Sponsored Instagram Content: The Effects of
Disclosure Warnings on Social Media Influencer
Credibility and Reactance
Student | Celine Beelen, BSc
Student number | 10737987
Graduate School of Communication
Master’s Program Communication Science Domain | Persuasive Communication
Supervisor | dr. Jeroen G.B. Loman
Date of completion | January 30th, 2020
Abstract
The use of influencer marketing has grown in popularity over the past few years. Regulatory
parties have increasingly interfered with brands and social media influencers (SMIs)
engaging in influencer marketing, highlighting the deceptive nature of this covert advertising
technique and providing guidelines on how to disclose sponsored content. Disclosure
warnings help consumers to recognise persuasive attempts and to activate their cognitive
defences. As a result, consumers may experience feelings of reactance towards the content
and the SMI. However, cognitive defences may not be activated if consumers do not
recognise the content’s persuasive intent. This study investigated the effects of disclosure warnings on the credibility of the SMI, recognition of persuasive intent, and experienced
negative affect towards sponsored content. Moreover, the moderating effect of type of SMI
was included in the study. The findings of an online experiment (N=147) indicate that the use
of disclosure warnings increases recognition of persuasive intent and the activation of
persuasion knowledge. Moreover, the type of SMI affects how consumers perceive the SMI’s credibility and how they perceive the credibility of the Instagram post. Furthermore, the study
found an interesting interaction effect for experienced negative affect as a result of SMIs
engaging in sponsored content. Whereas experienced negative affect decreases when
macro-influencers use disclosure warnings, the contradictory shows when micro-influencers
engage in disclosing sponsored content. Results are discussed in the light of SMI credibility
and reactance responses, and elaborating on these findings is recommended for future
Sponsored Instagram Content: The Effects of Disclosure Warnings on Social Media Influencer Credibility and Reactance
Over the past few years, advertising using social media influencers (SMIs), a strategy used
by marketers to employ influential individuals to recommend their products or services, has
grown in popularity. Moreover, it has shown to offer marketers fast and targeted access to
their audiences (Cheung, Luo, Sia, & Chen, 2009; Evans, Phua, Lim, & Jun, 2017). Reason
for the shift away from traditional media advertising is increasing consumer awareness
regarding persuasive advertising techniques employed by traditional media advertisers,
which has resulted in a lack of credibility and trust towards these traditional techniques
(Goldsmith & Clark, 2008; De Veirman, Cauberghe, & Hudders, 2017). Furthermore,
consumers have shown to value and trust the opinions of their peers over the opinions
provided in traditional advertising (Doh & Hwang, 2009; Goldsmith & Clarke, 2008; De
Veirman & Hudders, 2019). This fairly new advertising technique, also known as influencer
marketing, is effective due to the similarities between sponsored content by SMIs and their
genuine content, disguising its commercial nature (De Veirman et al., 2017). Consequently,
the disguised nature of the content makes it hard for consumers to recognise its persuasive
intent. Therefore, influencer marketing provides the perfect method for marketers to establish
effective advertising.
Regulatory institutions such as the Federal Trade Commission (FTC) have concerns
about the deceptive nature of this covert advertising technique, as many consumers fail to
recognise it as advertising (Federal Trade Commission [FTC], 2017). Disguising the
commercial intentions of sponsored content can be harmful to consumers as they are being
misled about the nature of the content, and therefore might be unable to activate cognitive
defences. In order to help consumers recognise commercial content, regulatory institutions
have provided guidelines about disclosing sponsored content. Notably, these guidelines state
their social media, they are obliged to disclose the sponsorship in their content (Federal
Trade Commission [FTC], 2017; Stichting Reclame Code, n.d.).
Prior research on disclosure warnings found that they activate consumer’s persuasion knowledge (PK) and help them to recognise sponsored content (Boerman, Willemsen, & Van Der Aa, 2017; De Veirman et al., 2017; Evans et al., 2017). As a consequence, feelings of
resistance may be evoked towards influencer marketing and might question the source’s credibility. Recent studies mainly focused on disclosure warning effects on ad recognition
and the activation of cognitive PK, i.e. the consumers’ recognition and understanding of persuasive content. However, little research has highlighted the activation of attitudinal PK,
which involves emotions that might surface as a result of disclosure warnings. In order to fill
this gap in literature, this study will include both cognitive as attitudinal PK to measure
reactance, a psychological process that occurs when individuals experience a threat to their
freedom of choice. Furthermore, little research included the role of SMI types when checking
for the effects of disclosure warnings. Therefore, this study will also investigate the role of the
SMIs’ number of followers. This study is scientifically relevant to gain a deeper understanding of how disclosure warnings affect consumers and their evaluations of sponsored content.
This information is essential for advertisers and SMIs, considering it provides them with
insight into how consumers receive sponsored content and how to alter sponsored content to
evade the (effects of) recognition of persuasive intent (RPI). Further, these insights are
important for regulatory institutions to examine the effectiveness of the current legislation and
to see where changes could be made to inform consumers. In order to gain a deeper
understanding of the effects of disclosure warnings, this study will aim to answer the
following research question:
‘To what extent does a ‘sponsored content’ disclosure warning on Instagram posts affect the perceived credibility of the social media influencer (SMI), recognition of persuasive intent and experienced negative affect, and are these effects moderated
Theoretical Background
Influencer Marketing and eWOM
Influencer marketing is a form of advertising in which advertisers use SMIs to
promote their products or services on social media platforms. Central in influencer marketing
are SMIs, individuals that generated a form of ‘celebrity’ capital by creating an authentic ‘personal brand’ on social media platforms, which in turn can be used for advertising purposes (Hearn & Schoenhoff, 2016). Over the past years, brands and advertisers have
increasingly used SMIs to reinforce brand messages. Nowadays, almost four out of five
brands (79%) use Instagram for influencer campaigns, a number that is set to rise in the
coming years (Influencer Marketing Hub, 2018; Schomer, 2019). The main source of the
popularity of influencer marketing is its effectiveness on consumers. There are two main
reasons for this effectiveness, related to source and content.
