Effects of sponsorship
disclosures on Instagram
The role of fit between SMI and product,
persuasion knowledge and source credibility
Master’s Thesis
Author:
Supervisor:
Roos Olthof
Dr E. de Waal
Graduate School of Communication
Master’s programme Communication Science
Student ID number: 10266623
1
Abstract
Social media influencers (SMIs) are often used to promote brands, and the sponsored content they create has to be properly disclosed according FTC rules. To find out how these
sponsorship disclosures affect persuasion knowledge and source credibility, both important factors in explaining attitudes and behavior, the presence and different types of sponsorship disclosures on Instagram SMI posts were investigated through an online experiment among 1210 Instagram-using women, aged 18 to 30. Moreover, the impact of fit between the SMI and product was studied. Results showed that adding a disclosure to sponsored Instagram content enhances persuasion knowledge and source trustworthiness, but the type of disclosure does not matter. Interestingly, negative effects of persuasion knowledge are diminished when people remember the (presence/type of) disclosure correctly, which was only 52.4% of the participants in this study. The fit between SMI and product has shown to have direct and indirect effects. Overall, this study contributes as confirmation of the Persuasion Knowledge Model (Friestad & Wright, 1994) and the Match-up Hypothesis (Kamins & Gupta, 1994) in the Instagram context, while adding evidence for theory about source credibility and effects on SMIs as well. Managerial implications are addressed and future research is discussed.
Keywords: sponsorship disclosure, Instagram, social media influencer, persuasion
knowledge, match-up hypothesis, fit, source credibility, SMI attitude, purchase intention, brand attitude
2
Introduction
Instagram, an online platform for sharing photos and videos, is booming; it had 800 million monthly active users in 2017 (Statista, 2017), providing new marketing possibilities for brands. Visuals are becoming more important for marketing and 63% of companies intend to use Instagram more in the future (Social Media Examiner, 2017). One way brands use Instagram is by partnering up with social media influencers (SMIs) or endorsers. SMIs create content involving brands, for which they are being paid or rewarded in another way. The SMIs have many followers, of which most are looking up to influencers, seeing them as role models. They try to come as close to their lifestyle as possible, for instance by buying the products SMIs use (Hung, 2014). SMIs are more powerful than actual celebrities in
influencing the consumer’s buying process, because people can easier relate to them, which makes them more credible (Djafarova & Rushworth, 2017). To reach customers, using SMIs to display products is rather cheap and effective for brands, without it being too obvious as advertisements. This is one of the reasons that this form of word-of-mouth generates more than twice the sales of paid advertising (Wong, 2014). People tend to deliberately ignore content by brands (Bakas, 2017), and the natural peer-to-peer aspect of word-of-mouth decreases chances that consumers will skip the advertisement (De Pelsmacker & Neijens, 2012). At the same time, these sponsorships are a large part of the SMI’s income, and sponsorship effects like positive brand attitudes can also spill over to the SMI; both the SMI and the brand can benefit from the collaboration.
There are rules for creating sponsored social media content like this. The Federal Trade Commission (FTC) says that “if there is a 'material connection' between an endorser and an advertiser […] that connection should be clearly and conspicuously disclosed” (Engle, 2017, p. 1). SMIs must use FTC guidelines to disclose sponsorships, but in only 7% of posts by the most popular Instagram influencers those guidelines are followed correctly (Mediakix,
3
2017). Often, they fail to implement correct placements or texts of disclosures. Most used sponsorship disclosures on Instagram are hashtags like #sponsored, but to better distinguish paid content from regular content, Instagram launched a new feature in June 2017: a label saying “Paid Partnership with [company X]”, placed between the username and picture. This makes it easier for SMIs to follow FTC rules, but businesses are concerned audiences will easier recognize the persuasion aspect of posts, which could cause reactance (Brehm, 1966; Miron & Brehm, 2006) and consequently more negative brand attitudes and purchase intentions.
However, for those expecting a paid partnership beforehand – those with high
persuasion knowledge – disclosures may not have such negative effects (Evans et al., 2017). Plus, it seems that influencer marketing is widely accepted by users, and transparency of SMIs in disclosing partnerships is most valued (Brooks, 2017). Transparency leads to
credibility, which – according to consumers – is the most important predictor of effectiveness of SMI marketing (Braun, 2017). Consumers believe that SMIs only partner up with brands that they love and fit with, because SMIs rely on their credibility (Brooks, 2017). The fit is highly important for effects on the brand as well as the SMIs (Till & Busler, 2000). Thus, when researching effects of celebrity (SMI) endorsements, it is highly valuable to look at the effects of the fit between the SMI and the product, since “studies that fail to include [fit] in their research efforts may produce inflated results” (Amos, Holmes and Strutton, 2008, p. 226).
Without knowledge of the fact that the content by the SMI is paid for by the brand, according to the FTC, positive brand or SMI related effects occur unethically. Therefore, this study considers persuasion knowledge and whether this is activated by sponsorship
disclosures. Following up on the release of the new disclosure option on Instagram, effects for different disclosures (“Paid Partnership with”-label versus “#sponsored”) are investigated too.
4
Because influencer marketing is relatively new, previous research mostly focused on effects on brands, rather than effects on SMIs (i.e., attitudes). This study’s main focus is the effect on source credibility, which has shown to be positively related to brand effects (Amos, Holmes and Strutton, 2008; Eisend, 2004; Wilson & Sherrell, 1993). Since SMIs are becoming brands themselves (Weinswig, 2016), investigating effects on SMIs is considered essential for
today’s marketing field as well. Therefore this study will attempt to answer the following research question:
To what extent do presence and type of sponsorship disclosure and fit between SMI and product affect source credibility and consequently SMI attitude (and brand effects), and what role does persuasion knowledge play in explaining the effects?
This study is important because it provides new evidence for effects of (different types of) disclosures on Instagram, with a revelatory focus on source credibility and the new
inclusion of effects of fit between SMI and product on these relationships. By giving insight in persuasion knowledge, this study can serve as grounded theory for advice for the FTC according sponsorship disclosure regulations, which can be used to help protect and educate consumers. Because this study puts its main focus on SMIs instead of brands, it adds a different point of view on SMI marketing on Instagram to the research field. Therefore, this study could not only help brands decide whether or not influencer marketing is a valuable strategy to start or maintain, but also SMIs can use this study for knowledge on implementing sponsorships and its disclosures on Instagram.
Theoretical framework
Presence of sponsorship disclosures and persuasion knowledge
As described in the Persuasion Knowledge Model, disclosures activate persuasion knowledge and recognition of advertising within the post (Friestad & Wright, 1994). In turn,
5
this can result in lower brand attitudes and purchase intentions (Evans, et al., 2017; Van Reijmersdal et al., 2016; Wei et al., 2008). Sponsored content is perceived as less irritating than posts by brands, because persuasion knowledge is less activated by it (Becker-Olsen, 2003; Tutaj & van Reijmersdal, 2012). In turn, based on the reactance theory, a warning of persuasive intent (i.e., disclosures) will result in counter arguing the message (Brehm, 1966; Miron & Brehm, 2006). Although disclosures are appreciated and wanted by audiences (Boerman, van Reijmersdal & Neijens, 2013), they negatively affect brand attitudes and purchase intentions, because people realize the advertising aspect (i.e., it activates persuasion knowledge), and therefore process content more critically (Boerman, van Reijmersdal & Neijens, 2012; 2013 2014; Evans et al., 2017; Kim & Song, 2017). Because it is important to understand why certain brand effects occur, the follow research question has been developed:
RQ1: To what extent is persuasion knowledge activated differently by the presence of
a disclosure versus no presence of a disclosure?
