Although this research showed that a clear and easy to understand form of sponsorship disclosure (#sponsored) caused respondents to come up with more less positive source derogations and a less positive brand attitude than respondents who saw an ambiguous form of disclosure (#gifted), marketers are still advised to use wording that is understandable for everyone. Using a hashtag with Gifted is less understandable as Sponsored, thus using such a form could still be considered as misinforming consumers.
Regarding consumers, the advice is to remain critical of the validity social influencers are posting on Instagram. As there is a fine line between honest word-of-mouth and paid advertising, sponsorship disclosures contain crucial information to assess the validity of a recommendation. Consumers should keep in mind that as long as it is not mandatory in every country to disclose sponsorship, to think about the intention of the influencer and what she is trying to accomplish with the post – e.g. persuading followers to buy certain products and brands.
Limitations and directions for future research
Some limitations should be mentioned. Namely, brand attitude was not measured before showing the Instagram post. Thus, no conclusions can be drawn about a potential change in opinion when seeing that the brand is involved with influencer marketing. Future research could include a before and after measure of attitude to better assess the influence of sponsorship disclosures.
Moreover, another limitation lies in the sample of the study. Namely, the sample consisted of a female only sample with an average age of 24 years and is only tested with one influencer and one type of product (earrings). Thus, generalizability should be questioned. While students heavily use Instagram, it thus is still unknown how other populations would respond.
Future studies could research with non-student samples and could test other types of products and use other influencers, such as micro influencers. Furthermore, the way counterarguing is measured could be a named a limitation as well. Namely, self-reported measures of counterarguing can be unreliable (Brusse et al., 2015). The use of other measures could give insights if the counterarguments relate to the source of the message, the sponsored brand, the advertisement or persuasion in general. Lastly, one should keep in mind that the results of this study could potentially only apply to Instagram. Therefore, future research should also test sponsorship disclosures on other social media channels, such as Facebook, Twitter and YouTube.
Conclusion
There remains criticism about the covert nature of advertisements of social influencers, as it’s hard to recognize the persuasive intent. Therefore, sponsorship disclosures should be used to inform consumers, especially when the paid nature is disguised. A lack of empirical research still left questions how consumers should be informed. This research showed that the use of sponsorship disclosures is not inevitably harmful for companies or social influencers. Moreover, the study showed that the use of clear, unambiguous language causes consumers to argue less positive source derogations, which in turn negatively influences brand attitude. To comply with regulations and to keep consumers well informed, unambiguous language use is still advised.
References
Boerman, S. C., & van Reijmersdal, E. A. (2016). Informing consumers about hidden advertising. A literature review of the effects of disclosing sponsored content. In P. De Pelsmacker (Ed.), Advertising in new formats and media: Current research and implications for marketers, (pp. 115-146). London, UK: Emerald Group Publishing.
Boerman, S. C., van Reijmersdal, E. A., & Neijens, P. C. (2012). Sponsorship disclosure: Effects of duration on persuasion knowledge and brand responses. Journal of
Communication, 62(6), 1047–1064.
Boerman, S. C., van Reijmersdal, E. A., & Neijens, P. C. (2013). Appreciation and effects of sponsorship disclosure. In S. Rosengren, M. Dahl�n, & S. Okazaki (Eds.), Advances in advertising research: Vol. IV. The changing roles of advertising (pp. 273-284). Wiesbaden, Germany: Springer.
Boerman, S. C., van Reijmersdal, E. A., & Neijens, P. C. (2014). Effects of sponsorship disclosure timing on the processing of sponsored content: A study on the effectiveness of European disclosure regulations. Psychology & Marketing, 31(3), 214-224.
Boerman, S. C., van Reijmersdal, E. A., & Neijens, P. C. (2015). How audience and
disclosure characteristics influence memory of sponsorship disclosure. International Journal of Advertising, 34(4), 576–592.
Boerman, S. C., Willemsen, L. M., van der Aa, E. P. (2017). “This post is sponsored”: Effects of sponsorship disclosure on persuasion knowledge and electronic word of mouth in the context of Facebook. Journal of Interactive Marketing, 38(1), 82-92.