Firstly, influencer marketing has shown to be effective because consumers feel
similar to SMIs and perceive them to be relatable and approachable (Djafarova, &
Rushworth, 2017; Schouten, Janssen, & Verspaget, 2019). Therefore, SMIs have the power
to influence consumers through sharing their opinions and information online, i.e. electronic
word-of-mouth (De Veirman et al., 2017). Electronic word-of-mouth (eWOM) can be defined
as “any positive or negative statement made by potential, actual, or former customers about a product or company, which is made available to a multitude of people and institutions via
the Internet” (Hennig-Thurau, Gwinner, Walsch, & Gremler, 2004, p.39). In context of consumer decision-making, eWOM has shown to be significantly more effective than
traditional advertising techniques. Since consumers perceive SMIs to be relatable and
approachable, the persuasive intent of their recommendations is disguised, resulting in
higher credibility of the source. Subsequently, the use of SMIs recommendations has shown
to increase consumers’ purchase intentions and positive brand attitudes (De Veirman et al., 2017, Doh & Hwang, 2009; Evans et al., 2017; Goldsmith & Clark, 2008).
Secondly, recommendations by SMIs are considered more credible because of the
content itself. Influencer campaigns are often seamlessly intertwined with the SMI’s own content and can be seen as a form of native advertising, i.e. the use of advertisements
altered to the specific forms and appearance of editorial content (De Veirman et al., 2017;
Wojdynski & Evans, 2016). As sponsored content by SMIs is very similar to their personal
content, its commercial nature is often disguised. Not only does that make it more difficult for
consumers to recognise the content’s persuasive intent, it is also what makes influencer marketing so highly effective (De Pelsmacker & Neijens, 2012; De Veirman et al., 2017; De
Veirman & Hudders, 2019).
Sponsorship disclosure warnings
The deceptive nature of influencer marketing and its effectiveness raises concerns.
Not recognizing influencer marketing as a form of advertising but rather as highly credible
eWOM endorsements might be harmful to consumers, and even raise ethical concerns,
since consumers are being misled about the true nature of the SMI's content. To help
consumers identify the commercial nature of sponsored content by SMIs, regulatory parties
have made an effort to inform consumers about the persuasive intent of sponsored content
and provide marketers and SMIs with guidelines on disclosing sponsored content (Federal
Trade Commission [FTC], 2017); Stichting Reclame Code, n.d.).
The primary technique to disclose sponsored content is by using disclosure warnings
(Boerman, Van Reijmersdal, & Neijens, 2012). Notably, disclosure warnings involve the use
of terms such as ‘advertising’, ‘ad’ or ‘sponsored’ in any form of sponsored content (Federal Trade Commission, 2019). Research has shown that when disclosure warnings are present,
recognition of persuasive intent (RPI) is more likely to occur. Subsequently, disclosure warnings increase ad recognition and the activation of persuasion knowledge (PK), i.e.
Willemsen, & Van Der Aa, 2017; De Veirman et al., 2017; Evans et al., 2017; Friestad &
Wright, 1994).
In literature, PK is believed to exist of two different dimensions: a cognitive and an
affective dimension (Boerman et al., 2012). The cognitive dimension of PK involves
consumers’ recognition and understanding of persuasive content, and can be activated by explicit cues, such as disclosure warnings. As a result, attitudinal PK is triggered, which, e.g.
includes the emotions that consumers experience when being exposed to advertising
(Boerman et al., 2012). However, PK may not be activated if the consumer does not
recognise the persuasion attempt as such (Evans et al., 2017).
To support consumers’ activation of PK, regulatory parties appealed for more standardization of disclosing sponsored content, resulting in the introduction of a
standardized format of disclosure on Instagram in 2017. The format discloses paid
sponsorships as follows: ‘Paid sponsorship with [brand]’, which is placed at the top of both Instagram posts and Instagram Stories (Boerman, 2020).
Effects of disclosure warnings on influencer credibility
An important factor in determining the effectiveness of sponsored content is the
credibility of its source. Source credibility is determined by an individual’s perception of the source’s expertise and the extent to which individuals perceive this person to be a reliable source (Goldsmith et al., 2000; Hovland, Janis, & Kelley, 1966). The more favourable the
consumer’s perception, the more the source is considered to be a credible source of product information (Seno and Lukas, 2007).
Prior studies found that explicit mentioning of sponsored content using disclosure
warnings negatively affects perceived source credibility (Uribe, Buzeta, & Velasquez, 2016;
Wojdynski & Evans, 2016). Furthermore, De Veirman and Hudders (2019) argue that
eWOM message, leaving the consumers to think that the influencer is biased, which affects
the perceived credibility towards the SMI. Accordingly, the following effect is expected:
H1a. The inclusion of a sponsored content disclosure warning with the Instagram post will decrease the perceived credibility of the SMI.
Effects of Disclosure Warnings on Reactance
Prior studies examining disclosure warnings in sponsored content have not explicitly
highlighted the activation of attitudinal PK in earlier studies. Yet, attitudinal PK is an essential
element when considering the experience of psychological reactance. According to Brehm
(1966), persuasive messages aimed to change one’s behaviour and attitudes can be perceived as a potential threat to freedom of choice. Reactance is a motivational state in
which an individual is threatened in its freedom of choice by a persuasive intent, resulting in
the individual trying to restore this freedom by resisting the persuasive message, in turn
reducing the effectiveness of the persuasive message (Brehm and Brehm, 1981). An
important trigger for resistance is the recognition of persuasive intent (RPI), as reactance
only occurs when the individual feels its freedom of choice is threatened or eliminated
(Bilandzic and Busselle, 2013; Evans et al., 2017).
The role of reactance in influencer marketing is interesting for two reasons. One of
the core reasons of the effectiveness of influencer marketing is that it decreases feelings of
reactance, which is feasible through its disguised nature and by building onto the parasocial
interaction SMIs have with their audience (Bilandzic & Busselle, 2013). On the other hand,
reactance can also be triggered to counteract the effectiveness of influencer marketing, as
psychological reactance can be aroused when subjects are forewarned about the source’s persuasive intent (Hass & Grady, 1975). Warning about upcoming persuasive attempts may
cause individuals to recognise its persuasive intent and experience a pressure to re-establish
attitudes towards the sender (Burgoon, Alvaro, Grandpre, & Voulodakis, 2002; De Veirman &
Hudders, 2019; Heilman & Toffler, 1976).
In the context of influencer marketing, disclosure warnings help consumers to
recognise sponsored content, which can trigger an individual’s PK, potentially evoking feelings of reactance towards the sponsored content (Evans et al., 2017). RPI is more likely
to happen when a disclosure warning is included in the message. Therefore, it is
hypothesized that:
H1b. The inclusion of a sponsored content disclosure warning with the Instagram post will increase the recognition of persuasive intent (RPI), as an indication of
increase reactance.