Type of sponsorship disclosure and persuasion knowledge
Showing disclosures prior to the content will lead to audiences recognizing the advertising aspect of the content, which activates critical content processing. This has a negative effect on brand attitudes (Boerman, van Reijmersdal & Neijens, 2013; 2014). The same effect occurs when disclosures are showed during display of content, but not when showed at the end, because that does not affect content processing. Because no different options in placement of disclosures on Instagram was possible until the introduction of the “Paid Partnership with”-label, no research about placement of disclosures prior or after the content in the Instagram context exists; all formerly effects researched are about the hashtags placed after the content. Therefore the following research question has been developed: RQ2: To what extent does the label “Paid Partnership with” activate persuasion
6 Presence of sponsorship disclosures, persuasion knowledge and source credibility
Interestingly, sponsorship disclosures are also found to have positive effects on source credibility (Carl, 2008; Chapple & Cownie, 2017). Nowadays, it is believed to be more common knowledge than before that celebrities (SMIs) are paid to endorse products. People believe that they are unlikely to abuse their position for money; SMIs will only promote products they believe in, to maintain their reputation and credibility (Djafarova & Rushworth, 2017). Source credibility is defined as the perception of the source according to one’s
attractiveness, trustworthiness and expertise within the area of the endorsed product (Ohanian, 1990; Reichelt, Sievert & Jacob, 2013).
Also, it seems like new (social) media have changed perceptions (of source credibility) to a point where transparency is valued more (Brooks, 2017). Realizing the sender’s
persuasive intent (i.e., persuasion knowledge) is then no longer a reason to instantly counter argue the message. Other negative effects of persuasion knowledge are found to be
diminished when persuasion tactics are perceived as appropriate (Wei, Fischer & Main, 2008). Disclosures could make endorsement more appropriate, resulting in less negative persuasive effects for people who are familiar with persuasion tactics: People seem to recognize advertised content anyway (Kim, Pasadeos & Barban, 2001). While effects of persuasion knowledge on brand effects have been researched often, the role of source credibility has not yet been fully covered. Although consumer reviews, such as Instagram posts, are perceived as more credible and less biased than when brands share the content (Hoffman & Daugherty, 2013), actual evidence for established theory in the Instagram context is lacking, which is why the following research questions have been developed:
RQ3a: What is the influence of the presence of a sponsorship disclosure on source
credibility?
7 Fit between SMI and product, presence of sponsorship disclosures and persuasion knowledge
To establish why brands use certain people to promote their product, the match-up hypothesis is fairly important, since high following numbers of SMISs alone are not enough to establish positive brand effects (De Veirman, Cauberghe & Hudders, 2017) The match-up hypothesis happens when “highly relevant characteristics of the spokesperson are consistent with highly relevant attributes of the brand” (Misra & Beatty, 1990, p. 160).
Good fit between SMI and product leads to more authenticity and a way of reaching right audiences for the product. It diminishes feelings of money-driven intentions of the SMI (De Veirman, Cauberghe & Hudders, 2017), but sponsorship disclosures can elect those feelings when the fit is bad, by showing the audience that the SMIs are paid to post the content. Therefore, the following research question has been developed:
RQ4: To what extent does fit between SMI and product affect the relationship between
sponsorship disclosure presence and persuasion knowledge?
Fit between SMI and product, brand attitude, purchase intention, source credibility and SMI attitude
The fit between SMI and product not only important because it can affect the relationship between sponsorship disclosures and persuasion knowledge, but also because good fit positively influences brand attitudes and purchase intentions (Kamins & Gupta, 1994; Kirmani & Shiv, 1998; Koernig & Boyd, 2009; Koo, Quarterman & Flynn, 2006; Olson, 2010; Olson & Thjømøe, 2011; Russel & Rasolofoarison, 2017). Brands should match their products with SMIs that fit well with the product, rather than just choosing attractive people (Till & Busler, 2000).
Effects on the SMI are also more positive when the fit with the product is high, because that generates altruistic motives: The SMIs are sooner believed to be driven by product quality – rather than by money – because of interest in the product, which makes
8
them more credible (De Veirman, Cauberghe & Hudders, 2017). This in turn leads to more positive brand attitudes and SMI attitudes (Bergkvist, Hjalmarson & Mägi, 2016; Jin & Phua, 2014). When endorsed products are perceived as invalid or false due to bad fit with the
endorser, the SMI is perceived as less credible (Cheung, Luo, Sia, & Chen, 2009). Comparing this to sponsorships of events, high fit between the sponsor and the event leads to better attitudes towards the sponsor (Gwinner & Eaton, 1999; Mazodier & Quester, 2014; Rifon et al., 2004; Smith, 2004), because of perceived altruistic motives, which is expected to also happen when SMIs promotes products that fit well, because they are believed to only promote products they appreciate themselves (Djafarova & Rushworth, 2017).
Previous research has offered plenty evidence for the positive effects of the match-up hypothesis on brand attitudes and purchase intentions, but fails to incorporate possible effects on source credibility. Also, insufficient evidence exists that the match-up hypothesis also applies to the Instagram context and SMIs themselves. Therefore, the following hypotheses and research questions have been developed:
H1: High fit between SMI and product positively influences (a) brand attitude, and (b)
purchase intention, as compared to low fit between SMI and product.
RQ5: What is the effect of fit between SMI and product on (a) source credibility, and
(b) SMI attitude?
Source credibility, SMI attitude, brand attitude and purchase intention
Investigating source credibility in this context is interesting because Instagram posts – sponsored or not – can be seen as a form of electronic word-of-mouth (eWOM) that relies on credibility to work and has shown to have impact on consumer (buying) behavior (Daugherty & Hoffman, 2014): SMIs share their experiences with products, on which other consumers can base their buying decisions (Moran & Muzellec, 2017). Because of social media’s “following” aspect, people tend to create feelings of a relationship with SMIs, which
9
strengthens this influence (Jin & Phua, 2014). However, when eWOM is mostly positive, consumers tend to get more skeptical about the credibility of the information and the source (Doh & Hwang, 2009).
While it is established that high levels of source credibility lead to high purchase intentions and positive brand attitudes in multiple media fields (Amos, Holmes and Strutton, 2008; Chaiken & Maheswaran, 1994; Lafferty, Goldsmith & Newell, 2014; Kamins et al., 1989; Pornpitakpan, 2004; Van Noort, Antheunis & Van Reijmersdal, 2012; Wu & Wang, 2011; Yoon, Kim & Kim, 1998), as well as the newly researched field Instagram (Djafarova & Rushworth, 2017), research about effects on SMI attitudes is lacking. The transfer theory, which states that during a celebrity (SMI) endorsement, endorser qualities are being
transferred to advertised products (Doss, 2011), could also apply to SMIs themselves. Therefore, the following research question and hypotheses have been developed:
H2: The higher someone’s evaluation of source credibility, the higher the (a) brand
attitude and (b) purchase intention.