Bohner, G., Ruder, M., & Erb, H. P. (2002). When expertise backfires: Contrast and assimilation effects in persuasion. British Journal of Social Psychology, 41(4), 495-519.
Brusse, E. D., Fransen, M. L., & Smit, E. G. (2015). The more you say, the less they hear: The effect of disclosures in Entertainment-Education. Journal of Media Psychology: Theories, Methods, and Applications, 27(4), 159-169.
Campbell, M. C., Mohr, G. S., & Verlegh, P. W. (2013). Can disclosures lead consumers to resist covert persuasion? The important roles of disclosure timing and type of response. Journal of Consumer Psychology, 23(4), 483–495.
Dekker, K., & van Reijmersdal, E. A. (2013). Disclosing celebrity endorsement in a television program to mitigate persuasion: How disclosure type and celebrity credibility
interact. Journal of Promotion Management, 19(2), 224-240.
Brown, D., &, Hayes, N. (2008). Influencer marketing: Who really influences your customers? Oxford, UK: Butterworth-Heinemann.
Evans, N. J., Phua, J., Lim, J., & Jun, H. (2017). Disclosing Instagram influencer advertising: The effects of disclosure language on advertising recognition, attitudes, and behavioral intent. Journal of Interactive Advertising, 17(2), 1-12.
Fransen, M. L., & Fennis, B. M. (2014). Comparing the implicit and explicit resistance induction strategies on message persuasiveness. Journal of Communication, 64(5), 915-934.
Friestad, M., & Wright, P. (1994). The Persuasion Knowledge Model: How people cope with persuasion attempts. Journal of Consumer Research, 21(1), 1–31.
Federal Trade Organization (2017, September). The FTC’s Endorsement Guides. Retrieved from https://www.ftc.gov/tips-advice/business-center/guidance/ftcs-endorsement-guides
Hoy, M. G., & Andrews, C. (2004). Adherence of prime-time televised advertising disclosures to the ‘clear and conspicuous’ standard: 1990 versus 2002. Journal of Public Policy and Marketing, 23(2), 170–182.
Ha, A. (2017, November 7). Instagram pushes more influencers to adopt its new format for sponsored posts. Retrieved from https://beta.techcrunch.com/2017/11/07/instagram-pushes-
more-influencers-to-adopt-its-new-format-for-sponsored-posts/18927851_106340099985831_759078861899038720_n/
Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling [White paper]. Retrieved from http://www.afhayes.com/ public/process2012.pdf
Hwang, Y., & Jeong, S. H. (2016). “This is a sponsored blog post, but all opinions are my own”: The effects of sponsorship disclosure on responses to sponsored blog posts. Computers in Human Behavior, 62(1), 528-535.
Instagram. (n.d.). Branded Content on Instagram. Retrieved from https://help.instagram.com/116947042301556
Jenkins, M. & Dragojevic, M. (2011). Explaining the process of resistance to persuasion: A politeness theory-based approach. Communication Research, 40(4) 559–590.
Lee, E., & Watkins, B. (2016). YouTube vloggers’ influence on consumer luxury brand perceptions and intentions. Journal of Business Research, 69(12), 5753–5760.
Lee, J., Kim, S., & Ham, C. (2016). A double-edged sword? Predicting consumers’ attitudes toward and sharing intention of native advertising on social media. American Behavioral Scientist, 60(12), 1425-1441.
Li, Y. M., Lee, Y. L., & Lien, N. J. (2014). Online social advertising via influential endorsers. International Journal of Electronic Commerce, 16(3), 119-153.
Ministerie van Justitie. (2001). Wet bescherming persoonsgegevens (Wbp). Retrieved from http://wetten.overheid.nl/BWBR0011468/2016-01-01
Moyer-��”, E., & Nabi, R. L. (2011). Comparing the effects of entertainment and
educational television programming on risky sexual behavior. Health Communication, 26(5), 416-426.