The experience of negative affect (ENA) characterizes the motivational state of
reactance, e.g. experiencing feelings of irritation, anger, annoyance and rage (Brehm, 1966;
Burgoon et al., 2002; Dillard and Shen, 2005). As disclosure warnings in sponsored content
are expected to increase reactance responses, it is hypothesized that:
H1c. The inclusion of a sponsored content disclosure warning with the Instagram post will increase experienced negative affect (ENA), as an indication of increased
reactance.
The Effect of Type of SMI
Prior studies have shown that SMI type can influence consumers’ perception of an SMI’s credibility and likability (De Veirman et al., 2017). SMIs are categorized by five types: nano, micro, meso, macro, and mega-influencers. There are different views in both literature
and practice what numbers indicate the limits for the distinction between the SMI types. In
literature, e.g., Boerman (2020) categorizes micro-influencers as SMIs with < 10,000
followers, meso-influencers as having 10,000 to one million followers and macro-influencers
micro- and macro-influencers, as micro-influencers are considered the largest group of SMIs
and macro-influencers are considered to have a diverse, broad reach (Boerman, 2020).
Micro-influencers will be defined as SMIs with 10,000 to 100,000 online followers on
Instagram, and macro-influencers as influencers with 100,000 to one million followers (Rose,
2019; White, 2019).
Marketers may have different motives to choose for certain types of SMIs for
influencer campaigns. On the one hand, micro-influencers tend to attract followers with
relatable content and are considered experts in their niche with valued opinions. Moreover,
consumers consider micro-influencers as more similar to themselves than macro-influencers
(Domingues Aguiar & Van Reijmersdal, 2018). Therefore, micro-influencers can be valuable
for brands to target specific audiences. On the other hand, macro-influencers attract a larger
audience; however, the engagement with their audience is weaker than the engagement
micro-influencers have with their audience. Macro-influencers can be valuable for brands to
reach a diverse, broad audience and are considered as an authority in their field of expertise.
Furthermore, consumers value macro-influencers to be more credible and popular than
micro-influencers (Boerman, 2020; Rose, 2019; White, 2019).
Even though micro- and macro-influencers' attributes are different, it is expected that
disclosure warnings will negatively affect both types of SMIs. Disclosure warnings explicitly
point out the SMI’s intentions for posting the content, leaving the consumer to think the SMI’s opinions are biased (De Veirman & Hudders, 2019). Therefore, when investigating the main
effect of disclosure warnings on the perceived SMI credibility, the following hypothesis will be
tested:
H2a. The effect of the inclusion of a sponsored content disclosure warning with the Instagram post on perceived credibility of the SMI will be unaffected by the number of
However, when evaluating the effects of disclosure warnings on RPI and ENA, there
is expected to be a difference between the types of SMIs. Although Boerman (2020) found
that effects of disclosure warnings showed no differences when comparing for different types
of SMIs, this study presumes that the number of followers might be important to consumers
when evaluating sponsored content. To help activate consumer’s PK, differences in the SMI’s number of followers may function as a peripheral cue (Boerman, 2020). Consumers’ PK might help consumers understand that an SMI’s reach is important to brands. For instance, consumers may understand that macro-influencers are individuals with a large
reach, making them more likely to be selected for influencer campaigns and to earn money
with their social media account (Domingues Aguiar & Reijmersdal, 2018). Consumers’ growing understanding of the importance of number of followers results in the expectation
that:
H2b. The SMI’s number of followers moderates the main effect of inclusion of a sponsored content disclosure warning with the Instagram post on recognition of
persuasive intent (RPI) and experienced negative affect (ENA). Consumers will experience more recognition of persuasive intent and negative affect towards macro
SMIs compared to micro SMIs.
The Current Experiment
In the current experiment, the hypotheses will be tested by exposing participants to
screenshots of a fictional Instagram profile of an SMI and its latest post. These materials will
be manipulated to contain a disclosure warning or not, and show different number of
followers of the SMI (macro vs. micro). Effect of the manipulations will be tested on the
participants’ perceived SMI credibility, RPI and ENA. Number of followers will be added as a possible moderator for the main effects. The hypotheses have been elaborated in the
Figure 1. Conceptual framework.
Method
Design and Participants
An online experiment was conducted with a 2 (disclosure: yes vs. no) x 2 (number of
followers: micro vs. macro) between-subjects factorial design, with perceived credibility of the SMI, recognition of persuasive intent (RPI) and experienced negative affect (ENA) as
dependent variables. A convenience sample of 163 Dutch participants participated and was
recruited via social media and interpersonal communication. Data collection ran from
December 4 until December 16, 2019. As an incentive for participating, two gift certificates
worth 15 euro were raffled for participants that voluntarily left their email address at the end
of the survey. Participants that failed to complete the study (N = 10), were younger than 18
years old (N = 3), or indicated they were familiar with the fictitious influencer (N = 3), were
excluded from further analysis, which led to a final sample of 147 participants.
Of all participants, 103 participants were female (70.1%), and the participants’ mean age was 26.67 years (SD= 10.51), which is a good resemblance of the average Instagram
user, as 35% of global Instagram audiences are aged between 25 and 34 years (Statistica,
indicated to use Instagram daily, 7.5% of the participants indicated never to use it, and 2.5%
indicated that he or she did not have an Instagram account.
Procedure
The experiment was conducted with an online survey, using the online research
program Qualtrics. Participants were sent links for participation. First, participants signed the
informed consent, and the questionnaire started with the participant’s demographics. Subsequently, participants were randomly assigned to one of the four experimental
conditions and exposed to an SMI's profile overview and its latest Instagram post for a
minimum of 10 seconds.
Afterwards, participants answered questions regarding SMI credibility, RPI, and ENA,
followed by PK. Next, participants indicated what they thought the goal of the experiment
was. Subsequently, a subset of two manipulation checks was employed. In order to mitigate
the effects of pre-existing adversities, participants were asked about their familiarity with the
brand Tommy Hilfiger, followed by their familiarity with the SMI. Finally, participants were
debriefed and thanked for their participation.
Materials and Measures
Stimulus materials.