10
Method
Design
To test the proposed hypotheses, a 3 (Sponsorship disclosure presence/type: no disclosure vs. label vs. hashtag) x 2 (Fit between SMI and product: low vs. high) between-subjects experimental design was employed in the Instagram context. Participants were randomly assigned to the conditions (see Table 1). All independent variables were manipulated and the rejection level for all analyses was set at p = .05.
Pretest
To establish the level of fit between SMI and product for different conditions, a pretest was conducted among 30 participants (different from the main study’s participants), collected through a convenience sample. The URL to the survey was posted in a women-only Facebook group of a fitness program. The average age was 24.17 (SD = 3.70), of which 83.9% is from The Netherlands. Prior to the pretest, various Instagram user-pages were reviewed to establish which products are promoted most by whom. For the pretest, an unknown SMI is being used with a fictitious username, “rosalinefitness”, which also puts emphasis on the SMI’s fitness identity. For the low-Fit condition, a teeth whitening kit had shown to be promoted often, and for the high-Fit condition a Branched-Chain Amino Acids (BCAAs) drink was chosen: a supplement which is often used by athletes or people who are into fitness in general.
11
Participants were randomly assigned to evaluate one of two pictures; the SMI promoting a teeth whitening kit, or the SMI promoting a BCAAs drink. In both pictures, the SMI was the same person, wearing a sports outfit, taking a ‘selfie’ in the mirror. The only difference is the product, which is showed in Appendix A (Figure 3 and 6).
To measure Fit between SMI and product, participants were asked to rate seven statements using seven-point scale items, ranging from 1) strongly disagree to 7) strongly agree (Olson & Thjømøe, 2011; Speed & Thompson, 2000), which can be found in the Appendix (B.1). Scales from both studies were adjusted to fit this study. Examples of statements are: “There is a logical connection between the brand and the endorser”, “The brand and the endorser fit together well” and “It makes sense to me that the endorser promotes this brand”. Principal component analysis with varimax rotation shows that the seven items form a single uni-dimensional scale: Only one component has an Eigenvalue above 1 (Eigenvalue is 5.239). The seven items explain 74.84 percent of the total variance. All items correlate positively with the first component, with good reliability (α = .942). Therefore it appears that the scale measures ‘fit between SMI and product’: the higher the score on the scale, the higher the fit between SMI and product.
A one-way between subjects ANOVA was conducted to compare the level of fit between SMI and product for the two different conditions. The Levene’s test showed no significance (F = 2.32, p = .139), which indicates that the groups differ from each other. Results showed a significantly lower fit between SMI and product for the teeth whitening kit (M = 3.08; SD = 1.18) than for the BCAAs drink (M = 5.2; SD = 0.81), F (1, 28) = 32.79, p < .001, 95% CI [1, 6.71]. Because of this successful pretest, the same products and username were used in the main study.
12 Participants
Participants were targeted through a convenience sample. Prerequisites for participants for the study were 1) women, 2) aged 18 to 30 years old, 3) with an active Instagram account. Reasons for these selection criteria were the fact that women between 18 and 30 are most likely to use Instagram (Sheldon & Bryant, 2016) and more interested in other women’s lifestyles, making them more susceptible to being influenced (Djafarova & Rushworth, 2017). An overall amount of 1430 participants was achieved, but 220 participants failed to respond (correctly) to all items, so their data was excluded from analysis. The total sample used consists of 1210 women, with a mean age of 21.64 (SD = 0.08), from 62 different countries worldwide. Countries with most participants were Australia (N = 126), Canada (N = 152), United Kingdom (N = 301) and United States (N = 399). Most participants were either employed full time (N = 410) or a student (N = 562). It can be stated that most of the
participants are experienced Instagram users and have a high level of fitness (knowledge) (see Appendix C). Also, the level of fitness is high among participants. The broad range of level of education and employment among participants is beneficial for the validity and
representativeness of this study, as is the broad representation of different nationalities.
Materials
The stimulus material consists of Instagram screenshots (see Appendix A, Figure 1-7), displayed as shown when someone is scrolling through their Instagram-feed, to strengthen the external validity. Participants are first shown a (fictitious) Instagram profile (Appendix A, Figure 1) of the SMI (rosalinefitness). All pictures on the profile are replicas of popular fitness influencer pictures and are therefore clearly fitness related. Real Instagram profiles have been observed, and amount of followers (485k), following (267) and number of pictures (1375) were chosen based on averages of 20 fitness SMIs, so the SMI will be more believable to be Instafamous. This way the stimulus material helps participants to imagine a real person.
13
Next a picture of the same SMI with the product (which depended on the condition) was shown. Fictitious brand names for both products (fit versus non-fit) were used to control pre-existing attitudes towards the brand (Till & Busler, 2000). Product’s names are “BiteWhite” for the teeth whitening kit and “Pro Extase” for the BCAAs drink. To manipulate sponsorship disclosure presence and type, three conditions were created: no sponsorship disclosure; a label; and a hashtag. All versions were identical, only the presence or type of disclosure differed. Specifically, the label said “Paid Partnership with Bite White/Pro Extase”, alike on Instagram, and was shown between the username and the picture. The hashtag used was “#sponsored” and was shown at the end of the text underneath the picture. This hashtag is an effective and clear sponsorship disclosure (Evans et al., 2017). The following texts, based on actual existing similar Instagram posts, were shown underneath the picture:
BiteWhite: Smile provided by the BiteWhite teeth whitening kit, giving you exactly
the shine you need for the perfect post-workout selfie
Pro Extase: Workout fueled by Pro Extase BCAAs, giving you exactly the pump you
need for the perfect post-workout selfie
Depending on the condition, the hashtag #sponsored followed directly after the sentence. No likes, comments and time stamps were shown for the post, to control other possible influential variables (see Appendix A, Figure 2-7).
Pilot test
In order to ensure smooth and understandable navigation through the online survey, a pilot test was conducted among 10 of this study’s researcher’s friends. They filled out the survey the same way as participants of the main study would, while sending comments or questions live through WhatsApp. Hereafter a few grammatical errors were fixed and text was added prior to the stimulus material, to better prepare participants to pay attention to the stimulus material.
14 Procedure
The study was conducted online, and the URL was posted in six different fitness-related Facebook groups, owned by fitness SMIs, to reach the relevant target group. The recruiting text is shown in Appendix F. In total, these Facebook groups combined count around 30.000 members. Each participant could complete the questionnaire in their own time and place, on the device of their choice (e.g., mobile phone or desktop). Participants had the option to leave their e-mail address at the end of the survey, to have a chance on winning a gift card for one of today’s most popular fitness clothing brands: “Gymshark”. Reasons for this are two-fold: to encourage possible participants to respond, and to prevent early drop-outs. The survey was active for a week.
When participants clicked on the URL for the survey, they first saw the factsheet “Instagram Influencers”. This name has been chosen to introduce the topic of the experiment, without priming participants towards the direction of marketing or advertisements. On the next page they had to agree to the Informed Consent form, after being reminded they could always decide not to participate. If they chose not to agree, they were sent to the end of the survey. The factsheet and Informed Consent form are shown in Appendices D and E.
After participants agreed, they were asked if they own and use an Instagram account. If not, it meant the end of the survey for them, because owning an active Instagram account is a prerequisite for this study. When continuing, participants were shown the first stimulus material: the Instagram profile page. Beforehand, participants had to imagine that they were scrolling through Instagram like they normally are and came across this profile. On the next page, they got a brief introduction for the second stimulus material: the Instagram post. They were randomly assigned to one of six versions of the post. For the profile page as well as the post, participants were asked to pay close attention before they were able to go to the next page, which was (automatically) allowed after ten seconds (to prevent click-through errors).
15
First, manipulation checks were conducted and thereafter participants were asked about their opinions about the influencer, which was measuring source credibility.
Measurements for purchase intention, brand attitude and SMI attitude followed subsequently. Persuasion knowledge was measured after the dependent variables, so effects of persuasion knowledge are because of the manipulation, and not because of the order of survey questions. As an additional manipulation check, fit between SMI and product was measured after. To determine possible control variables, questions about the participant’s level of (knowledge of) fitness were asked, followed by demographics (age, gender, whereabouts, education, and employment). After this, as additional control variables, participants were asked about their Instagram use, how they received the URL to the experiment and on which device they filled out the survey. On the last page participants had the option to leave their e-mail address to have a chance at winning a gift card. Here it was also mentioned they could send an e-mail to ask questions or deletion of answers, or interest in the results. This was the end of the survey.
Measures
Source credibility
Source credibility was measured using fifteen seven-point semantic differential items, covering three components source credibility consists of: attractiveness, trustworthiness and expertise (Ohanian, 1990). The items were mixed and divided in three parts, to eliminate possible recognition of intention to measure credibility for those who are acquainted with measurements like these. Participants were asked to give their opinion about the person in the Instagram post they just saw. Examples of items are unattractive/attractive, not classy/classy, undependable/dependable, dishonest/honest, not an expert/expert, inexperienced/experienced. A full list of items can be found in the Appendix (B.3).
Principal component analysis with varimax rotation shows two components with an Eigenvalue above 1 (Eigenvalues are 8.65 and 1.40). The fifteen items together explain 67
16
percent of the total variance. It appears that the scale measures ‘source credibility’ for all items: the higher the score on the scale, the higher the credibility, so the fifteen items will not be separated. Properties and Chronbach’s alpha of source credibility as well as the three components, attractiveness, trustworthiness and expertise separately can be found in Table 2. Analyses were done for source credibility as a whole as well as the three different components separately, to ensure the most specific results.
Brand attitude and SMI attitude
Brand attitude – someone’s internal evaluation of a brand (Mitchell & Olson, 2000) – and SMI attitude – someone’s internal evaluation of the SMI – were measured with five seven-point semantic differential items which have proved to measure brand attitude (Spears & Singh, 2004). Participants were asked to describe their overall feelings about the brand (for brand attitude) or person (SMI attitude) shown in the Instagram post they just saw. Items were unappealing/appealing, bad/good, unpleasant/pleasant, unfavorable/favorable, and
unlikable/likable, for both brand attitude and SMI attitude, but were asked separately. These items were considered appropriate for SMI attitude as well, considering that SMIs are practically becoming brands themselves and thus should be treated alike (Weinswig, 2016). Principal component analysis with varimax rotation shows that the five items for brand attitude form a single uni-dimensional scale: Only one component has an Eigenvalue above 1 (Eigenvalue is 4.01). The five items explain 80.15 percent of the total variance. It appears that the scale measures ‘brand attitude’: the higher the score on the scale, the higher the brand attitude.
For SMI attitude, principal component analysis also showed one component for the five items (Eigenvalue is 4.35), explaining 86.94 percent of the total variance. Therefore it appears that this scale measures ‘SMI attitude’: the higher the score on the scale, the higher the SMI attitude. See Table 2 for properties and Chronbach’s alpha of both variables.
17 Purchase intention
Purchase intention – (the probability of) someone’s willingness to buy the advertised product (Dodds, Monroe & Grewal, 1991) – was also measured with five seven-point semantic differential items (Spears & Singh, 2004). Participants were asked to what extent they intend to buy the brand shown in the Instagram post they just saw. The statements they had to rate can be found in the Appendix (B.2). Examples are “I would buy the brand (definitely not – definitely)”, “My purchase interest is (very low – very high)” and “I would buy the brand (probably not – probably)”.
Principal component analysis with varimax rotation shows that the five items form a single uni-dimensional scale: Only one component has an Eigenvalue above 1 (Eigenvalue is 3.93). The seven items explain 78.53 percent of the total variance. It appears that the scale measures ‘purchase intention’: the higher the score on the scale, the higher the purchase intention (see Table 2 for properties and Chronbach’s alpha).
Persuasion knowledge
Persuasion knowledge in this study is defined as whether or not someone recognizes that the SMI in the Instagram post is being paid. Participants were asked to answer a single item, with answering options: “Yes”, “No”, “I don’t know/Other” for the question: “Do you
18
think the influencer was paid to post the content?” Participants could add suggestions at the third option. This item indicates recognition of sponsored content. For analyses, the answer option “I don’t know” was declared as missing, leaving N = 1070. In total, 930 participants thought the SMI was paid (see Table 3). Mind that even with disclosures present, not all participants recognize the “paid” aspect of the post.
Control variables
To make sure that possible differences in dependent measures are because of the manipulation, multiple control variables have been tested. The level of (knowledge) of fitness was measured by asking participants if they are familiar with supplements like Creatine and BCAAs; if they had been working out in the last six months; and if yes, how much. Although actively using an Instagram account is a prerequisite for this study, different gradations of usage possibly exist. Participants were therefore asked how much time on average they are using Instagram per day. Also, an estimation of how long they have been active on Instagram and if they are familiar with the term “Instagram Influencer” was asked. If they were familiar with this term, the amount of (fitness-related) SMIs they are following was asked directly after. These variables are, together with the demographics age, gender, whereabouts, education and employment (see Appendix C), control variables in this study if they are not equally distributed across conditions.
19 Manipulation checks
Disclosure type, disclosure presence and fit between SMI and product were manipulated for this study. To test the manipulation of the sponsorship disclosures,
participants were asked whether or not the Instagram post was mentioned to be sponsored. They had four answering options: “Yes, a label that said Paid Partnership with…”, “Yes, a hashtag that said #sponsored”, “No, the post was not mentioned to be sponsored” or “I don’t know”. The last option was automatically declared missing. In total, 40% of the label
condition, 44.87% of the hashtag condition and 57.64% of the no disclosure condition passed the manipulation check, so only 52.4% of all participants remembered the disclosure
correctly. To take this into account for the results, the analyses have been repeated for participants who remembered the disclosure correctly only (N = 575). Participants who answered incorrectly are not completely restrained from analysis, since effects can occur unknowingly too.