Stimulus materials consisted of four screenshots: two overviews of an Instagram
profile, and two screenshots of an Instagram post, that belonged to a fictitious SMI named
Laura Rose. The screenshots contained two different experimental manipulations: the
inclusion of disclosure warnings and the number of followers. Using the profile layout and
pictures from an existing SMI account put the stimulus materials together. To guarantee the
privacy of the SMI and her followers, alterations were made to her name, the number of
followers, comments, account names of commenters and the short introduction on the profile
Manipulation number of followers.
The SMI’s profile overview was manipulated by number of followers, see Image 1. The micro influencer conditions displayed a number of followers of 22,400; the
macro-influencer conditions a number of 410,000. Besides the number of followers, the Instagram
profile overviews were kept identical for all conditions.
Image 1. Stimulus material micro-influencer (left) and stimulus material macro-influencer (right).
Image 2. Stimulus including disclosure (left) and stimulus with disclosure excluded (right).
Manipulation disclosure warnings.
The Instagram post was identical for all conditions and consisted of a full-body photo
common way of SMIs to disclose the brand they are wearing. The post included identical
comments for all conditions and an identical amount of likes (2.358). In the disclosure
condition, a sponsorship disclosure warning was included to indicate the content was
sponsored. The disclosure warning consisted of a badge with the text ‘Betaald partnerschap met tommyhilfiger’, and a hash tag at the end of the caption ‘#ad’, both as provided in the guidelines of the Dutch Social Media Advertising Code (Stichting Reclame Code, n.d.).
Perceived credibility of the SMI.
Four five-point semantic differential scales were used to measure the perceived SMI
credibility (adopted from: Ohanian, 1990). Participants were asked to indicate to what extent
they believed the SMI was: untrustworthy/trustworthy, dishonest/honest, unreliable/reliable,
and insincere/sincere. A principal component analysis with Varimax rotation confirmed a single underlying factor (Eigenvalue = 2.99, explained variance = 74.69%). Therefore, a
mean total score was calculated from the four items, with higher scores indicating higher
perceived credibility of the SMI (Cronbach’s α = .89; M = 3.52, SD = .74). Recognition of persuasive intent.
RPI was measured as an indication of reactance towards the sponsored content, by
assessing agreement to four statements (e.g., “The message tried to make a decision for me”, measured on a 5-point response scale ranging from 1 (strongly disagree) to 5 (strongly agree) (adopted from: Dillard & Shen, 2005). A principal component analysis with Varimax rotation confirmed a single underlying factor (Eigenvalue = 2.54, explained variance =
63.4%), therefore, a mean score was calculated from the four items (Cronbach’s α = .79; M = 2.29, SD = .87). Higher scores on the construct indicated higher RPI.
Experienced negative affect.
ENA was measured as an indication of reactance. Participants were asked to indicate
irritated, angry, aggravated and annoyed (adopted from: Dillard & Shen, 2005). To prevent that participants would understand that the researcher was interested in negative emotions,
the following four (bogus) positive emotions were added: happy, interested, enthusiastic, and
curious, which were excluded in further analysis. ENA was measured based on a five-point Likert scale ranging from 1 (totally not applicable) to 5 (totally applicable). A principal
component analysis with Varimax rotation confirmed a single underlying factor (Eigenvalue =
2.58, explained variance = 64.61%). Therefore, a mean score was calculated for the four
items (Cronbach’s α = .82; M = 1.81, SD = .83). Higher scores on the negative emotions indicated higher ENA.
Perceived credibility of the Instagram post.
Perceived credibility of the post was included as an explorative dependent variable, to
see to what extent disclosure warnings influence the credibility of the post. To measure the
perceived credibility of the Instagram post, the same scale was used as the scale to measure
perceived SMI credibility and wording of the question was altered. A principal component
analysis with Varimax rotation confirmed a single underlying factor (Eigenvalue = 2.57,
explained variance = 64.3%). Therefore, a mean score was calculated from the four items,
with higher scores indicating higher perceived credibility of the Instagram post (Cronbach’s α = .81; M = 3.38, SD = .70).
Persuasion knowledge.
It was intended to measure PK a priori as a control variable to check if groups
systematically differed on PK. However, the measurement was deployed after exposure to
the stimuli materials, which makes it unsuitable as a control variable, as it is likely that the
disclosure manipulation influenced the responses to the measurement. Therefore, it is more
appropriate to treat it as a dependent variable. Since no a priori hypotheses about this
(strongly disagree) to 5 (strongly agree), to what extent they believed that the Instagram post
by the SMI was an advertisement (adopted from Boerman et al., 2012; M = 4.26, SD = 1.06).
Demographics.
Participants were asked to indicate their age in years and their gender (0 = Male, 1=
Female, 2 = Other, namely, 3 = I would rather not say). Furthermore, participants’ highest (completed) level of education was measured using the following scale: 0 = Basisonderwijs,
1 = LBO / VBO / VMBO, 2 = MBO, 3 = HAVO / VWO, 4 = HBO, 5= Wetenschappelijk
onderwijs (WO) - Bachelor, 6 = Wetenschappelijk onderwijs (WO) - Master, 7 = Other,
namely. The participants’ frequency of Instagram usage was measured by asking them: ‘How often do you use Instagram?’ with the following answer options: 1 = Never, 2 = Yearly, 3 = Monthly, 4 = Weekly, 5 = Daily, or 6 = I don’t have Instagram (adopted from Boerman, 2020).
Control variables.
Participants were asked to indicate whether they were familiar with the brand Tommy
Hilfiger (1= No, 2 = Yes). Of all participants, 146 (99.3%) recognised the brand Tommy
Hilfiger, excluding the possibility that participants may not recognise the possible commercial
nature of the Instagram post. Furthermore, participants were examined about their familiarity
with Laura Rose, the SMI (1= No, 2 = Yes). Of all participants, 3 participants indicated to be
familiar with the SMI before participation in the study, and, therefore, were excluded from
further analyses.
Manipulation checks.
At the end of the survey, two different manipulation checks were conducted. First, a
check was employed to measure if the participants noticed the distinction between the micro
and macro-influencer. Participants were asked to indicate the SMI’s number of followers using two questions based on recall and recognition. For recall, participants filled in an open
followers > 100.000, adapted from Boerman, 2020). Only 42 participants (28.6%) succeeded
to recall the type of SMI correctly. To measure recognition, participants were asked to select
the number of followers of the SMI on a three-point nominal scale (0 = 22.400, 1 = 410.000,
2 = I don’t know). In the micro-influencer condition (N = 77), 77.9% succeeded to recognise the number of followers. This was just more than half (51.4%) for the macro-influencer
condition (N = 70). Therefore, the manipulation is considered fairly successful.
Secondly, to measure if the participants were correctly exposed to the disclosure
warning, participants were asked to indicate whether they had seen a disclosure warning in
the post (0 = No, 1 = Yes) (adopted from Boerman, 2020). In the non-disclosure conditions
(N = 73), 87.7% correctly indicate that they did not see a warning for sponsorship disclosure.
Nine participants marked to recognise seeing a disclosure warning, however, had not been
exposed to a disclosure warning. In the disclosure conditions (N= 74), 55 participants
(74.3%) succeeded to recognise the disclosure warning. Again, this manipulation is
considered fairly successful.
Strategy for Analyses
Descriptives and frequencies for various demographic characteristics were examined,
e.g. the mean age of the participants, to obtain a clear overview of the study’s sample,. Subsequently, randomization was checked for equal distribution of various demographic
variables, e.g. age and education level.
MANOVA analysis.
To test the study’s hypotheses, a MANOVA analysis was run with the inclusion of disclosure warnings and number of followers as independent variables and perceived SMI
credibility, RPI, and ENA as dependent variables. Within each of the conducted analyses, a
total sample of 147 participants was employed. No distinction was made whether or not
tested to see if the main results could hold up when checked for only successful
manipulations, using a filter in the dataset (N = 81).
Explorative, a MANOVA analysis was run with the inclusion of disclosure warnings
and number of followers as independent variables, and credibility of the Instagram post and
PK as dependent variables.
Results
Randomization Checks
Randomization checks indicate there were no significant differences in gender, χ2
(6) = 9.09, p = .169, age, F (3, 146) = .83, p = .478, level of education, Fisher-exact p = .251,
and Instagram usage across the four experimental conditions, F (3, 146) = .87, p = .456.
Therefore, it may be assumed that the variables age, gender, level of education and
Instagram usage are all equally distributed over the four conditions and that randomization
has been successful.
Main Analyses
A 2 (inclusion disclosure warnings) x 2 (number of followers) MANOVA with three
dependent variables (perceived SMI credibility, recognition of persuasive intent [RPI] and
experienced negative affect [ENA]) was employed to test the hypotheses. Levene's F test was conducted for all dependent variables to see whether there were equal variances
between the conditions, and yielded non-significant results for perceived SMI credibility (F (3,
143) = 1.665, p = .177), RPI (F (3, 143) = 1.05, p = .373), and ENA (F (3, 143) = 2.64, p =
.052). Therefore, it may be assumed that the groups in the population have the same
variance in the dependent variables.
The multivariate tests indicated a statistically significant main effect of disclosure
warnings on the combined dependent variables (F (3, 141) = 3.18, p = .026, Wilks' Λ = 0.937, partial η2 = .06), and a significant interaction effect between the inclusion of disclosure
warnings and number of followers on the combined dependent variables (F (3, 141) = 3.11, p
= .028; Wilks' Λ = .938, partial η2 = .06).
Table 1. Means of SMI credibility, RPI, and ENA for the conditions of disclosure warnings and influencer types.
Macro Micro Total
Discl. No discl. Discl. No discl. Discl. No discl.
Macro SMI Micro SMI SMI credibility 3.33 (.81) 3.32 (.75) 3.53 (.56) 3.75 (.80) 3.44 (.69) 3.60 (.79) 3.38 (.78) 3.64 (.70) RPI 2.56 (.94) 2.13 (.76) 2.44 (.83) 2.04 (.86) 2.50 (.88) 2.08 (.81) 2.35 (.88) 2.24 (.86) ENA 1.86 (.80) 2.04 (.84) 1.98 (.93) 1.41 (.60) 1.92 (.87) 1.70 (.78) 1.95 (.82) 1.69 (.83)
Note. Standard deviations between brackets.
Perceived SMI credibility.
The MANOVA found no statistically significant main effect of disclosure warnings on
the perceived SMI credibility, F (1, 143) = 1.71, p = .193. Repeating this analysis with only
successful manipulations did not change this result, F (1, 77) = 3.04, p = .085. In the sample,
the SMI credibility for participants in the non-disclosure condition was somewhat higher than
participants in the disclosure condition, see Table 1. However, these differences do not exist
for the overall population. Hypothesis 1a will be rejected.
Moreover, the MANOVA showed a statistically significant, small main effect of
number of followers on SMI credibility, F (1, 143) = 4.51, p = .035, η2 = .032. Repeating this analysis with only successful manipulations however yielded a non-significant result, F (1,
77) = .16, p = .694. Participants that were exposed to the micro influencer valued its
Finally, the MANOVA indicates that there is no statistically significant interaction
between the effects of the inclusion of a disclosure warning and number of followers of the
SMI on SMI credibility, F (1, 143) = .23, p = .632. Repeating this analysis with only
successful manipulations yielded a similar result, F (1, 77) = .02, p = .890. This means that
hypothesis 2a is supported. There are no differences in the population when controlling the
effect of disclosure warnings on perceived SMI credibility for number of followers, see Table
1.
Recognition of persuasive intent.
The MANOVA shows a statistically significant, small effect of disclosure warnings on
RPI, F (1, 143) = 8.79, p = .004, η2 = .058. Repeating this analysis with only successful manipulations did not change this result, F (1, 77) = 6.43, p = .013, η2= .077. Participants in the disclosure condition showed higher RPI than participants in the non-disclosure condition,
see Table 1. Hypothesis 1b is supported.
The main effect of number of followers on RPI yielded a statistically non-significant
result, F (1, 143) = .51, p = .475. Repeating this analysis with only successful manipulations
did not change this result, F (1, 77) = 1.72, p = .194. In the sample, participants in the
macro-influencer condition reported higher RPI than participants in the micro-macro-influencer condition ,
see Table 1. However, these differences are not found in the overall population.
Finally, there is no statistically significant interaction effect of number of followers on
the main effect of inclusion of disclosure warnings on RPI, F (1, 143) = .01, p = .920.
Repeating this analysis with only successful manipulations yielded a similar result, F (1, 77) =
.94, p = .335. Participants exposed to the macro-influencer scored higher on RPI than those
exposed to the micro influencer, see Table 1. These differences are not found within the
Experienced negative effect.