Even though manipulation checks for fit between SMI and product were already implemented in the pretest, analysis has been repeated for the main study. For this test, it is necessary to recognize the name of the Instagram account and product, otherwise data was restrained from analysis. A one-way between subjects ANOVA was conducted to compare the level of fit between SMI and product for the two different conditions. The Levene’s test showed significance (F = 119.94, p < .001), so the assumption of equal variances in the population has been violated. Therefore the confidence interval was set at p < .01 for this analysis. The results showed a significantly lower fit between SMI and product for the teeth whitening kit (M = 3.68; SD = 1.27) than for the BCAAs drink (M = 5.30; SD = 0.85), F (1, 1208) = 657.48, p < .001, 99% CI [1.00, 7.00]. Thus, in the main study, manipulation was also successful.
20
Results
Randomization
To confirm equal distribution of participants over the six conditions (see Appendix G), a Chi square test was implemented, which showed no significance for age, nationality,
education and employment. A One-Way ANOVA was used for Instagram use, (level of) fitness (knowledge) and following of influencers, which also resulted in no significance. Any effects of the experimental manipulations cannot be attributed to a-priori differences between participants. Therefore, no variables were included as control variables in the main analyses.
Main analyses
Presence of sponsorship disclosures and persuasion knowledge
To answer RQ1, “To what extent is persuasion knowledge activated differently by the presence of a disclosure versus no presence of a disclosure?”, a One-way between subjects Analysis of Variance (ANOVA) (n = 1070) was conducted to compare the effect of
Disclosure presence on persuasion knowledge in the “disclosure” versus “no disclosure” condition. Levene’s test showed significance (F(1, 1068) = 48.66, p < .001), so equal
variances are not assumed. Therefore, the confidence level was set at p < .01 for this analysis. The One-way ANOVA showed a significant effect (see Table 4) of disclosure presence on persuasion knowledge (F(1, 1068) = 12.88, p < .001). Repeating the analysis for participants who remembered the (lack of) disclosure correctly did not result in a different outcome (F(1, 518) = 48.03, p < .001). Thus, presence of a disclosure activates persuasion knowledge more than no presence of a disclosure.
Type of sponsorship disclosure and persuasion knowledge
21
persuasion knowledge differently than the hashtag “#sponsored”?”, a One-way between subjects ANOVA (n = 723) was conducted to compare the effect of disclosure type on persuasion knowledge in the label and hashtag condition. Levene’s test showed no
significance (p = .242), so equal variances are assumed. The ANOVA showed no significant effect of disclosure type on persuasion knowledge (F(1, 721) = .342, p = .559). Therefore, the label “Paid Partnership with” does not activate persuasion knowledge differently than the hashtag “#sponsored”. Repeating the analysis for participants who remembered the disclosure correctly (n = 320) only did not result in significant findings (F(1, 318) = 2.20, p = .139).
Presence of sponsorship disclosures, persuasion knowledge and source credibility
To answer RQ3a, “What is the influence of the presence of a sponsorship disclosure on source credibility?”, also a One-way ANOVA was conducted (n = 1210). The Levene’s test was not significant (F(1, 1208) = 1.46, p = .227). The results (see Table 4) showed no significant difference in source credibility for the conditions (F(1, 1208) = 2.47, p = .116), but a significant effect was found on the component trustworthiness (F(1, 1208) = 3.98, p = .045). When repeating the analysis for participants who remembered the presence of a disclosure correctly, Levene’s test showed no significance (F(1, 573) = .015, p = .903). The Multiple Analysis of Variance (MANOVA) showed that disclosure presence positively influences source credibility, attractiveness and trustworthiness (but not expertise) (see Table 4). To answer RQ3b, “To what extent does persuasion knowledge influence source credibility?”, a One-way between subjects ANOVA (n = 1070) was conducted with
persuasion knowledge (yes versus no) as between-subjects factor. The Levene’s test showed significance (F(1, 1068) = 7.70, p = .006), so the assumption of equal variances in the
population has been violated. Therefore, the confidence level was set at p < .01. The results in Table 5 showed that participants with persuasion knowledge score significantly lower on source credibility than participants without persuasion knowledge (F(1, 1068) = 11.73, p =
22
.001). A MANOVA was conducted for the three separate components of source credibility, which showed that equal variances have been violated for trustworthiness (F(1, 1068) = 5.52,
p= .019) and expertise (F(1, 1068) = 8.69, p = .003), so the confidence level was set at p < .01
for those analyses. Results showed a significant effect of persuasion knowledge on
trustworthiness (F(1, 1068) = 24.86, p < .001), and on expertise (F(1, 1068) = 8.60, p = .003) No significant effects on attractiveness were found (F(1, 1068) = 1.72, p = .190). Therefore, persuasion knowledge does not influence source credibility as a whole; only negatively influences trustworthiness and expertise of the source. The higher the persuasion knowledge, the lower the source trustworthiness and expertise.
Interestingly, when repeating the analysis only for participants who correctly
remembered the (lack of) disclosure, the effect of persuasion knowledge on source credibility (and the three components separately) was not significant anymore (F(1, 518) = 1.93, p = .166). Therefore, for participants who remembered the disclosure correctly, persuasion knowledge did not negatively affect source credibility.
23 Fit between SMI and product, presence of sponsorship disclosures and persuasion knowledge
As established before, disclosure presence positively affects persuasion knowledge: Persuasion knowledge is higher with a disclosure present than without a disclosure present. To establish the role of fit between SMI and product in this relationship (RQ4), a Multivariate Analysis of Covariance (n = 1070) with fit between SMI and product as covariate showed that Levene’s test was significant (F (3, 1066) = 115.40, p < .001) so the confidence level was set at p < .01. Results showed a significant interaction effect (Disclosure presence x Fit between SMI and product) on persuasion knowledge (F(3, 1066) = 31.08, p < .001, η² = .08).
Therefore, fit between SMI and product explains 8.0% of the total variance in persuasion knowledge: Participants score significantly lower on persuasion knowledge when the fit is high than when the fit is low, regardless of the presence or type of a disclosure (see Table 6). Results for participants who remembered the presence of disclosure correctly did not differ much (p < .001, η² = .058).
24 Fit between SMI and product, brand attitude, purchase intention, source credibility and SMI attitude
To test H1a and H1b stating that good fit between SMI and product has a positive effect on brand attitude and purchase intention, a MANOVA (n = 1210) was conducted with fit between SMI and product (high vs. low) as between-subjects factor. Mind that because disclosure presence and type were not included in this analysis, no distinction between sponsorship disclosures is made, so this analysis is based solely on the variable fit between SMI and product. The Levene’s test showed no significance for purchase intention (F = 3.03,
p = .082), indicating that those groups are equal. For brand attitude (F = 11.60, p = .001) no
equal variances are assumed, so the confidence level was set at p < .01. The results in Table 7 showed a significant positive effect of high fit between SMI and product, as compared to a low fit between SMI and product, on purchase intention (F(1, 1208) = 40.58, p < .001) and brand attitude (F(1, 1208) = 8.12, p = .004), accepting bot H1a and H1b.
Source credibility was added to the MANOVA, to answer RQ5a, “What is the effect of fit between SMI and product on source credibility?”. The Levene’s test showed
significance (F(1, 1208) = 4.29, p = .039), so the confidence level was set at p < .01 for the analysis. The results (see Table 7) showed that good fit between SMI and product does not positively influence source credibility significantly (F(1, 1208) = 3.39, p = .066), when
25
compared to bad fit.
This analysis has been repeated for the three components of source credibility separately. Levene’s test showed no significance for attractiveness (F(1, 1208) = 1.81, p = .179), trustworthiness (F(1, 1208) = 0.69, p = .407) and expertise (F(1, 1208) = 3.75, p = .053), so equal variances are assumed. As displayed in Table 7, the MANOVA showed a significant effect of fit between SMI and product on trustworthiness (F(1, 1208) = 4.14, p = .042) and expertise (F(1, 1208) = 6.60, p = .010), but not on attractiveness (F(1, 1208) = 0.12,
p = .733). Therefore the answer to RQ5a is: The fit between SMI and product does not
positively influence source credibility as a whole, but it does positively influence the trustworthiness and expertise of the source.
The variable SMI attitude was included in the MANOVA to answer RQ5b: “What is the effect of fit between SMI and product on SMI attitude?”. Levene’s test showed no
significance (F(1, 1208) = 1.57, p = .211). The results in Table 7 showed that fit between SMI and product does not significantly influence SMI attitude (F(1, 1208) = 0.51, p = .475).
Source credibility, SMI attitude, brand attitude and purchase intention
To test H2a and H2b, stating that the higher the source credibility, the higher the brand attitude and purchase intention, and RQ6: “To what extent does source credibility predict SMI
26
attitude?”, three separate single linear regressions were calculated for the three different dependent variables. Results (Table 8) showed that level of source credibility predicts purchase intention (F(3, 1206) = 72.68, p < .001, R² = 0.153), brand attitude (F(3, 1206) = 142.73, p < .001, R² = 0.255) and SMI attitude (F(3, 1206) = 487.97, p < .001, R² = 0.505). Only expertise does not significantly predict SMI attitude. Therefore H2a and H2b are accepted: the higher the source credibility, the higher the brand attitude and purchase
intention. In line with these findings, the answer to RQ6 is: the higher the source credibility, the higher the SMI attitude.
Conclusion and discussion
This study’s main focus was the effect of presence and type of sponsorship disclosure in the Instagram context on persuasion knowledge and source credibility. Findings are in line with the Persuasion Knowledge Model (Friestad & Wright, 1994); sponsorship disclosure presence indeed activates persuasion knowledge more when compared to absence of a disclosure, but disclosure type does not matter. The “Paid Partnership with” label activates persuasion knowledge just as much as the “#sponsored”-hashtag. In turn, persuasion knowledge has shown to negatively affect source credibility, but this effect does not exist for people who actually correctly notice the disclosure. Therefore it would be beneficial to educate Instagram
27
users about disclosures, ensuring that they pay attention to them more. Then disclosures lead to better persuasion knowledge and higher source credibility simultaneously, which is a win-win situation. Additionally, this study reinforces evidence for the match-up hypothesis (Kamins & Gupta, 1994; Till & Busler, 2000); good fit between SMI and product positively affects brand attitudes and purchase intentions. Building on this theory, good fit does lead to higher source trustworthiness and expertise and moreover influences the relationship between the presence of a disclosure and persuasion knowledge, but good fit alone is not enough to establish substantial positive SMI attitudes. Therefore, brands and SMIs that are collaborating should pursue good fit and high source credibility; that leads to high purchase intentions, brand attitudes and SMI attitudes. This is in line with previous research (Amos, Holmes and Strutton, 2008; Eisend, 2004; Wilson & Sherrell, 1993), although theory about the influence on the SMI on Instagram is a novel addition to the research field.
This study does not come without limitations. Firstly, although the sample of
respondents is quite big, with high internal validity, external validity is limited to women who are experienced in fitness and Instagram. Still, this group is interesting to investigate and target, because they are most likely to follow – and be influenced by – SMIs (Djafarova & Rushworth, 2017). Noteworthy is the high persuasion knowledge of participants of this study, especially since only 52.4% remembered the disclosure correctly, which indicates that not everyone pays attention (well) to disclosures. This is why analyses concerning only those who remembered the disclosure correctly have been added. It might be beneficial to first look at attentiveness towards sponsorship disclosures on Instagram, before attributing effects to such manipulations. No research about attentiveness towards the label existed, but future research should provide insight regarding this. On the other hand, effects do not always happen consciously, so usefulness of the results is not necessarily threatened by lack of attention towards disclosures. And although persuasion knowledge is already high in this sample of
28
participants, no gradations of persuasion knowledge have been investigated, which could provide more in-depth evaluations of the role persuasion knowledge plays in the relationship between sponsorship disclosures and source credibility: Because persuasion knowledge was a dichotomous variable in this study, a full mediation could not be established. Therefore, future research could broaden theory on this topic more insightfully.
Additionally, on Instagram, the role of attractiveness in source credibility seems doubtful. Although attractiveness does affect attitudes and purchase intentions, it did not significantly change due to the manipulations of this study. Traditional theories about (measurements of) source credibility do not seem to fully apply to the Instagram context; attractiveness does not play an important role in establishing credibility. Since attractiveness is based on appearance, chances are high people only follow SMIs on Instagram who they find attractive already, which would explain the high mean-score on attractiveness in all conditions. It is advised to treat attractiveness separately from trustworthiness and expertise when researching source credibility in the Instagram context, with trustworthiness as most important factor. This has shown to significantly change outcomes of this study. This also means that SMIs should focus on reaching trustworthiness and expertise – not attractiveness – when aspiring credibility.
References
Abendroth, L. J., & Heyman, J. E. (2013). Honesty is the best policy: The effects of
disclosure in word-of-mouth marketing. Journal of Marketing Communications, 19(4), 245-257.
Amos, C., Holmes, G., & Strutton, D. (2008). Exploring the relationship between celebrity endorser effects and advertising effectiveness: A quantitative synthesis of effect size. International Journal of Advertising, 27(2), 209-234.
29
Bakas, A. (2017). Influencers steeds belangrijker om jongeren te bereiken.
Communicatietrend. Retrieved from:
https://www.communicatietrends.nl/influencers-steeds-belangrijker-om-jongeren-te-bereiken/
Becker-Olsen, K. L. (2003). And now, a word from our sponsor--a look at the effects of sponsored content and banner advertising. Journal of Advertising, 32(2), 17-32.
Bergkvist, L., Hjalmarson, H., & Mägi, A. W. (2016). A new model of how celebrity
endorsements work: attitude toward the endorsement as a mediator of celebrity source and endorsement effects. International Journal of Advertising, 35(2), 171-184.
Boerman, S. C., 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 Advances in Advertising Research, 9, 273-284.
Boerman, S. C., 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.
Braun, J. (2017). Consumers: Credibility is king when it comes to influencer marketing. Izea. Retrieved from: https://izea.com/2017/07/18/credibility-king-influencer-marketing/
Brehm, J. W. (1966). A theory of psychological reactance. New York: Academic Press.