The MANOVA analysis yielded a non-significant main effect of disclosure warnings
on ENA, F (3, 143) = 2.22, p = .138. Repeating this analysis with only successful
manipulations yielded similar results, F (1, 77) = .319, p = .574. In the sample, participants
that were not exposed to a disclosure warning experienced less negative affect than
participants exposed to the disclosure warning, see Table 1. These differences are not found
within the population. Hypothesis 1c will be rejected.
Secondly, a non-significant trend was found for the effect of number of followers on ENA, F
(1, 143) = 3.67, p = .057, η2
= .023. Repeating this analysis with only successful
manipulations yielded a similar result, F (1, 77) = 3.68, p = .059, η2 = .046. In the sample, participants that were exposed to the macro-influencer experienced somewhat more
negative affect than those exposed to the micro influencer, see Table 1.
Figure 2. Interaction effect for number of followers on the main effect of disclosure warnings on ENA.
Subsequently, the study yielded a small, significant interaction effect for number of
followers on the main effect of inclusion of a disclosure warning on ENA, F (1, 143) = 7.94, p
yielded a non-significant trend, F (1, 77) = 3.16, p = .080. In the non-disclosure condition,
participants exposed to the macro-influencer experienced more negative affect than those
exposed to the micro influencer. However, the opposite was found in the disclosure
condition, see Table 1. Hypothesis 2b can therefore only be partially supported. When
Instagram posts by a macro-influencer do not include disclosure warnings, consumers
experience more negative affect.
Explorative Analyses
Levene's F test was conducted to see whether there were equal variances between
the conditions. The test was not significant for perceived credibility of the Instagram post (F
(3, 143) = 2.20, p = .091). Therefore, it may be assumed that the groups in the population
have the same variance in perceived credibility of Instagram post. However, Levene’s F test was significant for PK (F (3, 143) = 3.02, p = .032). Nevertheless, this violation of the
MANOVA criteria is not too serious, since the sample sizes can be regarded as equal.
The multivariate tests indicated that there was a statistically significant main effect of
the inclusion of disclosure warnings on perceived credibility of the Instagram post and PK, (F
(2, 142) = 6.92, p = 0.001), and a non-significant trend for the main effect of number of
followers on the combined dependent variables (F (2, 142) = 2.44, p = .091; Wilks' Λ = .97, partial η2 = .03).
Perceived credibility of the Instagram post.
The MANOVA yielded a small, significant result for the main effect of number of
followers on perceived credibility of the Instagram post, F (1, 143) = 4.86, p = .029, η2 = .033. Participants exposed to the micro-influencer reported higher credibility of the Instagram post
than those exposed to the macro-influencer.
Persuasion knowledge.
The MANOVA yielded a statistically medium, significant main effect of the inclusion of
disclosure warnings on PK, F (1, 143) = 13.37, p < .001, η2 = .09). Participants in the disclosure condition reported higher PK than those in the non-disclosure condition.
All other comparisons yielded non-significant results. See Table 2 for all means.
Table 2. Means of credibility of the Instagram post, and PK for the conditions of disclosure warnings and influencer types.
Macro Micro Total
Discl. No discl. Discl. No discl. Discl. No discl.
Macro SMI Micro SMI Credibility Instagram post 3.30 (0.60) 3.21 (.62) 3.36 (.63) 3.65 (.84) 3.33 (.61) 3.44 (.78) 3.25 (.61) 3.50 (.75) PK 4.67 (.63) 3.97 (1.31) 4.53 (1.03) 4.00 (.97) 4.59 (.86) 3.99 (1.14) 4.33 (1.07) 4.26 (1.03)
Note. Standard deviations between brackets.
Discussion
This study aimed to examine the impact of disclosure warnings in sponsored content.
In order to provide insights into the gap in research, this study answered the following
research question: ‘To what extent does a ‘sponsored content’ disclosure warning on Instagram posts affect the perceived credibility of the social media influencer (SMI), recognition of persuasive intent and experienced negative affect, and are these effects moderated by the number of followers of the SMI (i.e. micro vs. macro)? The study’s main results indicate that disclosure warnings increase recognition of persuasive intent (RPI) and
Results indicate that disclosure warnings do not negatively affect SMI credibility.
Surprisingly, these findings are inconsistent with prior research by De Veirman and Hudders
(2019). When including ad scepticism, they found that ad scepticism is very likely to occur as
a result of disclosure warnings, which in turn negatively affects consumers’ perceptions of the SMI’s credibility. Secondly, the findings do indicate that the SMI’s number of followers influences their credibility. This study found that consumers tend to consider
micro-influencers to be more credible than macro-micro-influencers, which is contradictory to prior
research by Boerman (2020) and De Veirman et al. (2017). They found that the more
followers an SMI had, the more credible they were perceived by consumers. This study’s contrasting finding may be explained due to consumers perceiving micro-influencers as more
similar to them than macro-influencers, which could result in consumers finding them more
credible (Domingues Aguiar & Van Reijmersdal, 2018). Furthermore, the study’s findings for the interaction effect on SMI credibility show non-significant results.
Secondly, the research found that disclosure warnings increase RPI, supporting
findings by Boerman (2020) and Evans et al. (2017). These are essential findings, as they
support the effectiveness of disclosure warnings and indicate consumers’ potential
experience of psychological reactance. Moreover, SMI’s number of followers did not increase the RPI, again supporting findings by Boerman (2020). These findings suggest that the
number of followers is no cue for RPI. Interestingly, persuasion knowledge (PK) was
relatively high throughout the experiment, suggesting high levels of consumer awareness
regarding influencer marketing. Therefore, consumers might keep in mind that SMIs,
regardless of which type, are likely to advertise. Furthermore, there was no significant
interaction effect on RPI, supporting findings by Boerman (2020) that SMIs’ number of followers do not function as a peripheral cue to help activate consumer’s PK. These findings indicate that the type of SMI is not related to the activation of PK.
Interestingly, when consumers are exposed to a disclosure warning, they experience
found if there was no disclosure warning present, a relatively new finding within disclosure
warning research. When micro-influencers use no disclosure warnings, consumers
experience less negative affect, which could be explained by consumers perceiving them as
authentic. However, when they engage in (disclosing) sponsored content, consumers may
perceive them as a sell-out, exchanging their authenticity for monetary benefits. On the other
hand, consumers’ overall affect towards macro-influencers is already fairly negative, as they are aware that macro-influencers are likely to engage in sponsored content (Domingues
Aguiar & Reijmersdal, 2018). When macro-influencers do not use disclosure warnings,
consumers may experience a feeling of envy towards them because they dislike the feeling
of being misled (De Veirman & Hudders, 2019). However, when macro-influencers do
include disclosure warnings, consumers appreciate them for being honest about the
intentions for posting the content.
Finally, the explorative analyses yielded some interesting results. On the one hand,
disclosure warnings do seem to activate (cognitive and attitudinal) PK, supporting results
from Boerman et al. (2012), Evans et al. (2017), and Hass and Grady (1975). On the other
hand, the SMI’s number of followers influences how consumers perceive the Instagram post’s credibility. Consumers tend to value posts by micro-influencers as more credible than those of macro-influencers.
Limitations and future research
The current study has some limitations. Firstly, a large number of participants failed to
recognise the disclosure warning or number of followers. These findings are not surprising,
as previous studies showed how little attention consumers pay to disclosure warnings
(Boerman et al., 2012; Boerman, 2020; Wojdynski & Evans, 2016). Moreover, the profile
overview did not seem sufficient for the participants to recognise the SMI’s number of followers. Nonetheless, this does not imply that the manipulations were not effective. In fact,
valid. In order to test this, future research can use eye-tracking devices to determine how
much attention is paid to disclosure warnings and the SMI’s number of followers when consumers evaluate SMIs. Though, the results of this study need to be treated with caution,
as it cannot entirely be ruled out that the manipulations caused differences between the
conditions.
Secondly, the overall study lacked statistical power. In the main analyses, effect sizes
for the effects found were fairly small. Additionally, the study found some interestingly
non-significant trends. Since the study’s sample was rather small, future research should include a larger sample to measure more accurately and to elaborate on these small effects.
Thirdly, the study intended to measure PK a priori as a control variable. However,
since the measurement was deployed after the manipulation, it was unsuitable for treating it
as a control variable. Notable, however, is that disclosure warnings do seem to activate PK.
Future research could build onto these results, by measuring pre-existing PK before
manipulations by asking participants about the specific features they believe to be good
indicators for sponsored content.
Finally, despite that the sample showed to be a good representation of the overall
population, the current experiment involved a convenience sample. Therefore, the results
found need to be treated with caution and cannot be fully generalised. Repeating the current
experiment with a simple random sample is recommended.
Moreover, future studies can elaborate on some interesting findings of this study.
Firstly, it is to interesting investigate the interaction effect on ENA further. This study found
that if SMIs engage in sponsored content, consumers experience less negative affect
towards macro-influencers compared to micro-influencers. The contrary was found when no
disclosure warnings were present. Future research could investigate what triggers
differences in consumers’ evaluations of micro- and macro-influencers engaging in sponsored content. Furthermore, it may be interesting to include non-sponsorship
disclosures in future studies, as prior studies found that consumers appreciate clear
non-sponsorship disclosures (De Veirman & Hudders, 2019). Moreover, Van Reijmersdal et al.
(2016) found that when non-sponsorship disclosures are used, this may, over time, result in
consumers appreciating and recognising transparency by the source, which in turn may even
soften consumers’ resistance to sponsored content.
Moreover, RPI seems to be mostly influenced by the presence of a disclosure
warning (cognitive cue), whereas the ENA seems to be affected by the characteristics of the
SMI (attitudinal cue). It may be interesting to find more support for these findings, as
reactance is desirable since it allows consumers to resist persuasion attempts. Furthermore,
these results confirm the idea of the cognitive and the attitudinal dimension of persuasion
knowledge (Boerman et al., 2012).
Practical implementations
Based on the results, some practical implementations can be drawn. Even though
results show how little attention consumers pay to disclosure warnings, as found in previous
research (Boerman et al., 2012; Boerman, 2020; Wojdynski & Evans, 2016), disclosure
warnings do increase RPI. This study offers regulatory parties valuable insights about how
consumers process sponsored content, as disclosure warnings show to be effective cues for
recognition of sponsored content and activation of consumers’ PK. Furthermore, the study’s findings confirm consumers’ development of PK regarding influencer marketing, a valuable finding for regulatory parties.
For brands and SMIs, the findings are valuable as they show that consumers still
struggle to recognise sponsored content. On the one hand, this is favourable for brands,
since this indicates that influencer marketing is still effective and marketers’ persuasive intent remain fairly disguised in influencer campaigns. On the other hand, these findings are
favourable for SMIs, as the findings suggest that disclosure warnings do not harm their
have to worry that engaging in influencer campaigns harms their credibility. Noteworthy,
however, are the findings that disclosing sponsored content may be harmful to consumers’ perceptions of micro-influencers; yet, it may be valuable for consumers’ perceptions of macro-influencers.
In conclusion, this study finds support for the effectiveness of disclosure warnings
and its potential to evoke reactance towards sponsored content. Notably, the high levels of
persuasion knowledge in the current study suggest that consumers are highly aware of the
persuasive intent of influencer marketing, indicating that consumers are developing their
persuasion knowledge on influencer marketing over time. Future research is recommended
to elaborate on disclosure warnings effects on SMI evaluations and consumers’ experience of reactance, and specifically the attitudinal dimension of reactance.