Brooks, A. (2017). Instagram's new influencer tools quash the question of credibility.
AdNews. Retrieved from:
30
Carl, W. J. (2008). The role of disclosure in organized word‐ of‐ mouth marketing programs. Journal of Marketing Communications, 14(3), 225-241.
Chaiken, S., & Maheswaran, D. (1994). Heuristic processing can bias systematic processing: Effects of source credibility, argument ambiguity, and task importance on attitude judgment. Journal of personality and social psychology, 66, 460-460.
Chapple, C., & Cownie, F. (2017). An Investigation into Viewers’ Trust in and Response Towards Disclosed Paid-for-Endorsements by YouTube Lifestyle Vloggers. Journal
of Promotional Communications, 5(2).
Cheung, M. Y., Luo, C., Sia, C. L., & Chen, H. (2009). Credibility of electronic word-of-mouth: Informational and normative determinants of online consumer
recommendations. International Journal of Electronic Commerce, 13(4), 9-38.
Daugherty, T., & Hoffman, E. (2014). eWOM and the importance of capturing consumer attention within social media. Journal of Marketing Communications, 20(1-2), 82-102.
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.
De Pelsmacker, P., & Neijens, P. C. (2012). New advertising formats: How persuasion
knowledge affects consumer responses. Journal of Marketing Communications, 18(1), 1-4.
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.
31
Dodds, W. B., Monroe, K. B., & Grewal, D. (1991). Effects of price, brand, and store information on buyers' product evaluations. Journal of marketing research, 307-319.
Doh, S. J., & Hwang, J. S. (2009). How consumers evaluate eWOM (electronic word-of-mouth) messages. CyberPsychology & Behavior, 12(2), 193-197.
Doss, S. (2011). The transference of brand attitude: the effect on the celebrity endorser. Journal of Management and Marketing Research, 7, 1.
Eisend, M. (2004). Is it still worth to be credible? A meta-analysis of temporal patterns of source credibility effects in marketing. ACR North American Advances, 352-357.
Engle, M. K. (2017). Sample educational letter to social media influencer. Federal Trade
Commision. Retrieved from:
https://www.ftc.gov/system/files/attachments/press- releases/ftc-staff-reminds-influencers-brands-clearly-disclose-relationship/influencer_template.pdf
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, 1-12.
Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope with persuasion attempts. Journal of Consumer Research, 21(1), 1-31.
Gwinner, K. P., & Eaton, J. (1999). Building brand image through event sponsorship: The role of image transfer. Journal of advertising, 28(4), 47-57.
Hoffman, E., & Daugherty, T. (2013). Is a picture always worth a thousand words? Attention to structural elements of eWOM for consumer brands within social media. ACR North
32
Hung, K. (2014). Why celebrity sells: A dual entertainment path model of brand endorsement. Journal of advertising, 43(2), 155-166.
Jin, S. A. A., & Phua, J. (2014). Following celebrities’ tweets about brands: The impact of Twitter-based electronic word-of-mouth on consumers’ source credibility perception, buying intention, and social identification with celebrities. Journal of
Advertising, 43(2), 181-195.
Kamins, M. A., & Gupta, K. (1994). Congruence between spokesperson and product type: A matchup hypothesis perspective. Psychology & Marketing, 11(6), 569-586.
Kim, B., Pasadeos, Y., & Barban, A. (2001). On the deceptive effectiveness of labeled and unlabeled advertorial formats. Mass Communication and Society, 4(3), 265-281. doi:10.1207/S15327825MCS0403_02
Kim, M., & Song, D. (2017). When brand-related UGC induces effectiveness on social media: the role of content sponsorship and content type. International Journal of Advertising, 1-20.
Kirmani, A., & Shiv, B. (1998). Effects of source congruity on brand attitudes and beliefs: the moderating role of issue-relevant elaboration. Journal of Consumer Psychology, 7(1), 25-47.
Koernig, S. K., & Boyd, T. C. (2009). To catch a tiger or let him go: The match-up effect and athlete endorsers for sport and non-sport brands. Sport Marketing Quarterly, 18(1), 25.
Koo, G. Y., Quarterman, J., & Flynn, L. (2006). Effect of perceived sport event and sponsor image fit on consumers' cognition, affect, and behavioral intentions. Sport Marketing
33
Lafferty, B. A., Goldsmith, R. E., & Newell, S. J. (2002). The dual credibility model: The influence of corporate and endorser credibility on attitudes and purchase
intentions. Journal of Marketing Theory and Practice, 10(3), 1-11.
Mazodier, M., & Quester, P. (2014). The role of sponsorship fit for changing brand affect: A latent growth modeling approach. International Journal of Research in
Marketing, 31(1), 16-29.
Mediakix. (2017). 93% of top celebrity social media endorsements violate FTC guidelines.
Mediakix. Retrieved 11/27/2017, from:
http://mediakix.com/2017/05/celebrity-social-media-endorsements-violate-ftc-instagram/
Miron, A. M., & Brehm, J. W. (2006). Reactance theory - 40 years later. Zeitschrift für
Sozialpsychologie, 37(1), 9-18.
Misra, S., & Beatty, S.E. (1990). Celebrity spokesperson and brand congruence: An assessment of recall and affect. Journal of Business Research, 21, 159–173.
Mitchell, A. A., & Olson, J. C. (2000). Are product attribute beliefs the only mediator of advertising effects on brand attitude?. Advertising & Society Review, 1(1).
Moran, G., & Muzellec, L. (2017). eWOM credibility on social networking sites: A framework. Journal of Marketing Communications, 23(2), 149-161.
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.
Olson, E. L. (2010). Does sponsorship work in the same way in different sponsorship contexts?. European Journal of Marketing, 44(1/2), 180-199.
34
Olson, E. L., & Thjømøe, H. M. (2011). Explaining and articulating the fit construct in sponsorship. Journal of Advertising, 40(1), 57-70.
Petty, R. E., & Cacioppo, J. T. (1979). Effects of forwarning of persuasive intent and involvement on cognitive responses and persuasion. Personality and Social
Psychology Bulletin, 5(2), 173-176..
Pornpitakpan, C. (2004). The effect of celebrity endorsers' perceived credibility on product purchase intention: The case of Singaporeans. Journal of International Consumer
Marketing, 16(2), 55-74.
Reichelt, J., Sievert, J., & Jacob, F. (2014). How credibility affects eWOM reading: The influences of expertise, trustworthiness, and similarity on utilitarian and social functions. Journal of Marketing Communications, 20(1-2), 65-81.
Rifon, N. J., Choi, S. M., Trimble, C. S., & Li, H. (2004). Congruence effects in sponsorship: The mediating role of sponsor credibility and consumer attributions of sponsor
motive. Journal of Advertising, 33(1), 30-42.
Russell, C. A., & Rasolofoarison, D. (2017). Uncovering the power of natural endorsements: a comparison with celebrity-endorsed advertising and product
placements. International Journal of Advertising, 36(5), 761-778.
Sheldon, P., & Bryant, K. (2016). Instagram: Motives for its use and relationship to narcissism and contextual age. Computers in Human Behavior, 58, 89-97.