References
Bilandzic, H., & Busselle, R. (2013). Narrative persuasion. The Sage Handbook of
Persuasion: Developments in Theory and Practice, 200-219. DOI: 10.4135/9781452218410.n13
Boerman, S. C. (2020). The effects of the standardized Instagram disclosure for
micro-and meso-influencers. Computers in Human Behavior, 103, 199-207. DOI:
10.1016/j.chb.2019.09.015
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. DOI: 10.1111/j.1460-2466.2012.01677.x
Brehm, J. W. (1966). A theory of psychological reactance. New York, NY: Academic
Brehm. S. S.. & Brehm. J. W. (1981). Psychological reactance: A theory of freedom
and control. San Diego. CA: Academic Press
Burgoon, M., Alvaro, E., Grandpre, J., & Voulodakis, M. (2002). Revisiting the Theory of Psychological Reactance: Communicating Threats to Attitudinal Freedom. The Persuasion
Handbook: Developments in Theory and Practice, 213–232. DOI: 10.4135/9781412976046
.n12
Cheung, M. Y., Luo, C., Sia, C. L., & Chen, H. (2009). Credibility of electronic
word-of-mouth: Informational and normative determinants of on-line consumer
Recommendations. International Journal of Electronic Commerce, 13(4), 9–38. DOI: 10.2753/JEC1086-4415130402
De Veirman, M., Cauberghe, V., & Hudders, L. (2017). Marketing through Instagram
influencers: The impact of number of followers and product divergence on brand
attitude. International Journal of Advertising, 36(5), 798–828. DOI: 10.1080/ 02650487.2017.1348035
De Veirman, M., & Hudders, L. (2019). Disclosing sponsored Instagram posts: the
role of material connection with the brand and message-sidedness when disclosing covert
advertising. International Journal of Advertising, 39(1), 94–130. DOI: 10.1080/02650 487.2019.1575108
De Pelsmacker, P., & Neijens, P. C. (2012). New advertising formats: How
persuasion knowledge affects consumer responses. Journal of Marketing Communications,
18(1), 1-4. DOI: 10.1080/13527266.2011.620762
Dillard, J. P., & Shen, L. (2005). On the nature of reactance and its role in persuasive
health communication. Communication Monographs, 72, 144-168. DOI:
Djafarova, E., & Rushworth, C. (2017). Exploring the credibility of online celebrities'
Instagram profiles in influencing the purchase decisions of young female users. Computers in
Human Behavior, 68, 1-7. DOI: 10.1016/j.chb.2016.11.009
Doh, S. J., & Hwang, J. S. (2009). How consumers evaluate eWOM (electronic
word-of-mouth) messages. CyberPsychology & Behavior, 12(2), 193-197. DOI:
10.1089/cpb.2008.0109
Domingues Aguiar, T., & Reijmersdal, E. (2018). Influencer Marketing. Amsterdam:
SWOCC.
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), 138-149. DOI:
10.1080/15252019.2017.1366885
Federal Trade Commission (FTC). (2019). Disclosures 101 for Social Media
Influencers. Retrieved December 10, 2019, from
https://www.ftc.gov/system/files/documents/plain-language/1001a-influencer-guide-508_1.pdf
Federal Trade Commission (FTC). (2017). FTC Staff Reminds Influencers and
Brands to Clearly Disclose. Retrieved December 9, 2019, from
https://www.ftc.gov/news-events/press-releases/2017/04/ftc-staff-reminds-influencers-brands-clearly-disclose.
Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people
cope with persuasion attempts. Journal of consumer research, 21(1), 1-31. DOI:
10.1086/209380
Goldsmith, R. E., & Clark, R. A. (2008). An analysis of factors affecting fashion
opinion leadership and fashion opinion seeking. Journal of Fashion Marketing and
Goldsmith, R. E., Lafferty, B. A., & Newell, S. J. (2000). The impact of corporate
credibility and celebrity credibility on consumer reaction to advertisements and brands.
Journal of Advertising, 29(3), 43-54. DOI :10.1080/00913367.2000.10673616
Hass, R. G., & Grady, K. (1975). Temporal delay, type of forewarning, and resistance
to influence. Journal of Experimental Social Psychology, 11(5), 459-469. DOI:
10.1016/0022-1031(75)90048-7
Heilman, M. E., & Toffler, B. L. (1976). Reacting to reactance: An interpersonal
interpretation of the need for freedom. Journal of Experimental Social Psychology, 12(6),
519-529. DOI: 10.1016/0022-1031(76)90031-7
Hennig-Thurau, T., Gwinner, K. P., Walsh, G., & Gremler, D. D. (2004). Electronic
word-of-mouth via consumer-opinion platforms: what motivates consumers to articulate
themselves on the internet?. Journal of Interactive Marketing, 18(1), 38-52. DOI:
10.1002/dir.10073
Hovland, C., Janis, I. and Kelley, H. (1966), Communication and Persuasion:
Psychological Studies of Opinion Change. Yale University Press, New Haven.
Influencer Marketing Hub. (2018, October 15). What is an Influencer? - Factors that
define a Social Media Influencer. Retrieved on November 19, 2019, from
https://influencermarketinghub.com/what-is-an-influencer/
Ohanian, R. (1990). Construction and validation of a scale to measure celebrity
endorsers' perceived expertise, trustworthiness, and attractiveness. Journal of Advertising,
19(3), 39-52. DOI: 10.1080/00913367.1990.10673191
Rose, J. (2019, June 28). Choosing Between Nano, Micro, or Macro-influencers for
Your Business. Retrieved October 22, 2019, from
Seno, D., & Lukas, B. A. (2007). The equity effect of product endorsement by
celebrities: A conceptual framework from a co-branding perspective. European Journal of
Marketing, 41(1/2), 121-134. DOI: 10.1108/03090560710718148
Schomer, A. (2019, December 17). Influencer Marketing: State of the social media
influencer market in 2020. Retrieved January 10, 2020, from
https://www.businessinsider.com/influencer-marketing-report?international=true&r=US&IR=T
Schouten, A. P., Janssen, L., & Verspaget, M. (2019). Celebrity vs. influencer
endorsements in advertising: The role of identification, credibility, and product-endorser fit.
International Journal of Advertising, 1-24. DOI: 10.1080/02650487.2019.1634898
Stichting Reclame Code. (n.d.). Archief Nederlandse Reclame Code. Retrieved
December 9, 2019, from https://www.reclamecode.nl/nrc/
Uribe, R., Buzeta, C., & Velásquez, M. (2016). Sidedness, commercial intent and
expertise in blog advertising. Journal of Business Research, 69(10), 4403-4410. DOI:
10.1016/j.jbusres.2016.04.102
Van Reijmersdal, E. A., Fransen, M. L., van Noort, G., Opree, S. J., Vandeberg, L.,
Reusch, S., & Boerman, S. C. (2016). Effects of disclosing sponsored content in blogs: How
the use of resistance strategies mediates effects on persuasion. American Behavioral
Scientist, 60(12), 1458-1474. DOI: 10.1177/0002764216660141
White, K. (2019, August 20). The Difference Between Mega, Macro, Micro and Nano
Influencers. Retrieved October 22, 2019, from
https://www.insightsforprofessionals.com/marketing/social-media/difference-mega-macro-micro-nano-influencers
Wojdynski, B. W., & Evans, N. J. (2016). Going native: Effects of disclosure position
and language on the recognition and evaluation of online native advertising. Journal of