Smith, G. (2004). Brand image transfer through sponsorship: A consumer learning perspective. Journal of marketing management, 20(3-4), 457-474.
35
Social Media Examiner. (2017). Social Media Marketing Industry Report. Social Media
Examiner. Retrieved 11/18/2017, from:
https://www.socialmediaexaminer.com/social-media-marketing-industry-report-2017/
Spears, N., & Singh, S. N. (2004). Measuring attitude toward the brand and purchase intentions. Journal of Current Issues & Research in Advertising, 26(2), 53-66.
Speed, R., & Thompson, P. (2000). Determinants of sports sponsorship response. Journal of
the Academy of Marketing Science, 28(2), 226-238.
Statista. (2017). Instagram monthly active users 2017. Statistic. Retrieved 11/6/2017, from: https://www.statista.com/statistics/253577/number-of-monthly-active-instagram-users/
Till, B. D., & Busler, M. (2000). The match-up hypothesis: Physical attractiveness, expertise, and the role of fit on brand attitude, purchase intent and brand beliefs. Journal of
Advertising, 29(3), 1-13. doi: 10.1080/00913367.2000.10673613
Tutaj, K., & Van Reijmersdal, E. A. (2012). Effects of online advertising format and
persuasion knowledge on audience reactions. Journal of Marketing Communications, 18(1), 5-18. doi:10.1080/13527266.2011.620765
Van Noort, G., Antheunis, M. L., & Van Reijmersdal, E. A. (2012). Social connections and the persuasiveness of viral campaigns in social network sites: Persuasive intent as the underlying mechanism. Journal of Marketing Communications, 18(1), 39-53.
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
36
Wei, M. L., Fischer, E., & Main, K. J. (2008). An examination of the effects of activating persuasion knowledge on consumer response to brands engaging in covert
marketing. Journal of Public Policy & Marketing, 27(1), 34-44.
Weinswig, D. (2016). Influencers are the new brands. Forbes. Retrieved from:
https://www.forbes.com/sites/deborahweinswig/2016/10/05/influencers-are-the-new-brands/#1a4140737919
Wilson, E. J., & Sherrell, D. L. (1993). Source effects in communication and persuasion research: A meta-analysis of effect size. Journal of the Academy of Marketing
Science, 21(2), 101-112.
Wong, K. (2014). The explosive growth of influencer marketing and what it means for you.
Forbes. Retrieved from:
https://www.forbes.com/sites/kylewong/2014/09/10/the-explosive-growth-of-influencer-marketing-and-what-it-means-for-you
Wu, P. C., & Wang, Y. C. (2011). The influences of electronic word-of-mouth message appeal and message source credibility on brand attitude. Asia Pacific Journal of
Marketing and Logistics, 23(4), 448-472.
Yoon, K., Kim, C. H., & Kim, M. S. (1998). A cross-cultural comparison of the effects of source credibility on attitudes and behavioral intentions. Mass Communication and
37
APPENDIX
A. Stimulus material
38 Different conditions:
Figure 2 Figure 3
39
Figure 4 Figure 5
40
Figure 6 ` Figure 7
41
B. Scales
1. Items for fit between SMI and product
There is a logical connection between the brand and the endorser
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
The brand and the endorser are similar
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
The brand and the endorser fit together well
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
The brand and the endorser stand for similar things
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
The image of the brand and the image of the endorser are similar
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
It makes sense to me that the endorser promotes this brand
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
The endorser and the product are similar
42 2. Items for Purchase intention
I would buy the brand
Never 1 2 3 4 5 6 7 Definitely
I intend to buy the brand
Definitely not 1 2 3 4 5 6 7 Definitely
I would buy the brand:
Probably not 1 2 3 4 5 6 7 Probably
My purchase interest is:
Very low 1 2 3 4 5 6 7 Very high
I would buy the brand:
43 3. Items for SMI Credibility
The influencer is: Attractiveness:
Unattractive 1 2 3 4 5 6 7 Attractive Not classy 1 2 3 4 5 6 7 Classy
Ugly 1 2 3 4 5 6 7 Beautiful
Plain 1 2 3 4 5 6 7 Elegant
Not sexy 1 2 3 4 5 6 7 Sexy
Trustworthiness: Undependable 1 2 3 4 5 6 7 Dependable Dishonest 1 2 3 4 5 6 7 Honest Unreliable 1 2 3 4 5 6 7 Reliable Insincere 1 2 3 4 5 6 7 Sincere Untrustworthy 1 2 3 4 5 6 7 Trustworthy Expertise:
Not an expert 1 2 3 4 5 6 7 Expert Inexperienced 1 2 3 4 5 6 7 Experienced Unknowledgeable 1 2 3 4 5 6 7 Knowledgeable Unqualified 1 2 3 4 5 6 7 Qualified Unskilled 1 2 3 4 5 6 7 Skilled
44 4. Items for persuasion knowledge
- When a post is sponsored, I always expect someone to mention that
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
- When a post is sponsored, I always assume someone received money
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
- When someone received free stuff, I always expect them to mention that
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
- When a post is NOT sponsored, I always expect someone to mention that
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
- Even without a disclosure, I know when a post is sponsored
I strongly disagree 1 2 3 4 5 6 7 I strongly agree
- When a brand is mentioned without a disclosure, I still expect the post to be sponsored I strongly disagree 1 2 3 4 5 6 7 I strongly agree
- When someone does NOT mention a sponsorship, I always assume the post is NOT sponsored
45
47
D. Factsheet “Instagram Influencers” Welcome!
You are invited to participate in a research project that is being carried out under the
responsibility of the Graduate School of Communication, which forms part of the University of Amsterdam.
The title of the research project for which we are requesting your assistance is “Instagram Influencers”. Women aged 18 to 30 years old, owning an Instagram account can participate in this project. The objective of the research is to gain insight in influencers on Instagram.
The experiment we ask you to participate in exists of a small survey for you to fill in. This will probably last approximately a maximum of 10 minutes. In recognition for your assistance, you have the chance of winning a Gymshark gift card if you leave your e-mail address at the end of the survey.
As this research is being carried out under the responsibility of the Graduate School of Communication, University of Amsterdam, we can guarantee that:
1. Your anonymity will be safeguarded, and that your answers or data will not be passed on to third parties under any conditions, unless you first give your express permission for this.
2. You can refuse to participate in the research or cut short your participation without having to give a reason for doing so. You also have up to 24 hours after participating to withdraw your permission to allow your answers or data to be used in the research.
3. Participating in the research will not entail your being subjected to any appreciable risk or discomfort, the researchers will not deliberately mislead you, and you will not be exposed to any explicitly offensive material.
48
4. No later than five months after the conclusion of the research, we will be able to provide you with a research report that explains the general results of the research.
For more information about the research and the invitation to participate, you are welcome to contact the project leader at any time: Roos Olthof. You can send an e-mail to
roos.olthof@student.uva.nl. Any complaints or comments will be treated in the strictest confidence.
We trust that we have provided you with sufficient information. We would like to take this opportunity to thank you in advance for your assistance with this research, which is greatly appreciated.
With kind regards,
Roos Olthof
(Don't forget to leave your e-mail address at the end of this survey to have a chance at winning the Gymshark gift card!)
